Metals Handbook: Corrosion

ASM INTERNATIONAL

®

Publication Information and Contributors

Corrosion was published in 1987 as Volume 13 of the 9th Edition Metals Handbook. With the fourth printing (1992), the series title was changed to ASM Handbook. The Volume was prepared under the direction of the ASM International Handbook Committee.

Volume Chairmen The Volume Chairmen were Lawrence J. Korb, Rockwell International and David L. Olson, Colorado School of Mines

Authors and Reviewers • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

H. Ackerman Edco Products, Inc. Donald R. Adolphson Sandia Laboratories D.C. Agarwal Haynes International, Inc. V.S. Agarwala Naval Air Development Center John D. Alkire Amoco Corporation John R. Ambrose University of Florida Albert A. Anctil Department of the Army Phillip J. Andersen Zimmer D.B. Anderson National Bureau of Standards Peter L. Andresen General Electric Research and Development Center Dennis M. Anliker Champion International Corporation Frank J. Ansuini Consulting Engineer A.J. Armini Surface Alloys Corporation William G. Ashbaugh Cortest Engineering Services Aziz I. Asphahani Haynes International, Inc. Terje Kr. Aune Norsk Hydro (Norway) Denise M. Aylor David Taylor Naval Ship Research & Development Center Robert Baboian Texas Instruments, Inc. C. Bagnall Westinghouse Electric Corporation V. Baltazar Noranda Research Centre (Canada) Edward N. Balko Englehard Corporation Calvin H. Baloun Ohio University R.C. Bates Westinghouse Electric Corporation Michael L. Bauccio The Boeing Company Charles Baumgartner General Electric Company Richard Baxter Sealand Corrosion Control, Ltd. R.P. Beatson Pulp and Paper Research Institute of Canada John A. Beavers Battelle Columbus Division T.R. Beck Electrochemical Technology, Inc. S. Belisle Noranda Inc. (Canada) Robert J. Bell Heat Exchanger Systems, Inc. B.W. Bennett Bell Communications Reseach David C. Bennett Champion International Corporation E.L. Bereczky Unocal Corporation Carl A. Bergmann Westinghouse Electric Corporation I.M. Bernstein Carnegie-Mellon University A.K. Bhambri Morton Thiokol Inc. Robert C. Bill Lewis Research Center National Aeronautics & Space Administration C.R. Bird Stainless Foundry & Engineering, Inc. Neil Birks University of Pittsburgh

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

R. Ross Blackwood Tenaxol, Inc. Malcolm Blair Delray Steel Casting, Inc. A.J. Blazewicz Babcock & Wilcox J. Blough Foster Wheeler Development Corporation Michael E. Blum FMC Corporation Bennett P Boffardi Calgon Corporation P.W. Bolmer Kaiser Aluminum & Chemical Corporation Rodney R. Boyer Boeing Commercial Airplane Company Samuel A. Bradford University of Alberta (Canada) Robert W. Bradshaw Sandia National Laboratories J.W. Braithwaite Sandia National Laboratories W.F. Brehm Westinghouse Hanford Company P. Bro Technical Consultant R. Brock Teledyne CAE Alan P. Brown Argonne National Laboratory M. Browning Technical Consultant S.K. Brubaker E.I. Du Pont de Nemours & Company, Inc. John C. Bruno J & L Specialty Products Corporation James H. Bryson Inland Steel Company R.J. Bucci Alcoa Laboratories Charles D. Bulla ICI Americas Inc. Donald S. Burns Spraymetal, Inc. H.E. Bush Corrosion Consultant Dwight A. Burford Colorado School of Mines J. Butler Platt Brothers & Company W.S. Butterfield Beloit Corporation L.E. Cadle Texas Eastern Products Pipeline Company John Campbell Quality Carbide, Inc. L.W. Campbell General Magnaplate Corporation Thomas W. Cape Chemfil Corporation Bernie Carpenter Colorado School of Mines Allan P. Castillo Sandusky Foundry & Machine Company Victor Chaker The Port Authority of New York and New Jersey George D. Chappell Nalco Chemical Company Robert S. Charlton B.H. Levelton & Associates, Ltd. (Canada) G. Dale Cheever General Motors Research Laboratories Newton Chessin Martin Marietta Aerospace Robert John Chironna Croll-Reynolds Company, Inc. Omesh K. Chopra Argonne National Laboratory Wendy R. Cieslak Sandia National Laboratories Ken Clark Fansteel--Wellman Dynamics Clive R. Clayton State University of New York at Stony Brook S.K. Coburn Corrosion Consultants, Inc. Robert Coe Public Service Company of Colorado B. Cohen Air Force Wright Aeronautical Laboratories Roland L. Coit Technical Consultant L. Coker Exxon Chemical Company N.C. Cole Combustion Engineering Inc. E.L. Colvin Aluminum Company of America J.B. Condon Martin Marietta Energy Systems, Inc. B. Cooley Hoffman Silo Inc. Richard A. Corbett Corrosion Testing Laboratories, Inc. B. Cox Atomic Energy of Canada Ltd. W.M. Cox Corrosion and Protection Centre University of Manchester (England)

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Bruce Craig Metallurgical Consultants, Inc. K.R. Craig Combustion Engineering Inc. William R. Cress Allegheny Power Service Corporation Paul Crook Haynes International, Inc. Thomas W. Crooker Naval Research Laboratory Ronald D. Crooks Hercules, Inc. Carl E. Cross Colorado School of Mines Robert Crowe Naval Research Laboratory J.R. Crum Inco Alloys International, Inc. Daniel Cubicciotti Electric Power Research Institute William J. Curren Cortronics, Inc. Michael J. Cusick Colorado School of Mines Carl J. Czajkowski Brookhaven National Laboratory Brian Damkroger Colorado School of Mines P.L. Daniel Babcock & Wilcox Joseph C. Danko American Welding Institute Vani K. Dantam General Motors Corporation C.V. Darragh The Timken Company Ralph M. Davison Avesta Stainless, Inc. Sheldon W. Dean Air Products and Chemicals, Inc. Terry DeBold Carpenter Technology Corporation Thomas F. Degnan Consultant James E. Delargey Detroit Edison Stephen C. Dexter University of Delaware Ronald B. Diegle Sandia National Laboratories J.J. Dillon Martin Marietta Energy Systems, Inc. Bill Dobbs Air Force Wright Aeronautical Laboratories R.F. Doelling The Witt Company James E. Donham Consultant R.B. Dooley Electric Power Research Institute D.L. Douglass University of California at Los Angeles Donald E. Drake Mobil Corporation L.E. Drake Stauffer Chemical Company Carl W. Dralle Ampco Metal Edgar W. Dreyman PCA Engineering, Inc. Barry P. Dugan St. Joe Resources Company Arthur K. Dunlop Corrosion Control Consultant Walter B. Ebner Honeywell Inc. G.B. Elder Union Carbide Corporation Peter Elliott Cortest Engineering Services Inc. Edward Escalante National Bureau of Standards Charles L.L. Faust Consultant R. Fekete Ford Motor Company Ron Fiore Sikorsky Aircraft S. Fishman Office of Naval Research W.D. Fletcher Westinghouse Electric Corporation Mars G. Fontana Materials Technology Institute F. Peter Ford General Electric Research & Development Center Robert Foreman Park Chemical Company L.D. Fox Tennessee Valley Authority Anna C. Fraker National Bureau of Standards David Franklin Electric Power Research Institute Douglas B. Franklin George C. Marshall Space Flight Center Administration

National Aeronautics & Space

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

David N. French David French Inc. R.A. French BASF Corporation R.E. Frishmuth Cortest Laboratories Allan Froats Chromasco/Timminco, Ltd. (Canada) P. Fulford Florida Power and Light Company J.M. Galbraith Arco Alaska Inc. J.W. Gambrell American Hot Dip Galvanizers Association S. Ganesh General Electric Company Richard P. Gangloff University of Virginia Thomas W. Gardega National Thermal Spray Company Warren Gardner Department of the Air Force Andrew Garner Pulp and Paper Research Institute of Canada D. Gearey Corrosion and Protection Centre University of Manchester (England) George A. Gehring, Jr. Ocean City Research Corporation Floyd Gelhaus Electric Power Research Institute Randall M. German Rensselaer Polytechnic Institute William J. Gilbert Croll-Reynolds Company, Inc. Paul S. Gilman Allied-Signal William Glaeser Battelle Columbus Division Samuel V. Glorioso Lyndon B. Johnson Space Center National Aeronautics & Space Administration Cluas G. Goetzel Stanford University Michael Gold Babcock & Wilcox Barry M. Gordon General Electric Company Gerald M. Gordon General Electric Company Andrew John Gowarty Department of the Army Robert Graf United Technologies Research Center Richard D. Granata Lehigh University Stanley J. Green Electric Power Reseach Institute C.D. Griffin Carbomedics, Inc. Richard B. Griffin Texas A&M University John Grocki Haynes International, Inc. Earl C. Groshart Boeing Aerospace Company V.E. Guernsey Electroplating Consultants International Ronald D. Gundry Buckeye Pipe Line Company S.Wm. Gunther Mangel, Scheuermann & Oeters, Inc. Jack D. Guttenplan Rockwell International H. Guttman Noranda Research Centre (Canada) J. Gutzeit Amoco Corporation Charles E. Guzi Procter and Gamble Company Harvey P. Hack David Taylor Naval Ship Research & Development Center J.D. Haff E.I. Du Pont de Nemours & Company, Inc. Christopher Hahin Materials Protection Associates William B. Hampshire Tin Research Institute, Inc. James A. Hanck Pacific Gas & Electric Company Paul R. Handt Dow Chemical Company Michael Haroun Oklahoma State University Charles A. Harper Westinghouse Electric Corporation J.A. Hasson E.F. Houghton & Company David Hawke Amax Magnesium Gardner Haynes Texas Instruments, Inc. F.H. Haynie Environmental Protection Agency Robert H. Heidersbach California Polytechnic State University C. Heiple Rockwell International

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Lawrence E. Helwig USX Corporation James B. Hill Allegheny Ludlum Corporation James Hillis Dow Chemical Company John P. Hirth Ohio State University Norris S. Hirota Electric Power Research Institute N.J. Hoffman Rockwell International E.H. Hollingsworth Aluminum Company of America (retired) A. Craig Hood ACH Technologies R.L. Horst Aluminum Company of America J.B. Horton J.B. Horton Company K. Houghton Wollaston Alloy Inc. Louis E. Huber, Jr. Technical Consultant F.J. Hunkeler NRC Inc. H.Y. Hunsicker Aluminum Company of America (retired) J.R. Hunter Pfizer Inc. Carl A. Hutchinson Federal Aviation Administration S. Ibarra Amoco Corporation N. Inoue Kubota America Corporation R.I. Jaffee Electric Power Research Institute J.F. Jenkins Naval Civil Engineering Loboratory James W. Johnson WKM--Joy Division Mark J. Johnson Allegheny Ludlum Corporation Philip C. Johnson Materials Development Corporation Otakar Jonas Consultant Allen R. Jones M&T Chemicals, Inc. L. Jones ERT, A Resource Engineering Company R.H. Jones Battelle Pacific Northwest Laboratories R.M. Kain LaQue Center for Corrosion Technology, Inc. Herbert S. Kalish Adamas Carbide Corporation M.H. Kamdar Benet Weapons Laboratory Russell D. Kane Cortest Laboratories A. Kay Akron Sand Blast & Metallizing Company T.M. Kazmierczak UGI Corporation J.R. Kearns Allegheny Ludlum Corporation Victor Kelly NDT International G.D. Kent Parker Chemical Company H. Kernberger Bohler Chemical Plant Equipment (Austria) George E. Kerns E.I. Du Pont de Nemours & Company, Inc. R.J. Kessler Department of Transportation Bureau of Materials Research Yong-Wu Kim Inland Steel Company Fraser King Whiteshell Nuclear Research Establishment (Canada) J.H. King Chrysler Corporation Thomas J. Kinstler Metalplate Galvanizing, Inc. W.W. Kirk LaQue Center for Corrosion Technology, Inc. Samuel Dwight Kiser Inco Alloys International, Inc. Erhard Klar SCM Metal Products D.L. Klarstrom Haynes International, Inc. D.T. Klodt Manville Corporation Gregory Kobrin E.L. Du Pont de Nemours & Company, Inc. G.H.Koch Battelle Columbus Division John W. Koger Martin Marietta Energy Systems, Inc. Thomas G. Kollie Martin Marietta Energy Systems, Inc. Juri Kolts Conoco Inc. Karl-Heintz Kopietz Henry E. Sanson & Sons, Inc.

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Karl A. Korinek Parker Chemical Company Curt W. Kovach Crucible Materials Corporation Peter Krag Colorado School of Mines H.H. Krause Battelle Columbus Division William D. Krippes J.M.E. Chemicals A.S. Krisher ASK Associates Clyde Krummel Morton Thiokol, Inc. Kenneth F. Krysiak Hercules, Inc. Paul Labine Petrolite Research & Development J.Q. Lackey E.I. Du Pont de Nemours & Company, Inc. G.Y. Lai Haynes International, Inc. F.K. Lampson Marquordt Corporation E.A. Lange Technical Consultant Bruce Lanning Colorado School of Mines John Larson Ingersoll-Rand Company S. Larson Sundstrand Aviation David S. Lashmore National Bureau of Standards R.M. Latanison Massachusetts Institute of Technology J.A. Laverick The Timken Company Herbert H. Lawson Armco, Inc. Harvey H. Lee Inland Steel Company T.S. Lee National Association of Corrosion Engineers Henry Leidheiser, Jr. Center for Surface and Coating Research G.L. Leithauser General Motors Corporation Jack E. Lemons University of Alabama School of Dentistry G.G. Levy Chrysler Corporation Richard O. Lewis University of Florida Barry D. Lichter Vanderbilt University E.L. Liening Dow Chemical Company Bernard W. Lifka Aluminum Company of America Stephen Liu Pennsylvania State University Carl E. Locke University of Kansas A.W. Loginow Consulting Engineer F.D. Lordi General Electric Company C. Lundin University of Tennessee R.W. Lutey Buckman Laboratories, Inc. Fred F. Lyle, Jr. Southwest Research Institute Richard F. Lynch Zinc Institute Inc. A.J. Machiels Electric Power Research Institute J. Lee Magnon Dixie Testing & Products Inc. Gregory D. Maloney Saureisen Cements Company Paul E. Manning Haynes International, Inc. Miroslav I. Marek Georgia Institute of Technology Christopher Martenson Sandvik Steel Company J.A. Mathews Duke Power Company S.J. Matthews Haynes International, Inc. D. Mattox Sandia National Laboratories Daniel J. Maykuth Tin Research Institute, Inc. Joseph Mazia Mazia Tech-Com Services, Inc. M.M. McDonald Rockwell International J.E. McLaughlin Exxon Research & Engineering Company David H. Meacham Duke Power Company David N. Meendering Colorado School of Mines Jay Mehta J&L Specialty Products Corporation

Lehigh University

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

R.D. Merrik Exxon Research & Engineering Company Thomas Metz Naval Air Propulsion Center Fred H. Meyer, Jr. Air Force Wright Aeronautical Laboratories K. Miles Pulp & Paper Research Institute of Canada G.A. Minick A.R. Wilfley & Sons, Inc. K.L. Money LaQue Center for Corrosion Technology, Inc. B.J.Moniz E.I. Du Pont de Nemours & Company, Inc. Raymond W. Monroe Maynard Steel Casting Company Jean A. Montemarano David Taylor Naval Ship Research & Development Center J.F. Montle Carboline Company P.G. Moore Naval Research Laboratory Robert E. Moore United Engineers and Constructors Hugh Morrow Zinc Institute Inc. Robert E. Moser Electric Power Research Institute Max D.Moskal Stone Container Corporation Herbert J. Mueller Corrosion Consultant John J. Mueller Battelle Columbus Division S.K. Murarka Abitibi-Price Inc. (Canada) Charles A. Natalie Colorado School of Mines J. Lawrence Nelson Electric Power Research Institute James K. Nelson PPG Industries, Inc. R.J. Neville Dofasco Inc. (Canada) Dale C.H. Nevison Zinc Information Center, Ltd. R.A. Nichting Colorado School of Mines R.R. Noe Public Service Electric and Gas Company Peter Norberg AB Sandvik Steel Company (Sweden) W.J. O'Donnell Public Service Electric and Gas Company Thomas G. Oakwood Inland Steel Reseach Laboratories D.L. Olson Colorado School of Mines William W. Paden Oklahoma State University T.O. Passell Electric Power Research Institute C.R. Patriarca Haynes International, Inc. David H. Patrick ARCO Resources Technology Steven J. Pawel University of Tennessee G. Peck Cities Service Oil & Gas Corporation Bruno M. Perfetti USX Corporation Sam F. Pensabene General Electric Company Jeff Pernick International Hardcoat, Inc. William L. Phillips E.I. Du Pont de Nemours & Company, Inc. Joseph R. Pickens Martin Marietta Laboratories Hugh O. Pierson Ultramet D.L. Piron École Polytechnique de Montreal (Canada) Patrick Pizzo San Jose State University M.C. Place, Jr. Shell Oil Company Frederick J. Pocock Babcock & Wilcox Ortrun Pohler Institut Straumann AG (Switzerland) Steven L. Pohlman Kennecott Corporation Charles Pokross Fansteel Inc. Ned W. Polan Olin Corporation D.H. Pope Rensselaer Polytechnic University A.G. Preban Inland Steel Company R.B. Priory Duke Power Company R.B. Puyear Monsanto Company M. Quintana General Dynamics Electric

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Christopher Ramsey Colorado School of Mines Robert A. Rapp Ohio State University Louis Raymond L. Raymond & Associates George W. Read, Jr. Technical Consultant J.J. Reilly McDonnell Douglas Corporation Roger H. Richman Daedalus Associates, Inc. R.E. Ricker National Bureau of Standards O.L. Riggs, Jr. Kerr McGee Corporation Blaine W. Roberts Combustion Engineering, Inc. J.T. Adrian Roberts Battelle Pacific Northwest Laboratories Charles A. Robertson Sun Refining & Marketing Company H.S. Rosenberg Battelle Columbus Division Philip N. Ross, Jr. Lawrence Berkeley Laboratory Gene Rundell Rolled Alloys S. Sadovsky Public Service Electric and Gas Company William Safranek American Electroplaters and Surface Finishers Society Headquarters Brian J. Saldanha Corrosion Testing Laboratories, Inc. William Scarborough Vickers, Inc. Glenn L. Scattergood Nalco Chemical Company L.R. Scharfstein Mobil Research and Development Company S.T. Scheirer Westinghouse Electric Corporation John H. Schemel Sandvik Specialty Metals Corporation George Schick Bell Communications Research Mortimer Schussler Fansteel Inc. (retired) Ronald W. Schutz TIMET Corporation B.J. Scialabba JME Chemicals John R. Scully David Taylor Naval Ship Research & Development Center J.J. Sebesta Consultant M. Sedlack Technicon Enterprises Inc. Ellen G. Segan Department of the Army R. Serenius Western Forest Products Ltd. (Canada) I.S. Shaffer Department of the Navy Sandeep R. Shah Vanderbilt University W.B.A. Sharp Westvaco Research Center C.R. Shastry Bethlehem Steel Corporation Barbara A. Shaw David Taylor Naval Ship Research & Development Center Robert A. Shaw Electric Power Research Institute Gene P. Sheldon Olin Corporation R.D. Shelton Champion Chemicals, Inc. T.S. Shilliday Battelle Columbus Division D.W. Shoesmith Atomic Energy of Canada Ltd. C.G. Siegfried Ebasco Services, Inc. W.L. Silence Haynes International, Inc. D.C. Silverman Monsanto Company G. Simard Reid Inc. (Canada) J.R. Simmons Martin Marietta Corporation Harold J. Singletary Lockheed-Georgia Company John E. Slater Invetech, Inc. J. Slaughter Southern Alloy Corporation George Slenski Air Force Wright Aeronautical Laboratories J.S. Smart III Amoco Production Company Albert H. Smith Charlotte Pipe and Foundry Company Dale L. Smith Argonne National Laboratory F.N. Smith Alcan International Ltd. (Canada)

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Gaylord D. Smith Inco Alloys International, Inc. Jerome F. Smith Lead Industries Association, Inc. Carlo B. Sonnino Emerson Electric Company Peter Soo Brookhaven National Laboratory N. Robert Sorenson Sandia National Laboratories C. Spangler Westinghouse Electric Corporation T.C. Spence The Duriron Company, Inc. Donald O. Sprowls Consultant Narasi Sridhar Haynes International, Inc. Stephen W. Stafford University of Texas at El Paso J.R. Stanford Nalco Chemical Company (retired) E.E. Stansbury University of Tennessee T.M. Stastny Amoco Corporation A.J. Stavros Union Carbide Corporation T. Steffans Anhauser-Busch Brewing Company, Inc. Robert Stiegerwald Bechtel National, Inc. Donald R. Stickle The Duriron Company, Inc. T.J. Stiebler Houston Light & Power Company John G. Stoecker III Monsanto Company Paul J. Stone Chevron U.S.A. M.A. Streicher University of Delaware John Stringer Electric Power Research Institute T.J. Summerson Kaiser Aluminum & Chemical Corporation M.D. Swintosky The Timken Company W.R. Sylvester Combustion Engineering, Inc. Barry C. Syrett Electric Power Research Institute Robert E. Tatnall E.I. Du Pont de Nemours & Company, Inc. Kenneth B. Tator KTA-Tator, Inc. George J. Theus Babcock & Wilcox David E. Thomas RMI Company C.B. Thompson Pulp & Paper Research Institute of Canada Norman B. Tipton The Singleton Corporation P.F. Tortorelli Oak Ridge National Laboratory Herbert E. Townsend Bethlehem Steel Corporation K.L. Tryon The Timken Company R. Tunder General Electric Company Arthur H. Tuthill Tuthill Associates Inc. John A. Ulam Clad Metals, Inc. Robert H. Unger TAFA Inc. William Unsworth Magnesium Elektron, Ltd. (England) T.K. Vaidyanathan N.Y.U. Dental Center Ralph J. Velentine VAL-CORR J.H. VanSciver Allied-Signal Corporation Ellis D. Verink, Jr. University of Florida R. Viswanathan Electric Power Research Institute Ray Wainwright Technical Consultant James Walker Federal Aviation Administration Donald Warren E.I. Du Pont de Nemours & Company, Inc. Ray Watts Quaker Petroleum Chemicals Company William P. Webb Failure Analysis Associates R.T. Webster Teledyne Wah Chang Albany John R. Weeks Brookhaven National Laboratory Lawrence J. Weirick Sandia National Laboratories Donald A. Wensley MacMillan Bloedel Research (Canada)

• • • • • • • • • • • • • • •

R.E. Westerman Pacific Northwest Laboratory Eddie White Air Force Wright Aeronautical Laboratories William E. White Petro-Canada Resources D. Whiting Portland Cement Association Ron Williams Air Force Wright Aeronautical Laboratories E.L. Williamson Southern Company Services G.G. Wilson Stora Forest Industries (Canada) G.C. Wood Corrosion and Protection Centre University of Manchester (England) Ian G. Wright Battelle Columbus Division T.E. Wright Alcan International Ltd. (Canada) (retired) B.A. Wrobel Northern Indiana Public Service Company B.S. Yaffe Diversey Wyandotte Metals T.L. Yau Teledyne Wah Chang Albany Ronald A. Yeske The Institute of Paper Chemistry Edward Zysk Englehard Corporation

Foreword Volume 13 of the Metals Handbook series was compiled in response to the demand from our membership for a detailed work on the multibillion-dollar problem that confronts nearly every design engineer in every industry: corrosion. It represents the culmination of three years of intensive planning, writing, editing, and production. The hard work has paid off. Corrosion is the largest, most comprehensive volume on a single topic ever published by ASM. We believe that our readers will find this Handbook useful, instructive, and enlightening. These pages cover every aspect of the subject: corrosion theory, forms of corrosion, testing and evaluation, design considerations, protection methods, and corrosion as it affects specific metals and alloys and specific industries. Our goal is to help you solve existing corrosion problems--and to help you prevent problems in the future. ASM INTERNATIONAL is indebted to Lawrence J. Korb, Co-chairman of the Handbook and the driving force behind the project, and to Co-Chairman David L. Olson. Their task of planning and coordinating this volume has been a yeoman's one, and they have been equal to it. Both Larry and Dave are Fellows of ASM and have served in leadership roles within the Society for many years--Larry as a past Chairman of the Publications Council and the Handbook Committee, and Dave as a past Chairman of the Joining Division Council and as a member of the Handbook Committee since 1982. They epitomize the vast pool of talent and energy made available to the Sociey by its dedicated members, without whom we could not survive. Thanks also go to the ASM Handbook Committee and to the ASM editorial staff for their tireless efforts. We are especially grateful to the nearly 500 authors and reviewers listed in the next several pages. Their generous commitment of time and expertise, their willingness to share their years of experience and knowledge, to write and rewrite, has made this Handbook a truly outstanding source of information. •

•

Raymond F. Decker President ASM INTERNATIONAL Edward L. Langer Managing Director ASM INTERNATIONAL

Preface The cost of corrosion to U.S. industries and the American public is currently estimated at $170 billion per year. Although corrosion is only nature's method of recycling, or of returning a metal to its lowest energy form, it is an insidious enemy that destroys our cars, our plumbing, our buildings, our bridges, our engines, and our factories. Corrosion can often be predictable, such as the uniform corrosion of steel ship hulls or tanks, or it can be totally unpredictable and catastrophic, such as the hydrogen embrittlement or stress corrosion of critical structural members and pressure vessels in the aerospace and chemical processing industries. While corrosion obeys well-known laws of electrochemistry and thermodynamics, the

many variables that influence the behavior of a metal in its environment can result in accelerated corrosion or failure in one case and complete protection in another similar case. We can no longer think of materials and environments as monolithic. It makes no sense to ask whether stainless steel is compatible with sulfuric acid. Rather, the question we must ask is which alloy of stainless steel, with which microstructure, with which design detail, is compatible with which sulfuric acid. What is the acid's temperature, concentration, pH, impurity level, types of trace species, degree of aeration, flow velocity, etc.? Avoiding detrimental corrosion requires the interdisciplinary approach of the designer, the metallurgist, and the chemist. Sooner or later, nearly everyone in these fields will be faced with major corrosion issues. It is necessary to learn to recognize the forms of corrosion and the parameters that must be controlled to avoid or mitigate corrosion. This Handbook was written with these three engineering disciplines in mind. We have attempted to put together a reference book that is well rounded and complete in its coverage--for we want this to be the first book you select when researching a corrosion problem. Each article is indexed to other appropriate sections of the Handbook, and each provides a road map to the thousands of individual bibliographical references that were used to compile the information. The Handbook is organized into eight major Sections. The first is a Glossary of metallurgical and corrosion terms used throughout the Volume. Nearly 600 terms are defined, selected from more than 20 sources. Of course, one of the most difficult terms to get corrosion experts to agree upon is a definition for "corrosion" itself, for where does one draw the line? Is not the hydride, which precipitates in a stressed titanium weld, a form of corrosion just as the hydrogen embrittlement of steel? And where does corrosion stop--with a metal, or is the environmental reaction of a ceramic or polymer also a form of corrosion? In this Handbook we have limited our discussion of corrosion to metals, by and large, but have included reactions with external environments which may diffuse inside a metal, leading to its destruction as an "internal environment." The second Section covers the theory of corrosion from the thermodynamic and kinetic points of view. It covers the principles of electrochemistry, diffusion, and dissolution as they apply to aqueous corrosion and high-temperature corrosion in salts, liquid metals, and gases. The effects of both metallurgical and environmental variables on corrosion in aqueous solutions are discussed in detail. The third Section describes the various forms of corrosion, how to recognize them, and the driving conditions or parameters that influence each form of corrosion, for it is the control of these parameters which can minimize or eliminate corrosion. For convenience, this Section is divided into articles on general corrosion, localized corrosion, metallurgically influenced corrosion, mechanically assisted degradation, and environmentally induced cracking. More than 20 distinct corrosion mechanisms are discussed. mIn the fourth Section, methods of corrosion testing and evaluation in the laboratory as well as in-place corrosion monitoring are discussed. For each major form of corrosion (pitting, stress-corrosion cracking, etc.), the existing techniques used in their evaluation are discussed along with the advantages and limitations of each particular test and the quality of the test data generated. The fifth Section looks at corrosion from the design standpoint. Which materials and design details minimize corrosion? What are the corrosion problems with weldments and how can they be addressed? Finally, how do you place an economic value on your selection of alternate materials or coatings? The next Section reviews the various methods used for corrosion protection. These include surface conversion coatings, anodizing, ceramic coatings, organic coatings, metallic coatings (both as barrier metals and as sacrificial coatings), thermal spray coatings, CVD/PVD coatings, and other methods of surface modification. It also discusses the principles of and the approaches to anodic and cathodic protection. Finally, the various types and uses of corrosion inhibitors are thoroughly discussed. The seventh Section covers the corrosion of 27 different metal systems, including all major structural alloy systems and precious metals, and relates the latest information on such topics as powder metals, cemented carbides, amorphous metals, metal matrix composites, hard chromium plating, brazing alloys, and clad metals. In many areas, complete articles have been written where only a few paragraphs were available in existing corrosion texts. For each metal system, the authors discuss the alloys available, the nature of the corrosion resistance film that forms on the metal, and the mechanisms of corrosion, including the metallurgical factors or elements that inhibit or accelerate corrosion. Various forms of corrosion are discussed as well as various environmental effects. The behavior of these metal systems in

atmospheres (rural, marine, industrial), in waters (fresh water and seawater), and in alkalies, acids, salts, organic chemicals, and gases is discussed. Methods of corrosion protection most applicable to each metal system are reviewed. The final Section of the Handbook is where all of this knowledge is put into practice. It vividly illustrates how far we've come in understanding and combating corrosion and how far we have yet to go. The corrosion experiences of experts from 20 major industries are covered in detail--from fossil fuels to nuclear power, from the chemical processing to the marine industries, from prosthetic devices to the space shuttle, from pharmaceuticals to electronics, from petroleum production and refining to heavy construction. The authors describe the corrosion problems they encounter, tell how they solve them, and present illustrated case histories. We think you will find this Handbook a broad-based approach to understanding corrosion, with sufficient data and examples to solve many problems directly, and references to key literature for further research into highly complex corrosion issues. There is no cookbook for corrosion avoidance! We hope this Volume with its road map of references will lead you to a better understanding of your corrosion problems and assist you in their solutions. This Handbook would not have been possible without the generous contributions of the nearly 500 leading corrosion experts who donated their expertise as authors and reviewers. They represent many of the leading industries and educational institutions in this country and abroad. The articles in this Handbook represent tremendous individual efforts. We are also grateful to the Handbook staff at ASM INTERNATIONAL and for the extremely valuable contributions of several technical societies and industrial associations, including the National Association of Corrosion Engineers, the American Society for Testing and Materials, the Electric Power Research Institute, the Pulp and Paper Research Institute of Canada, the Tin Research Institute, the Institute of Paper Chemistry, the American Hot Dip Galvanizers Association, and the Lead Industries Association. In addition, we particularly appreciate the efforts of those who took responsibility for coordinating authors and papers for many articles or entire Sections of this Volume: Dr. Miroslav Marek, Dr. Bruce Craig, Dr.Steven Pohlman, Mr. Donald Sprowls, Mr. James Lackey, Dr. Herbert Townsend, Dr. Thomas Cape, Mr. Kenneth Tator, Dr. Ralph Davison, Dr. Aziz Asphahani, Mr. R. Terrence Webster, Mr. Robert Charlton, Mr. James Hanck, and Mr. Fred Meyer, Jr. This has truly been a collective venture of the technical community. We thank those who willingly have shared their knowledge with all of us. • •

L.J. KorbCo-Chairman D.L. OlsonCo-Chairman

General Information Officers and Trustees of ASM International (1986-1987) • • • • • • • • • • • • • • •

Raymond F. Decker President and Trustee University Science Partners, Inc. William G. Wood Vice President and Trustee Kolene Corporation John W. Pridgeon Immediate Past President and Trustee Chemtech Ltd. Frank J. Waldeck Treasurer Lindberg Corporation Trustees Stephen M. Copley University of Southern California Herbert S. Kalish Adamas Carbide Corporation William P. Koster Metcut Research Associates, Inc. Robert E. Luetje Kolene Corporation Gunvant N. Maniar Carpenter Technology Corporation Larry A. Morris Falconbridge Limited Richard K. Pitler Allegheny Ludlum Corporation (retired) C. Sheldon Roberts Consultant Materials and Processes Klaus M. Zwilsky National Materials Advisory Board National Academy of Sciences Edward L. Langer Managing Director

Members of the ASM Handbook Committee (1986-1987)

• • • • • • • • • • • • • • • • • • •

Dennis D. Huffman (Chairman 1986-; Member 1983-) The Timken Company Roger J. Austin (1984-) Materials Engineering Consultant Peter Beardmore (1986-) Ford Motor Company Deane I. Biehler (1984-) Caterpillar Tractor Company Robert D. Caligiuri (1986-) SRI International Richard S. Cremisio (1986-) Rescorp International Inc. Thomas A. Freitag (1985-) The Aerospace Corporation Charles David Himmelblau (1985-) Lockheed Missiles & Space Company, Inc. John D. Hubbard (1984-) Hinderliter Heat Treating L.E. Roy Meade (1986-) Lockheed-Georgia Company Merrill L. Minges (1986-) Air Force Wright Aeronautical Laboratories David. V. Neff (1986-) Metaullics Systems David LeRoy Olson (1982-) Colorado School of Mines Paul E. Rempes (1986-) Champion Spark Plug Company Ronald J. Ries (1983-) The Timken Company E. Scala (1986-) Cortland Cable Company, Inc. David A. Thomas (1986-) Lehigh University Peter A. Tomblin (1985-) De Havilland Aircraft of Canada Ltd. Leonard A. Weston (1982-) Lehigh Testing Laboratories, Inc.

Previous Chairmen of the ASM Handbook Committee • • • • • • • • • • • • • • • • • • • • • •

R.S. Archer (1940-1942) (Member, 1937-1942) L.B. Case (1931-1933) (Member, 1927-1933) T.D. Cooper (1984-1986) (Member, 1981-1986) E.O. Dixon (1952-1954) (Member, 1947-1955) R.L. Dowdell (1938-1939) (Member, 1935-1939) J.P. Gill (1937) (Member, 1934-1937) J.D. Graham (1966-1968) (Member, 1961-1970) J.F. Harper (1923-1926) (Member, 1923-1926) C.H. Herty, Jr. (1934-1936) (Member, 1930-1936) J.B. Johnson (1948-1951) (Member, 1944-1951) L.J. Korb (1983) (Member, 1978-1983) R.W.E. Leiter (1962-1963) (Member, 1955-1958, 1960-1964) G.V. Luerssen (1943-1947) (Member, 1942-1947) G.N. Maniar (1979-1980) (Member, 1974-1980) J.L. McCall (1982) (Member, 1977-1982) W.J. Merten (1927-1930) (Member, 1923-1933) N.E. Promisel (1955-1961) (Member, 1954-1963) G.J. Shubat (1973-1975) (Member, 1966-1975 W.A. Stadtler (1969-1972) (Member, 1962-1972) R. Ward (1976-1978) (Member, 1972-1978) M.G.H. Wells (1981) (Member, 1976-1981) D.J. Wright (1964-1965) (Member, 1959-1967)

Staff ASM International staff who contributed to the development of the Volume included Joseph R. Davis, Senior Editor; James D. Destefani, Technical Editor; Heather J. Frissell, Editorial Supervisor; George M. Crankovic, Assistant Editor; Diane M. Jenkins, Word Processing Specialist; Robert L. Stedfeld, Director of Reference Publications; Kathleen M. Mills, Manager of Editorial Operations; with editorial assistance from J. Harold Johnson, Robert T. Kiepura, and Dorene A. Humphries Conversion to Electronic Files

ASM Handbook, Volume 13, Corrosion, was converted to electronic files in 1997. The conversion was based on the Fourth Printing (December 1992). No substantive changes were made to the content of the Volume, but some minor corrections and clarifications were made as needed. ASM International staff who contributed to the conversion of the Volume included Sally Fahrenholz-Mann, Bonnie Sanders, Scott Henry, Grace Davidson, Randall Boring, Robert Braddock, Kathleen Dragolich, and Audra Scott. The electronic version was prepared under the direction of William W. Scott, Jr., Technical Director, and Michael J. DeHaemer, Managing Director. Copyright Information (for Print Volume) Copyright © 1987 by ASM International All Rights Reserved. ASM Handbook is a collective effort involving thousands of technical specialists. It brings together in one book a wealth of information from world-wide sources to help scientists, engineers, and technicians solve current and long-range problems. Great care is taken in the compilation and production of this Volume, but it should be made clear that no warranties, express or implied, are given in connection with the accuracy or completeness of this publication, and no responsibility can be taken for any claims that may arise. Nothing contained in the ASM Handbook shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in the ASM Handbook shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement. Comments, criticisms, and suggestions are invited, and should be forwarded to ASM International. Library of Congress Cataloging-in-Publication Data (for Print Volume) ASM INTERNATIONAL Metals Handbook. Includes bibliographics and indexes. Contents: v. 1. Properties and selection

[etc.]

v. 9 Metallography and microstructures

[etc.]

v. 13. Corrosion.

1. Metals--Handbooks, manuals, etc. I. ASM International. Handbook Committee. II. Title: ASM Handbook. TA459.M43 1978 669 78-14934 ISBN 0-87170-007-7 (v.1) SAN 204-7586 Printed in the United States of America

Introduction Miroslav I. Marek, School of Materials Engineering, Georgia Institute of Technology

Introduction PERHAPS THE MOST STRIKING FEATURE of corrosion is the immense variety of conditions under which it occurs and the large number of forms in which it appears. Numerous handbooks of corrosion data have been compiled that list the corrosion effects of specific material/environment combinations; still, the data cover only a small fraction of the possible situations and only for specific values of, for example, the temperature and composition of the substances involved. To prevent corrosion, to interpret corrosion phenomena, or to predict the outcome of a corrosion situation for conditions other than those for which an exact description can be found, the engineer must be able to apply the knowledge of corrosion fundamentals. These fundamentals include the mechanisms of the various forms of corrosion, applicable thermodynamic conditions and kinetic laws, and the effects of the major variables. Even with all of the available generalized knowledge of the principles, corrosion is in most cases a very complex process in which the interactions among many different reactions, conditions, and synergistic effects must be carefully considered. All corrosion processes show some common features. Thermodynamic principles can be applied to determine which processes can occur and how strong the tendency is for the changes to take place. Kinetic laws then describe the rates of the reactions. There are, however, substantial differences in the fundamentals of corrosion in such environments as aqueous solutions, non-aqueous liquids, and gases that warrant a separate treatment in this Section.

Corrosion in Aqueous Solutions Although atmospheric air is the most common environment, aqueous solutions, including natural waters, atmospheric moisture, and rain, as well as man-made solutions, are the environments most frequently associated with corrosion problems. Because of the ionic conductivity of the environment, corrosion is due to electrochemical reactions and is strongly affected by such factors as the electrode potential and acidity of the solution. As described in the article "Thermodynamics of Aqueous Corrosion," thermodynamic factors determine under what conditions the reactions are at an electrochemical equilibrium and, if there is a departure from equilibrium, in what directions the reactions can proceed and how strong the driving force is. The kinetic laws of the reactions are fundamentally related to the activation energies of the electrode processes, mass transport, and basic properties of the metal/environment interface, such as the resistance of the surface films (see the article "Kinetics of Aqueous Corrosion" in this volume). The fundamental kinetics of aqueous corrosion have been thoroughly studied. The simultaneous occurrences of several electrochemical reactions responsible for corrosion have been analyzed on the basis of the mixed potential theory, which provides a general method of interpreting or predicting the corrosion potential and reaction rates. The actual corrosion rates are then strongly affected by the environmental and metallurgical variables, as discussed in the articles "Effects of Environmental Variables on Aqueous Corrosion" and "Effects of Metallurgical Variables on Aqueous Corrosion," respectively. Special conditions exist in natural order and some industrial systems where biological organisms are present in the environment and attach themselves to the structure. Corrosion is expected by the presence of the organisms and the biological films they produce, as well as the products of their metabolism, as described in the Appendix "Biological Effects" to the aforementioned article on environmental variables.

Corrosion in Molten Salts and Liquid Metals These are more narrow but important areas of corrosion in liquid environments. Both have been strongly associated with the nuclear industry, for which much of the research has been performed, but there are numerous nonnuclear applications as well. In molten-salt corrosion, described in the article "Fundamentals of High-Temperature Corrosion in Molten Salts," the mechanisms of deterioration are more varied than in aqueous corrosion, but there are many similarities and some interesting parallels, such as the use of the E - pO2- diagrams similar to the E - pH (Pourbaix) diagrams in aqueous corrosion. Preferential dissolution plays a stronger role in molten-salt corrosion than in aqueous corrosion. Corrosion testing presents special problems and is much more involved than the familiar aqueous testing, usually requiring

expensive circulation loops and purification of the salts. Although the literature on molten-salt corrosion is substantial, relatively few fundamental thermodynamic and kinetic data are available. Liquid-metal corrosion, discussed in the article "Fundamentals of High-Temperature Corrosion in Liquid Metals," is of great interest in the design of fast fission nuclear reactors as well as of future fusion reactors, but is also industrially important in other areas, such as metal recovery, heat pipes, and various special cooling designs. Liquid-metal corrosion differs fundamentally from aqueous and molten-salt corrosion in that the medium, except for impurities, is in a nonionized state. The solubilities of the alloy components and their variation with temperature then play a dominant role in the process, and preferential dissolution is a major form of degradation. Mass transfer is another frequent consequence of the dissolution process. At the same time, the corrosion is strongly affected by the presence of nonmetallic impurities in both the alloys and the liquid metals.

Corrosion in Gases In gaseous corrosion, the environment is nonconductive, and the ionic processes are restricted to the surface of the metal and the corrosion product layers (see the article "Fundamentals of Corrosion in Gases"). Because the reaction rates of industrial metals with common gases are low at room temperature, gaseous corrosion, generically called oxidation, is usually an industrial problem only at high temperatures when diffusion processes are dominant. Thermodynamic factors play the usual role of determining the driving force for the reactions, and free energy-temperature diagrams are commonly used to show the equilibria in simple systems, while equilibria in more complex environments as a function of compositional variables can be examined by using isothermal stability diagrams. In the mechanism and kinetics of oxidation, the oxide/metal volume ratio gives some guidance of the likelihood that a protective film will be formed, but the major role belongs to conductivity and transport processes, which are strongly affected by the impurities and defect structures of the compounds. Together with conditions of surface film stability, the transport processes determine the reaction rates that are described in general form by the several kinetic rate laws, such as linear, logarithmic, and parabolic. The most obvious result of oxidation at high temperatures is the formation of oxide scale. The properties of the scales and development of stresses determine whether the scale provides a continuous oxidation protection. In some cases of oxidation of alloys, however, reactions occur within the metal structure in the form of internal oxidation. Like corrosion in liquids, selective or preferential oxidation is frequently observed in alloys containing components of substantially different thermodynamic stability. Thermodynamics of Aqueous Corrosion

CORROSION OF METALS in aqueous environments is almost always electrochemical in nature. It occurs when two or more electrochemical reactions take place on a metal surface. As a result, some of the elements of the metal or alloy change from a metallic state into a nonmetallic state. The products of corrosion may be dissolved species or solid corrosion products; in either case, the energy of the system is lowered as the metal converts to a lower-energy form. Rusting of steel is the best known example of conversion of a metal (iron) into a nonmetallic corrosion product (rust). The change in the energy of the system is the driving force for the corrosion process and is a subject of thermodynamics. Thermodynamics examines and quantifies the tendency for corrosion and its partial processes to occur; it does not predict if the changes actually will occur and at what rate. Thermodynamics can predict, however, under what conditions the metal is stable and corrosion cannot occur. The electrochemical reactions occur uniformly or nonuniformly on the surface of the metal, which is called an electrode. The ionically conducting liquid is called an electrolyte. As a result of the reaction, the electrode/electrolyte interface acquires a special structure, in which such factors as the separation of charges between electrons in the metal and ions in the solution, interaction of ions with water molecules, adsorption of ions on the electrode, and diffusion of species all play important roles. The structure of this so-called double layer at the electrified interface, as related to corrosion reactions, will be described in the section "Electrode Processes" in this article. One of the important features of the electrified interface between the electrode and the electrolyte is the appearance of a potential difference across the double layer, which allows the definition of the electrode potential. The electrode potential becomes one of the most important parameters in both the thermodynamics and the kinetics of corrosion. The fundamentals will be discussed in the section "Electrode Potentials," and some examples of the calculations of the potential from thermodynamic data are show in the section "Potential Versus pH (Pourbaix) Diagrams."

The electrode potentials are used in corrosion calculations and are measured both in the laboratory and in the field. In actual measurements, standard reference electrodes are extensively used to provide fixed reference points on the scale of relative potential values. The use of suitable reference electrodes and appropriate methods of measurement will be discussed in the section "Potential Measurements With Reference Electrodes." One of the most important steps in the science of electrochemical corrosion was the development of diagrams showing thermodynamic conditions as a function of electrode potential and concentration of hydrogen ions. These potential versus pH diagrams, often called Pourbaix diagrams, graphically express the thermodynamic relationships in metal/water systems and show at a glance the regions of the thermodynamic stability of the various phases that can exist in the system. Their construction and application in corrosion, as well as their limitations, will be discussed in the section "Potential Versus pH (Pourbaix) Diagrams." Thermodynamics of Aqueous Corrosion

Electrode Processes Charles A. Natalie, Department of Metallurgical Engineering, Colorado School of Mines

In the discussion of chemical reactions and valence, the topic of electrochemical reactions is usually treated as a special case. Electrochemical reactions are usually discussed in terms of the change in valence that occurs between the reacting elements, that is, oxidation and reduction. Oxidation and reduction are commonly defined as follows. Oxidation is the removal of electrons from atoms or groups of atoms, resulting in an increase in valence, and reduction is the addition of electrons to an atom or group of atoms, resulting in the decrease in valence (Ref 1). Because electrochemical reactions or oxidation-reduction reactions can be represented in terms of an electrochemical cell with oxidation reactions occurring at one electrode and reduction occurring at the other electrode, electrochemical reactions are often further defined as cathodic reactions and anodic reactions. By definition, cathodic reactions are those types of reactions that result in reduction, such as: M(aq)2+ + 2e- → M(s)

(Eq 1)

Anodic reactions are those types of reactions that result in oxidation, such as: M(s) → M(aq)2+ + 2e-

(Eq 2)

Because of the production of electrons during oxidation and the consumption of electrons during reduction, oxidation and reduction are coupled events. If the ability to store large amounts of electrons does not exist, equivalent processes of oxidation and reduction will occur together during the course of normal electrochemical reactions. The oxidized species provide the electrons for the reduced species. The example stated above, like many aqueous corrosion situations, involves the reaction of aqueous metal species at a metal electrode surface. This metal/aqueous interface is complex, as is the mechanism by which the reactions take place across the interface. Because the reduction-oxidation reactions involve species in the electrolyte reacting at or near the metal interface, the electrode surface is charged relative to the solution, and the reactions are associated with specific electrode potentials. The charged interface results in an electric field that extends into the solution. This electric field has a dramatic effect on the solution near the metal. A solution that contains water as the primary solvent is affected by an electrical field because of its structure. The primary solvent--water--is polar and can be visualized as dipolar molecules that have a positive side (hydrogen atoms) and a negative side (oxygen atoms). In the electric field caused by the charged interface, the water molecules act as small dipoles and align themselves in the direction of the electric field.

Ions that are present in the solution are also charged because of the loss or gain of electrons. The positive charged ions (cations) and negative charged ions (anions) also have an electric field associated with them. The solvent (water) molecules act as small dipoles; therefore, they are also attracted to the charged ions and align themselves in the electric field established by the charge of the ion. Because the electric field is strongest close to the ion, some water molecules reside very close to an ionic species in solution. The attraction is great enough that these water molecules travel with the ion as it moves through the solvent. The tightly bound water molecules are referred to as the primary water sheath of the ion. The electric field is weaker at distances outside the primary water sheath, but it still disturbs the polar water molecules as the ion passes through the solution. The water molecules that are disturbed as the ion passes, but do not move with the ion, are usually referred to as the secondary water sheath. Figure 1 shows a representation of the primary and secondary solvent molecules for a cation in water. Because of their smaller size relative to anions, cations have a stronger electric field close to the ion, and more water molecules are associated in their primary water sheath. However, anodic species have few, if any, primary water molecules. A detailed description of the hydration of ions in solution is given in Ref 2.

Fig. 1 Schematic of the primary and secondary solvent molecules for a cation in water

Because of the potential and charge established at the metal/aqueous interface of an electrode, ions and polar water molecules are also attracted to the interface because of the strong electric field close to the interface. Water molecules form a first row at the metal/aqueous interface. This row of water molecules limits the distance that hydrated ions can approach the interface. Figure 2 shows a schematic diagram of a charged interface and the locations of cations at the surface. Also, the primary water molecules associated with the ionic species limit the distance the ions can approach. For example, the plane of positive charge of the cations that reside near the surface of a negatively charged interface is a fixed distance from the metal due to the water molecules that are between the surface and the ions. This plane of charge is referred to as the Outer-Heimholz Plane (OHP).

Fig. 2 Schematic of a charged interface and the locations of cations at the electrode surface

Because of the structure of the charged interface described above, it is often represented (Ref 2) as a charged capacitor (Fig. 3). The potential drop across the interface is also often simplified as a linear change in potential from the metal surface to the OHP.

Fig. 3 Simplified double layer at a metal aqueous interface

The significance of the electronic double layer is that it provides a barrier to the transfer of electrons. If there were no difficulty in the transfer of electrons across the interface, the only resistance to electron flow would be the diffusion of aqueous species to and from the electrode. The surface would be nonpolarizable, and the potential would not be changed until the solution was deficient in electron acceptors and/or donors. This is of particular interest when dealing with the kinetics at the interface (see the article "Kinetics of Aqueous Corrosion" in this Volume). The double layer results in an energy barrier that must be overcome. Thus, reactions at the interface are often dominated by activated processes, and activation polarization plays a significant role in corrosion. The key to controlling corrosion usually consists of minimizing the kinetics; this slows the reaction rates sufficiently that corrosion appears to be stopped.

Thermodynamics of Aqueous Corrosion

Electrode Potentials Charles A. Natalie, Department of Metallurgical Engineering, Colorado School of Mines

The object of chemical thermodynamics is to develop a mathematical treatment of the chemical equilibrium and the driving forces behind chemical reactions. The desire is to catalog known quantitative data concerning equilibrium that can be later used to predict equilibria (perhaps even equilibria that has never been investigated by experimentation). The driving force for chemical reactions has been expressed in thermodynamic treatments as the balance between the effect of energy (enthalpy) and the effect of probability. The thermodynamic property that relates to probability is called entropy. The idea of entropy has been expressed as thermodynamic probability and is defined as the number of ways in which microscopic particles can be distributed among states accessible to them (Ref 3). The thermodynamic probability is an extensive quantity and is not the mathematical probability that ranges between 0 and 1.

Free Energy The driving force for chemical reactions depends not only on chemical formulas of species involved but also on the activities of the reactants and products. Free energy is the thermodynamic property that has been assigned to express the resultant enthalpy of a substance and its inherent probability. At constant temperature, free energy can be expressed as:

∆G = ∆H-T∆S

(Eq 3)

where ∆G is the change in free energy (Gibbs free energy), ∆H is the change in enthalpy, T is the absolute temperature, and ∆S is the change in entropy. When reactions are at equilibrium and there is no apparent tendency for a reaction to proceed either forward or backward, it has been shown that (Ref 4):

∆G° = -RT ln Keq

(Eq 4)

where ∆G° is the free energy change under the special conditions when all reactants and products are in a preselected standard state, R is the gas constant, and Keq is the equilibrium constant. The standard free energy of formations for an extensive number of compounds as a function of temperature have been cataloged; this allows the prediction of equilibrium constants over a wide range of conditions. It is necessary only to determine the standard free energy change for a reaction (∆G°, Eq 4) by subtracting the sum of the free energy of formations of the products at constant temperature. If an electrochemical cell is constructed that can operate under thermodynamic reversible conditions (the concept of reversibility is described in more detail in the section) "Potential Measurements With Reference Electrodes" in this article and in Ref 4) and if the extent of reaction is small enough not to change the activities of reactants and products, the potential remains constant, and the energy dissipated by an infinitesimal passage of charge is given by:

|∆G| = charge passed · potential difference or

|∆G| = nF · |E|

(Eq 5)

where n is the number of electrons per atom of the species involved in the reaction, F is the charge of 1 mol of electrons, and E is the cell potential. Because free energy has a sign that denotes the direction of the reaction, a thermodynamic sign convention must be selected. The common U.S. convention is to associate a positive potential with spontaneous reactions; thus, the reaction becomes:

∆G = -nFE

(Eq 6)

If the reaction occurs under conditions in which the reactants and products are in their standard states, the equation becomes:

∆G° = -nFE°

(Eq 7)

Combination with Eq 4 results in:

(Eq 8)

thus allowing the prediction of equilibrium data for electrochemical reactions.

Cell Potentials and the Electromotive Force Series If a strip of zinc metal is placed in water, some zinc ions will be converted to aqueous zinc ions because of the relatively large tendency for zinc to oxidize. Because of the electrons remaining in the metal, the positively charged zinc ions will remain very close to the negatively charged zinc strip and thus will establish a double layer, as described in the section "Electrode Potentials" in this article (Fig. 1). The potential difference established between the solution and the zinc is of the order of 1 V, but because the double layer is very small, the potential gradient (change in potential with respect to distance) can be very high. A negative electrode potetial (with respect to the standard hydrogen electrode discussed below) exists for a zinc electrode in a solution of zinc ions. However, if a copper strip is placed in a solution containing copper ions, a positive potential is established between the more noble copper strip and the solution. If, however, a metal is placed in a solution containing metal ions of a different nature, the first metal may dissolve, while the second metal deposits from its ions. A common example of this is the metal displacement reaction between zinc metal and copper ions, for which the complete oxidation-reduction reaction is:

Zn(s) + Cu(aq)2+ → Zn(aq)2+ + Cu(s)

(Eq 9)

If the reverse procedure is tried, that is, copper metal placed in a solution containing zinc ions, no reaction will take place to any measurable extent. For example, if the solution containing zinc ions has no copper ions present initially, the reaction will occur to a very small extent, with the reaction stopping when a certain very small concentration of copper ions has been produced. In the opposite case, zinc metal will react with copper ions almost to completion; the reaction will stop only when the concentration of copper ions is very small. The above experiment can be repeated with many combinations of metals, and the ability of one metal to replace another ion from solution can be used as a basis for tabulating the metals in a series. The table formed would show the abilities of metals to reduce other metal ions from solution. This electromotive force (emf) series for some common metals is shown in Table 1. The potentials listed in Table 1 are measured values, which will be described below as well as in the section "Potential Measurements With Reference Electrodes" in this article.

Table 1 Electromotive force series See also Fig. 4, which shows a schematic of an electrochemical cell used to determine the potential difference between copper and zinc electrodes. Electrode reaction

Standard potential at 25 °C (77 °F), volts versus SHE

Au3+ + 3e-

→ Au

1.50

Pd2+ + 2e-

→ Pd

0.987

Hg 2+ +2e-

→ Hg

0.854

→ Ag

Ag+ + e-

Hg 22+ + 2e- → 2Hg → Cu

Cu+ + e-

0.800

0.789

0.521

Cu2+ + 2e-

→ Cu

0.337

2H+ + 2e-

→ H2

(Reference)

0.000

→ Pb

Pb2+ + 2e-

-0.126

Sn2 + 2e-

→ Sn

-0.136

Ni2+ + 2e-

→ Ni

-0.250

Co2+ + 2e-

→ Ni

-0.277

Tl+ + e-

→ Tl

In3+ + 3e-

→ In

-0.336

-0.342

Cd2+ + 2e-

→ Cd

-0.403

Fe2+ + 2e-

→ Fe

-0.440

Ga3+ + 3e-

→ Ga

-0.53

Cr3+ + 3e-

→ Cr

-0.74

Cr2+ + 2e-

→ Cr

-0.91

Zn2+ + 2e-

→ Zn

-0.763

Mn2+ + 2e-

→ Mn

-1.18

Zr4+ + 4e-

→ Zr

-1.53

Ti2+ + 2e-

→ Ti

-1.63

Al3+ + 3e-

→ Al

-1.66

Hf4+ + 4e-

→ Hf

-1.70

→U

U3+ + 3e-

-1.80

Be2+ + 2e-

→ Be

-1.85

Mg2+ + 2e-

→ Mg

-2.37

Na+ + e-

→ Na

Ca2+ + 2e-

→ Ca

-2.71

-2.87

K+ + e-

→K

-2.93

Li+ + e-

→ Li

-3.05

The reactions described in establishing an emf series are referred to as electrochemical reactions. Electrochemical reactions are those reactions that involve oxidation (increase in valence) and reduction (decrease in valence), as described in the section "Electrode Processes" in this article. For the example of copper metal deposition using zinc metal, the oxidation reaction for producing electrons is:

Zn(s) = Zn(aq)2+ + 2eElectrons are consumed by copper ion according to the following reduction reaction:

(Eq 10)

Cu(aq)2+ + 2e- → Cu(s)

(Eq 11)

To study the reactions discussed above (Eq 9, 10, 11), an electrochemical cell, such as the one shown and described in Fig. 4, can be constructed by using a copper electrode in a solution of copper sulfate as one electrode and a zinc electrode in a solution of zinc sulfate as the other electrode. If the external conduction path is short circuited, electrons will flow from the zinc electrode (anode) as zinc dissolves to the copper electrode (cathode); this causes the deposition of copper metal. This type of arrangement would demonstrate how some electrochemical reactions can take place with the reactants and products physically separated and how the overall process can be visualized as two separate reactions that occur together.

Fig. 4 Typical electrochemical cell (a) used to study the free energy change that accompanies electrochemical or corrosion reactions. In this example, the cell contains copper and zinc electrodes in equilibrium, with their ions separated by a porous membrane to mitigate mixing. For purposes of simplicity, the concentration of metal ions is maintained at unit activity; that is, each solution contains about 1 g atomic weight of metal ion per liter. The reactions taking place on each side of the cell are represented by Eq 10, and 11, and the half-cell reactions for copper and zinc electrodes are given in Table 1. The rates of metal dissolution and deposition must be the same as shown in (b), which illustrates copper atoms being oxidized to cupric ions and, at other areas, cupric ions being reduced to metallic copper. Equilibrium conditions dictate that the reaction rates r1 and r2 be equal. Source: Ref 5

The two reactions listed in Eq 10, and 11, and shown schematically in Fig. 4 are often referred to as half-cell reactions. This nomenclature is due to the requirement that oxidation and reduction occur simultaneously under equilibrium conditions. Therefore, the reaction given in Eq 10 is defined as an oxidation half-cell reaction, and the reaction given in Eq 11 is a reduction half-cell reaction. The reaction in Eq 9 can be referred to as the overall electrochemical reaction and is the sum of the half-cell reactions given in Eq 10 and 11. Because specific, or absolute, potentials of electrodes cannot be measured directly, an arbitrary half-cell reaction is used as a reference by defining its potential as 0. All other half-cell potentials can then be calculated with respect to this zero reference. As described in the following section "Potential Measurements With Reference Electrodes," the hydrogen ion reaction 2H+ + 2e → H2 (Table 1) is used as the standard reference point. It is not possible to make an electrode from hydrogen gas; therefore, the standard hydrogen electrode (SHE) potential is measured by using an inert electrode, such as platinum, immersed in a solution saturated hydrogen gas at 1 atm. All values of electrode potential, therefore, are with reference to SHE. The potentials given in Table 1 are specifically the potentials measured relative to an SHE at 25 °C (77 °F) when all concentrations of ions are 1 molal, gases are at 1 atm of pressure, and solid phases are pure. This specific electrode

potential is referred to as the standard electrode potential and is denoted by E°. The standard electrode potential for zinc-the accepted value for which is -0.763 (Table 1)--can be calculated by measuring the emf of a cell made up, for example, of a zinc and a hydrogen electrode in a zinc salt solution of known activity Zn2+ and H+ (Fig. 5). This procedure could be repeated by exchanging the zinc electrode with any other metal and by assigning the half-cell electrode potentials measured for the electrochemical cells to the proper reactions in Table 1. Changes in concentration, temperature, and partial pressure will change the electrode potentials and the position of a particular metal in the emf series. In a particular, the change in electrode potential as a function of concentration is given by the Nernst equation:

E = E° −

RT (ox) ln nF (red )

(Eq 12)

where E is the electrode potential, E° is the standard electrode potential, R is the gas constant (1.987 cal/K mol), T is the absolute temperature (in degrees Kelvin), n is the number of moles of electrons transferred in the half-cell reaction, F is the Faraday constant (F = 23,060 cal/volt equivalent), and (ox) and (red) are the activities of the oxidized and reduced species, respectively.

Fig. 5 Electrochemical cell containing a zinc electrode and hydrogen electrode

Electrode potentials, as described above, are always measured when zero current is flowing between the electrode and the SHE. The potential is thus a reversible measurement of the maximum potential that exists and an indication of the tendency for the particular reaction to occur. For example, metals listed in Table 1 above molecular hydrogen are more noble and less resistant to oxidation than the metals listed below hydrogen when standard-state conditions exist. This tendency is a thermodynamic quantity and does not take into account the kinetic factors that may limit a reaction because of such physical factors as protection by corrosion product layers. Care should be taken when using an emf series such as that shown in Table 1. These values are for a very specific condition (standard state) and may not apply to a specific corrosion environment. More complete emf series (Ref 6, 7) and potentials in other environments (Ref 8) are available. Returning to the example of an electrochemical cell with copper and zinc electrodes, it is apparent that the chemical energy that exists between the copper and zinc electrodes can be converted to electrical energy (as occurs in a battery). However, the external circuit can be replaced with a direct current (dc) power supply, which can be used to force electrons to go in a direction opposite to the direction they tend to go naturally. Both concepts are useful when dealing with corrosion because the oxidation of a metal will always be coupled to a cathodic reaction and because corrosion

reactions are similar to the galvanic-type cell. Also, application of external potentials can be used to protect metals, as in cathodic protection (see article "Cathodic Protection" in this Volume). Corrosion processes are often viewed as the partial processes of oxidation and reduction previously described. The oxidation reaction (anodic reaction) constitutes the corrosion of the metallic phase, and the reduction reaction (cathodic reaction) is the result of the environment. Several different cathodic reactions are encountered in metallic corrosion in aqueous systems. The most common are:

2H+ + 2e- → H2 Hydrogen ion reduction O2 + 4H+ + 4e- → 2H2O Reduction of dissolved oxygen (acid media) O2 + 2H2O + 4e- → 40HReduction of dissolved oxygen (basic media) M3+ + e- → M2+ Metal ion reduction M2+ + 2e- → M Metal deposition Hydrogen ion reduction is very common because acidic media is so often encountered, and oxygen reduction is very common because of the fact that aqueous solutions in contact with air will contain significant amounts of dissolved oxygen. Metal ion reduction and metal deposition are less common and are encountered most often in chemical process streams (Ref 9). All of the above reactions, however, share one attribute: they consume electrons. Potential Measurements With Reference Electrodes D.L. Piron, Department of Metallurgical Engineering, École Polytechnique de Montreal

Electrode potential measurement is an important aspect of corrosion prevention. It includes determination of the corrosion rate of metals and alloys in various environments and control of the potential in cathodic and anodic protection. Many errors and problems can be avoided by intelligently applying electrochemical principles in the use of reference electrodes. Among the problems are the selection of the best reference for a specific case and selection of an adequate method of obtaining meaningful results. It is important to note that many different reference electrodes are available, and others can be designed by the users themselves for particular problems. Each electrode has its characteristic rest potential value, which can be used to convert the results obtained into numbers expressed with respect to other references. These conversions are frequently required for comparison and discussion, and this involves use of E-pH (Pourbaix) diagrams, which will be discussed later in this article. The electrode selected must than be properly used, taking into account the stability of its potential value and the problem of resistance (IR) drop. Thermodynamics of Aqueous Corrosion

Electrode Potential Conventions The use of reference electrodes is based on two fundamental conventions. One of these conventions sets a zero reference point in the potential scale, and the other gives a meaningful sign to potential values. The Zero Convention. The potential of an electrode can be determined only with respect to another electrode, the reference electrode. As discussed previously in the section "Electrode Potentials" in this article, only the potential difference between two electrodes, each with its own specific potential, is measured.

The absolute value of the potential of a particular electrode cannot be obtained experimentally. One electrode, therefore, must be selected as 0 in the potential scale. By convention, the standard hydrogen electrode (SHE) was chosen, that is, the standard electrode potential for the reaction 2H+ + 2e- → H2 is made to equal 0. This zero convention makes it possible to assign numbers to electrode potentials on the scale of electrode potentials. The SHE is arbitrarily fixed as the zero level, and all other potentials are expressed with respect to this reference. Practical measurements are performed with various reference electrodes having known values with respect to the SHE. For example, the saturated calomel electrode potential is +240 mV versus SHE, and the copper sulfate/copper (CuSO4/Cu) electrode potential is +310 mV versus SHE. The Sign Convention (The Reduction or Stockholm Convention). Electrode reactions may proceed in two

opposite directions. For example, the Fe2+/Fe system may undergo oxidation (Fe → Fe2+ + 2e-) or reduction (Fe2+ + 2e→ Fe).

The potential of this iron electrode is expressed with respect to the SHE = 0. The coupling of these two systems (Fe2+/Fe and H+/H2), however, brings about the spontaneous oxidation of iron. The situation is entirely different with a Cu2+/Cu system. In this case, the reduction is spontaneous in an electrochemical cell with a hydrogen electrode. This difference in the spontaneous reaction direction with respect to hydrogen can be represented by a sign. This sign is also very useful in computing cell potentials from single electrode values. The choice of a conventional direction for the reaction imposes a sign to the free energy. For the oxidation of Fe2+/Fe, the ∆

is negative, because a spontaneous reaction liberates energy. The ∆

would be positive for the reduction

reaction. In the case of copper, however, thereduction of Cu2+/Cu is spontaneous, and ∆

is therefore negative.

At the International Union of Pure and Applied Chemistry meeting held in Stockholm in 1953, it was decided to choose as the conventional direction the reduction reaction:

ox + ne- → red where ox represents the oxidized species, n is the number of electrons e-, and red is the reduced species. The sign of the electrode can be determined by using the following reaction, which was discussed previously (see Eq 6):

∆G = -nFE As a result, the Fe2+/Fe system has a negative sign, and Cu2+/Cu has a positive sign. In this reduction convention, a negative sign indicates a trend toward corrosion in the presence of H+ ions. The ferrous cations have a greater tendency to exist in aqueous solution than the H+ cations. A positive sign indicates, on the contrary, that the H+ ion is more stable than Cu2+, for example. The reduction convention selects a conventional direction reduction for electrochemical reactions. It is because this conventional direction is not necessarily the natural spontaneous direction that a sign can be given to the electrode potential. Example of Potential Conversion. The need to be consistent in expressing electrode potentials versus references in

a specific problem (regardless of the actual reference used in the measurement) is illustrated in the following example. The electrode potential of buried steel pipe is measured with respect to a CuSO4/Cu electrode, and the value is 650 mV for a pH 4 environment. If that value is mistakenly placed in the iron E-pH diagram, it could be concluded that corrosion is not going to take place. This conclusion would, however, be incorrect, because the E-pH diagrams are always computed with respect to the SHE. It is then necessary to express the result of the measurement with respect to that electrode before consulting the E-pH diagram. The measured electrode potential then has to be expressed with respect to the SHE.

Because the CuSO4/Cu electrode potential is +310 mV versus SHE, the number that expresses the measured potential is 310 mV higher with the CuSO4/Cu electrode than with the SHE. As a result, VSHE should be -340 mV. The principle of this conversion is illustrated in Fig. 6 in an electrode potential reference conversion schematic.

Fig. 6 Electrode potential conversion diagram

The value of -340 mV placed in the E-pH diagram at a pH 4 clearly indicates a corrosion region for iron (Fig. 7). It would then be definitely necessary to consider the cost benefit of a protection system for the steel pipe.

Fig. 7 Iron E-pH diagram. Dashed lines a and b are explained in Fig. 18 and in the corresponding text.

The Three-Electrode System When a system is at rest and no significant current is flowing, the use of only one other electrode as a reference is sufficient to measure the test electrode potential. When a current is flowing spontaneously in a galvanic cell or is impressed to an electrolytic cell, reactions at both electrodes are not at equilibrium, and there is consequently an overpotential on each of them. The potential difference measured between these two electrodes then includes the value of the two overpotentials. The potential of only the test electrode cannot be determined from this measurement. To obtain this value, a third electrode, the auxiliary electrode, must be used (Fig. 8). In this way, the current flows only between the test and the auxiliary electrodes. A high-impedance voltmeter placed between the test and the reference prevents any significant current flow through the reference electrode, which then does not show any overpotential. Its potential remains at its rest value. The test electrode potential and its changes under electric current flow can then be measured with respect to a fixed reference potential (most references are not made to be polarized by a current flow). The three-electrode system is widely used in the laboratory and in field potential measurement.

Fig. 8 Potential measurement with a luggin capillary. V, voltmeter

Electrode Selection Characteristics Stable and Reproducible Potential. Electrodes used as references should offer an acceptably stable and reproducible potential that is free of significant fluctuations. To obtain these characteristics, it is advantageous, whenever possible, to use reversible electrodes, which can be easily made.

The CuSO4/Cu electrode is an excellent example of a good reversible electrochemical system; it is widely used as a reference electrode in the corrosion field. It can be easily made by immersing a copper rod in a saturated CuSO4 aqueous solution, as shown in Fig. 9.

Fig. 9 Schematic of a CuSO4/Cu reference electrode

This electrode is reversible, because a small cathodic current produces the reduction reaction (Cu2+ + 2e- → Cu), while an anodic current brings about the oxidation reaction (Cu → Cu2+ + 2e-). This is not a corroding system like that of immersed iron, which dissolves anodically into Fe2+, because the immersed copper system produces the hydrogen evolution reaction under a cathodic current. In the case of the CuSO4/Cu electrode, the rest potential is the equilibrium potential that can be E the Nernst equation (Eq. 12). At 25 °C (77 °F), it would be:

computed by

where is the Cu2+/Cu equilibrium potential, 0.34 is the standard potential, and a is the activity of Cu2+ in the aqueous solution. This system then provides a well-defined reversible system that is reliable and easy to build. In some practical cases, however, nonreversible electrodes are used. Although not as well defined, their potential stability in a particular environment is considered sufficient in certain applications. In the selection of reference electrodes, their durability, life expectancy, and price must also be considered. Low Polarizability. The polarization of reference electrodes introduces an error in the potential measurement. The

potential versus current density response, called a polarization curve, should show a low overpotential and a high exchange current, iex, as can be seen in Fig. 10 (line a). More detailed information on polarization curves can be found in the article "Kinetics of Aqueous Corrosion" and in the section "Electrochemical Methods of Corrosion Testing" of the article "Laboratory Testing" in this Volume.

Fig. 10 Polarization curve for a good reference electrode (line a) and a poor reference electrode (line b)

A poor polarization characteristics for a reference electrode is represented by the dashed line (b) in Fig. 10. In this case, a small current density i1 produces a significant potential change from Erev to . This results in a large change in the overpotential ηb. The electrode represented by line (a) offers much better polarization characteristics. Under the same current density i1, the observed electrode potential overpotential ηa is then negligible.

remains very close to the reversible value. The resulting

It is a question of judgment as to how polarizable the reference electrode can be. The answer depends on the precision required and on the impedance of the voltmeter used. A high-impedance voltmeter may provide acceptable results with a more polarizable electrode than a less expensive measuring instrument. The Liquid Junction Potential. Reference electrodes are usually made of metal immersed in a well-defined

electrolyte. In the case of CuSO4/Cu electrode, the electrolyte is a saturated CuSO4 aqueous solution; for the saturated calomel electrode, it is a saturated potassium chloride (KCl) solution. This electrolyte that characterizes the reference electrode comes into contact with the liquid environment of the test electrode (Fig. 11). There is then direct contact between different aqueous media. The difference in chemical composition produces a phenomenon of interdiffusion. In this process, except for a few cases such as KCl, the cations and anions move at different speeds. However, for hydrogen chloride (HCl) solution in contact with another media, the H+ ions move faster than the Cl- ions. As a result, a charge separation appears at the limit between the two liquids (the liquid junction); this produces a potential difference called the liquid junction potential. This liquid junction potential is included in the measured potential, as expressed in:

V = VT - VR + VLJP

where VT is the unknown voltage to be measured, VR is the reference electrode potential, and VLJP is the unknown liquid junction potential.

Fig. 11 Schematic of an electrochemical cell with liquid junction potential. P, interface

In order to determine VT, the liquid junction potential has to be eliminated or minimized. The best way, when it is possible, is to design a reference electrode using electrolyte (Fig. 11) identical to the solution in which the test electrode is immersed. This can be done in some cases, for example, in overpotential measurement in a copper electrowinning cell. The reference electrode can be a copper wire in a glass tube simply immersed in CuSO4 cell electrolyte. A simple CuSO4/Cu reference electrode can be made in this way. Most of the time, however, this ideal solution is not possible, and the best approach is to minimize the liquid junction potential by using a reference electrolyte with a chemical composition as close as possible to the corrosion environment. In some cases, the use of a solution of KCl (such as in the calomel electrode) offers a partial answer. The diffusion rates of potassium (K+) and chloride (Cl-) ions are similar. In contact with another electrolyte, a KCl solution does not produce much charge separation and, consequently, no significant liquid junction potential. The ions present in the other solution, however, also diffuse, and they may do so at different rates, thus producing some separation of charge at the interface P (Fig. 11). The remaining liquid potential, after minimization, constitutes an error that is frequently accepted in electrode potential measurements, especially when compared with results determined under similar experimental conditions. Liquid junction potentials have to be minimized as much as possible. There is no general solution for this; each individual case has to be well thought out.

Operating Conditions for Reference Electrodes When a reference has been selected for a particular application, its proper use requires caution, as well as measurement methods based on the same electrochemical principles. In a measurement of the potential of a polarized electrode, it is important not to polarize the reference electrode and to keep its reference value. Very Low Current Density. It is important to use a reference electrode that operates at its known open-circuit

potential and to avoid any significant overpotential. This is achieved by using a high-impedance voltmeter that has a negligible input current and, for polarized test electrodes, by using an auxiliary electrode in a three-electrode system. The requirement is shown in Fig. 10 on curve a. The current must be maintained lower than i1 to avoid a significant overpotential ηa. The value tolerated for ηa is a matter of judgment that depends on the accepted magnitude of error in the

particular case under investigation. The use of an electrometer or a high-impedance voltmeter usually fulfills this requirement. The existence of an overpotential ηa could result, however, when less expensive equipment is used, and in this case, electrodes similar to electrode b in Fig. 10 should be avoided. The IR Drop and Its Mitigation. The IR drop is an ohmic voltage that results from the electric current flow in ionic

solutions. Electrolytes have an ohmic resistance, and when a current passes through them, an IR voltage can be observed between two distinct points. When the reference electrode is immersed at some distance from a working test electrode, it is in the electric field somewhere along the current line. An electrolyte resistance exists along the line between the test and the reference electrode. Because a current flows through that resistance, an IR voltage appears in the potential measurement according to:

V = VT - VR + IR where VT is the test potential to be measured, VR is the reference electrode potential, and IR is the ohmic drop. In this case, the liquid junction potential has been neglected. The IR drop constitutes a second unknown value in a single equation. It must be eliminated or minimized. The Luggin capillary is a tube, usually made of glass, that has been narrowed by elongation at one end. The narrow

end is placed as close as possible to the test electrode surface (Fig. 8), and the other end of the tube goes to the reference electrode compartment. The Luggin capillary is filled with cell electrolyte, which provides as electric link between the reference and the test electrode. The use of a high-impedance voltmeter prevents the current from flowing into the reference electrode and consequently into the capillary tube between the test electrode and the reference electrode department (Fig. 8). This absence of current eliminates the IR drop, and the measurement of VT is then possible. A residual IR drop may, however, exist between the tip of the Luggin capillary and the test. This is usually negligible, however, especially in highconductivity media. The remote electrode technique can be used only for measurement in an electrolyte with very low resistivity,

usually in the laboratory. It is applicable, for example, in a molten salt solution, in which the ohmic resistance R is very small. In such a case, the reference electrode can be placed a few centimeters away from the test electrode because the product IR remains negligible. In other electrolytes (for example, in measurements in soils) the ohmic resistance is rather large, and the IR drop cannot be eliminated in this manner. The Current Interruption Technique. In this case, when the current is flowing the IR drop is included in the measurement. A recording of the potential is shown in Fig. 12. At time t1, the current is interrupted so that I =0 and IR = 0.

Fig. 12 The potential decay at current interruption. See text for discussion.

At the moment of the interruption, however, the electrode is still polarized, as can be seen at point P in Fig. 12. The progressive capacitance discharge and depolarization of the test electrode takes some time. The potential measured at the instant of interruption then represents the test electrode potential corrected for the IR drop. Precise measurements of the potential of P are obtained with an oscilloscope. Potential Versus pH (Pourbaix) Diagrams D. L. Piron, Department of Metallurgical Engineering, École Polytechnique de Montreal

Potential -pH diagrams are graphical representations of the domain of stability of metal ions, oxides, and other species in solution. The lines that show the limits between two domains express the value of the equilibrium potential between two species as a function of pH. They are computed from thermodynamic data, such as standard chemical potentials, by using the Nernst equation (Eq. 12). Potential -pH diagrams then provide a graphical expression of Nernst's law. These diagrams also give the equilibrium of acid-base reactions independent of the potentials. These equilibria are represented by vertical lines at specific pHs. Potential -pH diagrams organize many important types of information that are useful in corrosion and in other fields of practice. They make it possible to discern at a glance the stable species for specific conditions of potential and pH. When applied to a metal, the equilibrium potential line gives the limit between the domains of stability of the metal and its ions. For conditions of potential and pH corresponding to metal stability, corrosion cannot take place, and the system is in a region of immunity. However, when the potential and pH correspond to the stability of ions, such as Fe 2+, the metal is not stable, and it tends to oxidize into Fe2+. The system is then in a corrosion region of the diagram. In the case of iron, corrosion is deaerated water is expressed by the electrochemical reaction Fe → Fe2+ + 2e-, and the species Fe2+ and Fe are considered. The Nernst equation (Eq. 12) makes it possible to compute the equilibrium potential for the system Fe2+/Fe:

This equilibrium potential can be represented as a horizontal line in a partial E-pH diagram (see, for example, Fig. 7). The line indicates the potential at which Fe and Fe2+ of a given concentration are in equilibrium and can coexist with no net tendency to transform into the other. Above the line is a domain of stability for Fe2+; iron metal is not stable at these potentials and tends to dissolve as Fe2+ and thus increase the Fe2+ concentration. Below the equilibrium line, the stability of the metallic iron increases, and the equilibrium concentration of Fe2+ decreases; that is, the metal becomes immune. When the reaction of the metal with water produces an oxide that protects it, the metal is said to be passivated. Diagrams such as Fig. 7 were first made by M. Pourbaix (Ref 10, 11) and have proved to be very useful in corrosion as well as in many other fields, such as industrial electrolysis, plating, electro-winning and electrorefining of metals, primary and secondary electric cells, water treatment, and hydrometallurgy. It is very important to emphasize that these diagrams are based on thermodynamic computations for a number of selected chemical species and the possible equilibria between them. It is then possible to predict from an E-pH diagram if a metal will corrode or not. It is, however, not possible to determine from these diagrams alone how long a metal will resist perforation. Pourbaix diagrams offer a framework for kinetic interpretation, but they do not provide precise information on corrosion rates (Ref 12). Moreover, they are not a substitute for kinetic studies. Because each diagram is computed for a selected number of chemical species, the addition of one or more species to the system will introduce several new equilibria. Their representation in the E-pH diagram will produce a new diagram that is different from the previous one. For example, the simple diagram of gold in water does not show any possible solubility for that metal. The addition of cyanide ions to the system, however, makes possible the formation of a gold complex soluble in water. Gold that does not corrode in water can dissolve in the presence of cyanide. This property is the basis of gold plating and of the hydrometallurgy of that metal.

Computation and Construction of E-pH Diagrams As discussed in the introduction to this article, E-pH diagrams are based on thermodynamic computations. The equilibrium potentials and the pH lines that set the limits between the various stability domains are determined from the chemical equilibria between the chemical species considered. It is interesting and practical to realize here that there are three types of reactions to be considered. • • •

Electrochemical reactions of pure charge transfer Electrochemical reactions involving both electrons and H+ Pure acid-base reactions

As will be shown, graphical expressions at the Nernst equation (Eq 12) can be constructed from each of these reactions. Reactions of Pure Charge Transfer. These electrochemical reactions involve only electrons and the reduced and

oxidized species. They do not have protons (H+) as reacting particles; consequently, they are not influenced by pH. An example of a reaction of this type is:

Ni2+ + 2e- → Ni The equilibrium potential is given by the Nernst equation (Eq 12). In the case of the nickel reaction given above, it can be written:

(Eq 13)

where E is the equilibrium potential for Ni2+/Ni; E° is the standard potential for Ni2+/Ni; R, T, F, and n are defined in Eq 12; and (Ni2+) is the Ni2+ activity in the solution.

The value of the potential obtained depends on the Fe2+ activity, but not on H+ ions, which do not participate in the electrochemical reaction. The result is then independent of the pH, and it can be represented by the horizontal line in an E-pH diagram. In order to obtain this result, it is necessary to compute the value of the standard potential, given by:

(Eq 14)

where is the stoichiometric coefficients of oxidizing and reducing species and μis given below. In this case under consideration:

and μ°Ni are standard chemical potentials. By convention, the standard potential of a chemical element is where μ° 0. This gives μ°Ni = 0, and simplifies the above equation:

The value of μ° Consequently:

, which can be found in the Atlas of Electrochemical Equilibria (Ref 13), is 11,530 cal.

This result can be introduced into Eq 13 as follows:

where 2.3 R = 4.57 at a temperature of 25 °C (77 °F) or 298 °K. If 2.3 RT/F = 0.059 V, the equilibrium potential for Ni2+/Ni will be:

E = -0.25 + 0.03 log (Ni2+) As previously stated for this case, the potential depends only on the activity of (Ni2+), not on the pH. It is customary here to select four activities: 1 or 100, 10-2, 10-4, and 10-6 g ion/L. This will provide four horizontal lines, as shown in Fig. 13: •

At a concentration of 10° g ion/L, E

= E° = -0.25 V

•

With 10 g ion/L, E

= -0.25 + 0.03 log 10-2 = -0.31 V

•

With 10-4 g ion/L, E

= -0.37 V

•

-2

-6

With 10 g ion/L, E

= -0.43 V

Fig. 13 Partial E-pH diagram for the Ni2+ + 2e-

→ Ni reaction

For any activity of Ni2+ in the solution, a horizontal line represents the equilibrium potential, that is, the potential at which Ni2+ ions and Ni metal can coexist. Above the line is the region of stability of Ni2+ ions; nickel metal at these potentials will tend to corrode and produce Ni2+, the stable species. Below the line, metallic nickel is stable, and nickel in these conditions will not corrode. Reactions Involving Both Electrons and H+. Nickel can also react with water to form an oxide, according to the

electrochemical reaction:

Ni + H2O → NiO + 2H+ + 2eThe standard potential E° is given by:

The Nernst equation (Eq 12) can in this case be written as follows:

(Eq 15)

The NiO and Ni are solid phases, and they are considered to be pure; their activity is therefore 1. The activity of water in aqueous solutions is also assumed to be 1. Equation 15 can then be simplified to:

Eeq = +0.11 + 0.03 log (H+)2 Because pH = -log (H+), it is possible to write:

(Eq 16)

= 0.11 - 0.06 pH

In this case, the equilibrium potential is a decreasing function of pH, as represented in a partial E-pH diagram (Fig. 14).

Fig. 14 Partial E-pH diagram for the Ni + H2O

→ NiO + 2H+ + 2e- reaction

The diagonal line in Fig. 14 gives the value of the equilibrium potential for Ni and NiO at all pH values. Above the line, NiO is stable, and below it, nickel metal is stable. Potential-pH diagrams are very general and can also be applied to electrochemical reactions involving nonmetallic chemicals. An example involving the reduction of nitrite ( ) in ammonia ( ) will be given here. In this case, the metal of the electrode supports the reaction by giving or taking away electrons as follows:

+ 8H+ + 6e- →

+ 2H2O

It is only a supporting electrode. The Nernst equation (Eq 12) can then be written:

(Eq 17)

where (H2O) = 1. the equation for the standard potential is:

The equilibrium potential is then given by:

It can be represented in an E-pH diagram for equal activity in

Fig. 15 Partial E-pH diagram for the

Above the line, there is a region in which . Below the line,

and

by a decreasing line in Fig. 15.

+ 8H+ + 6e-

+ 2H2O reaction

is predominantly stable but in equilibrium with smaller activities of

is predominantly stable, with smaller quantities of

.

Pure Acid-Base Reactions. The previous computations showed that there are possible equilibria between the metal

and its ions (such as Ni2+/Ni) and between the metal and its oxide (NiO/Ni). In the case of cobalt it is possible, as shown in Fig. 16, to determine the equilibria for Co2+/Co and for CoO/Co. The two equilibrium potential lines meet at some point P, and above them are two domains of stability for Co2+ and CoO. These two species are submitted to an acid-base chemical reaction:

Co2+ + H2O € CoO + 2H+

(Eq 18)

which does not involve electrons. It does not depend, then, on the potential, and it will be represented by a vertical line. Point P in Fig. 16 is one point on that line located at a pH 6.3 for an activity of 1 in Co2+ ions.

Fig. 16 Partial E -pH diagram for Co2+/Co and CoO/Co

The pH value of that line can also be computed from the chemical equilibrium, with the general equation:

∆G° = -RT In K or

(Eq 19)

where K is a constant, νR is the stoichiometric coefficient of the reactants, νP is the stoichiometric coefficient of the product, μ°R is the standard chemical potential of the reactant, and μ°p is the standard chemical potential of the product. In this case, the equilibrium as given in Eq 18 can be written:

By assuming that (CoO) and (H2O) both have activities of 1 and by replacing the standard chemical potentials by their values given in the Atlas of Electrochemical Equilibria (Ref 13), it is possible to write:

(Eq 20)

and finally:

log (Co2+) = 12.6 - 2 pH

(Eq 21)

for (Co2+) = 1 or pH = 6.3. This verifies the value obtained by tracing the two equilibrium lines. Figure 17(a) shows a partial E-pH diagram in which only three chemical species--Co, Co2+, and CoO--are considered. There are, however, other possible chemical species, such as CoO2 and introduces new equilibria that modify the diagram to give Fig. 17(b).

Fig. 17(a) Partial E-pH diagram for cobalt

, that must be considered. This

Fig. 17(b) E-pH diagram for cobalt

The Water E-pH Diagram. Pourbaix diagrams are traced for equilibrium reactions taking place in water; consequently, the water E-pH diagram always must be considered at the same time as the system under investigation. Water can be decomposed into oxygen and hydrogen, according to the following reactions:

2H+ + 2e- → H2 and

H2O → O2 + 2H+ + 2eThere are then two possible electrochemical equilibria for which the equilibrium potential can be determined by using the Nernst equation (Eq 12). For hydrogen:

where (H+) is the activity of H+ in water, and p E°

is the pressure of hydrogen near the electrode. Because, by convention,

= O, the above equation can be rewritten as follows:

(Eq 22)

The equilibrium potential for the system H2/H2O can be represented in Fig. 18 by line a, which decreases with the pH.

Fig. 18 The water E-pH diagram of 1 atm

The equilibrium potential for the oxygen/water reaction is given by the Nernst equation:

where p equals the pressure of O2 near the electrode. The activity is, as usual , assumed to be 1, and the standard potential for 02/H2O is computed to be 1.23 V. The following can then be written:

(Eq 23)

Equation 23 is represented under 1 atm pressure by line b in Fig. 18.

It is interesting to note that the pressures of hydrogen and oxygen in the vicinity of the electrode are usually identical and nearly equal to the pressure that exists in the electrochemical cell. To be rigorous, the water vapor pressure should be taken into account, but it is frequently neglected as not being very significant. When the pressure increases, line b in Fig. 18 is displaced upward in the diagram, and line a is lowered. The result is that the domain of water stability increases with increasing pressure. The water diagram is so important for a good understanding of the corrosion behavior of a metal that it is usually represented by dotted lines in all Pourbaix diagrams (Ref 13).

Practical Use of E-pH Diagrams The E-pH diagram is an important tool for understanding electrochemical phenomena. It provides much useful thermodynamics information in a simple figure. A few cases are presented here to illustrate its practical use in corrosion. Acid Corrosion of Nickel. A rod of nickel is immersed in an aqueous deaerated acid solution with a pH of 1 that

contains 10-4 g ion/L of Ni2+ ions. The system is under 1 atm pressure. These conditions make it possible to simplify the E-pH diagram, as shown in Fig. 19.

Fig. 19 E-pH diagram for nickel

At the metallic nickel/water interface, two electrochemical reactions are possible, and their equilibrium potentials can be computed:

Ni → Ni2+ + 2e-

with ENi = -0.25 + 0.03 log a

, which for concentration a = 10-4 gives ENi = -0.37 V (see Fig. 13). It follows that:

2H+ + 2e- → H2 with EH = -0.06 pH. At pH = 1, EH = -0.06 V. The nickel equilibrium potential is then more active than that of hydrogen, and electrons tend to flow from the negative nickel to the more positive hydrogen. Because both reactions occur on the same electrode surface, the electrons can go directly from the nickel to the hydrogen. The two reaction then tend to proceed under a common electrode potential of mixed potential, with a value somewhere between the nickel and hydrogen equilibrium potentials. The mixed potential EM is then above the Ni2+/Ni equilibrium potential in the region of Ni2+ stability (Fig. 19). Nickel is then not stable at low pH in water, and it tends to oxidize or corrode, producing Ni2+, according to the reaction:

Ni → Ni2+ + 2eThis charge separation could stop the ionization reaction if there were not another chemical reaction--the reduction of H+. The mixed potential EM is located below the H+/H2 equilibrium potential in a region where H2 is stable (Fig. 19). As a result, H+ can accept the electrons and is reduced according to 2H+ + 2e- → H2, producing H2 gas. The Pourbaix diagram explains the tendency for nickel to corrode in strong acid solutions. It does not indicate the rate of corrosion, however. This important information has to be obtained from a kinetic experiment, for example, by measuring the corrosion current in a polarization experiment. The Pourbaix diagram can also show that when the pH increases the difference between the nickel and the hydrogen equilibrium potential decreases in magnitude and that, consequently, the corrosion tendency becomes less important. For pHs between 6 and 8, Fig. 19 shows that hydrogen is more active than nickel. Under this condition, H+/H2 can no longer accept the electrons from nickel. Moreover, the potential of the system is in this case below the equilibrium potential of nickel in the region of metal immunity. In pure water at room temperature, nickel does not corrode for pHs between 6 and 8. Moreover, an increase in pressure according to Eq 22 lowers the equilibrium line of H+/H2 and does not change the equilibrium line of nickel. As a result, an increase in pressure favors the corrosion resistance of nickel. This behavior of nickel makes the metal slightly noble, and it is expected from the diagram to resist corrosion better than iron or zinc. The presence of elements, such as chloride, not considered in Fig. 19 may increase the corrosion tendency of nickel. For pHs higher than 8, NiO, Ni(OH)2, or Ni3O4 can form, as can be seen in Fig. 19. These oxides may in some cases protect the metal by forming a protective layer that prevents or mitigates further corrosion. This phenomenon is called passivation (passivation is described in detail in the article "Kinetics of Aqueous Corrosion" in this Volume). The presence of chloride is dangerous here, because it may attack the protective layer and then favor corrosion. Figure 19 also illustrates that for very strong alkaline solutions nickel may corrode as

when the potential is made anodic.

Corrosion of Copper. Observation of the copper E-pH diagram in Fig. 20 immediately reveals that the corrosion of copper immersed in deaerated acid water is not likely to occur. The H+/H2 equilibrium potential represented by line a is always more active than the Cu2+/Cu equilibrium potential. The H+ ions are then always in contact with immune copper metal that cannot corrode.

Fig. 20 Partial E-pH diagram for copper

The presence of dissolved oxygen in nondeaerated solutions introduces another possible reaction--O2/H2O reduction, with an equilibrium potential more noble than that of Cu2+/Cu. The O2/H2O system is then a good acceptor for the electrons abandoned by copper oxidation. The two electrochemical reactions:

O2 + 2H+ + 2e- → H2O and

Cu → Cu2+ + 2etake place at the same metal/solution interface at a common mixed potential. This discussion assumes that the solution does not contain chloride or other compounds capable of forming soluble complexes with copper. In the presence of such impurities, another diagram must be traced for copper that in some conditions reveals different corrosion behavior. The diagram gives valuable information if all the substances present in the actual system under investigation are taken into account when it is traced.

References 1. 2. 3. 4. 5. 6. 7. 8.

L. Pauling, General Chemistry, W.H. Freeman, 1964, p 338-360 J. O'M. Bokris and A.K.N. Reddy, Modern Electrochemistry, Vol 1, Plenum Press, 1977 J.M. Smith and H.C. Van Hess, Introduction to Chemical Engineering Thermodynamics, McGraw-Hill, 1975, p 159-162 K. Denbigh, Principles of Chemical Equilibrium, 2nd ed., Cambridge Press, 1981, p 133-186 M.G. Fontana, Corrosion Engineering, 2nd ed., McGraw-Hill, 1978, p 297-303 A.J. Bard, R. Parsons, and J. Jordan, Standard Potentials in Aqueous Solutions, Marcel Dekker, 1985 W.M. Latimer, Oxidation Potentials, Prentice-Hall, 1964 F.L. La Que, Corrosion Handbook, H.H. Uhlig, Ed., John Wiley & Sons, 1948, p 416

9. M.G. Fontana, Corrosion Engineering, 2nd ed., McGraw-Hill, 1978, p 12 10. Thermodynamique des Solutions Aqueuses Diluées, Potentiel D'oxydo-Réduction; (résumé de conférence), Bull. Soc. Chim. Belgique, Vol 48, Dec 1938 11. M. Pourbaix, Thermodynamics of Dilute Aqueous Solutions, Arnold Publications, 1949 12. R.W. Staehle, Marcel J.N. Pourbaix--Palladium Award Medalist, J. Electrochem. Soc., Vol 123, 1976, p 23C 13. M. Pourbaix, Atlas of the Electrochemical Equilibria, NACE, 1974

Kinetics of Aqueous Corrosion D. W. Shoesmith, Fuel Waste Technology Branch, Atomic Energy of Canada Ltd.

Introduction THE AQUEOUS CORROSION of metal is an electrochemical reaction. For metal corrosion to occur, an oxidation reaction (generally a metal dissolution or oxide formation) and a cathodic reduction (such as proton or oxygen reduction) must proceed simultaneously. For example, the corrosion of iron in acid solutions is expressed as follows:

Oxidation (anodic) Fe → Fe2+ + 2e-

(Eq 1)

Reduction (cathodic) 2H+ + 2e → H2

(Eq 2)

Overall reaction Fe + 2H+ → Fe2+ + H2

(Eq 3)

As a second example, for the corrosion of iron in a solution containing dissolved oxygen, the following expressions are used:

Oxidation (anodic) Fe → Fe2+ + 2eReduction (cathodic) O2 + 4H+ + 4e- → 2H2O

(Eq 4)

Overall reaction 2Fe + O2 + 4H+ → 2Fe2+ + 2H2O The reaction for metal dissolution (M → Mn+) driven by the cathodic reaction O

M + O → Mn+ + R

(Eq 5) R, is:

(Eq 6)

where M is a metal, O is oxygen or another oxidizing reagent, n+ is the multiple of the charge, and R is the reduced species or reduction. The corrosion process has been written as two separate reactions occurring at two distinct sites on the same surface (Fig. 1a). These two sites are known as the anode, or metal dissolution site, and the cathode, or the site of the accompanying reduction reaction.

Fig. 1 Schematics of two distinct corrosion processes. (a) The corrosion process M + O → Mn+ + R showing the separation of anodic and cathodic sites. (b) The corrosion process involving two cathodic reactions.

As shown in Fig. 1(a), the corroding metal is equivalent to a short-circuited energy-producing cell in which the energy is dissipated during the consumption of cathodic reagent and the formation of corrosion products. To maintain a mass balance, the amount of cathodic reagent consumed must be equal, in chemical and electrochemical terms, to the amount of corrosion product formed. Because electrons are liberated by the anodic reaction and consumed by the cathodic reaction, corrosion can be expressed in terms of an electrochemical current. Expressing the mass balance requirement in electrochemical terms, it can be stated that the total current flowing into the cathodic reaction must be equal, and opposite in sign to, the current flowing out of the anodic reaction (Fig. 1b). If measurable, this current can be taken as a gage of the rate of the corrosion process and therefore the rate of metal wastage. The current, known as the corrosion current, icorr, and the amount of metal corroded are related by Faraday's law:

(Eq 7)

where icorr is expressed in amps; t is the time (in seconds) for which the current has flowed; nF is the number of coulombs (C) required to convert 1 mol of metal to corrosion product, where n is the number of electrons involved in the metal dissolution reaction (n = 2 for Eq 1), and F is the Faraday constant (96,480 C/mol); M is the molecular weight of the metal (in grams); and w is the mass of corroded metal (in grams). Two additional observations can be made with regard to Fig. 1(b). First, several cathodic reactions may simultaneously support the metal corrosion; for example, in oxygenated acidic solutions, iron corrosion (Eq 1) could be simultaneously driven by the proton reduction (Eq 2) and the oxygen reduction (Eq 4). When complex alloys are involved, the metal corrosion process may also be the sum of more than one dissolution process. The corrosion current then equals the sum of the component partial currents:

icorr =

ia = - ic

(Eq 8)

Second, the area of the anodic and cathodic sites (Aa and Ac) may be very different (Aa is shown smaller than Ac in Fig. 1b). Therefore, although the anodic and cathodic currents must be equal, the respective current densities need not be:

ia = -ic ; Aa

Ac

therefore:

(Eq 9)

The term i/A is a current density and will be designated I. This inequality can have serious implications. For a smooth, single-component metal surface the anodic and cathodic sites will be separated, at any one instant, by only a few nanometers. The areas will shift with time so that the surface reacts evenly, thus undergoing general corrosion. However, such a situation often does not apply, and the presence of surface irregularities, alloy phases, grain boundaries, impurity inclusions, residual stresses, and high-resistance oxide films can often lead to the stabilization of discrete anodic and cathodic sites. Under these circumstances, metal dissolution can be confined to specific sites, and corrosion is no longer general but localized. The specific combination of a small anode and large cathode confines metal dissolution to a small number of localized areas, each dissolving with a large current density. Such a situation exists during such processes as pitting or cracking. Aqueous corrosion is a complicated process that can occur in various forms and is affected by many chemical, electrochemical, and metallurgical variables, including: • • • •

The composition and metallurgical properties of the metal or alloy The chemical (composition) and physical (temperature and conductivity) properties of the environment. The presence or absence of surface films The properties of the surface films, such as resistivity, thickness, nature of defects, and coherence.

The thermodynamic feasibility of a particular corrosion reaction is determined by the relative values of the equilibrium potentials, Ee, of the reactions involved. These potentials can be determined from the Nernst equation. The thermodynamics of a particular metal/aqueous system can be summarized in a potential-pH, or Pourbaix, diagram, as discussed in the section "Potential Versus pH (Pourbaix) Diagrams" of the article "Thermodynamics of Aqueous Corrosion" in this Volume. However, for this discussion, it is sufficient to state that if the corrosion reaction is to proceed such that metal M corrodes as Mn+, then:

(Ee)

< (Ee)O/R

(Eq 10)

However, the most important questions for the corrosion engineer are, How fast does the corrosion reaction occur? Is it localized? Can it be prevented or at least slowed to an acceptable rate? To answer these questions and to determine a course of action, it is essential to have some knowledge of the steps involved in the overall corrosion process. The overall process could be controlled by any one of several reactions, as shown in Fig. 2. Either the anodic (reaction area 1, Fig. 2) or cathodic (area 2) electron transfer reactions could be rate controlling. Alternatively, if these reactions are fast and the concentration of the cathodic reagent is low, then the rate of transport of the reagent O to the cathodic site (area 3) could be rate limiting. This situation is quite common for corrosion driven by dissolved oxygen that has limited solubility. If the metal dissolution reaction is reversible--that is, the reverse metal deposition reaction Mn+ + ne- → M can also occur--then the rate of transport of Mn+ away from the anode (area 4) could also be the slow step.

Fig. 2 Schematic of corrosion process showing various charge-transfer, film formation, and transport processes. See text for explanation of numbered reaction areas.

The presence of corrosion films adds other complications. If the concentration of dissolved metal cations close to the electrode achieves a value at which oxides, hydroxides, or metal salts precipitate (area 5), then corrosion could become controlled by transport of Mn+ (or O) through these porous precipitates (area 6). Alternatively, when coherent surface films form spontaneously on the metal surface by solid-state, as opposed to precipitation, reactions, then ionic transport of Mn+, or O2-, to the film growth sites at the two interfaces (oxide/metal or oxide/solution) (area 7) will ensure very low corrosion rates. The presence of film defects in the form of pores and grain boundaries will affect the rates of these processes. Finally, it is possible, under certain circumstances, for the corrosion process to be controlled by the electronic conductivity of surface films (area 8) when the cathodic process occurs on the surface of the film. In light of these numerous possibilities for control of the corrosion process, the remainder of this article will discuss these individual processes and the laws that govern them.

Activation Control Activation control is the term used to describe control of the corrosion process shown in Fig. 2 by the electrochemical reactions given in Eq 1 and 2. The overall anodic reaction is the transfer of a metal atom from a site in the metal lattice to the aqueous solution as the cation Mn+ or as some hydrolyzed or complexed metal cation species:

Mlattice → →

→

(Eq 11)

These steps are not necessarily separable experimentally. Similarly, cathodic reaction--for example, oxygen reduction-consists of a number of steps:

O2 + 2H+ + 2e- → H2O2

(Eq 12)

H2O2 + 2H+ + 2e- → 2H2O

(Eq 13)

The overall reactions are known as charge transfers. Either the anodic or cathodic charge-transfer reaction can control the overall corrosion rate. Both the anodic and cathodic reactions can be individually studied by using electrochemical methods in which the electrical potential applied to the electrode (or the current flowing through it) is controlled and the resulting current (or electrode potential) measured. Thus, the current-potential, or polarization, curves for both anodic and cathodic reactions can be determined. An example of an anodic polarization curve is shown in Fig. 3. This curve, for the anodic reaction, follows the Butler-Volmer equation:

(Eq 14)

where R is the gas constant, T is the absolute temperature, and β the symmetry coefficient taken to be close to 0.5. The term η the overpotential, defined by:

η E - Ee

(Eq 15)

and is a measure of how far the reaction is from equilibrium. At equilibrium (E = Ee, η= 0), no measurable current flows. However, the equilibrium is dynamic, with the rate of metal dissolution, ia, equal to the rate of metal cation deposition, -ic:

ia = -ic = io

(Eq 16)

where io is the exchange current.

Fig. 3 Current-potential relationship for a metal dissolution (M

→ Mn+)/deposition (Mn+ → M) process

If the potential is made more positive (anodic) than the equilibrium potential, then ia > |ic| and metal dissolution proceeds. Similarly, for cathodic potentials, ia < |ic| and metal cation deposition proceeds (Fig. 3). Over a short potential range, the two reactions oppose each other, but for sufficiently large overpotentials (ηa, anodic, and ηc, cathodic), one reaction occurs at a negligible rate, and the overpotential is then in the Tafel region, as indicated by point 1 in Fig. 3. The last term in Eq 14 can then be dropped, and the metal dissolution current density is given by:

(Eq 17)

Taking logarithms and rearranging yields:

(Eq 18)

where ba is the Tafel coefficient given by:

(Eq 19)

and is obtained from the slope of a plot of ηa against log ia. The intercept of this plot yields a value for io. Similarly, at cathodic overpotentials, a Tafel coefficient can be obtained for the metal cation deposition:

(Eq 20) A similar analysis can be performed for the cathodic process (O + ne → R), and Fig. 4 shows the two current-potential (polarization) curves.

Fig. 4 Current-potential relationships for a metal dissolution/deposition and an accompanying redox reaction showing how the two reactions couple together at the corrosion potential, Ecorr

If the two reactions are to couple together as a corrosion process, then the anodic current flowing because of metal dissolution must be counter-balanced by an equal cathodic current due to the reduction of O to R:

ia = -ic = icorr

(Eq 21)

where icorr is the corrosion current. This condition can be achieved only at a single potential, the corrosion potential, Ecorr, which must lie between the two equilibrium potentials, thus satisfying Eq 10:

(Ee)a < Ecorr < (Ee)c

(Eq 22)

such that the metal dissolution reaction is driven by an anodic activation overpotential:

= Ecorr - (Ee)a

(Eq 23)

and the cathodic reaction is driven by a cathodic activation overpotential:

= Ecorr - (Ec)c

(Eq 24)

The activation overpotential is a measure of how hard the anodic and cathodic reactions must be driven to achieve the corrosion current. Two additional observations can be made with regard to Fig. 4. First the thermodynamic driving force for corrosion is equal to the difference in equilibrium potentials:

∆Etherm = (Ee)c - (Ee)a

(Eq 25)

Generally, ∆Etherm is large, and the reverse reactions (Mn+ → M, R → O) can be neglected. Consequently, Ecorr is in the Tafel regions for both reactions (assuming no complications due to the presence of films). Second, the two polarization curves are not necessarily symmetrical and are seldom identical (Fig. 4). The shape of a curve is determined by the exchange current and the Tafel coefficient. The latter is determined by n and β in Eq 19 and 20. As shown in Fig. 4, the metal dissolution/deposition reaction has a large io, and the anodic and cathodic branches are close to symmetrical, as expected for β close to 0.5. The consequence of a large io is that the current-potential curve is steep, and only small overpotentials are required to achieve large currents. By contrast, the current-potential relationship for the cathodic reaction is shallow due to a small io, and the anodic and cathodic branches are not symmetrical. Rather than attempt to interpret this lack of symmetry in terms of an ill-defined symmetry coefficient, it is sufficient for this discussion to know that it is taken care of in the values of the Tafel coefficients. Because both reactions are occurring on different sites on the same surface (Fig. 1a), the corrosion current cannot be measured by coupling the material to a current-measuring device. The corrosion potential can be measured against a suitable reference electrode by using a voltmeter with an input impedance high enough to draw no current in the measuring circuit. The actual value of Ecorr cannot be predicted from the equilibrium potentials and therefore has no basic thermodynamic meaning. Figure 4 shows that its value is determined by the shape of the current-potential relationship for the two reactions and therefore by the kinetic parameters (io, β, n) for the two reactions. Because its value is determined by the properties of more than one reaction, the corrosion potential is often termed a mixed potential. In the literature, diagrams such as Fig. 4 are often plotted in the form log i versus E. The algebraic sign of the cathodic current is neglected so that the anodic and cathodic currents can both be plotted in the same quadrant (Fig. 5). Such diagrams are generally called Evans diagrams. The two linear portions in the log |i| versus E curves are the Tafel regions with slopes given by Eq 19 and 20. The exchange currents for the two reactions can be obtained by extrapolating the Tafel lines back to the respective equilibrium potentials (Fig. 5). Whether such diagrams are plotted linearly (i versus E) or in the logarithmic form is simply a matter of convenience. Sometimes, in the logarithmic plots, the nonlinearity close to the equilibrium potentials is ignored, and the curves are plotted as totally linear.

Fig. 5 Evans diagram for the corrosion process M + O

→ Mn+ + R

The intersection of the two polarization curves in the Evans diagrams gives a value for the corrosion current. This is true whatever the shape of the curves and irrespective of the rate-determining process. Such diagrams can be used to illustrate the impact of a variety of parameters on the corrosion process. Thermodynamic Driving Force. Figure 6(a) shows the same dissolution process driven by two different cathodic

reactions. Recalling the definition of ∆Etherm from Eq 25, the following can be written:

∆E'therm < ∆E''therm

(Eq 26)

giving:

i'corr < i''corr

(Eq 27)

That is, the bigger the difference in equilibrium potentials, the larger the corrosion current. The anodic activation overpotential for the first reaction [E'corr - (Ee)a], is less than that for the second reaction [E''corr - (Ee)a], and therefore the corrosion current for the second reaction is larger than for the first reaction, as shown in Eq 27.

(

)'' > (

)'

(Eq 28)

Fig. 6(a) Evans diagram for a metal dissolution coupled separately to two cathodic reactions with distinctly different equilibrium potentials, (Ee)''c and (Ee)'c

Kinetics of the Charge Transfer Reactions. The value of ∆Etherm is not the only parameter controlling the

corrosion rate. Figure 6(b) shows the same situation as in Fig. 6(a) except the two cathodic reactions possess very different polarization characteristics. Despite the fact that (Ee)''c > (Ee)'c, the activation overpotential, (

)'' is less than

( )'; therefore, the corrosion couple with the largest thermodynamic driving force produces the lowest corrosion current. Figure 6(b) shows that this can be attributed to the differences in exchange current, io, and Tafel coefficient, bc, for the two cathodic reactions. This situation often occurs for the corrosion of a metal in acid compared to its corrosion in dissolved oxygen. Even though the thermodynamic driving force is greater for corrosion in dissolved oxygen, corrosion often proceeds more quickly in acid. This is due to the slowness of the kinetics of oxygen reduction and can be appreciated by comparing the kinetic characteristics for the two processes on iron. Thus, (Io) / = 10-3 to 10-2 A/m2 and (bc) / ~120 mV/decade compared to (Io) / ~10-10 A/m2 and (bc) / > 120 mV/decade.

Fig. 6(b) Evans diagram for a metal dissolution coupled separately to two cathodic reactions, in which the impact of relative kinetics is greater than the thermodynamic driving force, ∆Etherm

Rate Control by the Anodic or Cathodic Reaction. The overall rate of corrosion will be controlled by the

kinetically slowest reaction, that is, the one with the smallest exchange current, io, and/or largest Tafel coefficient. The significance of this point and its importance in determining which reaction is rate controlling can be appreciated from Fig. 6(c), in which (io)a > (io)c and ba < bc. This leads to a large difference in activation overpotentials with . This means the cathodic reaction is strongly polarized and must be driven hard to achieve the corrosion current. However, the anodic reaction remains close to equilibrium, requiring only a small overpotential to achieve the corrosion current. Under these conditions, the corrosion potential lies close to the equilibrium potential for the kinetically fastest reaction. If the cathodic reaction was the faster, then Ecorr → (Ee)c, and metal dissolution would be rate controlling. If the kinetics of the two reactions are close, the corrosion potential will be approximately equidistant between the two equilibrium potentials, and the corrosion reaction will be under mixed anodic/cathodic control, as shown in Fig. 5.

Fig. 6(c) Evans diagram showing the impact on the corrosion current, icorr, and potential, Ecorr, of varying the kinetics of a fast metal dissolution (A1, A2 or a slow cathodic process (C1, C2)

The corrosion of iron in dissolved oxygen can be used to illustrate this point. For the metal dissolution reaction, (Io) ~10-5 to 10-4 A/m2 and (ba) ~50 to 80 mV/decade, whereas for oxygen reduction, (Io) / -10 2 ~10 A/m and (bc) / > 120 mV/decade. Consequently, oxygen reduction should be rate controlling, and the corrosion potential would be expected to be close to the metal dissolution equilibrium potential. Figure 6(c) shows the effects of changing the kinetics of the two reactions. Changes in the kinetics of the fast anodic reaction are reflected in variations in the value of Ecorr but have little effect on icorr; however, changes in the kinetics of the slow cathodic reaction have a large impact on the corrosion current but have little effect on the corrosion potential. Such effects can sometimes be used as diagnostic tests for ascertaining the rate-determining step. The maximum benefit in attempting to slow corrosion can be gained by attending to the rate-determining reaction. However, such measurements may not be unequivocal in the presence of corrosion films.

Measurement of Corrosion Rates A common method of measuring corrosion rates is simply to expose a carefully weighed piece of the material to the corrosion environment for a known length of time, remove and reweigh it, and calculate the mass lost. This is not always convenient in industrial applications because of the difficulty in placing, removing, and replacing metal coupons. However, it is possible to make use of the fact that corrosion is electrochemical in nature and to employ electrochemical methods to measure the corrosion rate.

When attempting an electrochemical measurement of the corrosion rate, one of the problems encountered is the desire to measure the current flowing at the corrosion potential. At this potential, no current will flow through an external measuring device (as discussed above). Consequently, any electrochemical attempt to measure icorr will rely on current measurements at potentials other than the corrosion potential. An approximation or extrapolation is then made to estimate the current flowing internally at the corrosion potential. The Tafel Method. As mentioned in the discussion of Fig. 4, the corrosion potential is generally in the Tafel region, in

which the anodic and cathodic reactions are both proceeding under conditions appropriate for a Tafel analysis. Consequently, the polarization curves for both processes are determined by applying potentials well away from the corrosion potential, plotting the logarithm of the current against overpotential as for the Tafel analysis (Eq 18), and then extrapolating the currents in the two Tafel regions to the corrosion potential to obtain the corrosion current. The method is illustrated in Fig. 7, in which, as with the Evans diagram, both currents are plotted in the same quadrant. The polarization curve shown in Fig. 7 is in the form obtained experimentally; as a result, the current at Ecorr passes through 0. In the Evans diagrams shown in Fig. 5 and 6(a), 6(b), and 6(c), ia and ic are shown independently, although it is generally not possible to measure them experimentally. The current measured in the external circuit and plotted in Fig. 7 is always the sum ia + ic ( = 0 at Ecorr).

Fig. 7 Plot of the total current (iT = ia + ic) versus potential showing the extrapolation of the Tafel regions to the corrosion potential, Ecorr, to yield the corrosion current, icorr

A simplified application of this method can be used to estimate the corrosion current from a simple measurement of the corrosion potential, because the latter may be the only measurable parameter in an industrial system. In this case, values of the exchange current (io), Tafel coefficients (ba, bc), and equilibrium potential (Ee) for the metal dissolution reaction must be known from a previous experiment. Combining Eq 18 and 23 yields:

(Eq 29)

or

(Eq 30)

Complications arise when corrosion films are present or when corrosion is not uniform.

The linear polarization method is applicable when corrosion occurs under activation control. As opposed to the

Tafel method, in which a large potential perturbation is applied to the system and seriously disturbs it, the linear polarization method uses only a small potential perturbation, ± ∆E ( 10 mV), for the freely corroding situation that occurs at Ecorr. This small perturbation makes the method appropriate for in situ measurements. The current measured in the external circuit then equals the change in corrosion current, ±∆i, caused by the small perturbation. Because both reactions proceed in their respective Tafel regions and in the vicinity of the corrosion potential, the currents are exponentially dependent on potential. For a small enough potential range ( 20 mV), these exponentials can be linearized, giving an approximately linear current-potential relationship. The relationship between ∆i and ∆E can then be obtained geometrically, as indicated in Fig. 8, which shows an expansion of the current-potential relationships around the corrosion potential. The terms sa and sc are the respective slopes of the anodic and cathodic curves at E = Ecorr. Thus, the following can be written:

(Eq 31)

Rearranging and differentiating Eq 30 yields:

(Eq 32)

for the anodic reaction. A similar process for the cathodic reaction gives:

(Eq 33)

Substituting for sa and sc in Eq 31 yields:

(Eq 34) The quantity ∆E/∆i is termed the polarization resistance. Again, a knowledge of the Tafel coefficients is required before the method can be applied.

Fig. 8 Plot of the current-potential relationships, expanded around the corrosion potential, showing their linearization (for small values of ∆E) to obtain icorr by using the linear polarization technique

The corrosion current can be converted into a mass flux by the application of Faraday's law (Eq 7). If the density of the material is known, a penetration rate (distance/time) can be obtained. Corrosion rates are generally expressed in millimeters per year or mils per year.

Mass Transport Control It has been assumed in this article that the corrosion rate is controlled by either the anodic or cathodic charge-transfer process (reaction area 1 or 2, Fig. 2). However, if the cathodic reagent at the corrosion site is in short supply, then mass transport of this reagent could become rate controlling (reaction area 3, Fig. 2). Under these conditions, the cathodic charge-transfer process is fast enough to reduce the concentration of the cathodic reagent at the corrosion site to a value less than that in the bulk solution. Because the rate of the cathodic reaction is proportional to the surface concentration of reagent, the reaction rate will be limited (polarized) by this drop in concentration. For a sufficiently fast charge transfer, the surface concentration will fall to zero, and the corrosion process will be totally controlled by mass transport. Because the corrosion rate is now determined at least in part by the rate of transport (the flux) of reagent to the corrosion site, this flux needs to be calculated. This can be accomplished by using the Nernst diffusion layer treatment, a simplification of the Fick's diffusion law treatment. The model is illustrated in Fig. 9. Because the concentration of reagent O is lower at the surface than in the bulk of solution, O will be transported down the chemical gradient at a rate (the flux J) proportional to the gradient of the concentration-distance profile. This is a statement of Fick's first law, which applies under steady-state conditions, that is, surface concentration and concentration gradient constant with time:

(Eq 35)

where D is the proportionality constant known as the diffusion coefficient (the negative sign accounts for the fact that the flux is down the gradient), and Co is the reagent concentration at a point x. The solid line in Fig. 9 represents the concentration profile calculated from Fick's treatment.

Fig. 9 Concentration-distance profile for the cathodic reagent O, depleted at the metal surface. The solid line shows Fick's treatment, and the dashed-dotted line indicates the approximation known as the Nernst diffusion layer treatment.

A simpler analysis can be achieved by linearizing the profile according to the Nernst diffusion layer treatment, as represented by the dashed-dotted line in Fig. 9. The resistance to mass transport lies within this diffusion layer, and the linearization yields a demarcation line at a distance δfrom the surface such that, for x > δ, the bulk concentration is maintained by convective processes. By contrast, for x ≤δ, reagent O is transported to the surface by diffusion only. This solution layer is called the diffusion layer, and its thickness is determined by the solution velocity. Using this simplified treatment, Eq 35 can be written as:

(Eq 36)

where is the reagent concentration at the corroding surface (x = 0), and is the concentration for x . For the steady state to be maintained, all the reagent transported down the gradient must react electrochemically, giving a current:

(Eq 37)

where the term nF takes care of the chemical to electrochemical conversion (see Eq 7, Faraday's law). Under the limiting conditions

O, a limiting or maximum current is obtained:

(Eq 38)

Because this is the maximum cathodic current that can flow, it represents the maximum achievable corrosion rate:

(Eq 39)

When corrosion occurs at this limit, the corrosion rate can be increased or decreased only by varying the bulk concentration of reagent, be given by: Eq 37.

, or the diffusion layer thickness, . For nonlimiting conditions, the corrosion current will

The effect of the concentration polarization can be seen by considering Fig. 10. For small shifts from the equilibrium potential (point 1), = , there is no limitation on the reagant supply. Charge transfer is completely rate controlling, and the overpotential is purely an activation overpotential: T

A

=

(Eq 40)

For larger shifts from the equilibrium potential, < (point 2), and the current is correspondingly less than expected on the basis of activation control; that is, the current follows the solid line as opposed to the dashed-dotted line. The current is both activation and concentration polarized, and the overpotential is the sum of an activation and a concentration overpotential: T

=

A

+

C

(Eq 41)

For a sufficiently large shift from equilibrium, the current becomes independent of potential, and the concentration overpotential becomes infinite (point 3). The corrosion rate is now at a maximum given by Eq 39.

Fig. 10 Polarization curve for the cathodic process showing activation polarization (point 1), joint activationconcentration polarization (point 2), and transport-limited corrosion control (point 3)

The impact of various parameters on a corrosion process proceeding under mass transport or mixed activation-transport control can be assessed by the use of an Evans diagram, as shown in Fig. 11. Three situations are considered. For cathodic curve 1, corrosion occurs with the cathodic reaction totally mass transport controlled; that is, = 0. If the solution is now stirred or made to flow, the thickness of the diffusion layer (Fig. 9) will decrease, and the corrosion current, given

by Eq 39, will increase as shown (curve 2). The corrosion potential will shift to more positive values. This shift in Ecorr is a consequence of the decrease in overpotential for the cathodic reaction due to the decrease in concentration overpotential: T

= Ecorr - (Ee)c =

C

C

A

+

C

(Eq 42)

because:

( (

)2 < ( T )2 < (

)1 T )1

(Eq 43)

Fig. 11 Evans diagram for a corrosion process initially controlled by the transport of cathodic reagent to the corroding surface (line 1). Lines 2 and 3 show the effect of increasing the transport rate of reagent.

For more vigorous stirring, the concentration overpotential becomes zero because the flux of reagent O to the corroding surface is now fast enough to maintain the surface concentration equal to the bulk concentration. The reaction becomes activation controlled again (curve 3). Fluid velocity no longer affects corrosion rate. Such changes in Ecorr and icorr with stirring or solution velocity can be used to indicate whether mass transport control is operative. If the anodic, as opposed to the cathodic, reaction was mass transport controlled, Ecorr would shift to more cathodic (negative) values with increased stirring or flow rate. Equation 38 indicates for mass transport control by the cathodic reaction the rate is directly proportional to the concentration of cathodic reagent and is inversely proportional to the thickness of the diffusion layer, which is determined by the fluid velocity (assuming the solution properties do not change). Corrosion in dissolved oxygen often proceeds in this manner, because the concentration of oxygen in solution is limited. Using this situation to demonstrate how velocity affects corrosion in flowing environments, Eq 37 can be written as:

(Eq 44)

where mc is a mass transport coefficient and would be given by DO/ if the Nernst diffusion layer treatment had been employed. As discussed above, to maintain the steady state, all the oxygen reaching the corroding surface is consumed, and the corrosion rate is given by:

(Eq 45)

where kc is the potential-dependent rate constant for the electron transfer reaction. The relationship between kc and io, and ba can be appreciated by comparing Eq 45 and 30. Eliminating

between Eq 44 and 45 yields:

(Eq 46)

where the constant kc can be considered the activation control parameter, and mc can be considered the mass transport control parameter. Whether or not activation kinetics or mass transport is rate determining is determined by the relative values of mc and kc. If mc kc, the bracketed term in Eq 46 reduces to kc, and the corrosion current is activation controlled. For kc mc, the term reduces to mc, and the corrosion current becomes mass transport controlled. For mc kc, Eq 46 cannot be simplified, and corrosion would be under joint control. If mass transport is a contributor to corrosion control, then a knowledge of the dependence of mc on flow rate is required. This dependence is found experimentally, and its form varies, depending on the geometry of the system. In general, this dependence takes the form:

icorr

fn

(Eq 47)

where f is the flow rate, and n is a constant that depends primarily on the geometry of the system. Confining attention to flow over a flat plate, n is 0.33 for laminar (smooth, Re < 2200) flow and ~0.7 for turbulent (Re > 2200) flow, where Re is the Reynold's number (Eq 48). The variation of the diffusion layer thickness, with flow conditions depends on flow rate as well as on solution properties, such as the kinematic viscosity (v), the diffusion coefficient of the reagent (D), and the geometry of the system (L). These effects can be accounted for by introducing two dimensionless parameters, the Reynolds number, Re, and the Schmidt number, Sc, given by:

(Eq 48) (Eq 49) It can be shown that for flow over a smooth, flat, corroding surface:

(icorr)max = 0.62 nFDO

(Re)0.5(Sc)0.33

(Eq 50)

showing that the corrosion rate is proportional to f0.5. Laminar flow can be maintained only up to a certain Reynolds number (or flow rate if L and v are constant), beyond which the flow becomes turbulent and the dependence on flow rate increases. For still higher flow rates, the condition mc kc can be achieved, and the corrosion rate will become activation controlled and therefore independent of flow rate. This is equivalent to the situations discussed in Fig. 11, in which the corrosion rate (current) reached a constant value as the effect of concentration overpotential was removed. These three regions are shown schematically in Fig. 12. The solid line shows the effect of flow rate when the anodic corrosion reaction is fast (kc large) and a large flow rate is required to achieve activation control (mc kc). The dotted

line shows the behavior expected for a slow anodic reaction (kc small) when only a low flow rate is required to achieve activation control.

Fig. 12 Impact of flow rate on corrosion current showing the regions of laminar and turbulent flow and the switch from transport to activation control at high flow rates

Passivation Thus far in the discussion on transport effects, only the transport of the cathodic reagent (area 3, Fig. 2) has been considered. The transport of metal dissolution product (area 4, Fig. 2) also affects the corrosion rate but in a different way. If the corrosion product is allowed to build up at the surface, supersaturation with regard to solid oxides and hydroxides can occur, leading to film formation reactions (area 5, Fig. 2). The effects of film formation have been referred to above. With regard to the Evans diagrams shown in Fig. 5, 6(a), 6(b), 6(c), and 11, it can be seen that very substantial corrosion rates would be achieved if the kinetics of both the anodic and cathodic reactions were fast. Fortunately, in many cases, the metal dissolution rate decreases to low values once the potential is raised above a critical value. The metal is said to be passivated. Passivation can occur when the corrosion potential exceeds (becomes more positive than) the potential corresponding to equilibrium between the metal and one of its oxides/hydroxides:

Ecorr > (Ee)M/MO

(Eq 51)

Inspection of the Pourbaix diagram for the particular metal/metal oxide/aqueous solution system shows that this condition moves the potential into the oxide stability region (Fig. 13). For point 1, Ecorr < (Ee)M/MO and corrosion of bare metal is expected, but for point 2, Ecorr > (E)M/MO, the metal should be oxide covered and passive. Under passive conditions, the corrosion rate will be dependent on the oxide film properties.

Fig. 13 Pourbaix diagram for the iron/water/dissolved oxygen system showing the effect of potential in moving the system from a corrosive (active) region (point 1) to a passive region (point 2)

The current-potential, or polarization, curve for the anodic process is shown in Fig. 14 and can be divided into a number of regions. In region AB, the active region, metal dissolution occurs unimpeded by the presence of surface films. The current, ia, should conform to the Tafel relationship (Eq 17), and its extrapolation back to (Ee)a would yield a value of (io)a. At a potential B, shown in Fig. 14 to coincide with (Ee)M/MO, there is a departure from the Tafel relationship that becomes more pronounced as the potential increases, leading eventually to a decrease in current to a low value. The electrode is said to have undergone an active-passive transition and, by point C, has become passive. The potential at point B may or may not correspond to the potential (Ee)M/MO. Thermodynamics demands only that the condition given in Eq 51 be satisfied for passivation to occur. The maximum current achieved immediately before the transition is termed the critical passivating current density. This can be considered as the current density required to generate a sufficiently high surface concentration of metal cations such that the nucleation and growth of the surface film can proceed.

Fig. 14 Polarization curve for a metal/metal ion system that undergoes an active to passive transition. See text for details.

The potential at which the current falls to the passive value is called the passivation potential. It corresponds to the onset of full passivity and is sometimes called the Flade potential. In most cases, it has no thermodynamic significance. For gold, platinum, and silver, it is close to (Ee)M/MO, but for most other metals, the passivation potential is much more positive than this equilibrium value. For E > Epass, the metal is said to be in the passive region. In this region, the current is independent of potential, and metal dissolution occurs at a constant rate. Two possible explanations can be offered for this constancy. First, dissolution in the passive region occurs by the transport of ionic species through the film (reaction area 5, Fig. 2) under the influence of the electric field across the film. The increase in potential through the passive region is accompanied by a progressive thickening of the film such that the electric field within the oxide, and therefore the dissolution current, remain constant. Second, the current is controlled by the rate of dissolution of the film (a chemical, as opposed to an electrochemical, process) and is potential independent. The current is just sufficient to replace the dissolving film. For potentials greater than point E, oxygen evolution can occur on the outside of the oxide film by the reaction:

4OH- → O2 + 2H2O + 4e-

(Eq 52)

For this last reaction to occur the film must be electronically conducting. This is possible because the passive films formed are commonly thin (nanometers) and possess semiconducting or even metallic properties. The dashed-dotted line in Fig. 14, in the potential region D to E, corresponds to the phenomenon of transpassivity. In this region, the oxide film starts to dissolve oxidatively, generally as a hydrolyzed cation in a higher oxidation state. An example would be the further dissolution of the passive film on chromium. Cr2O3 with chromium in the +3 oxidation state, to chromate.

with chromium in the +6 state.

The current in the passive region, then, is very dependent on the physical (conductivity, defect structure) and chemical (oxidation state) properties of the oxide. If the oxide were not present, then the current at potentials in the region C to E would be given by values obtained from the extrapolation of the active dissolution region, that is, line AB. These values would be extremely large. Any disruption of the passive film is a dangerous situation, and film breakdown at localized points leads to the initiation of such localized corrosion processes as pitting and cracking. These processes are characterized by very high local rates of metal dissolution and can lead to very rapid penetration of metal structures. Such processes will be discussed in the Section "Forms of Corrosion" in this Volume. The following discussion will describe the properties of the cathodic reaction required to force the corrosion potential into the passive region, thus causing passivation and maintaining the corrosion current equal to the passive dissolution current. For passivation to occur, two conditions must be met: • •

The equilibrium potential for the cathodic reaction must be greater than Epass, the passivation potential The cathodic reaction must be capable of driving the anodic reaction to a current in excess of the critical passivation current, icrit

Three possible situations are shown in Fig. 15. The dashed-dotted line shows the anodic polarization curve for the metal dissolution (M → Mn+ + ne-), and lines 1, 2, and 3 show the cathodic polarization curves for three different cathodic processes (On + ne- → Rn).

Fig. 15 Impact of various cathodic reactions on the corrosion current and potential for a metal capable of undergoing an active-passive transition.

Consider cathodic reaction 1 (Fig. 15), in which (Ee)c1 < Epass. Because the corrosion potential must lie between (Ee)a and (Ee)c1 for the two reactions to form a corrosion couple (Eq 22), the required condition for passivation, Ecorr > Epass, cannot be achieved. Therefore, the corrosion potential stays in the active region, and the metal will actively corrode. For cathodic reaction 2 (Fig. 15), the condition (Ee)c2 > Epass is met, but the two polarization curves intersect at an anodic current less than icrit (icrit is the minimum current density required to supply a sufficient concentration of Mn+ at the surface to initiate film growth by supersaturation with respect to the passivating oxide). Again, Ecorr < Epass, and the metal corrodes in the active region at a higher corrosion current than before. For cathodic reaction 3 (Fig. 15), the conditions (Ee)c3 > Epass and i > icrit are both met. Therefore, Ecorr > Epass and the metal passivates, with the corrosion current decreasing to a low value equal to the passive dissolution current. Mild oxidizing agents (∆Etherm = (Ee)c - (Ee)a, small) will allow active corrosion, and strong oxidizing agents (∆Etherm large) are required to force the metal or alloy into the passive region. As an example, steel corrosion in strong acid may proceed in the active region at a high rate, but in dissolved oxygen, the steel will passivate and corrode passively at an insignificant rate.

References 1. L.S. Van Delinder, Ed., Corrosion Basics--An Introduction, National Association of Corrosion Engineers, 1984 2. L.L. Shrier, Corrosion, George Newnes Ltd., 1963 3. J.M. West, Electrodeposition and Corrosion Processes, 2nd ed., Van Nostrand Reinhold, 1970 4. J.M. West, Basic Corrosion and Oxidation, 2nd ed., Ellis Horwood, 1986 5. G. Wranglen, An Introduction to Corrosion and Protection of Metals, Institut für Metal-Iskydd, 1972

Effects of Environmental Variables on Aqueous Corrosion D.C. Silverman and R.B. Puyear, Monsanto Company

Introduction CORROSION involves the interaction (reaction) between a metal or alloy and its environment. Corrosion is affected by the properties of both the metal or alloy and the environment. In this discussion, only the environment variables will be addressed, the more important of which include: • • • • •

pH (acidity) Oxidizing power (potential) Temperature (heat transfer) Velocity (fluid flow) Concentration (solution constituents)

The influence of biological organisms on these environmental variables is also an important consideration, as explained in the Appendix "Biological Effects" in this article. Additional information is available in the references cited in this article and in the Section "Specific Alloy Systems" in this Volume. Before discussing the relationships, the expanded portion of the potential-pH diagram of iron at 25 °C (77 °F) shown in Fig. 1 should be considered. As discussed in the article "Thermodynamics of Aqueous Corrosion" (see the section "Potential Versus pH (Pourbaix) Diagrams") in this Volume, these diagrams are thermodynamic and show the most stable state of the metal in an aqueous solution. The dependence of iron corrosion on oxidizing power (emf), acidity (pH), temperature, and species concentration is illustrated in Fig. 1. For example, suppose the corrosion potential lies at -0.5 V (standard hydrogen electrode, SHE) at a pH of 8. The most stable state of iron is Fe2+, indicating that iron dissolution is possible. If the pH is increased to 10 (the acidity is decreased,) the most stable state becomes magnetite (Fe3O4), and most likely, iron corrosion would greatly decrease. If the pH is then decreased to about 8.5, the most stable state (Fe2+ or Fe3O4) becomes dependent on the concentration of the dissolved iron species. Thus, the corrosion rate may become dependent on the dissolved species. A change in temperature would change the entire diagram.

Fig. 1 Potential-pH (Pourbaix) diagram for iron at 25 °C (77 °F) in water. Ionic species are at activities of 10-6 and 10-4. Source: Ref 1

This simple example shows the dominating role that the environmental variables play in corrosion. Complex interrelationships can exist. The combined values of the variables pH, potential, concentration, and temperature not only affect corrosion but also affect the action of each variable. For example, with respect to Fig. 1, the effect of a pH change is dependent on the concentration of the dissolved species, and vice versa. Therefore, although the variables are discussed individually, the important point is to realize that the effect of one variable can be dependent on the magnitude of another. This point will be further discussed in this article.

Effect of pH (Acidity) The concept of pH is complex. It is related to, but not synonymous with, hydrogen concentration or amount of acid. Before discussing how the magnitude of pH affects corrosion, some fundamentals are required. The pH is defined as the negative of the base ten logarithm of the hydrogen ion activity (Ref 2). This latter quantity is related to the concentration or molality through an activity coefficient. The term is expressed as

pH = -log a

= -log

m

(Eq 1)

where a is the hydrogen ion activity, is the hydrogen ion activity coefficient, and m is the molality (mol/1000 cm3 of water). The value of the activity coefficient is a function of everything in the solution (ions, nonionized species, and so on).

The pH is usually measured with a pH meter, which is actually an electrometer. The voltage of a hydrogen ion specific electrode is measured relative to a reference electrode. This voltage is compared to the internally stored calibration obtained from a defined standard to yield the unknown pH. The actual hydrogen ion concentration (acidity level) can be calculated from this measured pH if the activity coefficient is known. Because the test solution usually has constituents that are far different from those of the buffer, the calculated hydrogen ion concentration is at best an estimate (Ref 3). Thus, the pH measured by a pH meter and the actual amount of acid as defined by the hydrogen ion concentration are related but not necessarily equal. The importance of the hydrogen ion lies in its ability to interact with an alloy surface. Many alloys of commercial interest form an oxidized surface region, the outer most atomic layer of which often contains hydroxide-like species when water is present. Such a structure would tend to have a dependence on hydrogen ion concentration, possibly through a reaction that can be one step in corrosion (Ref 4):

H2O

OHadsorbed + H+ + e-

(Eq 2)

Thus, under a number of conditions, the hydrogen ion concentration can influence corrosion through the equilibrium that exists among it, water, and the hydroxide ion formed on the alloy surface. This interaction often results in a corrosion rate dependence on hydrogen ion concentration in the form of:

r=k

(Eq 3)

is the hydrogen ion concentration, and n is an exponent. The where r is the corrosion rate, k is the rate constant, C value of n can be dependent on the hydrogen ion concentration. This type of dependence of the reaction rate on the hydrogen ion concentration is found in a number of systems, which are discussed below. This discussion is not meant to be all-encompassing, but is meant to provide a flavor for how this dependence is observed in practice. Strongly Acid Conditions (pH < 5). Iron or carbon steel shows a complex dependence of the corrosion rate on pH.

At low pH, the corrosion mechanism is dependent not only on the hydrogen ion concentration but also on the counter-ions present. Thus, all discussion must include the total constituency of the fluid. For example, the corrosion rate of iron in sulfuric acid (H2SO4) between a pH of less than 0 and about 4 tends to be limited by the diffusion of and saturation concentration of iron sulfate (FeSO4) (Ref 5, 6). The metal dissolution rate is so high that the corrosion rate is equal to the mass transfer rate of iron from the saturated film of FeSO4 at the metal surface. Because mass transfer rates are sensitive to fluid velocity, the corrosion rate is sensitive to fluid flow. This effect is well documented for concentrated H2SO4. Corrosion of iron in hydrochloric acid (HCl) follows a different mechanism, and pH has a different effect on corrosion. The rate of corrosion is rapid at all acidic concentrations of pH < 3. Unlike the sulfate ion in H2SO4, the chloride ion seems to participate in and accelerate the corrosion rate (Ref 7). The corrosion rate increases with hydrogen ion concentration (decreasing pH). These effects are reflected in Eq 3. This behavior indicates that in HCl hydrogen ion directly influences the reaction kinetics. The ion does not influence corrosion through mass transfer. Corrosion of iron in phosphoric acid (H3PO4) solution follows a similar mechanism but with a subtle twist. Again, no passive film exists on the surface; however, the corrosion rate, at least between a pH of 0.75 and 4, seems to be independent of phosphate ion concentration at constant pH (Ref 8). The important point is that the pH effect on corrosion of carbon steel at low pH is not simple. Knowledge of how pH affects corrosion in one acid does not necessarily translate to knowledge in another acid. Very little information is available on the effect of acid mixtures on corrosion. Ferritic iron-chromium alloys have been found to exhibit behavior in concentrated H2SO4 reminiscent of the behavior of carbon steel. A strong fluid velocity sensitivity has been noted in 1 M H2SO4 (5 to 10 wt%) (Ref 9) for those alloys with less than 12 wt% Cr and in the 68 to 93 wt% range (Ref 10) for E-Brite 26-1 (26 wt% Cr). The corrosion rate tends to be related to the rate of mass transfer of FeSO4 from a saturated film on the surface. The one difference is that the presence of oxygen may impart a pseudopassivity that can be unstable. The major point is that the presence of chromium may provide little benefit in this environment. Both chromium content and H2SO4 concentration must be considered simultaneously, especially because an 18 wt% Cr ferritic alloy tends to be under activation control in 1 M H2SO4 (Ref 9).

The addition of nickel to create austenitic alloys alters this behavior in H2SO4 and eliminates, or at least diminishes, this velocity sensitivity, especially in the pH range of -0.5 to 3. At lower pH, the higher acid concentrations may produce a velocity sensitivity (Ref 11). Unfortunately, data are sparse on the effect of pH on the low corrosion rates expected for many of these alloys in this low pH range of -0.5 to 3. At still lower pH, the behavior is complex, and the particular literature on the alloy should be consulted. Impurities in the H2SO4 can significantly alter the corrosion resistance. The behavior of austenitic alloys in HCl is far different from that in H2SO4, even at the same pH or hydrogen ion concentration. The change from sulfate to chloride anion tends to be detrimental. The presence of the chloride ion raises the possibility of localized attack, for example, crevice corrosion, pitting, and stress-corrosion cracking (SCC) (Ref 12). Once again, behavior with respect to pH is complex. The literature on the particular alloy should be consulted to determine the actual behavior as a function of pH in acidic solutions. Non Group VIII base alloys show different types of pH dependencies at low pH. For example, in HCl, titanium is passive to a pH of about 0 or slightly lower. Then, a fairly abrupt change in mechanism occurs at still lower pH. There, titanium begins to corrode rather rapidly (Ref 13). The hypothesis is that the titanium valence changes from +4 to +3 and that Ti3+ is soluble (Ref 13, 14). The behavior in H2SO4 is somewhat different. Other metals and alloys are affected by acidic pH in different ways. Unfortunately, mechanistic data are less plentiful than for iron-base alloys. A number of metals show a very strong dependence of corrosion on pH. With aluminum, the rate increases exponentially as pH decreases in the acidic region (Ref 15). Indeed, the corrosion rate tends to have a very sharp minimum at a pH of 7 to 9, with sharp corrosion rate increases with both increasing and decreasing pH (Ref 15, 16). A similar effect of a sharp decrease in corrosion rate with increasing pH for pH < 4 has been noted for both zinc in HCl and lead in nitric acid (HNO3) (Ref 17). Indeed, this type of behavior would be expected for any metal or alloy whose oxide is soluble in acids, such as zinc, aluminum, lead, tin, and copper. Near-Neutral Conditions (5 < pH < 9). Corrosion behavior and alloy-environment interactions in the near-neutral

pH region differ significantly from those under acidic conditions. In most cases, pH no longer plays a direct role in corrosion. Iron (as carbon steel) has been one of the most extensively studied metals in this environment. Under acidic conditions, the oxide or hydroxide layers tend to dissolve. However, in the higher pH range, especially above a pH of about 5, these layers tend to remain on the surface. These layers have significant structure, which tends to be determined by the anions present in the solution (Ref 18). In addition, the corrosion kinetics become independent of pH, and hydrogen ion reduction is no longer an important reaction (Ref 19). The major reaction governing corrosion in most practical applications is the reduction of oxygen present in solution. Magnetite (Fe3O4) can be formed, which will tend to passivate iron (Ref 18). Thus, pH in this range no longer plays a major direct role in corrosion of iron, although the pH can still affect the solubility and equilibrium of other ions, such as sequestering agents. These other components can play a major role in corrosion in cooling water. This characteristic of pH in the range of 5 < pH < 9 no longer playing a dominant role is found with other metals, such as zinc and lead (Ref 17). Aluminum shows a very sharp minimum in corrosion rate at about a pH of 7 to 9, with the minimum being somewhat dependent on the counter-ion in solution (Ref 15). Alloys such as the austenitic iron-base and nickel-base alloys, ferritic alloys, and duplex alloys also tend to have general corrosion rates that are independent of pH in this range. Indeed, in pure water, these alloys would be passive. The presence of other constituents, such as chloride ions and oxygen, plays a much more dominant role, possibly changing the mechanism from uniform corrosion to localized attack. Strongly Basic Conditions (pH > 9). Basic conditions offer yet another set of corrosion characteristics. In a number

of cases, corrosion rate increases with pH (decreasing hydrogen ion concentration) or at least remains finite. In other cases, the increase in pH causes corrosion to occur when none was present at lower pH. These two types of behavior seem to encompass most metals and alloys, and representative examples will be described to demonstrate this behavior. Iron corrosion persists even at high pH. This persistence is caused by soluble species (Fe potentials, Fe

or, at elevated

) being the most thermodynamically stable corrosion products (Ref 1, 16). Even though a number

of iron hydroxide species can be found that can create a porous barrier (Ref 20), corrosion still persists, although usually at a fairly low rate, until very high pH is reached (Ref 17). At very high pH and especially at somewhat elevated temperatures, carbon steel can undergo SCC (Ref 21). Some environments that can cause SCC at high pH are sodium hydroxide (NaOH) at very high pH, carbonates and bicarbonates at moderately basic pH values, and possibly amines, although this point is controversial. Steel can also suffer SCC at lower pH, but this behavior is less prevalent. Examples of these environments are hydrogen fluoride (HF) vapors and hydrogen sulfide (H2S). The mechanism in these cases may be one of hydrogen embrittlement (Ref 22). A number of metals exhibit a sharp increase in corrosion rate with increasing pH. Among these are aluminum, zinc, and lead (Ref 15, 17). Aluminum corrosion increases very dramatically, changing by almost two orders of magnitude between a pH of 8 and 10. This increase is virtually independent of counter-ion and can be attributed to the formation of soluble aluminum hydroxide products (Ref 16). Tantalum, which suffers virtually no corrosion under most acidic and neutral pH conditions, shows a significant increase in corrosion rate at high pH (Ref 23). The cause of this corrosion is believed to be a slow dissolution or flaking off of surface layers (Ref 23.) This dissolution is probably caused by the formation of soluble tantalum hydroxide corrosion products (Ref 13). Some metals, such as nickel and zirconium, are very resistant to corrosion at high pH. Possibly nickel and especially zirconium rely on the formation of insoluble oxides for their corrosion protection (Ref 24). The austenitic and ferritic alloys tend to be immune to corrosion until very high pH is reached. One reason is that chromium, which is included in many of these alloys and which tends to accumulate on the surface, forms a passive oxide. This oxide, for example, chromium oxide (Cr2O3), is insoluble under these conditions. However, changes in temperature can affect corrosion at high pH.

Oxidizing Power (Potential) Oxidizing power, or potential, relates to the ability to remove or add electrons from the metal so as to oxidize or reduce the surface. This variable is separated from the discussions on solution chemistry because such a potential can be applied by an external voltage source, by galvanic coupling of different metals, or by solution constituents. Practical applications include increasing passivity by altering the surface oxide (anodic protection) or preventing corrosion by supplying electrons to the metal that would normally be yielded by metal corrosion (cathodic protection). The anodic reaction rate is shifted or changed in the protected metal. The alteration of the surface state to impart passivity is normally accomplished by anodic polarization of the metal or alloy surface to a potential noble to the corrosion potential. If an external voltage source is used to change the voltage, the technique is known as anodic protection (see the article "Anodic Protection" in this Volume). Among the practical examples of using externally applied anodic potentials to mitigate corrosion are mild steel and type 304 stainless steel in concentrated H2SO4, and NaOH (Ref 22). The addition of constituents to the environment may alter the surface potential to create a passive film. In this case, the constituent reacts with the metal to form a tenacious metal-oxide compound that passivates the surface. There are several well-known examples of anodic polarization of the surface by changing the environment. For example, the addition of small amounts of ozone to water decreases the corrosion of carbon steel in water (Ref 25). The hypothesis is that the corrosion potential moves in a noble direction and the ozone reacts with the iron to create a more tenacious oxide. In another example, the addition of HNO3 to H2SO4 has been shown to retard the corrosion of stainless steels. The hypothesis is that the potential is forced in the noble direction and the surface oxide layer becomes more protective. Polarization of the surface potential in the active or cathodic direction can also be used to decrease corrosion. When the potential is lowered by means of an external voltage source, the technique is known as cathodic protection (Ref 22). Many practical examples exist, such as the protection of steel at coating defects in underground carbon steel pipelines (Ref 25) or the protection of ships hulls in seawater (see the articles "Cathodic Protection," "Marine Corrosion," and "Corrosion of Pipelines" in this Volume). The electrons are supplied from either an inert or active counterelectrode.

Direct electrical coupling of a metal to a more active metal is another example of using cathodic potentials to affect corrosion. Coupling zinc to steel to protect the steel is a major example. In this case, zinc corrosion liberates electrons to the steel, and the steel potential moves in an active direction (Ref 26). Such cathodic polarization can be produced by constituents in the solution. Oxygen tends to polarize carbon steel in a noble direction and increase its corrosion. The addition of such species as sulfite (

) or hydrazine tends to cause a

reaction with the oxygen and thus remove it (Ref 26). The effect of ( ) tends to be to move the surface potential in the active direction. Such movement of potential may decrease the corrosion of iron (Ref 17). However, any change in potential may be dependent on other constituents, especially if they can interact with the inhibitor and the metal surface. Also, if the alloy is passive and this passivity is maintained by the oxygen, this addition could increase corrosion by moving the alloy into an active corrosion region (Ref 17).

Temperature and Heat Transfer Temperature is a complex external variable. Temperature is analogous to potential. A potential difference creates a current flow, the objective of which is to eliminate the potential difference. In a similar manner, a temperature difference creates a heat flow, the objective of which is to eliminate the temperature difference. Both potential and temperature are measures of energy. Temperature can affect corrosion in a number of ways. If the corrosion rate is governed completely by the elementary process of metal oxidation, the corrosion rate increases exponentially with an increase in temperature. This relationship is reflected in the Arrhenius expression:

(Eq 4)

where r is the corrosion rate, A is a preexponential factor, E is an activation energy, R is the gas constant, and T is the absolute temperature. The effect of temperature on corrosion rate is shown by solving Eq 4 at two temperatures and taking the ratio of the rates:

(Eq 5)

where the subscripts 1 and 2 refer to the two temperatures and ∆T is the difference in temperature (T2 - T1). Equation 5 can be used to evaluate the effect of a temperature change on corrosion rate for this simple rate process. Examples of corrosion that follow this simple rate law are iron in HCl (Ref 27) and iron in sodium sulfate (Na2SO4) at a pH of about 2 (Ref 28). This situation is most common for corrosion under acidic conditions. The temperature of the metal and the temperature of the solution often cannot be discussed separately from other variables. If a constituent in the solution that is important in corrosion has limited solubility, a temperature change can alter the concentration of that constituent. This alteration can have a profound effect on corrosion. One classical example is the corrosion of iron in the presence of oxygen in systems both closed from the atmosphere and open to the atmosphere. The corrosion rate of iron in a system closed to the atmosphere has been shown to increase almost linearly with temperature from about 40 to 160 °C (105 to 320 °F). However, in the open system, the corrosion rate increases up to about 80 °C (175 °F) and then decreases (Ref 17). Oxygen mass transfer, which is proportional to the oxygen concentration in the liquid, controls the corrosion rate of steel in water. As temperature increases, oxygen solubility decreases so that the oxygen will tend to leave the liquid. In the closed system, the oxygen cannot escape from the vapor space above the liquid. As temperature increases, the water vapor pressure increases, which tends to maintain the oxygen concentration in the liquid. The corrosion rate (mass transfer rate) continues to increase with temperature because of temperature effects on viscosity, diffusivity, and so on. In the open systems, oxygen can escape from the

immediate vicinity of the liquid. The vapor pressure remains constant. Above a certain temperature, the liquid-phase oxygen concentration in equilibrium with oxygen in the atmosphere has decreased to the extent that the corrosion (mass transfer) rate decreases. Another point often overlooked is that the ionization constant of water increases with temperature. Pure water with pH of 7 at one temperature will have a lower pH at a higher temperature. Thus, an increase in temperature could affect corrosion by moving the pH from a neutral to an acidic value. Fluid temperature changes can affect the polarity in galvanic corrosion. The corrosion potential of the anode might be more sensitive to temperature than that of the cathode. The anode potential can actually become noble with respect to that of the cathode (Ref 17). An example is the iron-zinc couple, the polarity of which can reverse as temperature increases. Iron will actually protect the zinc. The temperature of this reversal is as low as 60 °C (140 °F), but there is some dependence of temperature on constituents (Ref 26). Solution temperature can also affect the onset of localized attack of passive alloys such as type 304 and 316 stainless steels. The solution usually contains a species, such as chloride ion, that aids in the initiation process (Ref 29). The time to initiation of crevice corrosion has been shown to be a function of temperature. There are indications that such initiation times do not always decrease with increasing temperature (Ref 29). In addition, a critical crevice temperature can be defined for many of these alloys (Ref 30). This critical temperature determines the temperature boundary at which crevice corrosion can initiate. Indeed, the critical crevice temperature has been shown to be a function of the chromium and molybdenum content of austenitic and ferritic alloys. In practice, elevated or depressed temperatures are often created by heat transfer through a metal wall. Thus, the metal wall can be at a temperature different from that of the bulk fluid. There is a controversy over whether corrosion in the absence of heat transfer is identical to corrosion in the presence of heat transfer even if the metal temperatures are identical in the two situations (Ref 28). The effect of a difference between wall and fluid temperatures on corrosion depends on the corrosion mechanism. If the corrosion rate is under activation control and follows Eq 4, the corrosion rate in the presence of heat transfer might be similar to that expected for corrosion at the same wall temperature in the absence of heat transfer. If the corrosion rate is controlled by the diffusion of a species, such as oxygen, to the surface then heat transfer may greatly change the corrosion rate. This effect has several possible causes (Ref 31, 32). First, a temperature difference between the wall and bulk solution can affect the solubility and diffusion coefficient of the diffusing species. Second, boiling near or on the wall can increase turbulence and possibly cause cavitation or increased diffusion (mass transfer). Third, heat transfer in the absence of fluid flow, as in stagnant tanks, can cause natural convection currents that can enhance mass transfer. Thus, if heat transfer is present, it must be considered an environmental variable.

Velocity/Fluid Flow Rate Fluid flow rate, or fluid velocity, is also a complex variable (Ref 33). Its influence on corrosion is dependent on the alloy, fluid constituents, fluid physical properties, geometry, and corrosion mechanism. These relationships are best discussed in terms of specific examples. In a number of instances, the corrosion rate is determined by the rate of transfer of a species between the surface and the fluid. This situation arises when the corrosion reaction itself is very rapid and one of the corrosion reactants or products has low solubility in the bulk fluid. The corrosion rate becomes a function of the concentration gradient and is expressed by:

r = k (CW - CB)

(Eq 6)

where r is the corrosion rate, k is a mass transfer coefficient, CW is the concentration of the rate-limiting species at the metal wall, and CB is the concentration of the rate-limiting species in the bulk fluid. The value of k can often be correlated with the dimensionless quantities Reynolds number (Re) and Schmidt number (Sc). The mass transfer coefficient is expressed in terms of the Sherwood number (Sh). These numbers are related to physical properties of the fluid and geometry by:

(Eq 7a)

(Eq 7b)

(Eq 7c)

where v is the fluid velocity, d is a characteristic length (for example, pipe diameter), v is the kinematic viscosity (absolute viscosity divided by density), and D is the diffusion coefficient. For many geometries, these quantities can be related by:

Sh = a Reb Scc

(Eq 8)

where a, b, and c are constants. Equations 6, 7a, 7b, 7c, and 8 indicate that the corrosion rate can be calculated if it depends on the mass transfer rate of a species from or to the bulk fluid. The only information required is the geometry, fluid velocity, and physical properties. There are a number of examples of corrosion that follow this behavior. The corrosion of carbon steel and E-Brite 26-1 in concentrated H2SO4 is governed by the rate of mass transfer of FeSO4 from a saturated layer on the surface (Ref 5, 6, 10). Carbon steel corrosion in water in the near-neutral pH range is governed by the rate of mass transfer of dissolved oxygen from the bulk fluid to the surface (Ref 34). If a porous surface hydroxide layer forms, the mass transfer rate might become limited by diffusion through the porous film. This effect of velocity has ramifications for localized attack, especially pitting and crevice corrosion. The presence of fluid flow can sometimes be beneficial in preventing or decreasing localized attack. For example, type 316 stainless steel has been shown to pit in quiescent seawater but not in moving seawater (Ref 35). When the seawater is moving, the mass transfer rate of oxygen is high enough to maintain a completely passive surface, but in the absence of flow, the mass transfer of oxygen is too slow and the surface cannot remain passive (Ref 36). This observation indicates that sometimes fluid velocity can be beneficial even if the corrosion rate involves the mass transfer of a reactant or product. The propensity for localized attack to occur can sometimes be decreased by maintaining sufficient fluid motion. Under other circumstances, fluid flow can cause a type of erosion of a surface through the mechanical force of the fluid itself. This common process is called impingement. The process involves the removal of metal or alloy by the high wall shear stress created by the flowing fluid. Examples of such erosion occur either where fluid is forced to turn direction, for example, at pipe bends (Ref 22), or where high surface shear stresses can exist, for example, on ship hulls (Ref 35). Evidence exists that a critical wall shear stress can be defined for an alloy above which impingement causes erosion and below which such erosion is absent (Ref 37, 38). Thus, shear stress can be translated to a maximum velocity. This phenomenon has been demonstrated for copper-nickel alloys and aluminum alloys in salt water (Ref 39). When solids are present in the liquid, they can cause wear or solid erosion corrosion (Ref 40). The wear is caused by the relative movement of the solids with respect to the surface. Again, such wear is more prevalent where fluid is forced to change direction or where high shear stresses occur. The particles must penetrate the laminar sublayer with enough force to remove the passive film on the alloy. Therefore, high shear stresses are often required for this type of erosion to occur. This problem can be significant in such systems as salt water carrying solids (for example, sand or coal) and carbon steel carrying air plus particulates.

Concentration The concentration of constituents within the fluid often influences how the other variables manifest themselves. This discussion will focus on how the concentration of constituents works with other variables to influence corrosion behavior. During the previous discussion, the point was made that pH plays a major role in corrosion. For iron, the corrosion rate is large at very low pH, is independent of pH in the neutral pH range, decreases with increasing pH, and finally increases again at very high pH. Additions of small amounts of other components can change this behavior. For example, additions of chloride to H2SO4 increase the corrosion rate of iron. This increase is reported to be proportional to the chloride ion concentration raised to about the 0.5 power (Ref 41). A similar effect is reported for

chloride ion in HCl (Ref 7). Thus, chloride ion accelerates the corrosion of ion in acidic solutions. However, bromide and iodide ions may inhibit corrosion (Ref 41), although this finding is controversial (Ref 17). The dependence of the corrosion rate of iron on chloride ion concentration significantly decreases in neutral solutions when oxygen is present (Ref 17, 42). Oxygen accelerates the cathodic reaction far more than chloride can accelerate the anodic reaction. As salt concentration increases, the oxygen solubility decreases, masking the effect of chloride ion. The chloride ion effect is dependent on the cation, with the rate increasing in the order lithium chloride (LiCl), sodium chloride (NaCl), and potassium chloride (KCl) partially because of differences in oxygen solubility in the presence of these salts (Ref 17). These results illustrate that the effect of the concentration of one component on corrosion is often dependent on other environmental variables. Small additions of certain inhibitors or passivators have a marked effect on corrosion. For example, as little as 0.0023 mol/L of sodium nitrite (NaNO2) or sodium sulfite (Na2SO3) can decease the pit initiation rate of aluminum. Little improvement is found at higher concentration for this system (Ref 43). This behavior is often found with many types of inhibitors. For example, small concentrations (10 ppm) of NaNO2 (a passivator) can drastically inhibit the corrosion of iron, with little further decrease in corrosion found at higher concentrations (Ref 44). However, the critical concentration can depend on pH and on the presence of other constituents. Much higher concentrations may be required, depending on the other constituents. The actual concentration needed for a given system must be determined experimentally. Similarly, many organic inhibitors cause a drastic decrease in corrosion rate at very low concentrations, especially for iron in acidic solutions, with no benefit observed upon increasing the inhibitor concentration (Ref 45). All of these inhibitors tend to interact with the surface in one of three ways: a gettering of a finite amount of impurity in the solution (hydrazine), oxidation (passivation) of the surface (nitrite or chromate), or adsorption on the finite surface area to block corrosion (many organic inhibitors in acid). However, although this type of behavior is common even for iron in neutral, aqueous environments, exceptions do exist. Sometimes, corrosion can increase with inhibitor concentration until a maximum is reached, followed by a rapid decrease with still further increases in concentration (Ref 44). An example is chromate ion. Chromate ion is normally considered to be an inhibitor. However, at very low concentrations and in the presence of strong activating ions such as chlorides in acidic media, chromate ion can actually accelerate corrosion until enough chromate is present. Also, synergistic action may be observed in which the efficacy of one inhibitor is dependent on the presence of another species, for example, oxygen or other oxidizing agents. A question often asked is, What is the amount of chloride that is allowable before localized corrosion (crevice corrosion, pitting, or SCC) can occur in austenitic alloys? The answer is not straightforward. Work with boiling, saturated magnesium chloride suggests that 42 wt% Ni in the alloy prevents SCC (Ref 22). However, this rule of thumb does not answer the question. The maximum chloride concentration is dependent on the pH, other constituents, temperature, and other variables. Guidelines are available (Ref 29), and the articles in this Volume on the resistance of individual alloys to localized attack should be consulted.

Effects of Environmental Variables on Aqueous Corrosion D.C. Silverman and R.B. Puyear, Monsanto Company

Appendix: Biological Effects Stephen C. Dexter, College of Marine Studies, University of Delaware

Biological organisms are present in virtually all natural aqueous environments. In seawater environments, such as tidal bays, estuaries, harbors, and coastal and open ocean seawaters, a great variety of organisms are present. Some of these are large enough to observe with the naked eye, while others are microscopic. In freshwater environments, both natural and industrial, the large organisms are missing, but there is still a great variety of microorganisms, such as bacteria and algae. In all of these environments, the tendency is for organisms in the water to attach to and grow on the surface of structural materials, resulting in the formation of a biological film, or biofilm. The film itself can range from a microbiological

slime film on freshwater heat transfer surfaces to a heavy encrustation of hard-shelled fouling organisms on structures in coastal seawater. There is a voluminous amount of literature on the formation of such films and their many adverse effects (Ref 46, 47, and 48). The biofilms that form on the surface of virtually all structural metals and alloys immersed in aqueous environments have the capability to influence the corrosion of those metals and alloys. This influence derives from the ability of the organisms to change the environmental variables discussed earlier in this article (pH, oxidizing power, temperature, velocity, and concentration). Thus, the value of a given parameter at the metal/water interface under the biofilm may be quite different from that in the bulk electrolyte away from the interface. The result can be the initiation of corrosion under conditions in which there would be none in the absence of the film, a change in the mode of corrosion (that is, from uniform to localized), or an increase or decrease in the corrosion rate. It is important to note, however, that the presence of a biofilm does not necessarily mean that there will always be a significant effect on corrosion. The purpose of this Appendix is to consider in general the characteristics of organisms that allow them to interact with corrosion processes and the general mechanisms by which organisms can influence the occurrence or rate of corrosion.

General Characteristics of Organisms The organisms that are known to have an important impact on corrosion are mostly microorganisms such as bacteria, algae, and fungi (yeasts and molds). In this section, the general characteristics of the microorganisms that facilitate their influence on the electrochemistry of corrosion will be discussed (Ref 49, 50, and 51). Information on the individual organisms can be found in the discussions of biological corrosion in the article "General Corrosion" and "Localized Corrosion" in this Volume. Physical Characteristics. Microorganisms range in length from 0.1 to over 5 μm (some filamentous forms can be several hundred micrometers long) and up to about 3 μm in width. Many of them are motile; that is, they can "swim" to a favorable, or away from an unfavorable, environment. Because of their small size, they can reproduce themselves in a short time. Under favorable conditions, it is common for bacterial numbers to double every 20 min or less. Thus, a single bacterium can produce a mass of over one million organisms in less than 7 h.

In addition to rapid reproduction, the bacteria as a group can survive wide ranges of temperature (-10 to > 100 °C, or 15 to 212 °F), pH (~0 to 10.5), dissolved oxygen concentration (0 to saturation), pressure (vacuum to > 31 MPa, or 4500 psi), and salinity (tolerances vary from the parts per billion range to about 30% salt). Despite these wide ranges of tolerance for the microorganisms as a whole, most individual species have much narrower ranges. Most bacteria that have been implicated in corrosion grow best at temperatures of 15 to 45 °C (60 to 115 °F) and a pH of 6 to 8. Oxygen requirements vary widely with species. Microbes may be obligate aerobes (require oxygen for growth), microaerophilic (require minute levels of oxygen for growth), facultative anaerobes (grow with or without oxygen), or obligate anaerobes (grow only in the complete absence of oxygen). Some microbes can produce spores that are resistant to a variety of environmental extremes, such as drying, freezing, and boiling. Spores have been known to survive for hundreds of years under arctic conditions and then to germinate and grow 0when conditions become favorable. Many microbes can quickly adapt to a wide variety of compounds as food sources. This gives them high survivability under changing environmental conditions. Metabolic Characteristics. Many of the microorganisms implicated in corrosion are able to have an influence on the

electrochemical reactions involved by virtue of the products produced by their metabolism. A large percentage of them can form extracellular polymeric materials termed simply polymer, or slime. The slime helps glue the organisms to the surface, helps trap and concentrate nutrients for the microbes to use as food, and often shields the organisms from the toxic effect of biocides. The slime film can influence corrosion by trapping or complexing heavy-metal ions near the surface. It can also act as a diffusion barrier for chemical species migrating to or from a metal surface, thus changing the concentrations and pH at the interface where the corrosion takes place. Some species of microbes can produce organic acids, such as formic and succinic, or mineral acids, such as H 2SO4. These chemicals are corrosive to many metals. One series of bacteria is involved in metabolizing nitrogen compounds. As a group, they can reduce nitrates (

) (often used as a corrosion inhibitor) to nitrogen (N2) gas. Others can convert

to nitrogen dioxide (NO2), or vice versa, or they can break it down to form ammonia (NH3). Still other series of

bacteria are involved in the transformation of sulfur compounds (Fig. 2). They can oxidize sulfur or sulfides to sulfates (

) (or H2SO4), or they can reduce

to sulfides, often producing corrosive H2S as an end product.

Fig. 2 The sulfur cycle showing the role of bacteria in oxidizing elemental sulfur to sulfate ( reducing sulfate to sulfide (S2-). Source: Ref 52

) and in

Organisms that have a fermentative type of metabolism produce carbon dioxide (CO2) and hydrogen (H2); others can utilize CO2 and H2 as sources of carbon and energy, respectively. Numerous species of bacteria and algae either produce or utilize oxygen. It is rare that a corrosion process would not depend on the concentration of at least one of these three dissolved gasses. Finally, some bacteria are capable of being directly involved in the oxidation or reduction of metal ions, particularly iron and manganese. Such bacteria can shift the chemical equilibrium between Fe, Fe2+, and Fe3+, which will often influence the corrosion rate. Community Structure. The ability of an organism to survive on a surface and to influence corrosion is often related to

associations between that organism and those of other species. The bacteria implicated in corrosion may begin their lives on a metal surface as a scatter of individual cells, as shown in Fig. 3(a). As the biofilm matures, however, the organisms

will usually be found in thick, semicontinuous films (Fig. 3b) or in colonies (Fig. 3c). It is in these latter two forms that there is the most potential for survival and growth of the organisms capable of influencing corrosion.

Fig. 3 Various forms of bacterial film that can influence corrosion. (a) Scatter of individual cells. 6050×. (b) Semicontinuous film of bacteria in slime. 3150×. (c) Bacterial cells in a colony. 2700×

For example, the sulfate-reducing bacteria (SRB) are implicated in the corrosion of iron-base alloys in a variety of environments (Ref 52, 53). Most sulfate-reducing bacteria are obligate anaerobes, yet they are known to accelerate corrosion in aerated environments. This becomes possible when aerobic organisms form a film or colony and then, through their metabolism, create an anaerobic microenvironment with the organic acids and nutrients necessary for growth of the sulfate-reducing bacteria (Fig. 4). Thus, the organisms influencing corrosion can often flourish at the corrosion site by associating with other organisms in a microbial colony or consortium, even when the bulk environment is not conducive to their growth.

Fig. 4 Variations through the thickness of a bacterial film. Aerobic organisms near the outer surface of the film consume oxygen and create a suitable habitat for the sulfate-reducing bacteria at the metal surface. Source: Ref 52.

It should be noted that the dynamics of fluid flow past the metal surface can alter the form of the biofilm or can even prevent its formation. This can result in acceleration or deceleration of corrosion, depending on the role of the biofilm.

General Mechanisms of Influence The presence of a biological film on a corroding metal surface does not introduce some new type of corrosion, but it influences the occurrence and/or the rate of known types of corrosion. These biological influences can be divided into three general categories: • • •

Production of differential aeration or chemical concentration cells Production of organic and inorganic acids as metabolic by-products Production of sulfides under oxygen-free (anaerobic) conditions

Oxygen/Chemical Concentration Cells. Any biofilm that does not provide for complete, uniform coverage of the entire immersed surface of a metal or alloy has the potential to form concentration cells. In aerated environments, uncovered areas of the metal surface, in contact with oxygenated electrolyte, will be cathodic relative to those areas under the biofilm. Beneath the film or colony, oxygen is depleted as it is used by the organisms in their metabolism. Oxygen from the bulk electrolyte is unable to replenish those areas because of a combination of effects. First, oxygen migration through the film is slowed by the diffusion barrier effect, and second, oxygen that does penetrate the film is immediately utilized by the microbial metabolism. Formation of such a corrosion cell, as shown in Fig. 5, causes a pit to form at the anodic area under the bacterial colony.

Fig. 5 Schematic of pit initiation and tubercule formation due to an oxygen concentration cell under a biological deposit. Source: Ref 53

As the pit grows, iron dissolves according to the anodic reaction:

Fe → Fe2+ + 2eThe cathodic reaction is reduction of dissolved oxygen outside the pit to form OH- according to:

O2 + 2H2O + 4e- → 4OHThe insoluble ferrous hydroxide corrosion product forms by the reaction:

3Fe2+ + 6OH- → 3Fe(OH)2 Corrosion products mingle with bacterial film to form a corrosion tubercule, which itself may cause a problem with obstruction of fluid flow in piping systems. In addition, if the above process takes place in the presence of bacteria capable of oxidizing ferrous ions to ferric ions, the corrosion rate will be accelerated because the ferrous ions are removed from solution as soon as they are produced. This depolarizes the anode and accelerates corrosion of iron under the deposit. The ferric ions form ferric hydroxide (Fe(OH)3), which contributes to the rapid growth of the tubercule. This process has been responsible for corrosion and plugging of iron water pipes. If chlorides are present in the system, the pH of the electrolyte trapped inside the tubercule may become very acid by an autocatalytic process similar to that described in the article "Localized Corrosion" in this Volume for crevice corrosion and pitting. Chloride ions from the environment combine with ferric ions produced by corrosion in the presence of the bacteria to form a highly corrosive, acidic ferric chloride solution inside the tubercule. This has been responsible for severe pitting of stainless steel piping systems, as described in the section "Localized Biological Corrosion" in the article "Localized Corrosion" in this Volume. Acid Production. The sulfur oxidizing bacteria can produce up to about 10% H2SO4. This mineral acid, with its accompanying low pH, is highly corrosive to many metals, ceramics, and concrete. Other species of bacteria produce organic acids that are similarly corrosive.

The acids produced by these organisms can also contribute to corrosion by aiding the breakdown of coatings systems. Alternatively, other organisms that have no direct influence on corrosion may be involved in the breakdown of coatings. The breakdown products are then sometimes usable as food by the acid-producing bacteria, ultimately leading to accelerated corrosion of the underlying metal.

Anaerobic Sulfide Production. The most thoroughly documented case in which microbes are known to cause

corrosion is that of iron and steel under anaerobic conditions in the presence of sulfate-reducing bacteria. Based on electro-chemistry, deaerated soils of near-neutral pH are not expected to be corrosive to iron and steel. However, if the soil contains sulfate-reducing bacteria and a source of sulfates, rapid corrosion has been found to occur. The classical mechanism originally proposed for this corrosion involved the removal of atomic hydrogen from the metal surface by the bacteria using the enzyme hydrogenase (Ref 54). The removed hydrogen was then supposedly utilized by the bacteria in the reduction of sulfates to sulfides. The following set of equations was proposed to explain this mechanism:

4Fe → 4Fe2+ + 8eAnodic reaction 8H2O → 8H+ + 8OHDissociation of water 8H+ + 8e- → 8H Cathodic reaction

(Eq 9) (Eq 10) (Eq 11) (Eq 12)

Fe2+ + S2- → FeS Corrosion product 3Fe2+ + 6OH- → 3Fe(OH)2 Corrosion product

(Eq 13) (Eq 14)

Without sulfate-reducing bacteria, the mechanism would stop after Eq 11, when the surface became covered by a monolayer of hydrogen. According to the theory, this hydrogen is stripped off by the bacteria, a process known as cathodic depolarization; this allows corrosion to continue. It is now recognized that this original mechanism, although it undoubtedly plays an important role, does not represent the entire process (Fig. 6). It has been shown that the iron sulfide (FeS) film produced is protective if continuous but that it causes galvanic corrosion of the bare iron underneath if defective. Other corrosive substances, such as H2S, can also be produced. The sulfate-reducing bacteria have been identified as contributors to the corrosion of stainless, copper, and aluminum alloys, but the details of the mechanism are still being debated (Ref 52, 53).

Fig. 6 Schematic of the anaerobic corrosion of iron and steel showing the action of sulfate-reducing bacteria in removing hydrogen from the surface to form FeS and H2S

Additional information on the organisms involved in corrosion and the industries, environments, and alloy-electrolyte systems in which they have been active can be found in the articles "General Corrosion" and "Localized Corrosion" in this Volume. Information on detecting and characterizing biological corrosion in the laboratory and in the field can be found in the article "Evaluation of Microbiological Corrosion." Information on controlling biological corrosion can be found in the article "Control of Environmental Variables in Water Recirculating Systems."

References 1.

D.C. Silverman, Presence of Solid Fe(OH)2 in EMF-pH Diagram for Iron, Corrosion, Vol 38 (No. 2), 1982, p 453 2. R.A. Robinson and R.H. Stokes, Electrolyte Solutions, 2nd ed., Academic Press, 1959 3. R.G. Bates, Electrometric pH Determinations, John Wiley & Sons, 1954 4. P. Lorbeer and W.J. Lorenz, The Kinetics of Iron Dissolution and Passivation in Solutions Containing Oxygen, Electrochim. Acta, Vol 25 (No. 4), 1980, p 375 5. B.T. Ellison and W.R. Schmeal, Corrosion of Steel in Concentrated Sulfuric Acid, J. Electrochem. Soc., Vol 125 (No. 4), 1978, p 524. 6. T.K. Ross, G.C. Wood, ad I.J. Mahmud, The Anodic Behavior of Iron Carbon Alloys in Moving Acid Media, J. Electrochem. Soc., Vol 113, 1966, p 334 7. R.J. Chin and K. Nobe, Electrodissolution Kinetics of Iron in Chloride Solutions, J. Electrochem. Soc., Vol 119 (No. 11), 1972, p 1457 8. A.N. Katrevick, G.M. Florianovich, and Ya. M. Kolotyrkin, Clarification of the Kinetic Parameters of the Reaction of the Active Dissolution of Iron in Phosphate Solutions, Zashch. Met., Vol 10 (No. 4), 1974, p 369 9. B. Alexandre, A. Caprani, J.C. Charbonnier, M. Keddam, and P.H. Morel, The Influence of Chromium on the Mass Transfer Limitation of the Anodic Dissolution of Ferritic Steels Fe-Cr in Molar Sulfuric Acid, Corros. Sci., Vol 21 (No. 11), 1981, p 765 10. D.C. Silverman and M.E. Zerr, Application of Rotating Cylinder Electrode--E-Brite 26-1/Concentrated Sulfuric Acid, Corrosion, Vol 42 (No. 11) 1986, p 633

11. "The Corrosion Resistance of Nickel Containing Alloys in Sulfuric Acid and Related Compounds," CEB1, International Nickel Company, Inc., 1983 12. "Resistance of Nickel and High Nickel Alloys to Corrosion by Hydrochloric Acid, Hydrogen Chloride, and Chlorine," CEB-3, International Nickel Company, Inc., 1974 13. D.C. Silverman, "Derivation and Application of EMF-pH Diagrams," Paper 66, presented at Corrosion/85, Boston, MA, National Association of Corrosion Engineers, 1985 14. A. Alon, M. Frenkel, and M. Schorr, Corrosion and Electrochemical Behavior of Titanium in Acidic, Oxidizing, Aqueous, and Alcoholic Systems, in Proceedings of the Fourth International Congress on Metallic Corrosion, National Association of Corrosion Engineers, 1969, p 636 15. V. Vujicic and B. Lovrecek, A Study of the Influence of pH on the Corrosion Rate of Aluminum, Surf. Technol., Vol 25, 1985, p 49 16. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions, National Association of Corrosion Engineers, 1974 17. N.D. Tomashov, Theory of Corrosion and Protection of Metals, B.H. Tytell, I. Geld, and H.S. Preiser, Trans., Macmillan, 1966 18. M. Fischer, Possible Models of the Reaction Mechanism and Discussion of the Functions of Passivating Additions During the Chemical Passivation of Iron in Weakly Acidic, Neutral and Basic pH Ranges, Werkst. Korros., Vol 29, 1978, p 188 19. B.J. Andrzejaczek, Influence of Oxygen on the Kinetics of Corrosion Processes on Iron in Water, Korrosion, Vol 15 (No. 5), 1984, p 239 20. R.N. Sylva, The Hydrolysis of Iron (III), Rev. Pure Appl. Chem., Vol 22, 1972, p 115 21. W. Frank and L. Graf, Mechanism of Intergranular Corrosion and Intergranular Stress Corrosion Cracking in Mild Steels, Res. Mech., Vol 13, 1985, p 251 22. M.G. Fontana and N.D. Green, Corrosion Engineering, 2nd ed., McGraw-Hill, 1978 23. D.R. Knittel, Zirconium: A Corrosion Resistant Material for Industrial Applications, Chem. Eng., Vol 87 (No. 11), 1980, p 98 24. M. Schussler, Corrosion Data Survey on Tantalum, Fansteel, Inc., 1972 25. D.T. Merrill and J.A. Drago, "Evaluation of Ozone Treatment in Air Conditioning Cooling Towers," EPRI FP-1178, Electric Power Research Institute, Sept 1979 26. H.H. Uhlig, Corrosion and Corrosion Control, John Wiley & Sons, 1963 27. F.N. Speller, Corrosion Causes and Prevention, McGraw-Hill, 1951 28. Ya. M. Kolotyrkin, V.S. Pakhomov, A.G. Parshin, and A.V. Chekhovskii, Effect of Heat Transfer on Corrosion of Metals, Khim, Neft. Mashinostr., Vol 16 (No. 12), 1980, p 20 29. J.W. Oldfield, T.S. Lee, and R.M. Kain, Avoiding Crevice Corrosion of Stainless Steels, in Proceedings of Stainless Steels '84, Institute of Metals, 1985, p 205 30. R.J. Brigham, The Initiation of Crevice Corrosion on Stainless Steels, Mater. Perform., Vol 24 (No. 12), 1985, p 44 31. A.G. Parshin, V.S. Parkhomov, and Ya. M. Kolotyrkin, The Influence of Heat Transfer on the Kinetics of Cathode Processes With Limiting Diffusion Steps, Corros. Sci., Vol 22 (No. 9), 1982, p 845 32. T.K. Ross, Corrosion and Heat Transfer--A Review, Br. Corros. J., Vol 2, 1967, p 131 33. U. Lotz and E. Heitz, Flow Dependent Corrosion, Werkst, Korros., Vol 34, 1983, p 451 34. B.K. Mahato, C.Y. Cha, and L.W. Shemilt, Unsteady State Mass Transfer Coefficients Controlling Steel Pipe Corrosion Under Isothermal Flow Conditions, Corros. Sci., Vol 20, 1980, p 421 35. H.R. Copson, Effects of Velocity on Corrosion, Corrosion, Vol 16 (No. 2), 1960, p 86t 36. T.R. Beck and S.G. Chan, Experimental Observations and Analysis of Hydrodynamic Effects on Growth of Small Pits, Corrosion, Vol 37 (No. 11), 1981, p 665 37. P.A. Lush, S.P. Hutton, J.C. Rowlands, and B. Angell, The Relation Between Impingement Corrosion and Fluid Turbulence Intensity, in Proceedings of the Sixth European Congress on Metallic Corrosion, 1977, p

38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

50. 51. 52. 53. 54.

137 D.C. Silverman, Rotating Cylinder Electrode for Velocity Sensitivity Testing, Corrosion, Vol 40 (No. 5), 1984, p 220 B.C. Syrett, Erosion-Corrosion of Copper-Nickel Alloys in Sea Water and Other Environments--A Literature Review, Corrosion, Vol 32 (No. 6), 1976, p 242 B. Vyas, Erosion-Corrosion, in Treatise on Materials Science and Technology, Vol 16, Academic Press, 1979, p 357 S.S. Abd. El Rehim, M. Sh. Shalaby, S.M. Abd. El Halum, Effect of Some Anions on the Anodic Dissolution of Delta-S2 Steel in Sulfuric Acid, Surf. Technol., Vol 24, 1985, p 241 E. McCafferty, Electrochemical Behavior of Iron Within Crevices in Nearly Neutral Chloride Solutions, J. Electrochem. Soc., Vol 121 (No. 9), 1974, p 1007 H. Boehni, Pitting and Crevice Corrosion, in Corrosion in Power Generating Equipment, Proceedings of the Eighth International Brown Boveri Symposium, 1984, p 29 M. Cohen, Dissolution of Iron, in Corrosion Chemistry, G.R. Brubaker and P.B.P. Phipps, Ed., ACS Symposium Series, 89, American Chemical Society, 1979, p 126 R. Hausler, Corrosion Inhibition and Inhibitors, in Corrosion Chemistry, G.R. Brubaker and P.B.P. Phipps, Ed., ACS Symposium Series, 89, American Chemical Society, 1979, p 262 J.D. Costlow and R.C. Tipper, Ed., Marine Biodeterioration: An Interdisciplinary Study, Proceedings of the Symposium, Naval Institute Press, 1984 D.C. Marshall, Interfaces in Microbial Ecology, Harvard University Press, 1976 D.C. Savage and M. Fletcher, Ed., Bacterial Adhesion, Plenum Press, 1985 D.H. Pope, D. Duquette, P.C. Wayner, and A.H. Johannes, Microbiologically Influenced Corrosion: A State-of-the-Art-Review, Publication 13, Materials Technology Institute of the Chemical Process Industries, Inc., 1984 D.H. Pope, "A Study of Microbiologically Influenced Corrosion in Nuclear Power Plants and a Practical Guide for Countermeasures," EPRI NP-4582, Final Report, Electric Power Research Institute, 1986 J.D.A. Miller, Ed., Microbial Aspects of Metallurgy, Elsevier, 1970 Microbial Corrosion, Proceedings of the Conference, National Physical Laboratory, The Metals Society, 1983 S.C. Dexter, Ed., Biologically Induced Corrosion, Proceedings of the Conference, National Association of Corrosion Engineers, 1986 C.A.H. Von Wolzogen Kuhr and L.S. Van der Vlugt, Water, Den Haag, Vol 18, 1934, p 147-165

Effects of Metallurgical Variables on Aqueous Corrosion D.W. Shoesmith, Fuel Waste Technology Branch, Atomic Energy of Canada Ltd.

Introduction THE STRUCTURE AND COMPOSITION of both metals and alloys are important in deciding their corrosion characteristics. Indeed, structure and composition are critical in many forms of localized corrosion. For a metal or alloy to corrode evenly, the anodic and cathodic sites must be interchangeable. This implies that every site on the surface is energetically equivalent and therefore equally susceptible to dissolution, but this is never the case. This article will provide an introduction to the effects of crystal structure, alloying, heat treatments, and the resulting microstructures on corrosion properties. Detailed information on these metallurgical variables for a wide variety of ferrous and nonferrous metals and alloys can be found in the Section "Specific Alloy Systems" in this Volume. Reference should also be made to Volumes 9 and 10 of ASM Handbook, formerly 9th Edition of Metals Handbook for supplementary data on crystallographic/microstructural analysis and interpretation.

Metals and Metal Surfaces Metals form as a series of irregular crystals. If these crystals or grains were perfect, the metal atoms would lie in regular close-packed planes. If this were true, the rate of metal dissolution would depend on which crystallographic planes were exposed to the corrosive environment. In addition to these perfect features, there are many sources of atomic disarray within the crystals that can lead to defects where they emerge at the surface. Some of the more significant crystal defects are described below. Additional information on defects can be found in the article "Crystal Structure of Metals" in Metallography and Microstructures, Volume 9 of ASM Handbook, formerly 9th Edition Metals Handbook. Stacking Faults. The atoms in metals form close-packed layers that stack in various sequences. The most common

crystal structures found in metals are the body-centered cubic (bcc), the face-centered cubic (fcc), and the hexagonal close-packed (hcp). The unit cells for these structures are shown schematically in Fig. 1 (a unit cell is a parallelepiped whose edges form the axes of a crystal; it is the smallest pattern of atomic arrangement). If the crystal structures are mixed, resulting in an error in the normal sequence of stacking of atomic layers, stacking faults are produced. These faults can extend for substantial distances through, and across, the crystal.

Fig. 1 Schematic of the unit cells for the most common crystal structures found in metals and alloys. (a) bcc. (b) fcc. (c) hcp

A slip plane is the lattice plane separating two regions of a crystal that have slipped relative to each other. Such

permanent displacements occur under the influence of plastic deformation, as described in the article "Plastic Deformation Structures" in Metallography and Microstructure, Volume 9 of ASM Handbook, formerly 9th Edition Metals Handbook. Dislocations. Slip regions can be caused by the movement of various small lattice dislocations, such as an additional

layer of atoms or a stacking fault on one side of the defect. Dislocations are defects that exist in nearly all real crystals. An edge dislocation, which is the edge of an incomplete plane of atoms within a crystal, is shown in cross section in Fig. 2. In this illustration, the incomplete plane extends partway through the crystal from the top down, and the edge dislocation (indicated by the standard symbol ) is its lower edge.

Fig. 2 Schematic of a section through on edge dislocation, which is perpendicular to the plane of the illustration and is indicated by the symbol

If forces are applied (arrows, Fig. 3) to a crystal, such as the perfect crystal shown in Fig. 3(a), one part of the crystal will slip. The edge of the slipped region, shown as a dashed line in Fig. 3(b), is a dislocation. The portion of this line at the left near the front of the crystal and perpendicular to the arrows (Fig. 3b) is an edge dislocation, because the displacement involved is perpendicular to the dislocation. The slip deformation in Fig. 3(b) has also formed another type of dislocation. The part of the slipped region near the right side, where the displacement is parallel to the dislocation, is termed a screw dislocation. In this part, the crystal no longer consists of parallel planes of atoms, but of a single plane in the form of a helical ramp (screw). As the slipped region spread across the slip plane, the edge-type portion of the dislocation moved out of the crystal, leaving the screw-type portion still embedded (Fig. 3c). When all of the dislocation finally emerged from the crystal, the crystal was again perfect but with the upper part displaced one unit from the lower part (Fig. 3d). Therefore, Fig. 3 illustrate the mechanism of plastic flow by the slip process, which is actually produced by dislocation movement. Point defects may be vacancies caused by the absence of one or more atoms in the crystal, impurity atoms of different

sizes, and interstitial atoms (small atoms in spaces between the lattice atoms). Points defects can affect significant volumes of the crystal. Grain Boundaries. The interface between grains is termed the grain boundary, and it is a region of major atomic

disarray at which many faults and dislocations congregate. This disarray makes it energetically easier for impurities to concentrate at grain boundaries as opposed to the grain interior, where the atomic arrangement is more regular.

The areas at which these defects emerge on the surface constitute sites of high energy. These energetic sites possess increased chemical activity because each contains atoms with an incomplete number of nearest neighbors (Fig. 4). Atom A, lying within a relatively perfectly closed-packed plane, is strongly coordinated on all but one side. Therefore, it has a lower chemical free energy compared to atom B in a step. When compared to atom C at a kink site, atom A has an even lower chemical free energy. Dissolution of kink sites, concentrated at dislocations or grain boundaries, will obviously be accompanied by a greater release of energy than dissolution of atoms from the planes.

Fig. 4 Schematic of a dislocation emerging at a surface. A, plane atom; B, step atom; C, kink atom

Anodic metal dissolution sites are more likely to be found at dislocations, and grain boundaries. In electrochemical terms:

(Ee)kink < (Ee)step < (Ee)plane

(Eq 1)

and because these dislocations tend to concentrate at grain boundaries:

(Ee)grain boun < (Ee)grain

(Eq 2)

where Ee is the equilibrium corrosion potential, as described in the article "Kinetics of Aqueous Corrosion" in this Volume. Therefore, the thermodynamic driving force. Etherm, for dissolution (the difference in cathodic and anodic equilibrium potentials):

∆Etherm = (Ee)c -- (Ee)a

(Eq 3)

will be greater for grain-boundary sites than for the grains themselves. This does not mean that grain boundaries will always corrode preferentially, because the initial etching of the grain boundary will produce an increased surface area in this region. The corresponding additional interfacial energy will decrease the total free energy of the site. The chemical or mechanical activation of the grain boundaries determines whether they will be more active than the grains.

Alloys and Their Surfaces Pure metals have a low mechanical strength and are rarely used in engineering applications. Stronger metallic materials, which are combinations of several elemental metals known as alloys, are most often used. Commonly used alloys have a good combination of mechanical, physical, fabrication, and corrosion qualities. The specific applications determines which of these qualities is deemed most important for alloy selection. Alloys can be single phase or polyphase, depending on the elements present and their mutual solubilities. For example, the addition of nickel to copper does not alter the fcc structure. The nickel occupies a lattice position within the copper host lattice, and the two metals are said to form a substitutional solid solution. By contrast, the alloying element can occupy an interstitial site in the host lattice and is said to form an interstitial solid solution. An example of such an alloy would be carbon steel, in which the small carbon atom is interstitially accommodated in the iron lattice.

It is often impossible to dissolve a large amount of one element in another. When this is attempted, two or more phases may form. The predominant phase is known as the primary phase, or matrix. The other, smaller phase is known as the secondary phase, or precipitate. Precipitates often contain the nonmetallic elements present in the alloy. If they are insoluble in the matrix, they concentrate, like impurities, as dislocations. This means they are often found at grain boundaries.

Phase Diagrams Graphs of phase stability as a function of temperature and composition are called phase diagrams. They are based on the equilibrium conditions in the alloy. The stable phases at each temperature and composition are shown on the diagram. If two metallic components are involved, the graph is termed a binary phase diagram. Ternary and quaternary diagrams are necessary for more complex, systems. Figure 5 shows a portion of the binary phase diagram for the Fe-C system. Even for this system, the diagram is complex. As discussed below, this complexity is the key to the wide range of steel properties available.

Fig. 5 Portion of the binary phase diagram for the Fe-C system

Iron and Steels As an example of the diversity of structures possible for a given material and its alloys, the structures and phases possible for iron, carbon steels, and stainless steels will be discussed. The primary purpose in this discussion will be to emphasize the principles, because these systems are covered in detail in the Section "Specific Alloy Systems" in this volume (especially the articles "Corrosion of Cast Irons," "Corrosion of Carbon Steels," "Corrosion of Alloy Steels," and "Corrosion of Stainless Steels"). Iron. Depending on the temperature of formation, iron can exist in three different phase modifications, or allotropes:

ferrite (α-iron), which is bcc; austenite (γ-iron), which is fcc; and δ-ferrite, which is bcc but of slightly different cell dimensions than normal ferrite. These allotropes are shown in Fig. 6.

Fig. 6 Phases of iron, carbon steel, and stainless steel

Cast Irons and Carbon Steels. The most important alloying element of steel is carbon, whose solubility is different

in the various phase modifications of iron. The small carbon atoms occupy interstitial positions between the iron atoms. Consequently, the less densely packed fcc lattice of austenite (γ-iron) can accommodate the carbon atom more readily

than the bcc ferrite γ-iron). Therefore, austenite formation is promoted by alloying with carbon. Carbon is said to be an austenite-stabilizing element. Carbon steels contain less than 2% C; cast irons contain more than 2% C. For steels, any composition can be heated until a homogeneous solid solution of austenite is obtained. This is apparent from the phase diagram shown in Fig. 5. Upon cooling, the four phases (ferrite, austenite, cementite, and martensite) can be formed. The relative proportions of these phases are determined by the carbon content, the rate of cooling, and any subsequent heat treatment. Cementite is an iron carbide containing 6.67% C (by weight) with the composition Fe 3C. It forms as a mixture with ferrite when cooling slowly from the austenite region of the phase diagram (Fig. 5). The mixture, know as pearlite, forms separate grains, along with ferrite grains, in plain carbon steels. It possesses a lamellar structure with alternate bands of ferrite and cementite. Bainite, an austenite transformation product, is a lathlike aggregate of ferrite and cementite that forms under conditions intermediate to those that result in the formation of pearlite and martensite. The way austenite is cooled determines the rate of segregation and the grain size of the ferrite and cementite phases. This provides the opportunity to produce a range of carbon steels, each having different mechanical properties. Very rapid cooling, or quenching, produces martensite. Under these conditions, the normal phase separation to produce ferrite and pearlite does not occur. A metastable forced solution of carbon in ferrite is obtained. The forcing of a martensite structure is known as hardening. Additional information on the microstructural constituents of carbon steels can be found in the article "Carbon and Alloy Steels" in Metallography and Microstructures, Volume 9 of ASM Handbook, formerly 9th Edition Metals Handbook. For a carbon content greater than 2%, the phase diagram in Fig. 5 shows that heating will not bring the mixture into the single-phase austenite region. In other words, a homogeneous solid solution cannot be achieved. Cast irons in this composition region are formed by casting from the molten state. They are used where hardness and corrosion resistance are required and where brittleness due to the cementite content poses no problem. Stainless Steels. The three major phases in carbon steels--ferrite, austenite and martensite--are also formed in stainless

steels. In addition, two other stainless steel categories, ferritic-austenitic (duplex) and precipitate-hardened, can be produced by specific heat treatments (Fig. 6). The microstructural characteristics of iron-chromium and iron-chromiumnickel alloys are discussed in the articles "Wrought Stainless Steels" and "Stainless Steel Casting Alloys" in Metallography and Microstructures, Volume 9 of ASM Handbook, formerly 9th Edition Metals Handbook. The stability and the mechanical and physical properties of the various phases depend on the combination of alloying elements present. Alloying elements can be divided into two categories: • •

Austenite stabilizers: carbon, nitrogen, nickel, and manganese Ferrite stabilizers: silicon, chromium, molybdenum, niobium, and titanium

The selection of a stainless steel for a particular engineering application depends on which mechanical or physical property is considered to be most important.

Effect of Alloying on Corrosion Resistance

One of the primary reasons for producing alloyed, or stainless, steels is to improve corrosion resistance. Alloying can affect corrosion resistance in many different ways. Increased Nobility. Alloying can have a genuinely thermodynamic effect on corrosion resistance by increasing the

nobility of the material. This is achieved by a decrease in ∆Etherm, which is expressed by Eq 3 and illustrated in the Evans diagram shown in Fig. 7. Thus: [(Ee)c -- {(Ee)a}M] > [(Ee)c -- {(Ee)a}A]

(Eq 4)

or: ( Etherm)M > ( Etherm)A

(Eq 5)

and: (icorr)M > (icorr)A

(Eq 6)

where icorr is the corrosion current, and the subscripts M and A indicate metal and alloy, respectively. The decrease in corrosion current is caused by an increase in the equilibrium potential for the anodic reaction. The equilibrium potential for the cathodic reaction may also change from metal to alloy, but this change is insignificant compared to the effect on the anodic reaction.

Fig. 7 Evans diagram for metal M in the pure form and incorporated into a more noble alloy

Noble systems of this intermetallic type are rare and generally produced only when the alloying element is a noble metal, such as gold, platinum, or palladium. For such alloys, the less noble constituent metal, for example, titanium in a Ti-2Pd alloy, often dissolves preferentially, leaving a protective film of noble metal on the alloying surface. Disruption of this surface film, which is often thin, will reinitiate corrosion. Formation of a Protective Film. The addition of controlled amounts of selected alloying elements can often improve

the stability and protectiveness of surface oxide films formed on the material surface. Thus, the addition of chromium to iron has a major effect on the corrosion resistance in acid. This can be appreciated by studying the anodic polarization curves for iron-chromium alloys shown in Fig. 8. The curves exhibit an active-passive transition, as discussed in the article "Kinetics of Aqueous Corrosion" in this Volume. The addition of chromium leads to decreases in the critical current for passivation (icrit), the passivation potential (Epass), and the passive current (ipass), as indicated by the arrows in Fig. 8. Therefore, for the cathodic reaction shown, passivation is achievable only in case 3 for a steel containing more than approximately 12% Cr. The improved resistance to corrosion is due to an increase in chromium content of the ironchromium oxide layer formed on the alloy surface.

Fig. 8 Schematic anodic polarization diagrams for stainless steels containing various amounts of chromium. (1) 3% Cr; (2) 10% Cr; (3) 14% Cr. The polarization curve for the cathodic reaction O + ne- → R is also shown. Arrows indicate the effect of chromium addition on icrit, Epass, ipass, and itrans.

The parameters icrit, ipass, and Epass can be decreased even further by the addition of up to 8% Ni. This is the reason for the extensive use of 18-8 stainless steel (Fe-18Cr-8Ni). Problems of Alloying. Alloying is not without its problems, one of which is illustrated in Fig. 8. As the chromium content is increased, the current due to transpassive dissolution, itrans, at positive potentials also increases. Thus, under very aggressive oxidizing conditions, steels with high chromium contents become less resistant to corrosion because of

oxidative dissolution of chromium (as Cr6+, that is,

) from the oxide (containing Cr3+).

Other problems can also be introduced if the alloy is carelessly heat treated in the range 420 to 700 °C (790 to 1290 °F). A high carbon content in chromium steels can lead to the formation of chromium carbide. The carbides separate to the grain boundaries, leaving the steel deficient in chromium close to the grain boundary. This region is consequently less noble and is preferentially attacked. The steel is said to be sensitized to intergranular (intercrystalline) attack. One method of counteracting this process is to make further alloying additions of niobium or titanium, either of which will stabilize the carbon and prevent chromium carbide separation. Impurities remaining after the fabrication process can also have a major effect on corrosion resistance. For example, sulfide inclusions in the form of conductive metal sulfides can act as local cathodes in steel and can promote corrosion. Sulfides catalyze the cathodic reactions, as shown by the Evans diagram in Fig. 9. The rate of the proton reduction reaction (determined by the exchange current, io, and the Tafel coefficient, bc, as described in the article "Kinetics of Aqueous Corrosion" in this Volume) is increased at the sulfide inclusions, leading to a higher corrosion potential and increased corrosion current. The corrosion tends to occur locally, and pitting is observed near the inclusion.

Fig. 9 Evans diagram for steel corrosion showing the increased corrosion rate when the cathodic reaction is catalyzed by a sulfide inclusion

The presence of precipitates with minor alloying elements and impurities can lead to problems, because phases with widely different electrochemical properties are then present. This can result in local variations in corrosion resistance. Also, the addition of alloying elements to improve the resistance to general, or uniform, corrosion may cause increased susceptibility to localized corrosion processes, such as pitting or intergranular corrosion.

Effect of Heat Treatment Many of the mechanical properties of materials are improved by various heat treatments. Unfortunately, such properties as hardness and strength are often achieved at the expense of corrosion resistance. For example, the hardness and strength of martensitic steels are counterbalanced by a lower corrosion resistance than for the ferritic and austenitic steels. The very high strengths achieved for precipitation-hardened steels are due to the secondary precipitates formed during the solution heat treating and aging process. As discussed above, precipitates with electrochemical properties distinctly different from those of the matrix have a deleterious effect on corrosion. Processes such as cold working, in which the material is plastically deformed into some desired shape, lead to the formation of elongated and highly deformed grains and a decrease in corrosion resistance. Cold working can also introduce residual stresses that make the material susceptible to stress-corrosion cracking. An improvement in corrosion resistance can be achieved by subsequently annealing at a temperature at which grain recrystallization can occur. A partial anneal leads to stress relief without a major effect on the overall strength of the material. From the corrosion viewpoint, welding is a particularly troublesome treatment. Because welding involves the local heating of a material, it can lead to phase transformations and the formation of secondary precipitates. It can also induce stress in and around the weld. Such changes can lead to significant local differences in electrochemical properties as well as the onset of such processes as intergranular corrosion. Therefore, the weld filler metal should be as close in electrochemical properties to the base metal as technically feasible, and the weld should be subsequently stress relieved. Detailed information on the corrosion problems associated with welded joints can be found in the article "Corrosion of Weldments" in this Volume.

References 1. R.M. Brick, R.B. Gordon, and A. Phillips, Structure and Properties of Alloys, McGraw-Hill, 1965

2. L.S. Van Delinder, Ed., Corrosion Basics--An Introduction, National Association of Corrosion Engineers, 1984 3. L.L. Shrier, Corrosion, George Newnes Ltd., 1963 4. G. Wranglen, An Introduction to Corrosion and Protection of Metals, Institut für Metal-lskydd, 1972

Fundamentals of High-Temperature Corrosion in Molten Salts John W. Koger, Martin Marietta Energy Systems, Inc.

Introduction MOLTEN SALTS, often called fused salts, can cause corrosion by the solution of constituents of the container material, selective attack, pitting, through electrochemical reactions, by mass transport due to thermal gradients, by reaction of constituents of the molten salt with the container material, by reaction of impurities in the molten salt with the container material, and by reaction of impurities in the molten salt with the alloy. Many hundreds of molten salt-metal corrosion studies have been documented, and predictions of corrosion are difficult if not impossible in engineering systems. The most prevalent molten salts in use are nitrates and halides. Other molten salts that have been extensively studied but are not widely used include carbonates, sulfates, hydroxides, and oxides. A somewhat general discussion of molten salt corrosion will be presented in this article, with emphasis on nitrates/nitrites and fluorides. Specific examples of results from experiments will be presented for some actual systems; these examples will indicate the scope of a program needed for a particular application.

Thermodynamics and Kinetics of Molten Salt Corrosion The chemistry of molten salts can be as complicated as one wishes to make it, based on the definition of a molten salt and whether or not the media may be wholly ionic. For simplicity, most of the processes considered in this article involve electrode processes. According to Inman and Lovering (Ref 1), except in rare cases in which hydrogen is a part of the molten salt or the melts are exposed to hydrogen atmosphere, the hydrogen ion plays a very small role. The oxygen ions are generally quite important in matters of corrosion. The function pO2- (equivalent to pH in aqueous environments) defines the oxide ion activity. The higher the value of pO2-, the more corrosion of metal will occur. Also, the concentration of oxide ions can influence the corrosive effects of certain nonoxygen-containing melts that have been subject to hydrolysis through contact with atmospheric moisture. In molten salt systems, corrosion is rarely inhibited because of the reactivity of the molten salts and the high temperatures. Molten salts often act as fluxes, thus removing oxide layers on container materials that generally might prove to be protective. Molten salts are generally good solvents for precipitates; therefore, passivation, because of deposits, generally does not occur. One of the most familiar mechanisms of corrosion arises from ions of metals more noble than the container material, that is, the metal being corroded. In some cases, the more noble metal can be a constituent part of the molten salt, and in others, it can occur as an impurity in the system. Another mechanism is best described by the example of silver in molten sodium chloride (NaCl). Thermodynamics would not predict a corrosion problem. However, the reaction occurs because sodium, as a result of the formation of silver chloride (AgCl), can dissolve in molten NaCl and distill out of the system. Thus, the reaction proceeds.

If a molten salt contains oxyanion constituents that can be reduced, oxide ions are released. Corrosion will occur on a metal in contact with the salt. Lastly, oxygen itself can be reduced to oxide ions. However, uncombined oxygen is rarely found in molten salts because of limited solubility. The potential, E, versus pO2- diagram is often used as the equivalent of the E versus pH (Pourbaix) diagrams for aqueous corrosion (Ref 2). Both of these diagrams are used to establish the stability characteristics of a metal in the respective media. A typical E versus pO2- diagram for iron in a molten salt at an elevated temperature is shown in Fig. 1. Areas of corrosion, immunity, passivation, and passivity breakdown are evident. Additional information on Pourbaix diagrams used in aqueous corrosion studies is available in the article "Thermodynamics of Aqueous Corrosion" in this Volume.

Fig. 1 Typical E versus pO2- diagram for iron in a molten salt at an elevated temperature

Actually, the E versus pO2- diagram is probably more useful than the Pourbaix diagram because of the absence of kinetic limitations at elevated temperatures. The following problems, however, do exist: • • • • •

Molten salt electrode reactions and the concomitant thermodynamic data are not readily available Products from the reactions are often lost by vaporization Diagrams based on pure component thermodynamic data are unrealistic because of departure from ideality Lack of passivity even where predictions would show passive behavior The stable existence of oxides other than the O2- species

Test Methods A number of kinetic and thermodynamic studies have been carried out in capsule-type containers. These studies can determine the nature of the corroding species and the corrosion products under static isothermal conditions and do provide some much-needed information. However, to provide the information needed for an actual flowing system, corrosion

studies must be conducted in thermal convection loops or forced convection loops, which will include the effects of thermal gradients, flow, chemistry changes, and surface area effects. These loops can also include electrochemical probes and gas monitors. An example of the types of information gained from thermal convection loops during an intensive study of the corrosion of various alloys by molten salts will be given below. A thermal convection loop is shown in Fig. 2.

Fig. 2 Natural circulation loop and salt sampler

Purification Molten salts, whether used for experimental purposes or in actual systems, must be kept free of contaminants. This task, which includes initial makeup, transfer, and operation, is specific for each type of molten salt. For example, even though the constituents of the molten fluoride salts used in the Oak Ridge Molten Salt Reactor Experiment were available in very pure grades, purification by a hydrogen/hydrogen fluoride (H2/HF) gas purge for 20 h was necessary (Ref 3). For nitrates with a melting point of approximately 220 °C (430 °F), purging with argon flowing above and through the salt at 250 to 300 °C (480 to 570 °F) removes significant amounts of water vapor (Ref 4). Another purification method used for this same type of salt consisted of bubbling pure dry oxygen gas through the 350- °C (660- °F) melt for 2 h and then bubbling pure dry nitrogen for 30 min to remove the oxygen (Ref 5). All metals that contact the molten salt during purification must be carefully selected to avoid contamination from transfer tubes, thermocouple wells, the makeup vessel, and the container itself. This selection process may be an experiment in itself.

Nitrates/Nitrites Nitrate mixtures have probably been studied and used more than any other molten salt group. This is perhaps because of the low operating temperatures possible (200 to 400 °C, or 390 to 750 °F). Steels of varying types are generally chosen for these systems. As shown in Fig. 1, the E versus pO2- diagram for iron indicates regions of corrosion, immunity, passivity, and passivity breakdown at temperatures of 240 to 400 °C (465 to 750 °F) (Ref 6). In general, the basicity of the melt prevents iron corrosion. Protection by passive films is less reliable, because oxide ion discharge may breakdown the passive film. Electropolished iron spontaneously passivates in molten sodium nitrate-potassium nitrite in the temperature range of 230 to 310 °C (445 to 590 °F) at certain potentials (Ref 7). A magnetite (Fe3O4) film is formed, along with a reduction of nitrite or any trace of oxygen gas dissolved in the melt. At higher potentials, all reactions occur on the passivated iron. Above the passivation potentials, dissolution occurs with ferric ion soluble in the melt. At even higher potentials, nitrogen oxides are evolved, and nitrate ions dissolve in the nitrite melt. At higher currents, hematite (Fe2O3) is formed as a suspension, and NO2 is detected. Carbon steel in molten sodium nitrate-potassium nitrate (NaNO3-KNO3) at temperatures ranging from 250 to 450 °C (480 to 840 °F) forms a passivating film consisting mainly of Fe3O4 (Ref 5). Iron anodes in molten alkali nitrates and nitrites at temperatures ranging from 240 to 320 °C (465 to 610 °F) acquire a passive state in both melts. In nitrate melts, the protective Fe3O4 oxidizes to Fe2O3, and the gaseous products differ for each melt (Ref 8). An interesting study was conducted on the corrosion characteristics of several eutectic molten salt mixtures on such materials as carbon steel, stainless steel, and Inconel in the temperature range of 250 to 400 °C (480 to 750 °F) in a nonflowing system (Ref 9). The salt mixtures and corrosion rates are given in Table 1. As expected, the corrosion rate was much higher for carbon steel than for stainless steel in the same mixture. Low corrosion rates were found for both steels in mixtures containing large amounts of alkaline nitrate. The nitrate ions had a passivating effect.

Table 1 Corrosion rates of iron-base alloys in eutectic molten salt mixtures Corrosion rate

Salt mixture

Carbon steel

Stainless steel

μm/yr

mils/yr

μm/yr

mils/yr

NaNO3-NaCl-Na2SO4 (86.3,8.4,5.3 mol%, respectively)

15

0.6

1

0.03

KNO3-KCl (94.6 mol%, respectively)

23

0.9

7.5

0.3

LiCl-KCl (58.42 mol%, respectively)

63

2.5

20

0.8

Electrochemical studies showed high resistance to corrosion by Inconel. Again, the sulfate-containing mixture caused less corrosion because of passivating property of the nitrate as well as the preferential adsorption of sulfate ions. Surface analysis by Auger electron spectroscopy indicated varying thicknesses of iron oxide layers and nickel and chromium layers. The Auger analysis showed that an annealed and air-cooled stainless steel specimen exposed to molten lithium chloride (LiCl)-potassium chloride (KCl) salt had corrosion to a depth five times greater than that of an unannealed stainless steel specimen. Chromium carbide precipitation developed during slow cooling and was responsible for the increased corrosion. The mechanism of corrosion of iron and steel by these molten eutectic salts can be described by the following reactions:

Fe

Fe2+ + 2e-

LiCl + H2O

(Eq 1)

LiOH + HCl

(Eq 2)

H2 H2O + 2eH+ + eO2- + H2 O2 + 2eFe3+ + e-

O2-

(Eq 4)

Fe2+

(Eq 5)

Fe2+ + O2-

FeO

3FeO + O2-

Fe3O4 + 2e-

2Fe3O4 + O2-

(Eq 3)

3Fe2O3 + 2e-

(Eq 6) (Eq 7) (Eq 8)

In an actual flowing operating system of KNO3-NaO2-NaO3 (53, 40, and 7 mol%, respectively) at temperatures to 450 °C (840 °F), carbon or chromium-molybdenum steels have been used (Ref 10). For higher temperatures and longer times, nickel or austenitic stainless steels are used. Weld joints are still a problem in both cases. Alloy 800 and types 304, 304L, and 316 stainless steels were exposed to thermally convective NaNO3-KNO3 salt (draw salt) under argon at 375 to 600 °C (705 to 1110 °F) for more than 4500 h (Ref 4). The exposure resulted in the growth of thin oxide films on all alloys and the dissolution of chromium by the salt. The weight change data for the alloys indicated that the metal in the oxide film constituted most of the metal loss, that the corrosion rate, in general, increased with temperature, and that, although the greatest metal loss corresponded to a penetration rate of 25 μm/yr (1 mil/yr), the rate was less than 13 μm/yr (0.5 mil/yr) in most cases. These latter rates are somewhat smaller than those reported for similar loops operated with the salt exposed to the atmosphere (Ref 11, 12), but are within a factor of two to five. Spalling had a significant effect on metal loss at intermediate temperatures in the type 304L stainless steel loop. Metallographic

examinations showed no evidence of intergranular attack or of significant cold-leg deposits. Weight change data further confirmed the absence of thermal gradient mass transport processes in these draw salt systems. Raising the maximum temperature of the type 316 stainless steel loop from 595 to 620 °C (1105 to 1150 °F) dramatically increased the corrosion rate (Ref 11, 12). Thus, 600 °C (1110 °F) may be the limiting temperature for use of such alloys in draw salt.

Fluorides Because of the Oak Ridge Molten Salt Reactor Experiment, a large amount of work was done on corrosion by molten fluoride salts (Ref 3). Because these molten salts were to be used as heat transfer media, temperature gradient mass transfer was very important. Very small amounts of corrosion can result in large deposits, given that the solubility of the corrosion product changes drastically in the temperature range in question. Many other variables can also cause this phenomenon. Thus, a corrosion rate in itself does not provide complete information about corrosion. Because the products of oxidation of metals by fluoride melts are quite soluble in the corroding media, passivation is precluded, and the corrosion rate depends on other factors, including the thermodynamic driving force of the corrosion reactions. The design of a practicable system utilizing molten fluoride salts, therefore, demands the selection of salt constituents, such as lithium fluoride (LiF), beryllium fluoride (BeF2), uranium tetrafluoride (UF4), and thorium fluoride (ThF4), that are not appreciably reduced by available structural metals and alloys whose components (iron, nickel, and chromium) can be in near thermodynamic equilibrium with the salt. A continuing program of experimentation over many years has been devoted to defining the thermodynamic properties of many corrosion product species in molten LiF-BeF2 solutions. Many of the data have been obtained by direct measurement of equilibrium pressures for such reactions as:

H2(g) + FeF2(d)

Fe(c) + 2HF(g)

(Eq 9)

BeF2(l) + H2O(g)

(Eq 10)

and

2HF(g) + BeO(c)

where g, c, and d represent gas, crystalline solid, and solute, respectively, using the molten fluoride (denoted 1 for liquid) as the reaction medium. All of these studies have been reviewed, and the combination of these data with those of other studies has yielded tabulated thermodynamic data for many species in molten LiF-BeF2 (Table 2). From these data, one can assess the extent to which a uranium trifluoride (UF3) bearing melt will disproportionate according to the reaction:

4UF3(d)

3UF4(d) + U(d) f

(Eq 11)

Table 2 Standard Gibbs free energies (∆G ) of formation for species in molten 2LiF-BeF2 Temperature range: 733-1000 K Material(a)

-∆Gf (kcal/mol)

-∆Gf (1000 K) (kcal/mol)

Lif(l)

141.8-16.6 × 10-3 T K

125.2

BeF2(l)

243.9-30.0 × 10-3 T K

106.9

UF3(d)

338.0-40.3 × 10-3 T K

99.3

UF4(d)

445.9-57.9 × 10-3 T K

97.0

ThF4(d)

491.2-62.4 × 10-3 T K

107.2

ZrF4(d)

453.0-65.1 × 10-3 T K

97.0

NiF2(d)

146.9-36.3 × 10-3 T K

55.3

FeF2(d)

154.7-21.8 × 10-3 T K

66.5

CrF2(d)

171.8-21.4 × 10-3 T K

75.2

MoF6(g)

370.9-69.6 × 10-3 T K

50.2

Source: Ref 13 (a) The standard state for LiF and BeF2 is the molten 2LiF-BeF2 liquid. That for MoF6(g) is the gas at 1 atm. That for all species with d is that hypothetical solution with the solute at unit mole fraction and with the activity coefficient it would have at infinite dilution.

For the case in which the total uranium content of the salt is 0.9 mol%, as in the Oak Ridge Molten Salt Reactor Experiment, the activity of metallic uranium (referred to the pure metal) is near 10-15 with 1% of the UF4 converted to UF3 and is near 2 × 10-10 with 20% of the UF4 so converted (Ref 14). Operation of the reactor with a small fraction (usually 2%) of the uranium present as UF3 is advantageous insofar as corrosion and the consequences of fission are concerned. Such operation with some UF3 present should result in the formation of an extremely dilute (and experimentally undetectable) alloy of uranium with the surface of the container metal. Operation with 50% of the uranium as UF3 would lead to much more concentrated (and highly deleterious) alloying and to formation of uranium carbides. All evidence to date demonstrates that operation with relatively little UF3 is completely satisfactory. The data gathered to date reveal clearly that in reactions with structural metals, M:

2UF4(d) + M(c)

2UF3(d) + MF2(d)

(Eq 12)

chromium is much more readily attacked than iron, nickel, or molybdenum (Ref 14, 15). Nickel-base alloys, more specifically Hastelloy N (Ni-6.5 Mo-6.9Cr-4.5Fe) and its modifications, are considered the most promising for use in molten salts and have received the most attention. Stainless steels, having more chromium than Hastelloy N, are more susceptible to corrosion by fluoride melts, but can be considered for some applications. Oxidation and selective attack may also result from impurities in the melt:

M + NiF2

MF2 + Ni

(Eq 13)

M + 2HF

MF2 + H 2

(Eq 14)

or oxide films on the metal:

NiO + BeF2

NiF2 + BeO

followed by reaction of nickel fluoride (NiF2) with M.

(Eq 15)

The reactions given in Eq 13, 14, and 15 will proceed essentially to completion at all temperatures. Accordingly, such reactions can lead (if the system is poorly cleaned) to a rapid initial corrosion rate. However, these reactions do not give a sustained corrosive attack. On the other hand, the reaction involving UF4 (Eq 12) may have an equilibrium constant that is strongly temperature dependent; therefore, when the salt is forced to circulate through a temperature gradient, a possible mechanism exists for mass transfer and continued attack. Equation 12 is of significance mainly in the case of alloys containing relatively large amounts of chromium. If nickel, iron, and molybdenum are assumed to form regular or ideal solid solutions with chromium (as is approximately true) and if the circulation rate is very rapid, the corrosion process for alloys in fluoride salts can be simply described. At high flow rates, uniform concentrations of UF3 and chromium fluoride (CrF2) are maintained throughout the fluid circuit. Under these conditions, there exists some temperature (intermediate between the maximum and minimum temperatures of the circuit) at which the initial chromium concentration of the structural metal is at equilibrium with the fused salt. This temperature, TBP, is called the balance point. Because the equilibrium constant for the chemical reaction with chromium increases with temperature, the chromium concentration in the alloy surface tends to decrease at temperatures higher than TBP and tends to increase at temperatures lower than TBP. At some point, the dissolution process will be controlled by the solid-state diffusion rate of chromium from the matrix to the surface of the alloy. In some melts (NaF-LiF-KF-UF4, for example), the equilibrium constant for Eq 12 with chromium changes sufficiently as a function of temperature to cause the formation of dendritic chromium crystals in the cold zone. For LiF-BeF2-UF4-type mixtures, the temperature dependence of the mass transfer reaction is small, and the equilibrium is satisfied at reactor temperature conditions without the formation of crystalline chromium. Thus, the rate of chromium removal from the salt stream by deposition at cold-fluid regions is controlled by the rate at which chromium diffuses into the cold-fluid wall; the chromium concentration gradient tends to be small, and the resulting corrosion is well within tolerable limits. A schematic of the temperature gradient mass transfer process is shown in Fig. 3.

Fig. 3 Temperature-gradient mass transfer

Lithium fluoride-beryllium fluoride salts containing UF4 or ThF4 and tested in thermal convection loops showed temperature gradient mass transfer, as noted by weight losses in the hot leg and weight gains in the cold leg (Fig. 4). Hastelloy N was developed for use in molten fluorides and has proved to be quite compatible. The weight changes of corrosion specimens increased with temperature and time (Fig. 5 and 6). Electrochemical methods were used to determine the oxidation potential of molten fluoride salts in thermal convection loops. The values obtained correlated well with specimen weight change data (Fig. 7).

Fig. 4 Weight changes of type 316 stainless steel specimens exposed to LiF-BeF2-ThF4-UF4 (68, 20, 11.7, and 0.3 mol%, respectively) as a function of position and temperature

Fig. 5 Weight changes of Hastelloy N specimens versus time of operation in LiF-BeF2-ThF4 (73, 2, and 25 mol%, respectively)

Fig. 6 Weight changes of Hastelloy N exposed to LiF-BeF2-ThF4-UF4 (68, 20, 11.7, and 0.3 mol%, respectively) for various times

Fig. 7 Uranium (III) in fuel salt

A type 304L stainless steel exposed to a fuel salt for 9.5 years in a type 304L stainless steel loop showed a maximum uniform corrosion rat of 22 μm/yr (0.86 mil/yr). Voids extended into the matrix for 250 μm (10 mils), and chromium depletion was found (Fig. 8 and 9).

Fig. 8 Weight changes of type 304L stainless steel specimens exposed to LiF-BeF2-ZrF4-ThF4-UF4 (70, 23, 5, 1, and 1 mol%, respectively) for various times and temperatures

Fig. 9 Chromium and iron concentration gradient in a type 304L stainless steel specimen exposed to LiF-BeF2ZrF4-ThF4-UF4 (70, 23, 5, 1, and 1 mol%, respectively) for 5700 h at 688 °C (1270 °F)

The corrosion resistance of a maraging steel (Fe-12Ni-5Cr-3Mo) at 662 °C (1224 °F) was better than that of type 304L stainless steel, but was worse than that of a Hastelloy N under equivalent conditions. As shown in Table 3, the uniform corrosion rate was 14 m (0.55 mil/yr). Voids were seen in the microstructure of the specimens after 5700 h, and electron microprobe analysis disclosed a definite depletion of chromium and iron.

Table 3 Comparison of weight losses of alloys at approximately 663 °C (1225 °F) in similar flow fuel salts in a temperature gradient system Alloy

Weight loss, mg/cm2

Average corrosion

mils/yr

2490 h

3730 h

Maraging steel

3.0

4.8

14

0.55

Type 304 stainless steel

6.5

10.0

28

1.1

Hastelloy N

0.4

0.6

1.5

0.06

m/yr

Type 316 stainless steel exposed to a fuel salt in a type 316 stainless steel loop showed a maximum uniform corrosion rate of 25 μm/yr (1 mil/yr) for 4298 h. Mass transfer did occur in the system. For selected nickel- and iron-base alloys, a direct correlation was found between corrosion resistance in molten fluoride salt and chromium and iron content of an alloy. The more chromium and iron in the alloy, the less the corrosion resistance.

References cited in this section

3. J.W. Koger, Report ORNL-TM-4286, Oak Ridge National Laboratory, Dec 1972 13. C.F. Baes, Jr., "The Chemistry and Thermodynamics of Molten Salt Reactor Fuels," Paper presented at the AIME Nuclear Fuel Reprocessing Symposium, Ames, IA, American Institute of Mining, Metallurgical, and Petroleum Engineers, Aug 1969; see also 1969 Nuclear Metallurgy Symposium, Vol 15, United States Atomic Energy Commission Division of Technical Information Extension 14. G. Long, "Reactor Chemical Division Annual Program Report," ORNL-3789, Oak Ridge National Laboratory, Jan 1965, p 65 15. J.W. Koger, "MSR Program Semiannual Progress Report," ORNL-4622, Oak Ridge National Laboratory, Aug 1970, p 170 Literature on Molten Salt Corrosion A vast number of publications are noteworthy in connection with molten salt corrosion. Those mentioned below represent sources of information that are particularly helpful. Janz and Tompkins provide an extensive bibliography with over 400 references (Ref 16). Inman and Lovering give an excellent survey of the field with over 200 references (Ref 1). Allen and Janz discuss safety and health hazards (Ref 17). Gale and Lovering provide an overview for researchers considering working with molten salts (Ref 18).

Summary In order to study the corrosion of molten salts or to determine what materials are compatible with a certain molten salt, the following questions must be answered. What is the purpose of the investigation? Is the researcher interested in basic studies, or is this work for information or work preliminary to assessment for a real system? For basic studies, capsule experiments or information from capsules is sufficient. Otherwise, flow systems or information from flow systems will be needed at some point to assess temperature gradient mass transfer. Salts to be used in either case need to be purified, and the same purity must be used in each experiment unless this factor is a variable. Analytical facilities must be used for the chemistry of the salt, including impurity content and surface analysis of the metals in question.

Vast amounts of useful information can be obtained from capsule and flow experiments. It is hoped that the preceding information on specific systems will provide an appreciation of the problems involved and the material that can be obtained from various experiments.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

14. 15. 16. 17. 18.

D. Inman and D.G. Lovering, Comprehensive Treatise of Electrochemistry, Vol 7, Plenum Publishing, 1983 R. Littlewood, J. Electrochem. Soc., Vol 109, 1962, p 525 J.W. Koger, Report ORNL-TM-4286, Oak Ridge National Laboratory, Dec 1972 P.F. Tortorelli and J.H. DeVan, Report ORNL-TM-8298, Oak Ridge National Laboratory, Dec 1982 A. Baraka, A.I. Abdel-Rohman, and A.A. El Hosary, Br. Corros. J., Vol 11, 1976, p 44 S.L. Marchiano and A.J. Arvia, Electrochim. Acta, Vol 17, 1972, p 25 A.J. Arvia, J.J. Podesta, and R.C.V. Piatti, Electrochim. Acta, Vol 16, 1971, p 1797 A.J. Arvia, J.J. Podesta, and R.C.V. Piatti, Electrochim. Acta, Vol 17, 1972, p 33 H.V. Venkatasetty and D.J. Saathoff, International Symposium on Molten Salts, 1976, p 329 Yu. I. Sorokin and Kh. L. Tseitlin, Khim. Prom., Vol 41, 1965, p 64 R.W. Bradshaw, "Corrosion of 304 Stainless Steel by Molten NaNO3-KNO3 in a Thermal Convection Loop," SAND-80-8856, Sandia National Laboratory, Dec 1980 R.W. Bradshaw, "Thermal Convection Loop Corrosion Tests of 316 Stainless Steel and IN800 in Molten Nitrate Salts," SAND-81-8210, Sandia National Laboratory, Feb 1982 C.F. Baes, Jr., "The Chemistry and Thermodynamics of Molten Salt Reactor Fuels," Paper presented at the AIME Nuclear Fuel Reprocessing Symposium, Ames, IA, American Institute of Mining, Metallurgical, and Petroleum Engineers, Aug 1969; see also 1969 Nuclear Metallurgy Symposium, Vol 15, United States Atomic Energy Commission Division of Technical Information Extension G. Long, "Reactor Chemical Division Annual Program Report," ORNL-3789, Oak Ridge National Laboratory, Jan 1965, p 65 J.W. Koger, "MSR Program Semiannual Progress Report," ORNL-4622, Oak Ridge National Laboratory, Aug 1970, p 170 G.J. Janz and R.P.T. Tompkins, Corrosion, Vol 35, 1979, p 485 C.B. Allen and G.T. Janz, J. Hazard. Mater., Vol 4, 1980, p 145 R.J. Gale and D.G. Lovering, Molten Salt Techniques, Vol 2, Plenum Press, 1984, p 1

Fundamentals of High-Temperature Corrosion in Liquid Metals P.F. Tortorelli, Oak Ridge National Laboratory

Introduction CONCERN ABOUT CORROSION of solids exposed to liquid-metal environments, that is, liquid-metal corrosion, dates from the earliest days of metals processing, when it became necessary to handle and contain molten metals. Corrosion considerations also arise when liquid metals are used in applications that exploit their chemical or physical properties. Liquid metals serve as high-temperature reducing agents in the production of metals (such as the use of molten magnesium to produce titanium) and because of their excellent heat transfer properties, they have been used or considered as coolants in a variety of power-producing systems. Examples of such applications include molten sodium for liquidmetal fast breeder reactors and central receiver solar stations as well as liquid lithium for fusion and space nuclear reactors. In addition, tritium breeding in deuterium-tritium fusion reactors necessitates the exposure of lithium atoms to fusion neutrons. Breeding fluids of lithium or lead-lithium are attractive for this purpose. Molten lead or bismuth can serve as neutron multipliers to raise the tritium breeding yield if other types of lithium-containing breeding materials are used. Liquid metals can also be used as two-phase working fluids in Rankine cycle power conversion devices (molten cesium or potassium) and in heat pipes (potassium, lithium, sodium, sodium-potassium). Because of their high thermal conductivities, sodium-potassium alloys, which can be any of a wide range of sodium-potassium combinations that are molten at or near room temperature, have also been used as static heat sinks in automotive and aircraft valves. Whenever the handling of liquid metals is required, whether in specific uses as discussed above or as melts during processing, a compatible containment material must be selected. At low temperatures, liquid-metal corrosion is often insignificant, but in more demanding applications, corrosion considerations can be important in selecting the appropriate containment material and/or operating parameters. Thus, liquid-metal corrosion studies in support of heat pipe technology and aircraft, space, and fast breeder reactor programs date back many years and, more recently, are being conducted as part of the fusion energy technology program. In this article, the principal corrosion reactions and important parameters that control such processes will be briefly reviewed for materials (principally metals) exposed to liquid-metal environments. Only corrosion phenomena will be covered; liquid-metal cracking and environmental effects on mechanical properties are described in the article "Environmentally Induced Cracking" in this Volume (see in particular the sections "Liquid-Metal Embrittlement" and "Solid-Metal Embrittlement"). Furthermore, the discussion will be limited to corrosion under single-phase (liquid) conditions.

Acknowledgements Research sponsored by the Office of Fusion Energy, U.S. Department of Energy under Contract No. DE-AC0584OR21400 with the Martin Marietta Energy Systems, Inc.

Corrosion Reactions in Liquid-Metal Environments Liquid-metal corrosion can manifest itself in various ways. In the most general sense, the following categories can be used to classify relevant corrosion phenomena: • • • •

Dissolution Impurity and interstitial reactions Alloying Compound reduction

Definitions and descriptions of these types of reactions are given below. However, it is important to note that this classification is somewhat arbitrary, and as will become clear during the following discussion, the individual categories are not necessarily independent of one another. Dissolution The simplest corrosion reaction that can occur in a liquid-metal environment is direct dissolution. Direct dissolution is the release of atoms of the containment material into the melt in the absence of any impurity effects. Such a reaction is a simple solution process and therefore is governed by the elemental solubilities in the liquid metal and the kinetics of the rate-controlling step of the dissolution reaction. The net rate, J, at which an elemental species enters solution, can be described as:

J = k(C - c)

(Eq 1)

where k is the solution rate constant for the rate-controlling step, C is the solubility of the particular element in the liquid metal, and c is the actual instantaneous concentration of this element in the melt. Under isothermal conditions, the rate of this dissolution reaction would decrease with time as c increases. After a period of time, the actual elemental concentration becomes equal to the solubility, and the dissolution rate is then 0. Therefore, in view of Eq 1, corrosion by the direct dissolution process can be minimized by selecting a containment material whose elements have low solubilities in the liquid metal of interest and/or by saturating the melt before actual exposure. However, if the dissolution kinetics are relatively slow, that is, for low values of k, corrosion may be acceptable for short-term exposures. The functional dependence and magnitude of the solution rate constant, k, depend on the rate-controlling step, which in the simplest cases can be a transport across the liquid-phase boundary layer, diffusion in the solid, or a reaction at the phase boundary. Measurements of weight changes as a function of time for a fixed C - c (see discussion below) yield the kinetic information necessary for determination of the rate-controlling mechanism. Corrosion resulting from dissolution in a nonisothermal liquid-metal system is more complicated than the isothermal case. Although Eq 1 can be used to describe the net rate at any particular temperature, the movement of liquid--for example, due to thermal gradients or forced circulation--tends to make c the same around the liquid-metal system. Therefore, at temperatures where the solubility (C) is greater than the bulk concentration (c), dissolution of an element into the liquid metal will occur, but at lower temperatures in the circuit where C < c, a particular element will tend to come out of solution and be deposited on the containment material (or it may remain as a suspended particulate). A schematic of such a mass transfer process is shown in Fig. 1. If net dissolution or deposition is measured by weight changes, a mass transfer profile such as the one shown in Fig. 2 can be established. Such mass transfer processes under nonisothermal conditions can be of prime importance when, in the absence of dissimilar-metal effects (see below), forced circulation (pumping) of liquid metals used as heat transfer media exacerbates the transport of materials from hotter to cooler parts of the liquidmetal circuit. Normally, the concentration in the bulk liquid, c, rapidly becomes constant with time such that, at given temperature, the concentration driving force (C - c, Eq 1) is then also constant. However, much more elaborate analyses based on Eq 1 are required to describe nonisothermal mass transfer precisely. Such treatments must take into account the differences in k around the circuit as well as the possibility that the rate constant for dissolution (or deposition) may not vary monotonically with temperature because of changes in the rate-controlling step within the temperature range of dissolution (deposition). The presence of more than one elemental species in the containment material further complicates the analysis; the transfer of each element typically has to be handled with its own set of thermodynamic and kinetic parameters. Although a thermal gradient increases the amount of dissolution, plugging of coolant pipes by nonuniform deposition of dissolved species in cold zones often represents a more serious design problem than metal loss from dissolution (which sometimes may be handled by corrosion allowances. The most direct way to control deposition, however, is usually to minimize dissolution in the hot zone by use of more corrosion-resistant materials and/or inhibition techniques.

Fig. 1 Schematic of thermal gradient mass transfer in a liquid-metal circuit. Source: Ref 1

Fig. 2 Mass transfer as characterized by the weight changes of type 316 stainless steel coupons exposed around a nonisothermal liquid lithium type 316 stainless steel circuit for 9000 h. Source: Ref 2

Mass transfer may even occur under isothermal conditions if an activity gradient exists in the system. Under the appropriate conditions, dissolution and deposition will act to equilibrate the activities of the various elements in contact with the liquid metal. Normally, such a process is chiefly limited to interstitial element transfer between dissimilar metals, but transport of substitutional elements can also occur. Elimination (or avoidance) of concentration (activity) gradients across a liquid-metal system is the obvious and, most often, the simplest solution to any problems arising from this type of mass transport process. Under certain conditions, dissolution of metallic alloys by liquid metals can lead to irregular attack (Fig. 3). Although such localized corrosive attack can often be linked to impurity effects (see discussion below) and/or compositional inhomogeneities in the solid, this destabilization of a planar surface is due to preferential dissolution of one or more elements of an alloy exposed to a liquid metal. Indeed, the type of attack illustrated in Fig. 3 is thought to be caused by the preferential dissolution of nickel from an Fe-17Cr-11Ni (wt%) alloy (type 316 stainless steel). As such, this process resembles the dealloying phenomenon sometimes observed in aqueous environments (see the article "Metallurgically Influenced Corrosion" in this Volume for a description of dealloying corrosion). In contrast, an Fe-12Cr-1Mo steel, which did not undergo preferential dissolution of any of its elements, corroded uniformly when exposed under the same environmental conditions (Fig. 4)

Fig. 3 Polished cross section of type 316 stainless steel exposed to thermally convective Pb-17at.%Li at 500 °C (930 °F) for 2472 h. Source: Ref 3

Fig. 4 Polished cross section of Fe-12Cr-1MoVW steel exposed to thermally convective Pb-17at.%Li at 500 °C (930 °F) for 2000 h. Source: Ref 3

Apart from possible effects on morphological development, the changes in surface composition due to preferential elemental dissolution from an alloy into a liquid metal are important in themselves. For example, in austenitic stainless steels exposed to sodium or lithium, the preferential dissolution of nickel causes a phase transformation to a ferritic structure in the surface region. In many cases, an equilibrium surface composition is achieved such that the net elemental fluxes into the liquid metal are in the same proportion as the starting concentrations of these elements in the alloy. Such a phenomenon has been rigorously treated and characterized for sodium-steel systems. Impurity and Interstitial Reactions

For this discussion, impurity or interstitial reactions refer to the interaction of light elements present in the containment material (interstitials) or the liquid metal (impurities). Examples of such reactions include the decarburization of steel in lithium and the oxidation of steel in sodium or lead of high oxygen activity. In many cases, when the principal elements of the containment material have low solubilities in liquid metals (for example, refractory metals in sodium, lithium, and lead), reactions involving light elements such as oxygen, carbon, and nitrogen dominate the corrosion process. Impurity or interstitial reactions can be generally classified into two types: corrosion product formation and elemental transfer of such species. Corrosion Product Formation. The general form of a corrosion product reaction is:

xL + yM + zI = LxMyIz

(Eq 2)

where L is the chemical symbol for a liquid-metal atom, M is one species of the containment material, and I represents an interstitial or impurity atom in the solid or liquid (x, y, z > 0). The LxMyIz corrosion product that forms by such a reaction may be soluble or insoluble in the liquid metal. If it is soluble, the I species would cause greater dissolution weight losses and would result in an apparently higher solubility of M in L (Eq 1). This is a frequent cause of erroneous solubility measurements and is a good illustration of how dissolution and impurity reactions can be interrelated. Furthermore, if a soluble corrosion product forms at selected sites on the surface of the solid, localized attack will result. Under conditions in which a corrosion product is insoluble, a partial or complete surface layer will form. However, this does not necessarily mean that it can be observed. The product may be unstable outside the liquid metal environment or may dissolve in the cleaning agent used to remove the solidified residue of liquid metal from the exposed containment material. A good example of the importance of impurity or interstitial reactions that form corrosion products can be found in the sodium-steel-oxygen system. It is thought that the reaction:

3Na2O(l) + Fe(s) = (Na2O)2 · FeO(s) + 2Na(l)

(Eq 3)

increases the apparent solubility of iron in sodium at higher oxygen activities, while the interaction of oxygen, sodium, and chromium can lead to the formation of surface corrosion products, for example:

2Na2O(l) + Cr(s) = NaCrO2(s) + 3Na(l)

(Eq 4)

This second type of reaction (Eq 4) is of primary importance in the corrosion of chromium-containing steels by liquid sodium. It can be controlled by reducing the oxygen concentration of the sodium to less than about 3 ppm and/or by modifying the composition of the alloy through reduction of the chromium concentration of the steel. Such corrosion product reactions can also be observed in lithium-steel systems, in which nitrogen can increase the corrosiveness of the liquid-metal environment. In particular, the reaction:

5Li3N(in l) + Cr(s) = Li9CrN5(s) + 6Li(l)

(Eq 5)

or an equivalent one with iron, can play an important role in corrosion by liquid lithium. The Li9CrN5 corrosion product tends to be localized at the grain boundaries of exposed steels. Such reaction products can probably also be formed when there is sufficient nitrogen in the solid; experimental observations have indicated that nitrogen can increase corrosion by lithium, whether it is in the liquid metal or in the steel. Corrosion product formation is also important when certain refractory metals are exposed to molten lithium. Despite their low solubilities in lithium, niobium and tantalum can be severely attacked when exposed to lithium if the oxygen activities of these metals are not low. At temperatures below approximately 900 °C (1650 °F), the lithium reacts rapidly with the oxygen and niobium or tantalum (and their oxides and suboxides) to form a ternary oxide corrosion product. Such reactions result in localized penetration along grain boundaries and selected crystallographic planes. This form of corrosive attack can be eliminated, however, by minimizing the oxygen concentration of these refractory metals (Fig. 5)

and by using alloying additions that form oxides that do not react with the lithium and that minimize the amount of uncombined oxygen in the material (1 to 2 at % Zr in niobium and hafnium in tantalum).

Fig. 5 Effect on initial oxygen concentration (150 to 1700 ppm) in niobium on the depth of attack by lithium. Polished and etched cross sections of niobium exposed to isothermal lithium at 816 °C (1500 °F) for 100 h. (a) 150 ppm. (b) 500 ppm. (c) 1000 ppm. (d) 1700 ppm. Etched with HF-HNO3-H2SO4-H2O. Source Ref 4

A final example of a corrosion product reaction that can occur in a liquid-metal environment is the oxidation of a solid metal or alloy exposed to molten lead. In some cases, this reaction may actually be beneficial by providing a protective barrier against the highly aggressive lead. This barrier can act in a manner analogous to the behavior observed for the protective oxides formed in high-temperature oxidizing gases. However, this surface product will form and then heal only when the oxygen activity of the melt is maintained at a high level or when oxide formers, such as aluminum or silicon, have been added to the containment alloy to promote protection by the formation of alumina- or silica-containing surface products. Furthermore, reactions of additives to the melt with nitrogen in steel to form nitride surface films are thought to be the cause of reduced corrosion in lead and lead-bismuth systems. Elemental Transfer of Impurities and Interstitials. The second general type of impurity or interstitial reaction is

that of elemental transfer. In contrast to what is defined as corrosion product formation, elemental transfer manifests itself as a net transfer of interstitials or impurities to, from, or across a liquid metal. Although compounds may form or dissolve as a result of such transfer, the liquid-metal atoms do not participate in the formation of stable products by reaction with the containment material. For example, because lithium is such a strong thermodynamic sink for oxygen, exposure of oxygen-containing metals and alloys to this liquid often results in the transfer of oxygen to the melt. Indeed, for oxygencontaminated niobium and tantalum, high-temperature lithium exposures result in the rapid movement of oxygen into the lithium. The thermodynamic driving force for light element transfer between solid and liquid metals is normally expressed in terms of a distribution (or partitioning) coefficient. This distribution coefficient is the equilibrium ratio of the concentration of an element, such as oxygen, nitrogen, carbon, or hydrogen, in the solid metal or alloy to that in the liquid. Such coefficients can be calculated from knowledge or estimates of free energies of formation and activities based on equilibrium between a species in the solid and liquid. An example of this approach is its application to decarburization/carburization phenomena in a liquid-metal environment. Carbon transfer to or from the liquid metal can cause decarburization of iron-chromium-molybdenum steels, particularly lower-chromium steels, and carburization of refractory metals and higher-chromium alloys. There have been many studies of such reactions for sodium-steel systems. Although less work has been done in the area of lithium-steel carbon transfer, the same considerations apply. Specifically,

the equilibrium partitioning of the carbon between the iron-chromium-molybdenum steel and the lithium can be described as:

(Eq 6)

where CC(s), CC(Li) is the concentration of carbon in the steel and lithium, respectively; aCr is the chromium activity of the steel; C°C(s), and C°C(Li) represent the solubilities of carbon in the steel and lithium, respectively; ∆F° represents the free energies of formation of the indicated compounds; x, y is the stoichiometry of the chromium carbide; R is the gas constant; and T is the absolute temperature (in degrees Kelvin). Equation 6 indicates that in order to decrease the tendency for decarburization of an alloy--that is, to increase the partitioning coefficient, CC(s)/CC(Li)--the chromium activity of the alloy must be increased or the free energy of formation of the matrix carbide(s) must be lowered (made more negative) by alloy manipulation or thermal treatment to form a more stable carbide dispersion. Experiments in lithium and sodium have shown that these factors have the desired effect. Tempering of iron-chromium-molybdenum steels to yield more stable starting carbides can significantly reduce decarburization by these two liquid metals. With very unstable microstructures, the steel can be severely corroded because of rapid lithium attack of the existing carbides. Furthermore, alloying additions, such as niobium, form very stable carbides and can dramatically reduce decarburization. In addition, as shown by Eq 6, increasing the chromium level of a steel effectively decreases the tendency for carbon loss. With higher-chromium steels, for example, austenitic stainless steels, carburization can then become a problem. If two dissimilar steels of significantly differing chromium activities and/or microstructures are exposed to the same liquid metal, the melt can act as a conduit for the relatively rapid redistribution of carbon between the two solids. Similar considerations would apply for any light element transfer across a liquid metal in contact with dissimilar materials; this can be further complicated by concentration (activity) gradient mass transfer of substitutional elements, as discussed above. Alloying Reactions between atoms of the liquid metal and those of the constituents of the containment material may lead to the formation of a stable product on the solid without the participation of impurity or interstitial elements:

xM + yL = MxLy

(Eq 7)

This is not a common form of liquid-metal corrosion particularly with the molten alkali metals, but it can lead to detrimental consequences if it is not understood or expected. Alloying reactions, however, can be used to inhibit corrosion by adding an element to the liquid metal to form a corrosion-resistant layer by reaction of this species with the contaminant material. An example is the addition of aluminum to a lithium melt contained by steel. A more dissolutionresistant aluminide surface layer forms, and corrosion is reduced. Compound Reduction Attack of ceramics exposed to liquid metals can occur because of reduction of the solid by the melt. In very aggressive situations, such as when most oxides are exposed to molten lithium, the effective result of such exposure is the loss of structural integrity by reduction-induced removal of the nonmetallic element from the solid. The tendency for reaction under such conditions can be qualitatively evaluated by consideration of the free energy of formation of the solid oxide relative to the oxygen/oxide stability in the liquid metal. Similar considerations apply to the evaluation of potential reactions between other nonmetallic compounds (nitrides, carbides, and so on) and liquid metals.

Considerations in Materials Selection The above types of corrosion reactions must be considered in materials selection for liquid-metal containment. In many cases, particularly at low temperature or with less aggressive liquids (such as molten steel), liquid-metal corrosion is not an important factor, and many materials, both metals and ceramics, would suffice. Under more severe conditions, however, an understanding of the various types of liquid-metal corrosion is necessary to select or develop a compatible containment material. For example, for applications in high-temperature molten lithium, most oxides would be unstable with respect to this liquid metal, low-chromium steels would decarburize, and alloys containing large amounts of nickel or manganese would suffer extensive preferential dissolution and irregular attack. Materials selection would then be limited to higher-chromium ferritic/martensitic steels or high-purity refractory metals and alloys. A general summary of the types of the most common corrosion reactions and guidelines for materials selection and/or development is given in Table 1, which also includes typical examples for each category. Because two or more concurrent corrosion reactions are possible, and because consideration of all of the applicable materials consequences may lead to opposite strategies, materials selection for liquid-metal environments can become quite complex and may require optimization of several factors rather than minimization of any particular one. In addition, an assessment of the suitability of a given material for liquid-metal service must be based on the knowledge of its total corrosion response. As in many corrosive environments, a simple numerical rate is not an accurate measurement of the susceptibility of a material when reaction with the liquid metal results in more than one of the modes of attack shown in Fig. 6 and discussed above. Under such circumstances, a measurement reflecting total corrosion damage is much more appropriate for judging the ability of a material to resist corrosion by a particular liquid metal. Additional examples of liquid-metal corrosion can be found in the article "General Corrosion" (see the section "Liquid-Metal Dissolution") in this Volume. Table 1 Guidelines for materials selection and/or alloy development based on liquid-metal corrosion reactions Corrosion reaction

Guidelines

Example

Direct dissolution

Lower activity of key elements.

Reduce nickel in lithium, lead, or sodium systems.

Corrosion formation

Lower activity of reacting elements.

Reduce chromium and nitrogen in lithium systems.

In case of protective oxide, add elements to promote formation.

Add aluminum or silicon to steel exposed to lead.

Increase (or add) elements to decrease transfer tendency.

Increase chromium content in steels exposed to sodium or lithium.

Minimize element being transferred.

Reduce oxygen content in metals exposed to lithium.

Avoid systems that form stable compounds.

Do not expose nickel to molten aluminum.

Promote formation of corrosion-resistant layers by alloying.

Add aluminum aluminides.

Eliminate solids that can be reduced by liquid metal.

Avoid bulk oxide-lithium couples.

product

Elemental transfer

Alloying

Compound reduction

to

lithium

to

form

surface

Fig. 6 Representative modes of surface damage in liquid-metal environments. IGA, intergranular attack. Source: Ref 5

References 1. J.E. Selle and D.L. Olson, in Materials Considerations in Liquid Metal Systems in Power Generation, National Association of Corrosion Engineers, 1978, p 15-22 2. P.F. Tortorelli and J.H. DeVan, J. Nucl. Mater., Vol 85 and 86, 1979, p 289-293 3. P.F. Tortorelli and J.H. DeVan, J. Nucl. Mater., Vol 141-143, 1986, p 592-598 4. J.R. DiStefano and E.E. Hoffman, Corrosion Mechanisms in Refractory Metal-Alkali Metal Systems, At. Energy Rev., Vol 2, 1964, p 3-33 5. J.H. DeVan and C. Bagnall, in Proceedings of the International Conference on Liquid Metal Engineering and Technology, Vol 3, The British Nuclear Energy Society, 1985, p 65-72 Selected References •

T.L. Anderson and G.R. Edwards, The Corrosion Susceptibility of 2 Cr-1 Mo Steel in a Lithium-17.6 Wt Pct Lead Liquid, J. Mater, Energy Syst., Vol 2, 1981, p 16-25 • R.C. Asher, D. Davis, and S.A. Beetham, Some Observations on the Compatibility of Structural Materials With Molten Lead, Corros. Sci., Vol 17, 1977, p 545-547 • M.G. Barker, S.A. Frankham, and N.J. Moon, The Reactivity of Dissolved Carbon and Nitrogen in Liquid Lithium, in Proceedings of the Third International Conference on Liquid Metal Engineering and Technology, Vol 2, The British Nuclear Energy Society, 1984, p 77-83 • M.G. Barker, P. Hubberstey, A.T. Dadd, and S.A. Frankham, The Interaction of Chromium With Nitrogen Dissolved in Liquid Lithium, J. Nucl. Mater., Vol 114, 1983, p 143-149 • N.M. Beskorovainyi, V.K. Ivanov, and M.T. Zuev, Behavior of Carbon in Systems of the Metal-Molten Lithium-Carbon Type, in High-Purity Metals and Alloys, V.S. Emel'yanov and A.I. Evstyukhin, Ed., Consultants Bureau, 1967, p 107-119 • N.M. Beskorovainyi and V.K. Ivanov, Mechanism Underlying the Corrosion of Carbon Steels in Lithium, in High-Purity Metals and Alloys, V.S. Emel'yanov and A.I. Evstyukhin, Ed., Consultants Bureau, 1967, p 120129 • O.K. Chopra, K. Natesan, and T.F. Kassner, Carbon and Nitrogen Transfer in Fe-9Cr-Mo Ferritic Steels Exposed to a Sodium Environment, J. Nucl. Mater., Vol 96, 1981, p 269-284 • O.K. Chopra and P.F. Tortorelli, Compatibility of Materials for Use in Liquid-Metal Blankets of Fusion Reactors, J. Nucl, Mater., Vol 122 and 123, 1984, p 1201-1212 • L.F. Epstein, Static and Dynamic Corrosion and Mass Transfer in Liquid Metal Systems, in Liquid Metals Technology--Part I, Vol 53 (No. 20), Chemical Engineering Progress Symposium Series, American Institute of Chemical Engineers, 1957, p 67-81

• J.D. Harrison and C. Wagner, The Attack of Solid Alloys by Liquid Metals and Salt Melts, Acta Metall., Vol 1, 1959, p 722-735 • E.E. Hoffman, "Corrosion of Materials by Lithium at Elevated Temperatures," ORNL-2674, Oak Ridge National Laboratory, March 1959 • A.R. Keeton and C. Bagnall, Factors That Affect Corrosion in Sodium, in Proceedings of the Second International Conference on Liquid Metal Technology in Energy Production, CONF-800401-P1, J.M. Dahlke, Ed., U.S. Department of Energy, 1980, p 7-18 to 7-25 • B.H. Kolster, The Influence of Sodium Conditions on the Rate for Dissolution and Metal/Oxygen Reaction of AISI 316 in Liquid Sodium, in Proceedings of the Second International Conference on Liquid Metal Technology in Energy Production, CONF-800401-P1, J.M. Dahlke, Ed., U.S. Department of Energy, 1980, p 7-53 to 7-61 • J. Konys and H.U. Borgstedt, The Product of the Reaction of Alumina With Lithium Metal, J. Nuclear Mater., Vol 131, 1985, p 158-161 • K. Natesan, Influence of Nonmetallic Elements on the Compatibility of Structural Materials With Liquid Alkali Metals, J. Nucl. Mater., Vol 115, 1983, p 251-262 • D.L. Olson, P.A. Steinmeyer, D.K. Matlock, and G.R. Edwards, Corrosion Phenomena in Molten Lithium, Rev. Coatings Corros./Int. Quart. Rev., Vol IV, 1981, p 349-434 • A.J. Romano, C.J. Klamut, and D.H. Gurinsky, "The Investigation of Container Materials for Bi and Pb Alloys, Part I. Thermal Convection Loops," BNL-811, Brookhaven National Laboratory, July 1963 • E. Ruedl, V. Coen, T. Sasaki, and H. Kolbe, Intergranular Lithium Penetration of Low-Ni, Cr-Mn Austenitic Stainless Steels, J. Nucl. Mater., Vol 110, 1982, p 28-36 • J. Sannier and G. Santarini, Etude de la Corrosion de Deux Aciers Ferritiques par le Plomb Liquide Circulant dans un Thermosiphon; Recherche d'un Modele, J. Nucl. Mater., Vol 107, 1982, p 196-217 • S.A. Shields, C. Bagnall, and S.L. Schrock, Carbon Equilibrium Relationships for Austenitic Stainless Steel in a Sodium Environment, Nucl, Technol., Vol 23, 1974, p 273-283 • S.A Shields and C. Bagnall, Nitrogen Transfer in Austenitic Sodium Heat Transport Systems, in Material Behavior and Physical Chemistry in Liquid Metal Systems, H.U. Borgstedt, Ed., Plenum Press, 1982, p 493501 • R.N. Singh, Compatibility of Ceramics With Liquid Na and Li, J. Amer. Ceram. Soc., Vol 59, 1976, p 112115 • D.L. Smith and K. Natesan, Influence of Nonmetallic Impurity Elements on the Compatibility of Liquid Lithium With Potential CTR Containment Materials, Nucl. Technol., Vol 22, 1974, p 392-404 • A.W. Thorley, Corrosion and Mass Transfer Behaviour of Steel Materials in Liquid Sodium, in Proceedings of the Third International Conference on Liquid Metal Engineering and Technology, Vol 3, The British Nuclear Energy Society, 1985, p 31-41 • P.F. Tortorelli and J.H. DeVan, Mass Transfer Kinetics in Lithium-Stainless Steel Systems, in Proceedings of the Third International Conference on Liquid Metal Engineering and Technology, Vol 3, The British Nuclear Energy Society, 1985, p 81-88 • P.F. Tortorelli and J.H. DeVan, Effects of a Flowing Lithium Environment on the Surface Morphology and Composition of Austenitic Stainless Steel, Microstruct. Sci., Vol 12, 1985, p 213-226 • J.R. Weeks and H.S. Isaacs, Corrosion and Deposition of Steels and Nickel-Base Alloys in Liquid Sodium, Adv. Corros. Sci. Technol., Vol 3, 1973, p 1-66

Fundamentals of Corrosion in Gases Samuel A. Bradford, Department of Mining, Metallurgical and Petroleum Engineering, University of Alberta

Introduction ENGINEERING METALS react chemically when exposed to air or to other more aggressive gases. Whether they survive or not depends on how fast they react. For a few metals, the reaction is so slow that they are virtually unattacked, but for others, the reaction can be disastrous. High-temperature service is especially damaging to most metals because of the exponential increase in reaction rate with temperature. The most common reactant is oxygen in the air; therefore, all gas-metal reactions are usually referred to as oxidation, using the term in its broad chemical sense whether the reaction is with oxygen, water vapor, hydrogen sulfide (H2S), or whatever the gas might be. Throughout this article, the process will be called oxidation, and the corrosion product will be termed oxide. Corrosion in gases differs from aqueous corrosion in that electrochemical principles do not help greatly in understanding the mechanism of oxidation. For gaseous reactions, a fundamental knowledge of the diffusion processes involved is much more useful. The principles of high-temperature oxidation began to be understood only in the 1920s, whereas electrochemistry and aqueous corrosion principles were developed approximately 100 years earlier. The first journal devoted to corrosion in gases (Oxidation of Metals) began publication less than 20 years ago. In this article, a short summary of thermodynamic concepts is followed by an explanation of the defect structure of solid oxides and the effect of these defects on conductivity and diffusivity. The commonly observed kinetics of oxidation will be described and related to the corrosion mechanisms. These mechanisms are shown schematically in Fig. 1. The gas first adsorbs on the metal surface as atomic oxygen. Oxide nucleates at favorable sites and most commonly grows laterally to form a complete thin film. As the layer thickens, it provides a protective scale barrier to shield the metal from the gas. For scale growth, electrons must move through the oxide to reach the oxygen atoms adsorbed on the surface, and oxygen ions, metal ions, or both must move through the oxide barrier. Oxygen may also diffuse into the metal.

Fig. 1 Schematic illustration of the principal phenomena taking place during the reaction of metals with oxygen. Source: Ref 1

Growth stresses in the scale may create cavities and microcracks in the scale, modifying the oxidation mechanism or even causing the oxide to fail to protect the metal from the gas. Improved oxidation resistance can be achieved by developing better alloys and by applying protective coatings. The basic principles of alloy oxidation, discussed in the section "Alloy Oxidation: The Doping Principle" in this article, are applicable to both alloy development and use of metallic coatings for protection against corrosive gases. Fundamental Data. Essential to an understanding of the gaseous corrosion of a metal are the crystal structure and the molar volume of the metal on which the oxide builds, both of which may affect growth stresses in the oxide. For hightemperature service the melting point of the metal, which indicates the practical temperature limits, and the structural changes that take place during heating and cooling, which affect oxide adherence, must be known. These data are presented in Table 1 for pure metals. For the oxides, their structures, melting and boiling points, molar volume, and oxide/metal volume ratio (Pilling-Bedworth ratio) are shown in Table 2. The structure data were taken from many sources.

Table 1 Structures and thermal properties of pure metals

Metal

Structure(a)

Transformation temperature

°C

°F

Volume change upon cooling(b), %

Molar volume(c)

Melting point

°C

°F

cm3

in.3

Aluminum

fcc

...

...

...

660.4

1220.7

10.00

0.610

Antimony

rhom

...

...

...

630.7

1167.3

18.18

1.109

Arsenic

rhom

...

...

...

Sublimation 615

1139

12.97

0.791

Barium

bcc

...

...

...

729

1344

39

2.380

) hcp

1250

2282

...

...

...

4.88

0.298

...

...

-2.2

1290

2354

4.99

0.304

Beryllium

(

( ) bcc

Bismuth

rhom

...

...

...

271.4

520.5

21.31

1.300

Cadmium

hcp

...

...

...

321.1

610

13.01

0.793

) fcc

448

838

...

...

...

25.9

1.581

( ) bcc

...

...

-0.4

839

1542

...

...

( ) fcc

726

1339

...

...

...

20.70

1.263

( ) bcc

...

...

...

798

1468

...

...

Cesium

bcc

...

...

...

28.64

83.55

70.25

4.287

Chromium

bcc

...

...

...

1875

3407

7.23

0.441

) hcp

417

783

...

...

...

6.67

0.407

( ) fcc

...

...

-0.3

1495

2723

6.70

0.408

fcc

...

...

...

1084.88

1984.78

7.12

0.434

Calcium

Cerium

Cobalt

Copper

(

(

Metal

Structure(a)

Transformation temperature

Volume change upon cooling(b), %

Molar volume(c)

Melting point

°C

°F

cm3

in.3

...

...

...

19.00

1.159

...

-0.1

1412

2573

18.98

1.158

...

...

...

1529

2784

18.45

1.126

bcc

...

...

...

822

1512

28.98

1.768

) hcp

1235

2255

...

...

...

19.90

1.214

...

...

-1.3

1312

2394

20.16

1.230

ortho

...

...

...

29.78

85.60

11.80

0.720

diamond fcc

...

...

...

937.4

1719.3

13.63

0.832

fcc

...

...

...

1064.43

1947.97

10.20

0.622

) hcp

1742

3168

...

...

...

13.41

0.818

( ) bcc

...

...

...

2231

4048

...

...

Holmium

hcp

...

...

...

1474

2685

18.75

1.144

Indium

tetr

...

...

...

156.63

313.93

15.76

0.962

Iridium

fcc

...

...

...

2447

4437

8.57

0.523

) bcc

912

1674

...

...

...

7.10

0.433

( ) fcc

1394

2541

1.0

...

...

7.26

0.443

( ) bcc

...

...

-0.52

1538

2800

7.54

0.460

°C

°F

1381

2518

( ) bcc

...

Erbium

hcp

Europium

Dysprosium

Gadolinium

(

(

) hcp

( ) bcc

Gallium

Germanium

Gold

Hafnium

Iron

(

(

Metal

Lanthanum

Structure(a)

cm3

in.3

...

...

...

22.60

1.379

...

0.5

...

...

22.44

1.369

...

...

-1.3

918

1684

23.27

1.420

...

...

...

327.4

621.3

18.35

1.119

-193

-315

...

180.7

357.3

12.99

0.793

) hex

330

626

( ) fcc

865

( ) bcc

fcc

( ) bcc

Molar volume(c)

Melting point

°F

°F

(

Volume change upon cooling(b), %

°C

°C

Lead

Lithium

Transformation temperature

Lutetium

hcp

...

...

...

1663

3025

17.78

1.085

Magnesium

hcp

...

...

...

650

1202

13.99

0.854

) cubic

710

1310

...

...

...

7.35

0.449

( ) cubic

1079

1974

-3.0

...

...

7.63

0.466

( ) tetr

...

...

-0.0

1244

2271

7.62

0.465

Manganese

(

Mercury

rhom

...

...

...

-38.87

-37.97

14.81

0.904

Molybdenum

bcc

...

...

...

2610

4730

9.39

0.573

) hex

863

1585

...

...

...

20.58

1.256

( ) bcc

...

...

-0.1

1021

1870

21.21

1.294

Neodymium

(

Nickel

fcc

...

...

...

1453

2647

6.59

0.402

Niobium

bcc

...

...

...

2648

4474

10.84

0.661

Osmium

hcp

...

...

...

8.42

0.514

2700

4890

Metal

Structure(a)

Transformation temperature

°C

°F

Volume change upon cooling(b), %

Molar volume(c)

Melting point

°C

°F

cm3

in.3

Palladium

fcc

...

...

...

1552

2826

8.85

0.540

Platinum

fcc

...

...

...

1769

3216

9.10

0.555

120,210,315

248,410,599

...

...

...

452,480

846,896

...

640

1184

bcc

...

...

...

63.2

145.8

45.72

2.790

) hex

795

1463

...

...

...

20.80

1.269

( ) bcc

...

...

-0.5

931

1708

21.22

1.295

Rhenium

hcp

...

...

...

3180

5756

8.85

0.540

Rhodium

fcc

...

...

...

1963

3565

8.29

0.506

Rubidium

bcc

...

...

...

38.89

102

55.79

3.405

Ruthenium

hcp

...

...

...

2310

4190

8.17

0.499

) rhom

734

1353

...

...

...

20.00

1.220

( ) hcp

922

1692

...

...

...

20.46

1.249

( ) bcc

...

...

...

1074

1965

20.32

1.240

(

1337

2439

...

...

...

15.04

0.918

( ) bcc

...

...

...

1541

2806

...

...

( ) hex

209

408

...

217

423

16.42

1.002

Plutonium

Potassium

Praseodymium

Samarium

Scandium

Selenium

(

(

,

,

,

',

) hcp

12.04

14.48

0.735

0.884

Metal

Silicon

Structure(a)

diamond fcc

Silver

fcc

Transformation temperature

Volume change upon cooling(b), %

Molar volume(c)

Melting point

°C

°F

cm3

in.3

...

1410

2570

12.05

0.735

...

...

961.9

1763.4

10.28

0.627

°C

°F

...

...

...

Sodium

( ) bcc

-237

-395

...

97.82

208.08

23.76

1.450

Strontium

(

) fcc

557

1035

...

...

...

34

2.075

( ) bcc

...

...

...

768

1414

34.4

2.099

Tantalum

bcc

...

...

...

2996

5425

10.9

0.665

Tellurium

hex

...

...

...

449.5

841.1

20.46

1.249

) hcp

1289

2352

...

...

...

19.31

1.178

( ) bcc

...

...

...

1356

2472.8

19.57

1.194

(

) hcp

230

446

...

...

...

17.21

1.050

( ) bcc

...

...

...

303

577

...

...

(

1345

2453

...

...

...

19.80

1.208

( ) bcc

...

...

...

1755

3191

21.31

1.300

hcp

...

...

...

1545

2813

18.12

1.106

Terbium

Thallium

Thorium

(

Thulium

) fcc

Tin

( ) bct

13.2

55.8

27

231.9

449.4

16.56

1.011

Titanium

(

882.5

1621

...

...

...

10.63

0.649

...

...

...

1668

3034

11.01

0.672

) hcp

( ) bcc

Metal

Structure(a)

Transformation temperature

Volume change upon cooling(b), %

Molar volume(c)

Melting point

°C

°F

cm3

in.3

...

3410

6170

9.55

0.583

1222

...

...

...

12.50

0.763

769

1416

-1.0

...

...

13.00

0.793

( ) bcc

...

...

-0.6

1900

3452

8.34

0.509

Ytterbium

( ) fcc

7

45

0.1

819

1506

24.84

1.516

Yttrium

(

1478

2692

...

...

...

19.89

1.214

( ) bcc

...

...

...

1522

2772

20.76

1.267

hcp

...

...

...

420

788

9.17

0.559

) hcp

862

1584

...

...

...

14.02

0.856

( ) bcc

...

...

...

1852

3366

15.09

0.921

°C

°F

bcc

...

...

) ortho

661

( ) complex tetr

Tungsten

Uranium

(

Zinc

Zirconium

(

) hcp

Source: Ref 2 (a) fcc, face-centered cubic; rhom, rhombohedral; bcc, body-centered cubic; hcp, hexagonal close-packed; ortho, orthorhombic; tetr, tetragonal; hex, hexagonal; bct, body-centered tetragonal.

(b) Volume change upon cooling through crystallographic transformation.

(c) Molar volume at 25 °C (77 °F) or at transition temperature for structures not stable at 25 °C (77 °F).

Table 2 Structures and thermal properties of selected oxides Oxide

Structure

-Al2O3

D51 (corundum)

-Al2O3

(defect-spinel)

Melting point

Boiling or decomposition, d.

Molar volume(a)

Volume ratio

°C

°F

°C

°F

cm3

in.3

2015

3659

2980

5396

25.7

1.568

1.28

...

...

...

26.1

1.593

1.31

3632

26.8

1.635

0.69

d.1472

34.1

2.081

0.87

8.3

0.506

1.70

BaO

B1 (NaCl)

1923

3493

BaO2

Tetragonal (CaC2)

450

842

BeO

B4 (ZnS)

2530

4586

CaO

B1 (NaCl)

2580

4676

2850

5162

16.6

1.013

0.64

CaO2

C11 (CaC2)

...

...

d.275

d.527

24.7

1.507

0.95

CdO

B1 (NaCl)

d.900

d.1652

18.5

1.129

1.42

Ce2O3

D52 (La2O3)

...

...

47.8

2.917

1.15

CeO2

C1 (CaF2)

...

...

24.1

1.471

1.17

CoO

B1 (NaCl)

1935

3515

...

...

11.6

0.708

1.74

Co2O3

Hexagonal

...

...

d.895

d.1643

32.0

1.953

2.40

Co3O4

H11 (spinel)

...

...

39.7

2.423

1.98

Cr2O3

D51 (

Cs2O

1400

1692

2600

2552

3078

4712

CoO

2000

d.800

3900

7052

2435

4415

4000

7232

29.2

1.782

2.02

Hexagonal (CdCl2)

...

...

d.400

d.752

66.3

4.046

0.47

Cs2O3

Cubic (Th3P4)

400

752

650

1202

70.1

4.278

0.50

CuO

B26 monoclinic

1326

2419

...

...

12.3

0.751

1.72

Al2O3)

Cu2O

C3 cubic

1235

2255

d.1800

d.3272

23.8

1.452

1.67

Dy2O3

Cubic (Tl2O3)

2340

4244

...

...

47.8

2.917

1.26

Er2O3

Cubic (Tl2O3)

...

...

...

...

44.3

2.703

1.20

FeO

B1 (NaCl)

1420

2588

...

...

12.6

0.769

1.78 on

-iron

D51 (hematite)

1565

2849

...

...

30.5

1.861

2.15 on

-iron

...

...

...

...

...

...

...

1.02 on Fe3O4

D57 cubic

1457

2655

...

...

31.5

1.922

2.22 on

-iron

H11 (spinel)

...

...

d.1538

d.2800

44.7

2.728

2.10 on

-iron

...

...

...

...

...

...

...

Ga2O3

Monoclinic

1900

3452

...

31.9

1.947

1.35

HfO2

Cubic

2812

5094

21.7

1.324

1.62

HgO

Defect B10(SnO)

...

...

d.500

d.932

19.5

1.190

1.32

In2O3

D53(Sc2O3)

...

...

d.850

d.1562

38.7

2.362

1.23

IrO2

C4(TiO2)

...

...

d.1100

d.2012

19.1

1.166

2.23

K2O

C1(CaF2)

...

...

d.350

d.662

40.6

2.478

0.45

La2O3

D52 hexagonal

2315

4199

4200

7592

50.0

3.051

1.10

Li2O

C1 (CaF2)

1200

2192

14.8

0.903

0.57

MgO

B1 (NaCl)

2800

5072

3600

6512

11.3

0.690

0.80

MnO

B1 (NaCl)

...

...

...

...

13.0

0.793

1.77

MnO2

C4 (TiO2)

...

...

d.535

d.995

17.3

1.056

2.37

-Fe2O3

-Fe2O3

Fe3O4

1700

3092

5400

9752

1.2 on FeO

Mn2O3

D53 (Sc2O3)

...

...

d.1080

d.1976

35.1

2.142

2.40

H11 (spinel)

1705

1301

...

...

47.1

2.874

2.14

MoO3

Orthorhombic

795

1463

...

...

30.7

1.873

3.27

Na2O

C1 (CaF2)

Sublimation 1275

2327

...

...

27.3

1.666

0.57

Nb2O5

Monoclinic

1460

2660

...

...

59.5

3.631

2.74

Nd2O3

Hexagonal

...

...

46.5

2.838

1.13

NiO

B1 (NaCl)

1990

3614

...

...

11.2

0.683

1.70

OsO2

C4 (TiO2)

...

...

d.350

d.662

28.8

1.757

3.42

PbO

B10 tetragonal

888

1630

...

...

23.4

1.428

1.28

Pb3O4

Tetragonal

...

...

d.500

d.932

75.3

4.595

1.37

PdO

B17 tetragonal

870

1598

...

...

14.1

0.860

1.59

PtO

B17 (PdO)

...

...

d.550

d.1022

14.2

0.867

1.56

Rb2O3

(Th3P4)

489

912

...

...

62.0

3.783

0.56

ReO2

Monoclinic

...

...

d.1000

d.1832

19.1

1.166

2.16

Rh2O3

D51 (

...

...

d.1100

d.2012

31.0

1.892

1.87

SiO

Cubic

1880

3416

20.7

1.263

1.72

-Mn3O4

SiO2

-Al2O3)

cristobalite C9

1900

1700

3452

3092

1713

3115

2230

4046

25.9

1.581

2.15

SnO

B10 (PbO)

...

...

d.1080

d.1976

20.9

1.275

1.26

SnO2

C4 (TiO2)

1127

2061

...

...

21.7

1.324

1.31

SrO

B1 (NaCl)

2430

4406

22.0

1.343

0.65

3000

5432

Ta2O5

Triclinic

1800

3272

...

...

53.9

3.289

2.47

TeO2

C4 (TiO2)

733

1351

1245

2273

28.1

1.715

1.38

ThO2

C1 (CaF2)

3050

5522

4400

7952

26.8

1.635

1.35

TiO

B1 (NaCl)

1750

3182

3000

5432

13.0

0.793

1.22

TiO2

C4 (rutile)

1830

3326

2700

4892

18.8

1.147

1.76

Ti2O3

D51 (

...

...

d.2130

d.3866

31.3

1.910

1.47

Tl2O3

D53 (Sc2O3)

717

1323

d.875

d.1607

44.8

2.734

1.30

UO2

C1 (CaF2)

2500

4532

...

...

24.6

1.501

1.97

U3O8

Hexagonal

...

...

d.1300

101.5

6.194

2.71

VO2

C4 (TiO2)

1967

3573

...

...

19.1

1.166

2.29

V2O3

D51 (

1970

3578

...

...

30.8

1.879

1.85

V2O5

D87 orthorhombic

690

1274

d.1750

d.3182

54.2

3.307

3.25

WO2

C4 (TiO2)

1430

2606

17.8

1.086

1.87

32.4

1.977

3.39

29.8

1.819

3.12

-Al2O3)

-Al2O3)

1550

2822

Orthorhombic

1473

W2O5

Triclinic

Sublimation,

Y2O3

D53 (Sc2O3)

2410

4370

...

...

45.1

2.752

1.13

ZnO

B4 (wurtzite)

1975

3587

...

...

14.5

0.885

1.58

ZrO2

C43 monoclinic

2715

4919

...

...

22.0

1.343

1.57

-WO3

...

850

1562

1530

...

2786

Source: Ref 3 (a) Molar volume at 25 °C (77 °F) or at transition temperature for structures not stable at 25 °C (77 °F).

Thermodynamics of High-Temperature Corrosion in Gases Free Energy of Reaction. The driving force for reaction of a metal with a gas is the Gibbs energy change, ∆G. For the

usual conditions of constant temperature and pressure, ∆G is described by the Second Law of Thermodynamics as:

∆G = ∆H – T∆S

(Eq 1)

where ∆H is the enthalpy of reaction, T is the absolute temperature, and ∆S is the entropy change. No reaction will proceed spontaneously unless ∆G is negative. If ∆G = 0, the system is at equilibrium, and if ∆G is positive, the reaction is thermodynamically unfavorable; that is, the reverse reaction will proceed spontaneously. The driving force ∆G for a reaction such as aA + bB = cC + dD can be expressed in terms of the standard Gibbs energy change, ∆Go by:

(Eq 2)

where the chemical activity, a, of each reactant or product is raised to the power of its stoichiometric coefficient, and R is the gas constant. For example, in the oxidation of a metal by the reaction:

M is the reacting metal, MxOy is its oxide, and x and y are the moles of metal and oxygen, respectively, in 1 mol of the oxide. The Gibbs energy change for the reaction is:

(Eq 3)

In most cases, the activities of the solids (metal and oxide) are invarient; that is, their activities = 1 for pure solids, and for the relatively high temperatures and moderate pressures encountered in oxidation reactions, a can be approximated by its pressure. Therefore, at equilibrium where ∆G = 0:

(Eq 4)

where p

is the partial pressure of oxygen.

In solid solutions, such as an alloy, the partial molar Gibbs energy of a substance is usually called its chemical potential . If 1 mol of pure A is dissolved in an amount of solution so large that the solution concentration remains virtually unchanged, the Gibbs energy change for the mole of A is:

∆

A

=

A

(Eq 5)

- °A = RT ln aA

where μ°A is the chemical potential of 1 mol of pure A, the chemical potential activity of A in the solution.

A

is the value in the solution, and aA is the

Metastable Oxides. Thermodynamically unstable oxides are often formed in corrosion by gases. The Gibbs energy of

formation of the oxide, ∆G, is less negative than for a stable oxide, but in fact an unstable oxide can often exist indefinitely with no measurable transformation. A common example is wustite (FeO), which is formed during the hot rolling of steel. Thermodynamically, it is unstable below 570 °C (1060 °F), but it remains the major component of mill scale at room temperature because the decomposition kinetics is extremely slow. As another example, rapid kinetics can favor the formation of less stable oxide on an alloy. An alloy AB could oxidize to form oxides AO and BO, but if BO is more stable than AO, then any AO formed in contact with B should in theory convert to BO by the reaction:

B + AO

BO + A

Nevertheless, if AO grows rapidly compared with BO and the conversion reaction is slow, AO can be the main oxide found on the alloy. Thermodynamically unstable crystal structures of oxides are also sometimes found. A growing oxide film tends to try to align its crystal structure in some way with that of the substrate from which it is growing. This epitaxy can cause the formation of an unstable structure that fits the substrate best. For example, cubic aluminum oxide (Al2O3) may form on aluminum alloys instead of the stable rhombohedral Al2O3. Free Energy-Temperature Diagrams. Metal oxides become less stable as temperature increases. The relative

stabilities of oxides are usually shown on a Gibbs energy-temperature diagram, sometimes called an Ellingham diagram (Fig. 2), for common metals in equilibrium with their oxides. Similar diagrams are available for sulfides, nitrides, and other gas-metal reactions. In Fig. 2, the reaction plotted in every case is:

That is, 1 mol of O2 gas is always the reactant so that:

∆Go = RT ln p

(Eq 6)

For example, the Gibbs energy of formation of Al2O3 at 1000 °C (1830 °F), as read from Fig. 2, is approximately -840 kJ (-200 kcal) for

mol of Al2O3.

Fig. 2 Standard Gibbs energies of formation of selected oxides as a function of temperature. Source: Ref 4

The equilibrium partial pressure of O2 is:

(Eq 7)

and can also be read directly from Fig. 2 without calculation by use of the p scale along the bottom and right side of the diagram. A straight line drawn from the index point labeled O at the upper left of the diagram, through the 1000 °C (1830 °F) point on the Al/Al2O3 line intersects the p scale at approximately 10-35 atm, which is the O 2 partial pressure in equilibrium with aluminum and Al2O3 at 1000 °C (1830 °F). This means that any O2 pressure greater than 10-35 atm tends to oxidize more aluminum, while Al2O3 would tend to decompose to Al + O2 only if the pressure could be reduced to below 10-35 atm. Obviously, Al2O3 is an extremely stable oxide.

The oxidation of a metal by water vapor can be determined in the same way. The reaction is:

xM + yH2O = MxOy + yH2 The equilibrium p /p ratio for any oxide at any temperature can be found by constructing a line from the H index point on the left side of Fig. 2. For example, for the reaction:

2Al(l) + 3H2O(g) = A2O3(s) + 3H2(g) at 1000 °C (1830 °F), the equilibrium H2/H2O ratio is 1010. A ratio greater than this will tend to drive the reaction to the left, reducing Al2O3 to the metal. A ratio less than 1010 produces more oxide. Similarly, the oxidation of metals by carbon dioxide (CO2) is also shown on Fig. 2. For the reaction:

xM + yCO2 = MxOy + yCO the equilibrium carbon monoxide (CO)/CO2 ratio is found from the index point marked C on the left side of the diagram. Oxidation of aluminum by CO2 has an equilibrium CO/CO2 ratio approximately 1010 at 1000 °C (1830 °F). Isothermal Stability Diagrams. For situations that are more complicated than a single metal in a single oxidizing

gas, it is common to fix the temperature at some practical value and plot the other variables of gas pressures or alloy composition against each other. This produces isothermal stability diagrams, or predominance area diagrams, which show the species that will be most stable in any set of circumstances. One Metal and Two Gases. These diagrams, often called Kellogg diagrams, are constructed from the standard Gibbs

energies of formation, ∆Go, of all elements and compounds likely to be present in the system. For example, for the Ni-OS system, the ∆Go values of nickel monoxide (NiO) (s), nickel monosulfide (NiS) (l), nickel sulfate (NiSO4 (s), sulfur dioxide (SO2 (g), sulfur trioxide (SO3 (g), and S (l) are needed. In Fig. 3, the boundary between the Ni (s) and NiO (s) regions represents the equilibrium Ni (s) + O2 (g) = NiO (s); therefore, the diagram shows that at 1250 K any O2 pressure above about 10-11 atm will tend to form NiO from metallic nickel if p is low. Similarly, S2 gas pressure greater than about 10-7 atm will form NiS from nickel at low p . Also a mixed gas of 10-5 atm each of S2 and O2 should form nearly the equilibrium ratio of NiO (s) and NiSO4 (s).

Fig. 3 The Ni-O-S system at 1250 K. Source: Ref 5

If the principal gases of interest were SO2 and O2, the same ∆Go data could be used to construct a diagram of log p versus p , or as in Fig. 3, p isobars can be added to the figure (the dotted lines). Thus, a mixed gas of 10-5 atm each of SO2 and O2 will form only NiO at 1250 K, with neither the sulfide nor sulfate being as stable. When nickel metal is heated to 1250 K in the open air with sulfur-containing gases, pSO2 + p situation is shown by the dashed line in Fig. 3 labeled p = 0.2 atm.

+p

0.2 atm. The

An Alloy System and a Gas. Isothermal stability diagrams for oxidation of many important alloy systems have been

worked out, such as that for the Fe-Cr-O system shown in Fig. 4. In this diagram, the mole fraction of chromium in the alloy is plotted against log p so that for any alloy composition the most stable oxide or mixture of oxides is shown at any gas pressure.

Fig. 4 Stability diagram for the Fe-Cr-O system at 1300 °C (2370 °F)

For an alloy system in gases containing more than one reactive component, the pressures of all but one of the gases must be fixed at reasonable values to be able to draw an isothermal stability diagram in two dimensions. Figure 5 shows an example of such a situation: the Fe-Zn system in equilibrium with sulfur and oxygen-containing gases with SO2 pressure set at 1 atm and temperature set at 1164 K.

Fig. 5 The Fe-Zn-S-O system for p

= 1 atm at 1164 K. Source: Ref 6

Limitations of Predominance Area Diagrams. Isothermal stability diagrams, like all predominance area diagrams,

including Pourbaix potential-pH diagrams, must be read with an understanding of their rules: •

•

Each area on the diagram is labeled with the predominant phase that is stable under the specified conditions of pressure or temperature. Other phases may also be stable in that area, but in smaller amounts The boundary line separating two predominance areas shows the conditions of equilibrium between the two phases.

Also, the limitations of the diagrams must be understood to be able to use them intelligently: • • • •

The diagrams are for the equilibrium situation. Equilibrium may be reached quickly in high-temperature oxidation, but if the metal is then cooled, equilibrium is often not reestablished Microenvironments, such as gases in voids or cracks, can create situations that differ from the situations expected for the bulk reactant phases The diagrams often show only the major components, omitting impurities that are usually present in industrial situations and may be important The diagrams are based on thermodynamic data and do not show rates of reaction

Kinetics of Corrosion in Gases Mechanisms of Oxidation. In 1923, N.B. Pilling and R.E. Bedworth classified oxidizable metals into two groups: those that formed protective oxide scales and those that did not (Ref 7). They suggested that unprotective scales formed if the volume of the oxide layer was less than the volume of metal reacted. For example, in the oxidation of aluminum:

2Al +

Al2O3

the Pilling-Bedworth ratio is:

where the volumes can be calculated from molecular and atomic weights and the densities of the phases. If the ratio is less than 1, as is the case for alkali and alkaline earth metals, the oxide scales are usually unprotective, with the scales being porous or cracked due to tensile stresses and providing no efficient barrier to penetration of gas to the metal surface. If the ratio is more than 1, the protective scale shields the metal from the gas so that oxidation can proceed only by solid-state diffusion, which is slow even at high temperatures. If the ratio is much over 2 and the scale is growing at the metal/oxide interface, the large compressive stresses that develop in the oxide as it grows thicker may eventually cause the scale to spall off, leaving the metal unprotected. Exceptions to the Pilling-Bedworth theory are numerous, and it has been roundly criticized and rejected by many. Its main flaw is the assumption that metal oxides grow by diffusion of oxygen inward through the oxide layer to the metal. In fact, it is much more common for metal ions to diffuse outward through the oxide to the gas. Also, the possibility of plastic flow by the oxide or metal was not considered. Nevertheless, historically, Pilling and Bedworth made the first step in achieving understanding of the processes by which metals react with gases. And although there may be exceptions, the volume ratio, as a rough rule-of-thumb, is usually correct. The Pilling-Bedworth volume ratios for many common oxides are listed in Table 2. Defect Structure of Ionic Oxides. Ionic compounds can have appreciable ionic conductivity due to Schottky defects and/or Frenkel defects. Schottky defects are combinations of cation vacancies and anion vacancies in the proper ratio necessary to maintain electrical neutrality. Figure 6(a) illustrates a Schottky defect in a stoichiometric ionic crystal. With Schottky defects, the ions must diffuse into the appropriate adjacent vacancies to allow mass transfer and ionic electrical conductivity.

Fig. 6 Defects in ionic crystals. (a) Schottky defect. (b) Frenkel defect. Vacancies are indicated by open squares. Interstitial ion is shown as shaded circle.

Frenkel defects are also present in ionic crystals in such a way that electrical neutrality and stoichiometry are maintained (Fig. 6b). This type of defect is a combination of a cation vacancy and an interstitial cation. Metal cations are generally much smaller than the oxygen anions. Limited ionic electrical conductivity is possible in such crystals by diffusion of cations interstitially and by diffusion of cations into the cation vacancies. Metallic oxides are seldom, if ever, stoichiometric and cannot grow by mere diffusion by Schottky and Frenkel defects. For oxidation to continue when a metal is protected by a layer of oxide, electrons must be able to migrate from the metal, through the oxide, to adsorbed oxygen at the oxide/gas interface. Nevertheless, Schottky and Frenkel defects may provide the mechanism for ionic diffusion necessary for oxide growth. Defect Structure of Semiconductor Oxides. Oxides growing to provide protective scales are electronic semiconductors that also allow mass transport of ions through the scale layer. They may be conveniently categorized as ptype, n-type, and amphoteric semiconductors. Examples of the three types are listed in Table 3 (Ref 8).

Table 3 Classification of electrical conductors: oxides, sulfides, and nitrides Metal-excess semiconductors (n-type)

BeO, MgO, CaO, SrO, BaO, BaS, ScN, CeO2, ThO2, UO3, U3O8, TiO2, TiS2, (Ti2S3), TiN, ZrO2, V2O5, (V2S3), VN, Nb2O5, Ta2O5, (Cr2S3), MoO3, WO3, WS2, MnO2, Fe2O3, MgFe2O4, NiFe2O4, ZnFe2O4, ZnCo2O4, (CuFeS2), ZnO, CdO, CdS, HgS(red), Al2O3, MgAl2O4, ZnAl2O4, Tl2O3, (In2O3), SiO2, SnO2, PbO2

Metal-deficit semiconductors (p-type)

UO2, (VS), (CrS), Cr2O3, (1250 °C, or 2280 °F), MoO2, FeS2, (OsS2), (IrO2), RuO2, PbS

Source: Ref 8 (a) Metallic conductors.

The p-type metal-deficit oxides are nonstoichiometric with cation vacancies present. They will also have some

Schottky and Frenkel defects that add to the ionic conductivity. A typical example is NiO, a cation-deficient oxide that provides the additional electrons needed for ionic bonding and electrical neutrality by donating electrons from the 3d subshells of a fraction of the nickel ions. In this way, for every cation vacancy present in the oxide, two nickelic ions (Ni3+) will be present (Fig. 7). Each Ni3+ has a low-energy positively charged electron hole that electrons from other nickelous ions (Ni2+) can easily move into. The positive or p-type semiconductors carry most of their current by means of these positive holes.

Fig. 7 Illustration of the ionic arrangement in p-type NiO scale. Cation vacancies are indicated as open squares. The N3+ cations are shaded.

Cations can diffuse through the scale from the Ni/NiO interface by cation vacancies, to the NiO/ gas interface where they react with adsorbed oxygen. Electrons migrate from the metal surface, by electron holes, to the adsorbed oxygen atoms,

which then become oxygen anions. In this way, while Ni2+ cations and electrons move outward through the scale toward the gas, cation vacancies and electron holes move inward toward the metal. Consequently, as the scale thickens, the cation vacancies tend to accumulate to form voids at the Ni/NiO interface. The n-type semiconductor oxides have negatively-charged free electrons as the major charge carriers. They may be

either cation excess or anion deficient. Beryllium oxide (BeO) typifies the cation-excess oxides because the beryllium ion (Be2+) is small enough to move interstitially through the BeO scale. Its structure is shown in Fig. 8.

Fig. 8 Illustration of the ionic arrangement in n-type cation-excess BeO. Interstitial cations are shaded; free electrons are indicated as e-.

Oxygen in the gas adsorbs on the BeO surface and picks up free electrons from the BeO to become adsorbed O2- ions, which then react with excess Be2+ ions that are diffusing interstitially from the beryllium metal. The free electrons coming from the metal surface as the beryllium ionizes travel rapidly through vacant high-energy levels. As with p-type oxides, the cation-excess n-type oxides grow at the oxide/gas interface as cations diffuse outward through the scale. Another group of n-type semiconducting oxides is anion deficient, as exemplified by zirconium dioxide (ZrO2). In this case, although most of the cations are contributing four electrons to the ionic bonding, a small fraction of the zirconium cations only contributes two electrons to become the zirconium ion Zr2+. Therefore, to maintain electrical neutrality, an equal number of anion vacancies must be present in the oxide. This arrangement is shown in Fig. 9. The oxide grows at the metal/oxide interface by inward diffusion of O2- through the anion vacancies in the oxide.

Fig. 9 Illustration of the ionic arrangement in n-type anion-deficient ZrO2. Anion vacancies are indicated as open squares; Zr2+ ions are shaded.

Amphoteric Oxides. A number of compounds can be nonstoichiometric with either a deficiency of cations or a

deficiency of anions. An example is lead sulfide (PbS), which has a minimum in electrical conductivity at the stoichiometric composition. Thus, if the composition is Pb 0, and a and b are constants that were derived from the slope and the y-intercept of a straight line curve obtained when the logarithms of the mean pit depth for successively increasing areas on the pipe were plotted against the

logarithms of the corresponding areas. The dependence on area is attributed to the increased chance for the deepest pit to be found when the size of the sample of pits is increased through an increased area of corroded surface. The maximum pit depth D of aluminum exposed to various waters was found to vary as the cube root of time t, as shown in Eq 4 (Ref 10, 15):

D = Kt1/3

(Eq 4)

where K is a constant that is dependent on the composition of the water and the alloy. Equation 4 has been found to apply to several aluminum alloys exposed to different waters. Extreme value probability statistics (Ref 16, 17) have been successfully applied to maximum pit depth data to estimate the maximum pit depth of a large area of material on the basis of the examination of a small portion of that area (Ref 8, 10, 15). The procedure consists of measuring maximum pit depths on several replicate specimens and then arranging the pit depth values in order of increasing rank. A plotting position for each order of ranking is obtained by substitution in the relation M/(n + 1), where M is the order of ranking of the specimen in question, and n is the total number of specimens or values. For example, the plotting position for the second value out of 10 would be 2/(10 + 1) = 0.1818. These values are plotted on the ordinate of extreme value probability paper versus their respective maximum pit depths. A straight line indicates that extreme value statistics are applicable. Extrapolation of the straight line can be used to determine the probability that a specific pit depth will occur or the number of observations that must be made to find a particular pit depth. Loss in Mechanical Properties. If pitting is the predominant form of corrosion and if the density of pitting is

relatively high, the change in a mechanical property can be used to advantage for evaluation of the degree of pitting. The typical properties considered for this purpose are tensile strength, elongation, fatigue strength, impact resistance, and burst pressure (Ref 18, 19). The precautions that must be taken in the application of these mechanical test procedures are covered in most standard methods. However, it must be stressed that the exposed and unexposed specimens should be as close to replicate as possible. Therefore, consideration should be given to such factors as edge effects, direction of rolling, and surface conditions. Representative specimens of the metal are exposed to the same conditions except for the corrosive environment. The mechanical properties of the exposed and unexposed specimens are measured after the exposure, and the difference between the two results is attributed to corrosion damage. Some of these methods are better suited to the evaluation of other forms of localized corrosion, such as intergranular or stress corrosion. Therefore, their limitations must be considered. The often erratic nature of pitting and the location of pits on the specimen can affect results. In some cases, the change in mechanical properties due to pitting may be too small to provide meaningful results. Perhaps one of the most difficult problems is to separate the effects due to pitting from those caused by some other form of corrosion.

References 1.

2. 3. 4.

"Standard Test Methods for Pitting and Crevice Corrosion Resistance of Stainless Steels and Related Alloys by the Use of Ferric Chloride Solution," G 48, Annual Book of ASTM Standards, American Society for Testing and Materials "Standard Test Method for Pitting or Crevice Corrosion of Metallic Surgical Implant Materials," F 746, Annual Book of ASTM Standards, American Society for Testing and Materials "Standard Practice for Conducting Cyclic Potentiodynamic Polarization Measurements for Localized Corrosion," G61, Annual Book of ASTM Standards, American Society for Testing and Materials M. Hubbell, C. Price, and R. Heidersbach, Crevice and Pitting Corrosion Tests for Stainless Steels: A Comparison of Short-Term Tests With Longer Exposures, in Laboratory Corrosion Tests and Standards, STP 866, G.S. Haynes and R. Baboian, Ed., American Society for Testing and Materials, 1985, p 324-336

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

19.

B.E. Wilde, Critical Appraisal of Some Popular Laboratory Tests for Predicting the Localized Corrosion Resistance of Stainless Alloys in Sea Water, Corrosion, Vol 28 (No. 8), Aug 1972, p 283 F.L. LaQue and H.H Uhlig, An Essay on Pitting, Crevice Corrosion and Related Potentials, Mater. Perform., Vol 22 (No. 8), Aug 1983, p 34 "Standard Recommended Practice for Examination and Evaluation of Pitting Corrosion," G 46, Annual Book of ASTM Standards, American Society for Testing and Materials F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1985, p 197 "Standard Recommended Practice for Applying Statistics to Analysis of Corrosion Data," G 16, Annual Book of ASTM Standards, American Society for Testing and Materials B.R. Pathak, Testing in Fresh Waters, Handbook on Corrosion Testing and Evaluation, W.H. Ailor, Ed., John Wiley & Sons, 1971, p 553 P.M. Aziz and H.P. Godard, Influence of Specimen Area on the Pitting Probability of Aluminum, J. Electrochem. Soc., Vol 102, Oct 1955, p 577 G.N. Scott, Adjustment of Soil Corrosion Pit Depth Measurements for Size of Sample, in Proceedings of the American Petroleum Institute, Vol 14, Section IV, American Petroleum Institute, 1934, p 204 M. Romanoff, Underground Corrosion, National Bureau of Standards Circular 579, U.S. Government Printing Office, 1957, p 71 I.A. Denison, Soil Exposure Tests, in The Corrosion Handbook, H.H. Uhlig, Ed., John Wiley & Sons, 1948, p 1048 H.P. Godard, The Corrosion Behavior of Aluminum in Natural Waters, Can J. Chem. Eng., Vol 38, Oct 1960, p 1671 E.J. Gumbel, Statistical Theory of Extreme Values and Some Practical Applications, Applied Mathematics Series 33, U.S. Department of Commerce, 1954 P.M. Aziz, Application of the Statistical Theory of Extreme Values to the Analysis of Maximum Pit Depth Data for Aluminum, Corrosion, Vol 12, Oct 1956, p 495 T.J. Summerson, M.J. Pryor, D.S. Keir, and R.J. Hogan, Pit Depth Measurements as a Means of Evaluating the Corrosion Resistance of Aluminum in Sea Water, in Metals, STP 196, American Society for Testing and Materials, 1957, p 157 R. Baboian, "Corrosion Resistant High-Strength Clad Metal System for Hydraulic Brake Line Tubing," SAE Preprint No. 740290, Society for Automotive Engineers, 1972

Evaluation of Galvanic Corrosion Harvey P. Hack, David Taylor Naval Ship Research and Development Center

Introduction GALVANIC CORROSION, although listed as one of the forms of corrosion, should instead be considered as a type of corrosion mechanism, because any of the other forms of corrosion can be accelerated by galvanic effects. Therefore, any of the tests used for the more conventional forms of corrosion, such as uniform attack, pitting, or stress corrosion, can be used, with modifications, to determine galvanic-corrosion effects. The modifications can be as simple as connecting a second metal to the system or as complex as necessary to evaluate the appropriate parameters. A change in the method of data interpretation is often all that is needed to convert conventional test methods into galvanic-corrosion tests. This article will discuss component, model, electrochemical, and specimen tests. Additional information on galvanic corrosion can be found in the article "General Corrosion" in this Volume.

Component Testing Component testing is an especially useful technique for galvanic corrosion prediction. The materials in a system are often selected primarily for reasons other than galvanic compatibility. In complex components, such as valves or pumps, many

different materials can be used in a geometric configuration that is extremely difficult to model. In more complicated cases, even the most basic prediction, such as which materials will suffer increased corrosion due to galvanic effects, may not be possible from simple laboratory tests. Therefore, component testing becomes the best method for predicting material behavior in complex systems. Conducting component tests for galvanic corrosion is similar to conducting component tests for any other type of corrosion. The same care must be taken to ensure that the materials, the operation of the component, and the environment are similar to those in service. However, one important difference with regard to galvanic corrosion is the relationship between the component being tested and the other elements of the system. For example, it would be a waste of effort to expose a complicated piece of machinery in order to look for galvanic corrosion when the whole device is cathodically protected as a result of being attached to a protected structure. Alternatively, incorrect results would be obtained for the exposure of an isolated bronze mixed-material valve when the ultimate use was in a piping system made of a more noble metal that could accelerate the corrosion of the entire valve galvanically. When outside interactions of this type are possible, the interacting materials must be made part of the corrosion system by exposing the appropriate surface area of those materials electrically connected to, and in the same electrolyte as, the component being tested. The principal advantages of component testing are ease of interpretation of results and the lack of scaling or modeling uncertainties. The disadvantages include high cost and the need for extremely sensitive measures of corrosion damage to obtain results within reasonable time periods.

Modeling Even when the galvanic behavior of panels of the materials of interest is known, the geometrical arrangement of these materials may make galvanic corrosion prediction difficult because of the effects of solution (electrolyte) resistance on the corrosion rates. An example of this is a heat-exchanger tube in a tubesheet. Assuming the tube to be anodic to the tubesheet, areas of the tube near the tubesheet will have low solution resistance to the cathode and will corrode rapidly, but areas away from the tubesheet will have a large solution resistance to the cathodic metal and will therefore corrode more slowly. In the reverse case, in which the tubesheet is anodic to the tube, the areas of the cathodic tube near the tubesheet will drive the galvanic corrosion of the tubesheet much more than distant areas will. Computer Modeling. Geometrical effects can be modeled in computers by using such techniques as finite elements,

boundary elements, and finite differences. The best computer models solve a version of the Laplace equation for the electrolyte surrounding the corroding materials and use the polarization behavior of the material in question as boundary conditions at the metal/electrolyte interface. The analysis is similar to the heat flow analysis, with potential analogous to temperature, current analogous to heat flux, and the polarization boundary condition analogous to a special nonlinear type of temperature-dependent convective flux. The only data this type of model requires are the geometry, electrolyte conductivity, and polarization characteristics of the materials involved. The program generates potentials and current densities as a function of location, either of which can be related back to corrosion rate. The nonlinear boundary conditions make this type of computer modeling difficult to perform unless a large mainframe computer with sufficient computational capabilities is available. Computer modeling provides an excellent predictive tool for geometrical effects; however, it is still seen as less satisfying than physical scale model exposures. Physical scale modeling must model the solution resistance effects and the relative effects of polarization resistance

and solution resistance to obtain accurate geometrical predictive capability. When solution resistance is important, the best type of scale modeling is the scaled conductivity exposure. In this type of exposure, the model is reduced in size by some factor from the original. To maintain proper potential and current distribution scaling, the electrolyte conductivity must also be reduced by the same factor. Any resistive coatings, such as paints, must also have their conductivity scaled similarly. In the case of paints, this can be accomplished by applying a thinner layer, by the same scaling factor used for size, than the thickness used in practice. For example, a one-tenth scale model of a heat exchanger designed to operate in seawater with a conductivity of 4 mho/cm should be placed in seawater diluted to a conductivity of 0.4 mho/cm. In this case, the observed potential and current distributions will be the same between the model and the full-scale heat exchanger. For physical scale modeling, measurements that can be taken include potential distribution by the use of a movable reference electrode, corrosion depth as a function of location, and, if the model design permits, current to different parts of the structure and mass loss of certain model components.

Although less expensive than full-scale component testing, physical scale modeling has many of the disadvantages of component testing. In addition, a great inaccuracy in conductivity scaling stems from the fact that the polarization resistance of the materials in the system of interest is often a function of solution conductivity. Thus, changing solution conductivity may influence polarization resistance sufficiently to make the results of the model inaccurate. There is no experimental way to avoid this shortcoming.

Laboratory Testing Laboratory tests fall into two categories: electrochemical tests, in which the data are analyzed and reported in a way that assists galvanic-corrosion predictions, and specimen exposures, which may or may not be electrochemically monitored. Electrochemical Tests The use of electrochemical techniques to predict galvanic corrosion is summarized in Ref 1. The details that relate to testing techniques are discussed below. Galvanic Series. When the only information needed is which of the materials in the system are possible candidates for

galvanically accelerated corrosion and which will be unaffected or protected, the information from a galvanic series in the appropriate media is useful. Such a series is a list of freely corroding potentials of the materials of interest in the environment of interest arranged in order of potential (Fig. 1). The galvanic series is easy to use and is often all that is required to answer a simple galvanic-corrosion question. The material with the most negative, or anodic, corrosion potential has a tendency to suffer accelerated corrosion when electrically connected to a material with a more positive, or noble, potential. The disadvantages include: • • • •

No information is available on the rate of corrosion Active-passive metals may display two, widely differing potentials Small changes in electrolyte can change the potentials significantly Potentials may be time dependent

Fig. 1 Galvanic series for seawater. Dark boxes indicate active behavior of active-passive alloys.

Creating a galvanic series is a matter of measuring the corrosion potential of various materials of interest in the electrolyte of interest against a reference electrode half-cell, such as saturated calomel. This procedure is described in Ref 2. The details of such factors as meter resistance, reference cell selection, and measurement duration are also addressed in Ref 2.

There is little difference from a normal reading of corrosion potential except for the measurement duration and the creation of a list ordered by potential. To prepare a galvanic series that will be valid for the materials and environment of interest in service, all of the factors that affect the potential of those materials in that environment must be accounted for. These factors include material composition, heat treatment, surface preparation (mill scale, coatings surface finish, etc.), environmental composition (trace contaminants, dissolved gases, etc.), temperature, and flow rate. One important effect is exposure time, particularly on materials that form corrosion product layers. All of the precautions and warnings regarding the generation and use of a galvanic series are given in Ref 2. Polarization Curves. More useful information on the rate of galvanic corrosion can be obtained by investigating the polarization behavior of the materials involved. This can be done by generating stepped potential or potentiodynamic polarization curves or by obtaining potentiostatic information on polarization behavior. The objective is to obtain a good indication of the amount of current required to hold each material at a given potential. Because all materials in the galvanic system must be at the same potential in systems with low solution resistivity, such as seawater, and because the sum of all currents flowing between the materials must equal 0 by Kirchoff's Law, the coupled potential of all materials and the galvanic currents flowing can be uniquely determined for the system. The corrosion rate can then be related to galvanic current by Faraday's Law if the resulting potential of the anodic materials is well away from their corrosion potential, or the corrosion rate can be found as a function of potential by independent measurement.

Potentiodynamic polarization curves are generated by connecting the specimen of interest to a scanning potentiostat. This device applies whatever current is necessary between the specimen and a counter electrode to maintain that specimen at a given potential versus a reference electrode half-cell placed near the specimen. The current required is plotted as a function of potential over a range that begins at the corrosion potential and proceeds in the direction (anodic or cathodic) required by that material. Such curves would be generated for each material of interest in the system. Additional information on the method for generating these curves is available in the article "Laboratory Testing" in this Volume and in Ref 3. The scan rate for potential must be chosen such that sufficient time is allowed for completion of electrical charging at the interface. Potentiodynamic polarization is particularly effective for materials with time-independent polarization behavior. It is fast, relatively easy, and gives a reasonable, quantitative prediction of corrosion rates in many systems. However, potentiostatic techniques are preferred for time-dependent polarization. To establish polarization characteristics for timedependent polarization, a series of specimens is used, each held to one of a series of constant potentials with a potentiostat while the current required is monitored as a function of time. After the current has stabilized or after a pre-selected time period has elapsed, the current at each potential is recorded. Testing of each specimen results in the generation of one potential/current data pair, which gives a point on the polarization curve for that material. The data are then interpolated to trace out the full curve. This technique is very accurate for time-dependent polarization, but is expensive and time consuming. The individual specimens can be weighed before and after testing to determine corrosion rate as a function of potential, thus enabling the errors from using Faraday's Law to be easily corrected. The process of predicting galvanic corrosion from polarization behavior can be illustrated by the example of a steelcopper system. Steel has the more negative corrosion potential and will therefore suffer increased corrosion upon coupling to copper, but the amount of this corrosion must be predicted from polarization curves. If the polarization of each material is plotted as the absolute value of the log of current density versus potential and if the current density axis of each of these curves is multiplied by the wetted surface area of that material in the service application, then the result will be a plot of the total anodic current for steel and the total cathodic current for copper in this application as a function of potential (Fig. 2).

Fig. 2 Prediction of coupled potential and galvanic current from polarization diagrams, i, current; io, exchange current; Ecorr, corrosion potential

Furthermore, when the two metals are electrically connected, the anodic current to the steel must be supplied by the copper; that is, the algebraic sum of the anodic and cathodic currents must equal 0. If the polarization curves for the two materials, normalized for surface area as above, are plotted together, this current condition is satisfied where the two curves intersect. This point of intersection allows for the prediction of the coupled potential of the materials and the galvanic current flowing between them from the intersection point. This procedure works if there is no significant electrolyte resistance between the two metals; otherwise, this resistance must be taken into account in a complex manner that is beyond the scope of this article. Specimen Exposures Specimens for galvanic-corrosion testing include panels, wires, pieces of actual components, and other configurations of the materials of interest that are exposed in a process stream, a simulated service environment, or the actual environment. Specimens of the materials of interest are usually exposed in the same ratios of wetted or exposed areas as in the service application. The different materials are either placed in physical contact to provide electrical connection or are wired together such that the current between the materials can be monitored, usually as a function of time. Seldom can the effects of electrolyte resistance be included in this type of test, and the resistance is usually kept extremely low by appropriate relative placement of the materials. Immersion. There are virtually no standardized tests for galvanic corrosion under immersion conditions, partly because the type of information needed, the extent of modeling of the service situation, and the type of system studied vary widely. This makes development of a standard test difficult. However, some general guidelines for galvanic-corrosion specimen testing in liquid electrolytes are given in Ref 4.

Immersion testing always involves an electrical connection between at least two dissimilar metals. This is usually accomplished with a wire, as in Fig. 3, although threaded mounting rods have also been used successfully for electrical connection, such as the assembly shown in Fig. 4. Soldered or brazed connections have the best electrical integrity.

Fig. 3 Typical galvanic-corrosion immersion test setup using wire connections

Fig. 4 Typical galvanic-corrosion test specimen using a threaded rod for mounting and electrical connection

The electrolyte must be excluded from the contact area by applying a sealant, such as silicone or epoxy; by keeping the joint area out of the electrolyte by partial immersion of the specimen, in which case a waterline area is created; or by use of a tube and gasket or O-ring seal in the case of a threaded mounting rod. Mounting the specimen in a specially formulated epoxy has been found to be effective in minimizing crevice corrosion while maintaining a dry electrical connection. However, selection of the best epoxy formulation is difficult. Care must be taken that the sealant or gasketing material is stable in the electrolyte being studied. Almost any sealing procedure will create a potential area for crevice corrosion; thus, it is very difficult to study galvanic behavior independent of crevice corrosion behavior (see the article "Evaluation of Crevice Corrosion" in this Volume). Control specimens may be run with similar crevices and no electrical connection, but because the reproducibility of crevice corrosion behavior is not good, data scatter will be large. Under some circumstances, the galvanic effect of importance may be the acceleration of crevice corrosion attack. The relative wetted surface areas of the materials being tested will have an important effect on the magnitude of the galvanic attack. The larger the cathode-to-anode area ratio is, the larger the attack will be; therefore, it would at first seem

reasonable to accelerate the test by using a large ratio. This should not be done, because accelerating the attack may also change the mechanism of the attack, which would lead to erroneous conclusions. It is far more appropriate to use more accurate measurement techniques to determine the extent of the attack over a short period than to accelerate the test to obtain measurable attack quickly. If soldered or brazed connections are used for electrical connection, subsequent evaluation by weight loss is difficult; therefore, if weight loss is to be used to measure attack, threaded and sealed connections are preferred. Measurement of the electrical current flowing between the metals can give a very sensitive indication of the extent of the galvanic attack and will allow the attack to be monitored over time. Coupled potential is another parameter that is useful to follow during the course of the exposure. The effect of exposure time on the rate of attack should be properly considered. Initially high rates of galvanic attack may decay to acceptable levels in a short period of time, or initially low attack rates may increase to unacceptable levels over time. Current can be measured by inserting a resistor of 1 to 10 in the current circuit and measuring the potential decrease across this resistor with a voltmeter having a resistance of at least 1000 . The resistor should be sized such that the voltage across it does not exceed 5 mV; thus, the resistor will not significantly impede the current flow. Alternatively, a zero-resistance ammeter can be used instead of the resistor. This device is an operational amplifier connected to maintain 0 V across its input terminals (Fig. 5). A current-measuring resistor, placed in the feedback circuit, may be as large as the amplifier will allow, thus enabling currents in the nanoampere range to be easily measured. One simple way of creating a zero-resistance ammeter is by using a potentiostat with the counter electrode and reference electrode leads shorted together and set to a working electrode potential of 0 V (Fig. 6).

Fig. 5 Basic circuit for a zero-resistance ammeter

Fig. 6 Conversion of a potentiostat into a zero-resistance ammeter, WE, working electrode; CE, counter electrode; RE, reference electrode

The importance of electrolyte flow in galvanic corrosion should not be overlooked in establishing the test procedure. A test apparatus should be used that reproduces the flow under service conditions. If this is not possible and flow must be scaled, the exact scaling method will depend on the assumed corrosion processes. Cathodic reactions, such as oxygen reduction, that are controlled by diffusion through a fluid boundary layer are likely to be properly scaled by reproducing the hydrodynamic boundary layer of the service application in the test. This should reproduce the diffusion boundary layer that controls the reaction. Alternatively, the rate of reactions controlled by films such as anodic brightening of copper alloys, other passivation-type reactions, or control by calcareous deposit formation in seawater, may depend more on the shear stress at the surface

required to strip off the film. In this case, surface shear stress may be a better hydrodynamic parameter to reproduce in the test. Many different types of flow apparatus have been used, such as concentric tubes, in-line tubes, rotating cylinders, rotating ring-disks, rectangular flow channels with specimens mounted in the walls, and circulating foils. Each of these has its own hydrodynamic peculiarities, but one common area of concern is the leading edge of the specimen. It is difficult, even for specimens recessed in the walls of a flow channel, to avoid a step or gap that can create unexpected hydrodynamic conditions at the specimen surfaces downstream. Also, mounting to allow electrical connection must be considered, and crevice effects are essentially impossible to eliminate. Atmospheric Tests. General testing guidelines become more complex when considering atmospheric or cabinet

exposures. Testing in these environments differs markedly from immersion tests in a number of ways, most of which involve the insufficiency of electrolyte. Many of the variables that influence the behavior of specimens in the atmosphere are discussed in Ref 5. The thinness of the electrolyte film and the normally low conductivity of the electrolyte combine to limit galvanic effects to within about 5 mm (0.2 in.) of the dissimilar-metal interface. Thus, area ratio effects become relatively unimportant. Sealing the electrical connections becomes relatively less important than in immersion testing if the connections are more than 5 mm (0.2 in.) from the area to be evaluated and if corrosion products will not interfere with the continuity of the connection. Periodic checks of electrical continuity in atmospheric galvanic-corrosion tests are recommended. Geometrical effects also become unimportant, except as they relate to the entrapment of moisture. However, specimen orientation effects must be considered. The behavior of the galvanic couples will depend on whether they are exposed on the top or the bottom of the panel, whether they are sheltered or not, or other considerations, such as the effect of specimen mass on condensation. Because there are no standardized tests for galvanic corrosion immersed in electrolytes, it is somewhat surprising that several standard tests have emerged for atmospheric galvanic corrosion, even though less testing has been done in this area. One of these tests is an International Organization for Standardization (ISO) standard (Ref 6) and is also being developed by the American Society for Testing and Materials (ASTM). This test uses a 100- × 400-mm (4- × 16-in.) panel of the anodic material to which a 50- × 100-mm (2- × 4-in.) strip of the cathodic material is bolted (Fig. 7). After exposure, the anodic material can be evaluated for material degradation by weight loss and other corrosion measurements as well as by degradation of such mechanical properties as ultimate tensile strength.

Fig. 7 Specimen configuration for the ISO test for atmospheric galvanic corrosion. 1, anodic plate, 1 piece; 2, cathodic plate, 2 pieces; 3, microsection, 2 pieces; 4, tensile test specimen; 5, bolt, 8 × 40 mm, 2 pieces; 6 washers, 1 mm thick, 16 mm diam, 4 pieces; 7, insulating washers, 1 to 3 mm thick, 18 to 20 mm diam, 4 pieces; 8, insulating sleeve, 2 pieces; 9, nut, 2 pieces. Dimensions given in millimeters

This test is relatively easy to perform, but requires the availability of plate of the materials of interest and a prior knowledge of which material is anodic. Like any atmospheric galvanic-corrosion test, crevice effects cannot be adequately separated from galvanic effects in some cases; therefore, a coating is sometimes applied between the anode and cathode plates. The disadvantage of this test is the time required to obtain results; for systems with moderate corrosion rates, exposures of 1 to 5 years are not unusual. Another commonly used atmospheric galvanic-corrosion test is the wire-on-bolt test, sometimes referred to as the CLIMAT test (Ref 7, 8, 9). In this test, a wire of the anodic material is wrapped around a threaded rod of the cathodic material (Fig. 8). Because corrosion can be rapid in this test, exposure duration should usually be limited. This makes the test ideal for measuring atmospheric corrosivity as well as material corrosion properties. Not all materials of interest are available in the required wire and threaded rod forms, and analysis is usually restricted to weight loss measurement and observation of pitting. When the required materials are available, this test is less expensive and easier to conduct than the ISO plate test.

Fig. 8 Specimen configuration for the wire-on-bolt test for atmospheric galvanic corrosion

A third atmospheric galvanic-corrosion test has been used extensively by ASTM, but has not been standardized. This test (Ref 10) involves the use of 25-mm (1-in.) diam washers of the materials of interest bolted together as shown in Fig. 9. The bolt that holds the washers together can also be used to secure the assembly in position. After exposure, the washers can be disassembled for weight loss determination. The materials needed for this test are not as large as those for the ISO plate test, but it can take as long and cannot provide mechanical properties data.

Fig. 9 Specimen configuration for the washer test for atmospheric galvanic corrosion

References R. Baboian, Electrochemical Techniques for Predicting Galvanic Corrosion, in Galvanic and Pitting Corrosion--Field and Laboratory Studies, STP 576, American Society for Testing and Materials, 1976, p 5-19 2. "Standard Guide for Development and Use of a Galvanic Series for Predicting Galvanic Corrosion Performance," G 82, Annual Book of ASTM Standards, American Society for Testing and Materials 3. "Standard Reference Test Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements," G 5, Annual Book of ASTM Standards, American Society for Testing and Materials 4. "Standard Guide for Conducting and Evaluating Galvanic Corrosion Tests in Electrolytes," G 71, Annual Book of ASTM Standards, American Society for Testing and Materials 5. "Standard Practice for Conducting Atmospheric Corrosion Tests of Metals," G 50, Annual Book of ASTM Standards, American Society for Testing and Materials 6. "Corrosion of Metals and Alloys--Determination of Bi-Metallic Corrosion in Outdoor Exposure Corrosion Tests," ISO 7441, International Standards Organization 7. K.G. Compton, A. Mendizza, and W.W. Bradley, Atmospheric Galvanic Couple Corrosion, Corrosion, Vol II, 1955, p 383 8. H.P. Godard, Galvanic Corrosion Behavior of Aluminum in the Atmosphere, Mater. Prot., Vol 2 (No. 6), 1963, p 38 9. D.P. Doyle and T.E. Wright, Rapid Methods for Determining Atmospheric Corrosivity and Corrosion Resistance, in Atmospheric Corrosion, W.H. Aylor, Ed., John Wiley & Sons, 1982, p 227 10. R. Baboian, Final Report on the ASTM Study: Atmospheric Galvanic Corrosion of Magnesium Coupled to Other Metals, in Atmospheric Factors Affecting the Corrosion of Engineering Metals, STP 646, S.K. Coburn, Ed., American Society for Testing and Materials, 1978, p 17-29 1.

Evaluation of Intergranular Corrosion Richard A. Corbett and Brian J. Saldanha, Corrosion Testing Laboratories, Inc.

Introduction IN THE ARTICLE "Localized Corrosion" in this Volume, intergranular corrosion is defined and the mechanisms are described. It is the purpose of this article to discuss when to evaluate for susceptibility to intergranular attack and how to determine which of the various evaluation tests are applicable. However, it may first be necessary to review the methodology of intergranular corrosion and its effect on the various alloy families. Most alloys are susceptible to intergranular attack when exposed to specific environments. This is because grain boundaries are sites for precipitation and segregation, which makes them chemically and physically different from the grains themselves. Intergranular attack is defined as the selective dissolution of grain boundaries, or closely adjacent regions, without appreciable attack of the grains themselves. This is caused by potential differences between the grainboundary region and any precipitates, intermetallic phases, or impurities that form at the grain boundaries. The actual mechanism differs with each alloy system. Precipitates that form as a result of the exposure of metals at elevated temperatures (for example, during production, fabrication, and welding) often nucleate and grow preferentially at grain boundaries. If these precipitates are rich in alloying elements that are essential for corrosion resistance, the regions adjacent to the grain boundary are depleted of these elements. The metal is thus sensitized and is susceptible to intergranular attack in a corrosive environment. For example, in austenitic stainless steels such as AISI type 304, the cause of intergranular attack is the precipitation of chromium-rich carbides [(Cr, Fe)23C6] at grain boundaries. These chromium-rich precipitates are surrounded by metal that is depleted in chromium; therefore, they are more rapidly attacked at these zones than on undepleted metal surfaces. Impurities that segregate at grain boundaries may promote galvanic action in a corrosive environment by serving as anodic or cathodic sites. Therefore, this would affect the rate of dissolution of the alloy matrix in the vicinity of the grain boundary. An example of this is found in aluminum alloys when they contain intermetallic compounds, such as Mg5Al8 and CuAl2, at the grain boundaries. During exposures to chloride solutions, the galvanic couples formed between these precipitates and the alloy matrix can lead to severe intergranular attack. Susceptibility to intergranular attack depends on the corrosive solution and on the extent of intergranular precipitation, which is a function of alloy composition, fabrication, and heat treatment parameters. Corrosion tests for evaluating the susceptibility of an alloy to intergranular attack are typically classified as either simulated-service or accelerated tests. The first laboratory tests for detecting intergranular attack were simulated-service exposures. These were first observed and used in 1926 when intergranular attack was detected in an austenitic stainless steel in a copper sulfate-sulfuric acid (CuSO4-H2SO4) pickling tank (Ref 1). Another simulated-service test for alloys intended for service in nitric acid (HNO3) plants is described in Ref 2. In this case, for accelerated results, iron-chromium alloys were tested in a boiling 65% HNO3 solution. Over the years, specific tests have been developed and standardized for evaluating the susceptibility of various alloys to intergranular attack. For example, tests for the low-alloy austenitic stainless steels have been standardized by the American Society for Testing and Materials (ASTM) in Standard A 262, with its various practices (A to E). Practice A is a screening test that uses an electrolytic oxalic acid etch combined with metallographic examination. The other practices involve exposing the material (possibly after a sensitizing treatment) to boiling solutions of 65% HNO3, acidified ferric sulfate (Fe2(SO4)3) solution, nitric-hydrofluoric acid (HNO3-HF) solution, or acidified CuSO4 solution, depending on the specific alloy and its application. Similar ASTM tests have been developed for other higher-alloyed stainless steels, ferritic stainless steels, high nickel-base alloys, and aluminum alloys (Table 1).

Table 1 Appropriate evaluation tests and acceptance criteria for wrought alloys UNS number

Alloy name

Applicable tests (ASTM standards)

Sensitizing treatment

Exposure time, h

Criteria for passing, appearance or maximum allowable corrosion rate, mm/month (mils/month)

S43000

Type 430

Ferric sulfate (A 763X)

None

24

1.14 (45)

S44600

Type 446

Ferric sulfate (A 763X)

None

72

0.25 (10)

S44625

26-1

Ferric sulfate (A 763X)

None

120

0.05 (2) and no significant grain dropping

S44626

26-1S

Cupric sulfate (A 763Y)

None

120

No significant grain dropping

S44700

29-4

Ferric sulfate (A 763X)

None

120

No significant grain dropping

S44800

29-4-2

Ferric sulfate (A 763X)

None

120

No significant grain dropping

S30400

Type 304

Oxalic acid (A 262-A)

None

...

(a)

120

0.1 (4)

...

(a)

240

0.05 (2)

Ferric sulfate (A 262B)

S30403

Type 304L

Oxalic acid (A 262-A)

1 h at 675 °C (1250 °F)

Nitric acid (A 262-C)

S30908

Type 309S

Nitric acid (A 262-C)

None

240

0.025 (1)

S31600

Type 316

Oxalic acid (A 262-A)

None

...

(a)

120

0.1 (4)

...

(a)

Ferric sulfate (A 262B)

S31603

Type 316L

Oxalic acid (A 262-A)

1 h at 675 °C (1250 °F)

Ferric sulfate (A 262B)

S31700

Type 317

Oxalic acid (A 262-A)

None

Ferric sulfate (A 262B)

S31703

Type 317L

Oxalic acid (A 262-A)

1 h at 675 °C (1250 °F)

Ferric sulfate (A 262B)

120

0.1 (4)

...

(a)

120

0.1 (4)

...

(a)

120

0.1 (4)

S32100

Type 321

Nitric acid (A-262-C)

1 h at 675 °C (1250 °F)

240

0.05 (2)

S34700

Type 347

Nitric acid (A 262-C)

1 h at 675 °C (1250 °F)

240

0.05 (2)

N08020

20Cb-3

Ferric sulfate (G 28A)

1 h at 675 °C (1250 °F)

120

0.05 (2)

N08904

904L

Ferric sulfate (G 28A)

None

120

0.05 (2)

N08825

Incoloy 825

Nitric acid (A 262-C)

1 h at 675 °C (1250 °F)

240

0.075 (3)

N06007

Hastelloy G

Ferric sulfate (G 28A)

None

120

0.043 (1.7) sheet, plate, and bar; 0.05 (2) pipe and tubing

N06985

Hastelloy G-3

Ferric sulfate (G 28A)

None

120

0.043 (1.7) sheet, plate, and bar; 0.05 (2) pipe and tubing

N06625

Inconel 625

Ferric sulfate (G 28A)

None

120

0.075 (3)

N06690

Inconel 690

Nitric acid (A 262-C)

1 h at 540 °C (1000 °F)

240

0.025 (1)

N10276

Hastelloy C-276

Ferric sulfate (G 28A)

None

24

1 (40)

N06455

Hastelloy C-4

Ferric sulfate (G 28A)

None

24

0.43 (17)

N06110

Allcorr

Ferric sulfate (G 28B)

None

24

0.64 (25)

N10001

Hastelloy B

20% Hydrochloric acid

None

24

0.075 (3) sheet, plate, and bar; 0.1 (4) pipe and tubing

N10665

Hastelloy B-2

20% Hydrochloric acid

None

24

0.05 (2) sheet, plate, and bar; 0.086 (3.4) pipe and tubing

A95005A95657

Aluminum Association 5xxx alloys

Concentrated nitric acid (G 67)

None

24

(b)

(a) See A 262, practice A.

(b) See G 67, section 4.1.

The Purpose of Testing There is a perception in much of the industry that testing for susceptibility to intergranular attack is equivalent to evaluating the resistance of the alloy to general and localized corrosion. Although the tests used for evaluating susceptibility to intergranular attack are severe, they are not intended to duplicate conditions for the wide range of chemical exposures present in an industrial plant, even though some of these tests simulate service conditions. Testing for susceptibility to intergranular attack, however, is useful for determining whether the correct material, in the proper metallurgical condition, has been supplied by a vendor. There are some problems associated with quality assurance programs for purchased materials. Such programs are sometimes based on faith in what is supplied by a vendor or production mill and what is certified in the documentation sent along with the material. However, such confidence may be misplaced. For example, there have been a number of accounts in which alloys have been substituted, resulting in premature failure. In one case, this occurred when Hastelloy B valves were substituted for the Hastelloy C-276 valves that were ordered to handle a hypochlorite solution. The Hastelloy B valves failed in about 3 months. In addition, there are many examples in which the material supplied does not conform to its certified analysis. The problem of getting reliable certified analyses increases when documentation goes from a mill to an alloy supplier. In one case, for example, AISI type 304L stainless steel valves were ordered, but the vendor, having few orders for this alloy, substituted type 316L stainless steel valves and sent certifications that purposely omitted the molybdenum analysis. Normally, this would have been a good substitution for improved corrosion resistance at a bargain price, but these valves were destined for hot, concentrated HNO3 service and failed prematurely. These are just two examples of using a material that is incorrect or is not in the proper metallurgical condition; such problems, of course, are not limited to stainless steels. It should be realized that errors do occur and that for critical service the specified alloys must be in optimum metallurgical condition to resist intergranular attack and other forms of corrosion associated with precipitates at the grain boundaries.

Tests for Stainless Steels and Nickel-Base Alloys The austenitic and ferritic stainless steels, as well as most nickel-base alloys, are generally supplied in a heat-treated condition such that they are free of carbide precipitates that are detrimental to corrosion resistance. However, these alloys are susceptible to sensitization from welding, improper heat treatment, and service in the sensitizing temperature range. The phenomenon of sensitization of these alloys is discussed further in the article "Corrosion of Stainless Steels," "Corrosion of Weldments," and "Corrosion of Nickel-Base Alloys" in this Volume. The theory and application of acceptance tests for detecting the susceptibility of stainless steels and nickel-base alloys to intergranular attack are extensively reviewed in Ref 3 and 4. It would be repetitive to review this work other than to

discuss why and when it is necessary to evaluate the susceptibility of alloys to this form of attackand to discuss acceptable criteria for the tests used. Because sensitized alloys may inadvertently be used, acceptance tests are implemented as a quality control check to evaluate stainless steels and nickel-base alloys when: • • •

Different alloys, or regular carbon types of the specified alloy, are submitted for the low-carbon grades (for example, type 316 substituted for type 316L) and are involved in welding or heat treating An improper heat treatment during fabrication results in the formation of intermetallic phases The specified limits for carbon and/or nitrogen contents of an alloy are inadvertently exceeded

Some standard tests include acceptance criteria, but others do not (Ref 3). Some type of criterion in needed that can clearly separate material susceptible to intergranular attack from that resistant to attack. Table 1 lists evaluation tests and acceptance criteria for various stainless steels and nickel-base alloys that have been used by the DuPont Company, the U.S. Department of Energy, and others in the chemical-processing industry. Identifying such rates still leaves the buyer and seller free to agree on a rate that meets their particular needs.

Tests for Aluminum Alloys The electrochemically active paths at the grain boundaries of aluminum alloy materials can be either the solid solution or closely spaced anodic second-phase particles. The identities of the specific active paths vary with the alloy composition and metallurgical condition of the product, as discussed in the article "Corrosion of Aluminum and Aluminum Alloys" in this Volume and in Ref 5 and 6. The most serious forms of such structure-dependent corrosion are stress-corrosion cracking (SCC) and exfoliation. Stress-corrosion cracking requires the presence of a sustained tensile stress, and exfoliation occurs only in wrought products with a directional grain structure. Not all materials that are susceptible to intergranular attack, however, are susceptible to SCC or exfoliation. Therefore, specific tests are required for the latter (see the article "Evaluation of Exfoliation Corrosion" in this Volume). Strain-Hardened 5xxx Alloys. Alloys in this series that contain more than about 3% Mg are rendered susceptible to

intergranular attack (sensitized) by certain manufacturing conditions or after being subjected to elevated temperatures up to about 175 °C (350 °F). This is the result of the continuous grain-boundary precipitation of the highly anodic Mg2Al3 phase, which corrodes preferentially in most corrosive environments. The ASTM standard G 67 is a method that provides a quantitative measure of the susceptibility to intergranular attack of these alloys (Ref 7). This method consists of immersing test specimens in concentrated HNO3 at 30 °C (85 °F) for 24 h and determining the mass loss per unit area as the measure of intergranular susceptibility. When this second phase is precipitated in a relatively continuous network along grain boundaries, the preferential attack of the network causes whole grains to drop out of the specimens. Such dropping out causes relatively large mass losses of the order of 25 to 75 mg/cm2, although specimens of intergranular-resistant materials lose only about 1 to 15 mg/cm2. Intermediate mass losses occur when the precipitate is randomly distributed. The parallel relationship between the susceptibility to intergranular attack and to SCC and exfoliation of these particular alloys makes ASTM G 67 a useful screening test for alloy and process development studies. A problem arises, however, in selecting a pass-or-fail value in relation to the performance of intermediate materials in environments other than HNO3. Heat-Treated High-Strength Alloys. Materials problems caused by SCC, exfoliation, or corrosion fatigue of the

early 2xxx (aluminum-copper) alloys were identified with intergranular corrosion, and the blame came to be associated with improper heat treatment. In 1944, an accelerated test for detecting susceptibility to intergranular corrosion was incorporated into a U.S. Government specification for the heat treatment of aluminum alloys. This specification has been superseded by the current Military Specification MIL-H-6088F. Tests are required for periodic monitoring of 2xxx and 7xxx series alloys in all rivets and fastener components as well as sheet, bar, rod, wire, and shapes under 6.4 mm (0.25 in.) thick. Specimen preparation, test procedure, and evaluation criteria are detailed in Ref 8. Other Tests for Aluminum Alloys. The volume of hydrogen evolved upon immersion of etched 2xxx series (aluminum-copper-magnesium) aluminum alloys in a solution containing 3% sodium chloride (NaCl) and 1% hydrochloric acid (HCl) for a stipulated time has been used as a quantitative measure of the severity of intergranular

attack. A problem with this approach (which is quite valid) was that the correlation between the amount (or the rate) of hydrogen evolved is influenced by a number of factors, including alloy composition, temper, and grain size (Ref 9, 10). Applied current or potential in neutral chloride solutions (for example, 0.1 N NaCl) provides another direct method of assessing the degree of susceptibility to intergranular attack when accompanied by a microscopic examination of metallographic sections (Ref 9, 11, 12). More sophisticated electrochemical approaches for studying systems involving active-path corrosion use potentiodynamic methods. Tests for SCC are discussed in the article "Evaluation of StressCorrosion Cracking" in this Volume.

Tests for Other Alloys Although intergranular corrosion is present to some extent in alloys other than stainless and aluminum alloys, incidences of attack associated with this form of corrosion are few and are generally not of practical importance. Therefore, no attempts have been made to develop and standardize specific tests for detecting susceptibility to intergranular corrosion in these alloys. However, certain media have been commonly used for evaluating the susceptibility to intergranular corrosion of magnesium, copper, lead, and zinc alloys (Ref 13). These media are listed in Table 2. The presence or absence of attack in these tests is not necessarily a measure of the performance of the material in other corrosive environments. Table 2 Media for testing susceptibility to intergranular corrosion Alloy

Medium

Concentration, %

Temperature, °C ( °F)

Magnesium alloys

Sodium chloride plus hydrochloric acid

...

Room

Copper alloys

Sodium chloride plus sulfuric or nitric acid

1 NaCl, 0.3 acid

40-50 (105-120)

Lead alloys

Acetic acid or hydrochloric acid

...

Room Room

Zinc alloys

Humid air

100% relative humidity

95 (205)

Source: Ref 13 Magnesium Alloys. There are rare instances of reported intergranular corrosion of magnesium alloys, as in the case of chronic acid contaminated with chlorides or sulfates. The copper alloys that appear to be the most susceptible to intergranular corrosion are Muntz metal, admiralty metal,

aluminum brasses, and silicon bronzes. Admiralty alloys have been observed to suffer intergranular corrosion upon exposure to saline cooling waters, although the incidence of attack is very low. The antimonial grades are reportedly superior to the arsenical grades in this respect. Similarly, arsenical and phosphorized grades of aluminum brass have been observed to suffer intergranular corrosion in seawater-type exposures. Zinc die casting alloys have reportedly suffered intergranular corrosion in certain steam atmospheres. A laboratory

test for simulating service failures, and particularly for alloy development work, has been is use for testing the susceptibility of zinc-base die castings to intergranular corrosion (Ref 14). The test consists of exposing samples to air at 95 °C (205 °F) and 100% relative humidity for 10 days under conditions permitting condensation of hot water on the metal. Susceptibility to intergranular corrosion is assessed by the effect on mechanical properties, such as impact strength. Experience has shown that castings with mechanical properties and dimensions that are not significantly altered by the 10-day exposure in this test will not suffer intergranular attack in atmospheric service.

References 1. 2. 3. 4. 5.

6. 7. 8. 9. 10. 11. 12. 13. 14.

W.H. Hatfield, J. Iron Steel Inst., Vol 127, 1933, p 380-383 W.R. Huey, Trans. Am. Soc. Steel Treat., Vol 18, 1930, p 1126-1143 M.A. Streicher, in Intergranular Corrosion of Stainless Alloys, STP 656, American Society for Testing and Materials, 1978, p 3-84 M. Henthorne, in Localized Corrosion--Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 66-119 T.J. Summerson and D.O. Sprowls, Corrosion Behavior of Aluminum Alloys, in Aluminum Alloys: Their Physical and Mechanical Properties, Vol III, E.A. Starke, Jr. and T.H. Sanders, Jr., Ed., Engineering Materials Advisory Services Ltd., 1986, p 1576-1662 B.W. Lifka and D.O. Sprowls, in Localized Corrosion--Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 120-144 H.L. Craig, Jr., in Localized Corrosion--Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 17-37 "Heat Treatment of Aluminum Alloys," Military Specification MIL-H-6088F, United States Government Printing Office F.A. Champion, Corrosion Testing Procedures, 2nd ed., John Wiley & Sons, 1965, p 365, 366 G.J. Schafer, J. Appl. Chem., Vol 10, 1960, p 138 S. Ketcham and W. Beck, Corrosion, Vol 16, 1960, p 37 M.K. Budd and F.F. Booth, Corrosion, Vol 18, 1962, p 197 F.A. Champion, Corrosion Testing Procedures, John Wiley & Sons, 1964 H.H. Uhlig, Corrosion Handbook, John Wiley & Sons, 1948

Evaluation of Exfoliation Corrosion Donald O. Sprowls, Consultant

Introduction EXFOLIATION is a structure-dependent form of localized (usually) intergranular corrosion that is most familiar in certain alloys and tempers of aluminum. The mechanism of exfoliation is described in the article "Corrosion of Aluminum and Aluminum Alloys" in this Volume. The occurrence of exfoliation in susceptible materials is influenced to a marked degree by environmental conditions. Figure 1 illustrates the broad range of behavior in different types of atmospheres. For example, forged truck wheels made of an aluminum-copper alloy (2024-T4) give corrosion-free service for many years in the warm climates of the southern and western United States, but they exfoliate severely in only 1 or 2 years in the northern states, where deicing salts are used on the highways during the winter months.

Fig. 1 Comparison of exfoliation of aluminum alloy 2124 (heat treated to be susceptible; EXCO ED rating) in various seacoast and industrial environments. Specimens were 13-mm (

-in.) plate. Source: Ref 1

Accelerated laboratory tests do not precisely predict long-term corrosion behavior; however, answers are needed quickly in the development of new materials. For this reason, accelerated tests are used to screen candidate alloys before conducting atmospheric exposures or other field tests. They are also sometimes used for quality control tests. Several new laboratory tests for exfoliation corrosion have been standardized in recent years under the jurisdiction of American Society for Testing and Materials (ASTM) Committee G-1 on the Corrosion of Metals.

Spray Tests Three cyclic acidified salt spray tests have been widely used in the aluminum and aircraft industries. These are covered by the procedures described in Annexes A2, A3, and A4 of ASTM G 85 (Ref 2). This standard does not prescribe the particular practice, test specimen, or exposure period to be used for a specific product, nor does it define the interpretation to be given to the test results. These considerations are prescribed by specifications covering the material or product being tested or by agreement between the purchaser and the seller. Annex A2 describes a cyclic salt spray test that uses a 5% sodium chloride (NaCl) solution acidified to pH 3 with acetic

acid in a spray chamber at a temperature of 49 °C (120 °F). This test is applicable for exfoliation testing of 2xxx (drybottom operation) and 7xxx (wet-bottom operation; that is, with approximately 25 mm, or 1 in., of water present in the bottom of the test chamber) aluminum alloys with a test duration of 1 to 2 weeks. Results with 7075 and 7178 alloys in various metallurgical conditions have been shown to correlate well with results obtained in a seacoast atmosphere (4-year exposure at Point Judith, RI) (Ref 3). Annex A3 describes another cyclic-salt spray test that uses a 5% synthetic sea salt solution acidified to pH 3 with acetic

acid in a spray chamber at a temperature of 49 °C (120 °F). The test is applicable to the production control of exfoliationresistant tempers of the 2xxx, 5xxx, and 7xxx aluminum alloys (Ref 4, 5). Wet-bottom operating conditions are recommended with test durations of 1 to 2 weeks. Annex A4 describes a salt-sulfur dioxide (SO2) spray test that uses either 5% NaCl or 5% synthetic sea salt solution in a

spray chamber at a temperature of 35 °C (95 °F). The spray may be either cyclic or constant; this, along with the type of salt solution and the test duration, is subject to agreement between the purchaser and the seller. The test is applicable for 2xxx and 7xxx aluminum alloys. Test duration is 2 to 4 weeks (Ref 1).

Spray Tests Three cyclic acidified salt spray tests have been widely used in the aluminum and aircraft industries. These are covered by the procedures described in Annexes A2, A3, and A4 of ASTM G 85 (Ref 2). This standard does not prescribe the particular practice, test specimen, or exposure period to be used for a specific product, nor does it define the interpretation to be given to the test results. These considerations are prescribed by specifications covering the material or product being tested or by agreement between the purchaser and the seller. Annex A2 describes a cyclic salt spray test that uses a 5% sodium chloride (NaCl) solution acidified to pH 3 with acetic

acid in a spray chamber at a temperature of 49 °C (120 °F). This test is applicable for exfoliation testing of 2xxx (drybottom operation) and 7xxx (wet-bottom operation; that is, with approximately 25 mm, or 1 in., of water present in the bottom of the test chamber) aluminum alloys with a test duration of 1 to 2 weeks. Results with 7075 and 7178 alloys in various metallurgical conditions have been shown to correlate well with results obtained in a seacoast atmosphere (4-year exposure at Point Judith, RI) (Ref 3). Annex A3 describes another cyclic-salt spray test that uses a 5% synthetic sea salt solution acidified to pH 3 with acetic

acid in a spray chamber at a temperature of 49 °C (120 °F). The test is applicable to the production control of exfoliationresistant tempers of the 2xxx, 5xxx, and 7xxx aluminum alloys (Ref 4, 5). Wet-bottom operating conditions are recommended with test durations of 1 to 2 weeks. Annex A4 describes a salt-sulfur dioxide (SO2) spray test that uses either 5% NaCl or 5% synthetic sea salt solution in a

spray chamber at a temperature of 35 °C (95 °F). The spray may be either cyclic or constant; this, along with the type of salt solution and the test duration, is subject to agreement between the purchaser and the seller. The test is applicable for 2xxx and 7xxx aluminum alloys. Test duration is 2 to 4 weeks (Ref 1).

Visual Assessment of Exfoliation One of the problems in evaluating the extent of damage due to exfoliation corrosion is the lack of a generally acceptable numerical measure of the corrosion. Therefore, the usual practice, as noted above for ASTM G 34 and G 66, is to assign visual ratings reference to standard photographs, as shown in Fig. 2, 3, 4, and 5. The use of such ratings requires the inspector to make a judgment; because of this, the ratings are subject to variation among different inspectors. Further, the lack of numerical measures of the corrosion damage hampers analysis of test results when a number of test materials must be compared. One approach is to assign numbers as substitutes for the letters. It is proposed for this purpose that a geometric scale (such as EA = 1, EB = 2, EC = 4, ED = 8) would be more consistent with the differences in damage illustrated by the standard photos than successive numbers would be (that is, 1, 2, 3, 4).

References 1.

2. 3.

4. 5.

6.

S.J. Ketcham and E.J. Jankowsky, Developing an Accelerated Test: Problems and Pitfalls, in Laboratory Corrosion Tests and Standards, STP 866, G.S. Haynes and R. Baboian, Ed., American Society for Testing and Materials, 1985, p 14-23 "Standard Practice for Modified Salt Spray (Fog) Testing," G 85, Annual Book of ASTM Standards, American Society for Testing and Materials B.W. Lifka and D.O. Sprowls, Relationship of Accelerated Test Methods for Exfoliation Resistance in 7xxx Aluminum Alloys with Exposure to a Seacoast Atmosphere, in Corrosion in Natural Environments, STP 558, American Society for Testing and Materials, 1974, p 306-333 H.B. Romans, An Accelerated Laboratory Test to Determine the Exfoliation Corrosion Resistance of Aluminum Alloys, Mater. Res. Stand., Vol 9 (No. 11), 1969, p 31-34 S.J. Ketcham and P.W. Jeffrey, Exfoliation Corrosion Testing of 7178 and 7075 Aluminum Alloys, in Localized Corrosion--Cause of Metal Failure, STP 516, American Society for Testing and Materials, 1972, p 273-302 D.O. Sprowls, J.D. Walsh, and M.B. Shumaker, Simplified Exfoliation Testing of Aluminum Alloys, in Localized Corrosion--Cause of Metal Failure, STP 516, American Society for Testing and Materials,

1972, p 38-65 7. "Visual Assessment of Exfoliation Corrosion Susceptibility of 5xxx-Series Aluminum Alloys (ASSET Test)," G 66, Annual Book of ASTM Standards, American Society for Testing and Materials 8. T.J. Summerson, Interim Report, Aluminum Association Task Group on Exfoliation and Stress-Corrosion Cracking of Aluminum Alloys for Boat Stock, in Proceedings of the Tri-Service Corrosion Military Equipment Conference, Technical Report AFML-TR-75-42, Vol II, Air Force Materials Laboratory, 1975, p 193-221 9. "Standard Specification for Aluminum and Aluminum-Alloy Sheet and Plate," B 209, Annual Book of ASTM Standards, American Society for Testing and Materials 10. "Standard Test Method for Exfoliation Corrosion Susceptibility in 2xxx and 7xxx-Series Aluminum Alloys (EXCO Test)," G 34, Annual Book of ASTM Standards, American Society for Testing and Materials 11. D.O. Sprowls, T.J. Summerson, and F.E. Loftin, Exfoliation Corrosion Testing of 7075 and 7178 Aluminum Alloys--Interim Report on Atmospheric Exposure Tests (Report of ASTM G01.05.02 Interlaboratory Testing Program in Cooperation with the Aluminum Association), in Corrosion in Natural Environments, STP 558, American Society for Testing and Materials, 1974, p 99-113 12. B.W. Lifka, Corrosion Resistance of Aluminum Alloy Plate in Rural, Industrial, and Seacoast Atmospheres, Mater. Prot., to be published

Evaluation of Stress-Corrosion Cracking Donald O. Sprowls, Consultant

Introduction THERE ARE A NUMBER of corrosion-related causes of the premature fracture of structural components. The most common of these are compared in Fig. 1. Cracking due to corrosion fatigue occurs only under cyclic or fluctuating operating loads, while failure resulting from the other processes shown occurs under static or slowly rising loads. With certain alloy systems, hydrogen embrittlement (see the articles "Environmentally Induced Cracking" and "Evaluation of Hydrogen Embrittlement" in this Volume) may have a contributory role in each of these failure processes. Appropriate tests for the different failure modes are discussed in other articles in this Section.

Fig. 1 Causes of premature fracture influenced by the corrosion of a structural component

This article will follow the broad outline listed below: • • • • • • •

General state of the art Static loading of smooth specimens Static loading of precracked specimens Dynamic loading: slow strain rate testing Selection of test environments Appropriate tests for various alloy systems Interpretation of test results

General State of the Art Most stress-corrosion cracking (SCC) testing is performed either to determine the best material for a specific application or to compare the relative behaviors of material and environmental variations. Test conditions for the former should be representative of the most severe conditions anticipated in the intended service. For the latter, test conditions are usually chosen to produce various degrees of cracking in a reasonable time (Ref 1). The primary challenge in both cases is expressed well in the following statement, which was written a generation ago: "While it is relatively easy to determine if a product is susceptible to SCC, it is far more difficult to determine if it possesses a `degree of susceptibility' which will restrict its general usefulness" (Ref 2). Historically, service failures due to SCC have been identified with sustained tensile stress; thus, SCC testing has developed around the use of static loading. In some situations, it is advantageous to use an actual structural component for testing. However, this is usually not practical; more often, it is necessary to select smaller specimens that afford the required predictive capability. Before about 1965, only constant-load or constant-strain test of smooth and notched test specimens of various configurations were used to assess SCC. More test methods are currently available than ever before. During the 1960s, two accelerated testing techniques based on different mechanical approaches emerged. One technique tests and analyzes statically loaded, mechanically precracked test specimens by using linear elastic fracture mechanics concepts. The second technique consists of constant (slow) strain rate tests on smooth or precracked specimens. Laboratory testing with these techniques has frequently produced SCC, when the older, traditional tests have not. Initiation and Propagation of SCC The process of SCC is frequently discussed in terms of initiation (incubation and nucleation and propagation, and illustrations similar to Fig. 2 can be found in the literature. However, an accepted model has not been established. There may be a gradual transition from localized corrosion to crack initiation and growth with no separation of stages, or there may be a repeated succession of short steps of initiation and growth. In any event, from an engineering standpoint, it is convenient to hypothesize the process in two generic stages: initiation and propagation. This terminology will be used throughout this article.

Fig. 2 The relative influences of electrochemical and mechanical factors in the corrosion and SCC damage of a

susceptible material. The shaded area represents the transition of driving force from dominance by electrochemical factors to chiefly mechanical factors. Precise separation of initiation and propagation stages is experimentally difficult. Stimulation of cracking by atomic hydrogen may also become involved in this transition region.

Two basic corrosion reactions, anodic and cathodic, dominate the SCC process in conjunction with mechanical stress. The chemical composition of the environment, including pH and the presence of hydrogen recombination poisons that affect the cathode reaction product, and the composition and metallurgical condition of the metal determine which of the two partial corrosion reactions is dominant. Anodic SCC (active path corrosion) involves the dissolution of metal during the initiation and propagation of cracks. Cathodic SCC (embrittlement by corrosion product hydrogen) involves the deposition of hydrogen at cathodic sites on the metal surface or on the walls of a fissure or crack and its subsequent absorption into the metal lattice. More information on the mechanisms of SCC is availablein the article "Environmentally Induced Cracking" in this Volume. Figure 2 also suggests the relative influences of the electrochemical and mechanical driving forces in the SCC process. Figure 2 indicates a change as SCC proceeds, with the role of stress being negligible at first and then becoming dominant as subcritical cracking advances. Environmental action must always be involved, although it may be dominant only at first. The preexistence of a mechanical flaw or crack in the stressed metal may of course alter the initiation stage. Application of the fracture mechanics based stress intensity factor (J for elastic-plastic fracture mechanics; K for linear elastic fracture mechanics) as a driving force for the propagation of SCC is illustrated schematically in Fig. 2 and 3 (Ref 3, 4).

Fig. 3 Effect of stress intensity on the kinetics of SCC. Stages I and II may not always be straight lines but may be strongly curved, and one or the other may be absent in some systems. Stage III is of little interest and is generally absent in K-decreasing tests.

Standardization of Tests Standardization of SCC test methods in the United States was initiated in the 1960s by the American Society for Testing and Materials (ASTM), the National Association of Corrosion Engineers (NACE), and the federal government. Standard tests have also been developed in Europe (Ref 5), and uniform testing methods are currently under development on a broader basis through the International Organization for Standardization (ISO). Reference will be made throughout this article to available standards and to useful publications for details on the test methods.

There are several essential factors that must be given carefully consideration in the design of all types of SCC tests: • • • •

The composition of the test environment must remain constant throughout the test, unless changes are a part of the corrosion system of interest The materials used for SCC test fixtures must resist attack Stressing fixtures must remain dimensionally stable so as not to affect the stress placed on specimens during the test Galvanic action between the test specimens and ancillary equipment must be avoided; such action, if present, can either accelerate or retard SCC, depending on whether there is anodic or cathodic control

Static Loading of Smooth Specimens Tests for predicting the stress-corrosion performance of an alloy in a particular service application should be conducted with a stress system similar to that anticipated in service. Table 1 lists the numerous sources of sustained tension that are known to have initiated SCC in service and the applicable methods of stressing. Most of the SCC service problems involve tensile stresses of unknown magnitude that are usually very high. The tests that incorporate a high total strain are usually the most realistic in terms of duplicating service. Table 1 Stressing methods applicable to various sources of sustained tension in service Source of sustained tension in service

Constant strain

Constant load

Quenching after heat treatment

X

...

Forming

X

...

Welding

X

...

Misalignment (fit-up stresses)

X

...

Interference fasteners

X

...

Rigid

X

...

Flexible

...

X

Flareless fittings

X

...

Clamps

X

...

Residual stress

Interference bushings

Hydraulic pressure

X

X

Dead weight

...

X

Faying surface corrosion

X

X

Note: The greatest hazard arises when residual, assembly, and operating stresses are additive.

The results are strongly influenced by the mechanical aspects of the tests, such as method of loading and specimen size. These mechanical aspects can have variable effects on the initiation and propagation lifetimes and can influence estimates of a threshold stress. Therefore, an apparent threshold stress for SCC is not a material property, and threshold estimates must be qualified with regard to the test conditions and the significance level. Constant-Strain Versus Constant-Load Tests Constant-strain (fixed-displacement) tests are widely used, primarily because a variety of simple and inexpensive stressing jigs can be devised. However, there is poor reproducibility of the exposure stress with some of these techniques. Therefore, sophisticated procedures have been developed to improve this facet of testing. Constant-strain tests are sometimes called decreasing-load tests, because after the onset of SCC in small test specimens the gross section exposure stress decreases. This results from the opening of the crack (or cracks) under the high stress concentration at the crack tip (or tips) and causes some of the applied elastic strain to change to plastic strain, with an attendant reduction in the initial load (Ref 6, 7). Such trends in changing stress during crack growth are shown in Fig. 4.

Fig. 4 Schematic comparison of changing stress during initiation and growth of isolated SCC in constant-strain and constant-load tests of a uniaxially loaded tension specimen. (a) Constant-strain test. (b) Constant-load test. M is the maximum stress at crack tip, N is the average stress in the net section, and G is the applied stress to the gross section. Source: Ref 7

Comparison of the stress trends for a constant-strain test (Fig. 4a) with those for a constant-load test (Fig. 4b) reveals that neither method of loading provides a constant-stress test after growth of microcracks has occurred. True constant-load (dead-load) tests result in increasing stress levels as cracking progresses, and are more likely to lead to earlier failure with complete fracture and lower estimates of a threshold stress than constant-strain tests. Figures 4(a) and 4(b) illustrate basic trends that may be applied to all types of test specimens, including precracked specimens. Specific curves, however, will differ depending on other test conditions. The stiffness of the combined stressing frame/test specimen system can have a significant effect on materials evaluation if identical test procedures are not used (Ref 6). Many so-called constant-strain tests, particularly if a spring is included in the stressing system, are not actually constant-strain tests, because a significant amount of elastic strain energy may be contained in the stressing system. Depending on the "softness" of the spring or the elasticity of the stressing jig, the stiffness (compliance) of the stressing system can be varied greatly between zero stiffness (dead load) and infinite stiffness (true constant total strain). Figure 5 shows the typical change in net section stress with the onset of SCC in an intermediate-stiffness stressing frame.

Fig. 5 Effect of loading method and extent of cracking or corrosion pattern on average net section stress in a uniaxially loaded tension specimen. Behavior is generally representative, but curves will vary with specific alloys and tempers. (a) Localized cracking. (b) General cracking. Source: Ref 8 (ASTM G 49)

The corrosion pattern on the test specimen, particularly the number and distribution of cracks, can impair the precision of results obtained by either constant-strain or constant-load tests. When isolated stress-corrosion cracks propagate in a specimen stressed by either method, the average tensile stress on the net section increases rapidly until the notch fracture strength is reached and the specimen breaks (Fig. 5a). Less penetration is required for fracture of specimens under dead load; this indicates that specimen life is shorter with lower-stiffness stressing frames. When microcracks initiate close to one another, their individual stress concentrations interact and are relaxed. Consequently, there may not be a sufficient stress concentration in the true constant-strain test to propagate further SCC, and the specimen will not break (Fig. 5b). Under a constant load, however, the growth of many cracks continues, and the specimen ultimately breaks. With general cracking, crack propagation can be strongly influenced by frame stiffness. Therefore, SCC comparison of specimens tested at stress levels just above their thresholds is complicated by random variations in the cracking pattern, particularly when tested with relatively stiff stressing system. Although constant-load stressing appears to be advantageous for testing materials with relatively high resistance to SCC, difficulties arise when small-diameter specimens are utilized to avoid the use of massive loads or lever systems. In some test environments, highly stressed specimens may fail from general or pitting corrosion and an attendant increase in the effective stress. Such non-SCC failures complicate interpretation of test results, unless failure by SCC is confirmed by metallographic examination. Such extraneous failures are less likely to occur with specimens loaded under constant strain. Therefore, small test specimens, which are generally preferred for laboratory screening tests and research studies, must be used with caution when estimates of serviceability are required. To determine serviceability, larger specimens should be used, and a stressing system should be selected that best duplicates the anticipated service conditions. Bending Versus Uniaxial Tension Historically, the most extensively used stressing systems have incorporated constant-deformation specimens stressed by bending. This method is versatile because of the variety of simple techniques that can be used to test most metal products in all types of corrosive environments. The state of stress in a bend specimen, however, is much more complex than in a tension specimen. Theoretically, tensile stress is uniform throughout the cross section in the tension specimen, except at corners in rectangular sections, but the tensile stress in bend specimens varies through the specimen thickness.

Tensile stress is at a maximum on the convex surface and decreases steeply to zero at the neutral axis. It then changes to a compressive stress, which reaches a maximum on the concave surface. Thus, only about 50% of the metal surface is under tension, and stress can vary from maximum to zero, depending on the stressing system. As SCC penetrates the metal, the stress gradient through the section thickness produces changes in stresses and strains that are different from those in a uniaxial tension specimen. This tendency yields significantly different SCC responses for the two types of stressing (Fig. 6).

Fig. 6 Comparison of the SCC response with bending versus direct tension stressing under constant load for Al5.3Zn-3.7Mg-0.3Mn-0.1Cr T6 temper alloy steel. Tested to failure in 3% sodium chloride plus 0.1% hydrogen peroxide. Source: Ref 9

Bending stress specimens experience other sources of variability in stress that are not present with direct tension stressing. Variations occur in the principal longitudinal stress across the width of the specimen as well as with the presence of biaxial stresses, both of which are influenced by the design of the specimen. Therefore, just as in the case of constant-load stressing, optimal control of stress and more severe testing conditions are provided by uniaxial tension stressing. Statically loaded, smooth test specimens for SCC tests can be divided into three general categories: elastic strain specimens, plastic strain specimens, and residual stress specimens. The commonly used specimen geometries for each of these categories are discussed below. Elastic Strain Specimens To control the surface tensile stress applied by deformation loading, strain is usually restricted to the elastic range for the test material. The magnitude of the applied stress can then be calculated from the measured strain and modulus of elasticity. In constant-load stressing, the load typically is measured directly, and the stress is calculated by using the appropriate formula for the specimen configuration and the method of loading. Load cells or calibrated springs may be useful for applying and monitoring possible changes in load during the test. The commonly used types of specimens for tests under elastic-range stress are described below. Bent-beam specimens can be used to test a variety of product forms. The bent-beam configuration is primarily used

for sheet, plate, or flat extruded sections, which conveniently provide flat specimens of rectangular cross section, but it is also used for cast materials, rod, pipe, or machined specimens of circular cross section. This method is applicable to specimens of any metal that are stressed to levels less than the elastic limit of the material; therefore, the applied stress can be calculated or measured accurately (ASTM G 39) (Ref 8). Stress calculations by this method are not applicable to plastically stressed specimens. Bent-beam specimens are usually tested under constant-strain conditions, but constant-load conditions can also be used. In either case, local changes in the curvature of the specimen when cracking occurs result in changes in stress and strain during crack propagation. The "test stress" is taken as the highest surface tensile stress existing at the start of the test, that is, before the initiation of SCC. Several configurations of bent-beam specimens and stressing systems are illustrated in Fig. 7 and are described in detail in ASTM G 39 (Ref 8). When specimens are tested at elevated temperatures, the possibility of stress relaxation should be

investigated. More information on stress relaxation is available in the article "Creep, Stress-Rupture, and StressRelaxation Testing" in Mechanical Testing, Volume 8 of ASM Handbook, formerly 9th Edition Metals Handbook.

Bolt loaded double-beam specimen dimensions for various plate thicknesses t

a

b

L

S

mm

in.

mm

in.

mm

in.

mm

in.

mm

in.

3.2

0.125

100

4.0

50

2.0

250

10.0

305

12.0

6.4

0.25

100

4.0

50

2.0

250

10.0

305

12.0

9.5

0.375

120

4.75

90

3.5

330

13.0

380

15.0

13

0.5

120

4.75

90

3.5

330

13.0

380

15.0

19

0.75

140

5.5

150

6.0

430

17.0

480

19.0

25

1.0

150

6.0

200

8.0

510

20.0

560

22.0

38

1.5

165

6.5

305

12.0

635

25.0

685

27.0

Fig. 7 Schematic specimen and holder configurations for bent-beam stressing. (a) Two-point loaded specimen. (b) Three-point loaded specimen. (c) Four-point loaded specimen. (d) Welded double-beam specimen. (e) Boltloaded double-beam specimen. Formula for stressing specimen (e): d = 2fa/3Et(3L - 4a), where d is deflection (in inches), f is nominal stress (in pounds per square inch), and E is modulus of elasticity (in pounds per square inch). Source: Ref 10

Two-point loaded specimens can be used for materials that do not deform plastically when bent to (L - H)/H = 0.01.

The specimens should be approximately 25- × 250-mm (1- × 10-in.) flat strips cut to appropriate lengths to produce the

desired stress after bending, as shown in Fig. 7(a). The maximum stress occurs at the midlength of the specimen and decreases to zero at specimen ends. Three-point loaded specimens are flat strips that are typically 25 to 51 mm (1 to 2 in.) wide and 127 to 254 mm (5 to 10 in.) long. The thickness of a specimen is usually dictated by the mechanical properties of the material and the available product form. The specimen should be supported at the ends and bent by forcing a screw (equipped with a ball or knife-edge tip) against it at a point halfway between the end supports, as shown in Fig. 7(b). In a three-point loaded specimen, the maximum stress occurs at the midlength of the specimen and decreases linearly to zero at the outer supports.

Two- and four-point loaded specimens are often preferred over the three-point loaded specimen, because crevice corrosion often occurs at the central support of the three-point loaded specimen. Because this corrosion site is very close to the point of highest tensile stress, it may cathodically protect the specimen and prevent possible crack formation, or it may cause hydrogen embrittlement. Furthermore, the pressure of the central support at the point of highest load introduces biaxial stresses at the area of contact and can introduce tensile stresses where compressive stresses are normally present. Four-point loaded specimens are flat strips that are typically 25 to 51 mm (1 to 2 in.) wide and 127 to 254 mm (5 to

10 in.) long. The thickness of a specimen is usually dictated by the mechanical properties of the material and the available product form. The specimen is supported at the ends and is bent by forcing two inner supports against it, as shown in Fig. 7(c). The two inner supports are located symmetrically around the midpoint of the specimen. In a four-point loaded specimen, the maximum stress occurs between the contact points of the inner supports; the stress is uniform in this area. From the inner supports, the stress decreases linearly toward zero at the outer supports. The fourpoint loaded specimen is preferred over the three-point and two-point loaded specimens, because it provides a large area of uniform stress. Welded double-beam specimens consist of two flat strips 25 to 51 mm (1 to 2 in.) wide and 127 to 254 mm (5 to 10

in.) long. The strips are bent against each other over a centrally located spacer until both ends touch. The strips are held in position by welding the ends together, as shown in Fig. 7(d). In a welded double-beam specimen, the maximum stress occurs between the contact points of the spacer; the stress is uniform in this area. From contact with the spacer, the stress decreases linearly toward zero at the ends of the specimen. A bolt-loaded double-beam specimen is shown in Fig. 7(e), along with suggested specimen dimensions for various

thicknesses of plate and the formula for stressing such specimens (Ref 10). The beam deflections required to develop the intended tensile stress are calculated with the formula and are then applied by bolting the ends of the beams together. The deflections are measured with a dial gage to within ±0.0127 mm (±0.0005 in.). Thus, the error in stress application--if the beams are of homogeneous material and if the cross sections are uniform--is within 2%. The precision of the deflection measurement is within 0.5%, and the error in determining the modulus of elasticity, E, is within 1%. Constant-moment beam specimens are designed such that a constant moment exists from one end to the other

when the specimen is bent in the manner shown in Fig. 8. This bending produces equal stress along the length of the specimen. The width-to-thickness ratio is less than 4 so that biaxial stresses are eliminated.

Fig. 8 Bent beam designed to produce pure bending. Source: Ref 11

This type of specimen offers the advantage of a relatively large area of material under a uniform stress. Such specimens can be used when the dimensions of the specimen are too small for other bent-beam specimens--for example, when specimens are taken in the short-transverse direction in plate (see Fig. 18c). The elastic stress in the convex surface is calculated by using:

(Eq 1)

where h is the distance between inner edges of the supports, y is the maximum deflection between inner edges of the supports, t is the thickness of the specimen, and E is the modulus of elasticity. C-Ring Specimens. As discussed in ASTM G 38, (Ref 8), the C-ring is a versatile, economical specimen for

quantitatively determining the susceptibility to SCC of all types of alloys in a wide variety of product forms. It is particularly well suited for testing tubing and for making short-transverse tests on various product forms, as shown in Fig. 9. The sizes of C-rings can be varied over a wide range, but rings with outside diameters less than about 16 mm ( are not recommended because of increased difficulties in machining and decreased precision in stressing.

in.)

Fig. 9 Sampling procedure for testing various products with C-rings. (a) Tube. (b) Rod and bar. (c) Plate

The C-ring is typically a constant-strain specimen with tensile stress produced on the exterior of the ring by the tightening of a bolt centered on the diameter of the ring. However, an almost constant load can be developed by placing a calibrated spring on the loading bolt. C-rings can also be stressed in the reverse direction by spreading the ring and creating a tensile stress on the inside surface. These methods of stressing are illustrated in Fig. 10.

Fig. 10 Methods of stressing C-rings. (a) Constant strain. (b) Constant load. (c) Constant strain. (d) Notched C-ring; a similar notch could be used on the side of (a), (b), or (c).

Circumferential stress is of principal interest in the C-ring specimen. This stress is not uniform (Ref 12), as discussed previously in the section "Elastic Strain Specimens" in this article. The stress varies around the circumference of the Cring from zero at each bolt hole to a maximum at the middle of the arc opposite the stressing bolt. In a notched C-ring, a triaxial stress state is present adjacent to the root of the notch (Ref 13). For all notches, the circumferential stress at the root of the notch is greater than the nominal stress and can generally be expected to be in the plastic range. Generally, the C-ring can be stressed with high precision. The most accurate stressing procedure consists of attaching circumferential and transverse electrical strain gages to the surface stressed in tension, followed by tightening the bolt until the strain measurements indicate the desired circumferential stress. The amount of compression required on the C-ring to produce elastic straining and the degree of elastic strain can be predicted theoretically. Therefore, C-rings can be stressed by calculating the deflection required to develop a desired elastic stress (ASTM G 38) (Ref 8). In notched specimens, a nominal stress is estimated using a ring outside diameter measured at the root of the notch and by taking into consideration the stress-concentration factor, Kt, for the specific notch. O-ring specimens (Fig. 11) are used to develop a hoop stress in a particular part--for example, a cylindrical die forging

in which a critical end-grain structure associated with the parting plane of the forging exists only at the surface of the forging. A relatively large surface area of metal is placed under a uniform tensile stress, and the O-ring stressing plug assembly simulates service conditions in structures containing interference-fit components. Stressed O-rings have also been used to evaluate protective treatments for the prevention of SCC (Ref 14).

Fig. 11 O-ring SCC test specimen (a) and stressing plug (b). The O-ring is stressed by pressing it onto the plug, as shown in (c).

An O-ring is stressed by pressing it onto an oversized plug that is machined to a predetermined diameter to develop the desired stress at the outside surface of the ring. The nominal dimensions of this specimen can be varied to suit the part being tested, but certain characteristics should be observed to achieve adequate control of the stresses. The ring width should not be more than four times the wall thickness in order to ensure maximum uniformity of the hoop stress from the centerline to the edges of the ring. The tensile stress varies through the thickness of the ring and is highest at the inside surface. Interference required for stressing an O-ring can be calculated by using:

(Eq 2)

where I is the interference (on the diameter) between the O-ring and the plug, E is the modulus of elasticity, ID is the inside diameter, OD is the outside diameter, and F is the circumferential stress desired on the outside surface. Additional information regarding the design and stressing of O-ring specimens is given in Ref 15.

Tension Specimens. Specimens used to determine tensile properties in air are well suited and easily adapted to SCC,

as discussed in ASTM G 49 (Ref 8). When uniaxially loaded in tension, the stress pattern is simple and uniform, and the magnitude of the applied stress can be accurately determined. Specimens can be quantitatively stressed by using equipment for application of either a constant load, a constant strain, or an increasing load or strain. This type of test is one of the most versatile methods of SCC testing because of the flexibility permitted in the type and size of the test specimen, the stressing procedures, and the range of stress level. It allows the simultaneous exposure of unstressed specimens (no applied load) with stressed specimens and subsequent tension testing to distinguish between the effects of true SCC and mechanical overload. A wide range of test specimen sizes can be used, depending primarily on the dimensions of the product to be tested. Stress-corrosion test results can be significantly influenced by the cross section of the test specimen. Although large specimens may be more representative of most structures, they often cannot be prepared from the available product forms being evaluated. They also present more difficulties in stressing and handling in laboratory testing. Smaller cross-sectional specimens are widely used. They have a greater sensitivity to SCC initiation, usually yield test results rapidly, and permit greater convenience in testing. However, the smaller specimens are more difficult to machine, and test results are more likely to be influenced by extraneous stress concentrations resulting from nonaxial loading, corrosion pits, and so on. Therefore, use of specimens less than about 10 mm (0.4 in.) in gage length and 3 mm (0.12 in.) in diameter is not recommended, except when testing wire specimens. Tension specimens containing machined notches can be used to study SCC and hydrogen embrittlement. The presence of a notch induces a triaxial stress state at the root of the notch, in which the actual stress will be greater by a concentration factor that is dependent on the notch geometry. The advantages of such specimens include the localization of cracking to the notch region and acceleration of failure. However, unless directly related to practical service conditions, the results may not be relevant. Tension specimens can be subjected to a wide range of stress levels associated with either elastic or plastic strain. Because the stress system is intended to be essentially uniaxial (except in the case of notched specimens), great care must be exercised in the construction of stressing frames to prevent or minimize bending or torsional stresses. The simplest method of providing a constant load consists of a dead weight hung on one end of the specimen. This method is particularly useful for wire specimens. For specimens of larger cross section, however, lever systems such as those used in creep-testing machines are more practical. The primary advantage of any dead-weight loading device is the constancy of the applied load. A constant-load system can be modified by the use of a calibrated spring, such as that shown in Fig. 12. The proving ring, as used in the calibration of tension testing machines, has also been adapted to SCC testing to provide a simple, compact, easily operated device for applying axial load (Fig. 13). The load is applied by tightening a nut on one of the bolts and is determined by carefully measuring the change in ring diameter.

Fig. 12 Spring-loaded fixture used to stress 3.2-mm (0.125-in.) thick sheet tensile specimens in direct tension. Source: Ref 10

Fig. 13 Ring-stressed tension specimen for field testing. Source: Ref 1

Constant-strain SCC tests are performed in low-compliance tension-testing machines. The specimen is loaded to the required stress level, and the moving beam is then locked in position. Other laboratory stressing frames have been used, generally for testing specimens of smaller cross section. Figure 14(a) shows an exploded view of such a stressing frame, and Fig. 14(b) illustrates a special loading device developed to ensure axial loading with minimal torsion and bending of the specimen.

Fig. 14(a) Constant-strain SCC testing frame. Exploded view (left) showing the 3.2-mm (0.125-in.) diam tension specimen and various parts of the stressing frame. Final stressed assembly (right). Source: Ref 16

Fig. 14(b) Synchronous loading device used to stress specimens. The specimen is loaded to a prescribed strain value determined from a clip-on gage. The applied stress is given by the product of the strain and the material elastic modulus. A stressed assembly and one assembled finger-tight ready for stressing are shown.

For stressing frames that do not contain any mechanism for the measurement of load, the stress level can be determined from measurement of the strain. However, only when the intended stress is below the elastic limit of the test material is the average linear stress ( ) proportional to the average linear strain ( ), / = E, where E is the modulus of elasticity. When tests are conducted at elevated temperatures with constant-strain loaded specimens, consideration should be given to the possibility of stress relaxation. When stress relaxation or creep occurs in the test specimen, some of the elastic strain is converted to plastic strain and the nominal applied test stress is reduced. This effect is particularly important when the coefficients of thermal expansion are different for the specimen and stressing frame. Frequently, nonmetallic (plastic) insulators are used between the specimen and stressing frame to avoid galvanic action. If such plastic insulators are part of the stress-bearing system, creep (even at room temperature) can significantly alter the applied load on the specimen. Even though eccentricity in loading can be minimized to levels acceptable for tension-testing machines, tensile stress around the circumference of round specimens and at the corners of sheet-type specimens varies to some extent. Several factors may introduce bending moments on specimens, such as longitudinal curvature and misalignment of threads on threaded-end round specimens. These factors have a greater effect on specimens with smaller cross sections. Tests should

be made on specimens with strain gages affixed to the specimen surface around the circumference of 90° or 120° intervals to verify strain and stress uniformity and to determine if machining practices and stressing jigs are of adequate tolerance and quality. When SCC occurs, it generally results in complete fracture of the specimen, which is easy to detect. However, when testing relatively ductile materials at stress levels close to the threshold of susceptibility, fracture may not occur during the period of exposure. The presence of SCC in such cases must be determined by mechanical tests or by metallographic examination, as discussed previously. To study trends in SCC susceptibility, such as in alloy development research, it is often necessary to detect small differences in susceptibility. For this purpose, it is advantageous to use replicate sets of specimens stressed at several levels, including zero applied stress. The sets are then removed for metallographic examination or tension tests after appropriate periods of exposure. Figure 15 illustrates the use of this procedure with samples of 7075 aluminum alloy that have been given different thermal treatments to decrease susceptibility to SCC. Analysis of these breaking stress data by extreme value statistics enables calculation of survival probabilities and the estimation of a threshold stress, without depending on failures during exposure. By using an elastic-plastic fracture mechanics model, an effective flaw size is calculated from the mean breaking stress, the strength, and the fracture toughness of the test material. The effective flaw size corresponds to the weakest link in the specimen at the time of the tension test, and it therefore represents the maximum penetration of the SCC. An advantage to using flaw depth to examine SCC performance is that the effects of specimen size and alloy strength and toughness can be normalized. In contrast, the specimen lifetime and breaking strength are biased by those mechanical (non-SCC) factors.

Fig. 15 Mean breaking stress versus exposure time for short-transverse 3.2-mm (0.125-in.) diam aluminum alloy 7075 tension specimens tested according to ASTM G 44 at various exposure stress levels. Each point represents an average of five specimens. Source: Ref 3

Mean trends in the 207-MPa (30-ksi) exposure data for the three temper variants of aluminum alloy 7075 examined in Fig. 15 are shown in Fig. 16. These results clearly illustrate that the thermal treatments used to reduce the SCC susceptibility of the 7075-T651 decreased the SCC penetration (Ref 17). The equivalent performance of the 7075-T7X1 3.2- and 5.7-mm (0.125- and 0.225-in.) diam specimens is evident. In contrast, Fig. 17 shows the specimen biases in SCC ratings obtained by traditional pass-fail methods (Ref 18).

Fig. 16 Effect of temper on SCC performance of aluminum alloy 7075 subjected to alternate immersion in 3.5% NaCl solution at a stress of 207 MPa (30 ksi). Mean flow depth was calculated from the average breaking strength of five specimens subjected to identical conditions. Source: Ref 17

Fig. 17 Influence of specimen configuration on SCC test performance (alternate immersion in 3.5% sodium chloride per ASTM G 44). Aluminum alloy 7075-T7X51 specimens stressed 310 MPa (45 ksi); each point represents 60 to 90 specimens. Source: Ref 18

Tuning fork specimens are special-purpose specimens with numerous modifications (Fig. 18). In Europe, the metal is strained into the plastic range, and stresses and strains are usually not measured (Ref 19, 21). In the United States, however, these specimens have been used with measured strains in the elastic and plastic ranges. Specimens of the type shown in Fig. 18(b) are convenient when a small self-contained specimen is required that will afford some insight into the applied stresses. Such a specimen is particularly well suited for testing thin plate material in the longitudinal or longtransverse direction while keeping the original mill-finished surface intact.

Fig. 18 Typical tuning fork SCC test specimens. (a) Source: Ref 19. (b) Source: Ref 1. (c) Source: Ref 20

Tuning fork specimens are stressed by closing the specimen tines and restraining them in the closed position with a bolt placed at the tine ends. The amount of closures is determined from Eq 3, which was derived from the data obtained with strain gages placed at the base of the tines on calibration specimens (Ref 1):

S=A t

(Eq 3)

where S is the maximum tension stress in the outer fiber of either tine, A is the calibration constant, of closure at the tine ends, and t is the thickness of the tines.

is the total amount

The stress on tuning forks with straight tines is greatest in a small area at the base of the tines. In tuning forks with tapered tines, the maximum stress extends uniformly along the tapered section. Tuning forks must be given the same consideration with regard to biaxial stresses as other flexurally loaded specimens. The miniature tuning fork shown in Fig. 18(c) was devised to conduct short-transverse tests on sections that are too thin for tensile specimens or C-rings to be obtain (Ref 20). As with other tuning fork specimens, the relationship between strain on the grooved surface and the deflection at the ends of the legs can be determined through the use of strain gages. Plastic Strain Specimens Many accelerated SCC tests are performed with plastically deformed specimens, because these specimens are simple and economical to manufacture and use. These specimens are convenient for multiple replication tests of self-stressed (fixeddeflection) specimens in all environments. Because they usually contain large amounts of elastic and plastic strain, they provide one of the most severe tests available for smooth SCC test specimens. Generally, the stress conditions are not known precisely. However, the anticipated high level of stress can be obtained consistently only if the precautions described for each type of specimen are observed. Another consideration is that the cold work required to form the test specimen can change the metallurgical condition and the SCC behavior of certain alloys. Tests of this type are primarily used as screening tests to detect large differences between the SCC resistance of one alloy in several environments, one alloy in several metallurgical conditions in a given environment, and different alloys in the same environment. These tests are sometimes claimed to be too severe and therefore unsuitable for many applications, but the stress conditions are nevertheless representative of the high locked-in fabrication and assembly stresses frequently responsible for SCC in service. U-bend specimens are rectangular strips bent approximately 180° around a predetermined radius and maintained in

this plastically (and elastically) deformed condition during the test. Standardized test methods for this type of specimen are described in ASTM G 30 (Ref 8). Bends slightly less than or greater than 180° are also used, but the term U-bend is generally applied to test specimens that are bent beyond their elastic limits. Figure 19 illustrates typical U-bend configurations showing several different methods of maintaining the applied stress.

Alternative size

L

M

W

t

D

X

Y

R

mm

in.

mm

in.

mm

in.

mm

in.

mm

in.

mm

in.

mm

in.

mm

in.

A

80

3.2

50

2.0

20

0.8

2.5

0.098

10

0.4

32

1.26

15

0.55

5

0.2

B

100

4.0

90

3.5

9

0.35

3.0

0.12

7

0.28

25

0.98

38

1.50

16

0.6

C

120

4.7

90

3.5

20

0.8

1.5

0.06

8

0.31

35

1.4

35

1.4

16

0.6

D

130

5.1

100

4.0

15

0.6

3.0

0.12

6

0.24

45

1.77

32

1.26

13

0.51

E

150

5.9

140

5.5

15

0.6

0.8

0.03

3

0.12

61

2.40

20

0.8

9

0.35

F

310

12.2

250

9.8

25

0.98

13.0

0.51

13

0.51

105

4.13

90

3.5

32

1.26

G

510

20.1

460

18.1

25

0.98

6.5

0.26

13

0.51

136

5.35

165

6.5

76

3.0

Note:

= 1.57 rad

Fig. 19 Typical U-bend SCC specimens. (a) Various methods of stressing U-bends. (b) Typical U-bend specimen dimensions

U-bend specimens can be used for all materials sufficiently ductile to be formed into a U-configuration without cracking. A U-bend specimen is most easily made from strips of sheet, but specimens can be machined from plat, bar, wire, castings, and weldments. Of primary interest in U-bend specimens is circumferential stress, which is not uniform, as discussed previously in the section on "Bent-Beam Specimens" in this article. Stress distribution in the U-bend specimen is discussed in detail in Ref 22. A good approximation of applied strain can be obtained by:

(Eq 4)

where t is the specimen thickness, and R is the radius of curvature at the point of interest. Knowledge of the stress-strain curve is necessary to determine the stress. When a U-bend specimen is formed, the material in the outer fibers of the bend is strained into the plastic portion of the true stress/true strain curve, such as in section AB in Fig. 20(a). Several other stress-strain relationships that can exist in the outer fibers of a stressed U-bend test specimen are shown in Fig. 20(b) through (e). The actual relationship obtained depends on the method of stressing used.

Fig. 20 True stress/true strain relationships for stressed U-bends. See text for discussion of (a) to (e).

Stressing is usually achieved by a one- or two-stage operation. Single-stage stressing is accomplished by bending the specimen into shape and maintaining it in that shape. The two types of stress conditions that can be obtained by singlestage stressing are defined by point X in Fig. 20(b) and 20(c). In Fig. 20(c), some elastic strain relaxation has occured by allowing the U-bend legs to spring back slightly at the end of the stressing sequence. Two-stage stressing involves forming the approximate U-shape and then allowing the elastic strain to relax completely before the second stage of applying the test stress. The applied test strain can be a percentage (from 0 to 100% ) of the

tensile elastic strain that occurred during preforming (Fig. 20d) or can involve additional plastic strain (Fig. 20e). The convex specimen surface is stressed in tension in the region 0NM (Fig. 20d), and the concave surface is in compression. In the region MP, the situation is reversed; that is, compression is on the convex surface, and tension is on the concave surface. The slope MN of the curve shown in Fig. 20(d) is steep. Therefore, it is often difficult to apply reproducibly a constant percentage of the total elastic prestrain, and the specimen surface may remain under compressive stress. Therefore, because they result in a more severe test (that is, higher applied stress), the stress conditions in Fig. 20(b) and 20(e) are recommended. Thus, the final applied strain prior to testing consists of plastic and elastic strain. To achieve the conditions illustrated in Fig. 20(b) and 20(e), springback of the U-bend legs after achieving the final plastic strain must be avoided. For materials with relatively low creep resistance, there will be some strain relaxation. Residual Stress Specimens Most industrial SCC problems are associated with residual stresses developed in the metal during such processes as heat treatment, fabrication, and welding. Therefore, residual stress specimens simulating anticipated service conditions are useful for assessing the SCC performance of some materials in particular structures and in specific environments. Plastic Deformation Specimens. Residual stresses resulting from such fabricating operations as forming, straightening, and swaging that involve localized plastic deformation at room temperature can exceed the elastic limit of the material. Examples of specimens of this type that have been used are shown in Fig. 21 and 22. Other specimen types used include panels with sheared edges, punched holes, or stamped identification numbers and specimens that show evidence of other practical fabricating operations.

Fig. 21 SCC test specimens containing residual stresses from plastic deformation. (a) Cracked cup specimen (Ericksen impression). Source: Ref 1. (b) Joggled extrusion containing SCC in the plastically deformed region. Source: Ref 9

Fig. 22 SCC test specimens containing residual stresses from plastic deformation. Shown are 12.7-mm (0.5in.) diam stainless steel tubular specimens after SCC testing. (a) and (b) Annealed tubing that was cold-formed before testing. (c) Cold worked tubing tested in the as-received condition. Source: Ref 1

Weld Specimens. Residual stresses developed in and adjacent to welds are frequently a source of SCC in service.

Longitudinal stresses in the vicinity of a single weld are unlikely to be as large as stresses developed in plastically deformed weldments, because stress in the weld metal is limited by the yield strength of the hot metal that shrinks as it cools. High stresses can be built up, however, when two or more weldments are joined into a more complex structure. Test specimens containing residual welding stresses are shown in Fig. 23. In fillet welds, residual tensile stress transverse to the weld can be critical, as indicated in Fig. 23(a) for a situation in which the tension stress acts in the short-transverse direction in an aluminum-zinc-magnesium alloy plate.

Fig. 23 SCC test specimen containing residual stresses from welding. (a) Sandwich specimen simulating rigid structure. Note SCC in edges of center plate. Source: Ref 10. (b) Cracked ring-welded specimen. Source: Ref 1

Static Loading of Precracked (Fracture Mechanics) Specimens The use of precracked (fracture mechanics) specimens is based on the concept that large structures with thick components are apt to contain cracklike defects. After a stress-corrosion crack begins to grow, or if the specimen is provided with a mechanical precrack, classical stress analysis is inadequate for determining the response of the material subjected to stress in the presence of a corrodent. The mechanical driving force for cracks can be measured with linear elastic fracture mechanics theory in terms of the crack-tip stress intensity factor, K, which is expressed in terms of the remotely applied loads, crack depth, and test specimen geometry. At or above a certain level of K, SCC in a susceptible material will initiate and grow in certain environments, but below that level no measurable propagation is observed (Ref 23). The apparent threshold stress intensity for the propagation of SCC (assuming that crack nuclei form in a manner that cannot be described by fracture mechanics, such as localized corrosion) is designated KISCC (or Kth). Therefore, in terms of linear elastic fracture mechanics theory, for a surface crack in a large plat remotely loaded in tension, the shallowest crack (of a shape that is long compared to its depth) that will propagate as a stress-corrosion crack is acr = 0.2(KISCC/TYS)2, where TYS is tensile yield strength. Thus, a crack that is shallower than this critical value will not propagate under the given environmental conditions. The value of acr incorporates the SCC resistance, KISCC, and the contribution of stress levels (of the order of the yield strength) to SCC due to residual or assembly stresses in thick component sections (Ref 4). Therefore, the application of fracture mechanics does not provide independent information about SCC; it simply provides a usable method for treating the stress factor in the presence of a crack. When the rate of SCC propagation is determined and plotted as a function of Kt (the crack opening mode), the test results for a highly susceptible alloy will exhibit the general trend shown in Fig. 3. Actual curves vary depending on the SCC resistance and fracture toughness of the alloy. Although precracking may shorten or modify the initiation period, it does not circumvent it. Therefore, this method of testing also requires arbitrary and sometimes long exposure periods. Test Specimen Selection Almost all standard plane-strain fracture toughness test specimens can be adapted to SCC testing. These standard configurations should be used to ensure valid fracture analyses. Comprehensive discussions on SCC testing with precracked specimens can be found in Ref 24, 25, 26. Precracked specimens are illustrated schematically in Fig. 24 where they are classified with respect to loading methods and the relationship with the stress intensity factor as stress-corrosion cracking propagates. Proportional dimensions and tolerances per ASTM E 399 (Ref 27) for the more commonly used specimens are given in 25(a), 25(b), Fig. 25(c). Minor modifications to accommodate different loading arrangements and to facilitate mechanical precracking can be made to these configurations without invalidating the plane-strain constraints on the specimens. Figure 26 illustrates alternative chevron-notch and face-groove designs.

Fig. 24 Classification of precracked specimens for SCC testing. Asterisks denote commonly used configurations. Source: Ref 26

Fig. 25(a) Proportional dimensions and tolerances for cantilever bend test specimens. Width = W; thickness (B) = 0.5W; half loading span (L) = 2W; notch width (N) = 0.065W maximum if W > 25 mm (1.0 in.); N = 1.5 mm (0.06 in.) maximum if W 25 mm (1.0 in.); effective notch length (M) = 0.25 to 0.45W; effective crack depth (a) = 0.45 to 0.55W

Fig. 25(b) Proportional dimensions and tolerances for modified compact specimens. Surfaces should be perpendicular and parallel as applicable to within 0.002H TIR. The bolt centerline should be perpendicular to the

specimen centerline within 1°. Bolt of material similar to specimen where practical; fine threaded, square or Allen head. Thickness = B; net width (W) = 2.55B; total width (C) = 3.20B; half height (H) = 1.24B; hole diameter (D) = 0.718B + 0.003B; effective notch length (M) = 0.77B; notch width (N) = 0.06B; thread diameter (T) = 0.625B

Fig. 25(c) Proportional dimensions and tolerances for double-beam specimens. "A" surfaces should be perpendicular and parallel as applicable to within 0.002H TIR. At each side, the point "B" should be equidistant from the top and bottom surfaces to within 0.001H. The bolt centerline (load line) should be perpendicular to the specimen centerline to within 1°. Bolt of material similar to specimen where practical; fine threaded, square or Allen head. Half height = H; thickness (B) = 2H; net width (W) = 10H minimum; total width (C) = W + T; thread diameter (T) = 0.75H minimum; notch width (N) = 0.14H maximum; effective notch length (M) = 2H

Fig. 26 Alternative chevron notch (a) and face grooves (b) for single-edge cracked specimens

Standards for SCC tests using precracked specimens have not yet been developed, although recommended test procedures have been published for certain uses (Ref 28). The best state-of-the-art stress intensity and compliance calibration relationships and guidelines for testing the specimens illustrated in 25(a), 25(b), and Fig. 25(c) are discussed below. Standard names for these specimens and methods of loading per ASTM E 616, "Standard Terminology Relating to Fracture Testing," are used in 25(a), 25(b), and Fig. 25(c) and in paragraph headings for the following discussion. Reference is also made to familiar names used in the literature that may appear elsewhere in this article. Cantilever bend specimens (Fig. 25(a), sometimes referred to as single-edge-notched cantilever bend specimens, have been used in constant-load tests (K-increasing) for characterizing high-strength steels and titanium alloys (Ref 28). Equations 5, 6, and 7 are recommended (Ref 29, 30):

(Eq 5)

(Eq 6)

(Eq 7)

where e is the base of natural logarithm (2.718), x = [0.1426 + 11.92(a/W) - 17.42(a/W)2 + 15.84(a/W)3 - 2.235(a/W)4], y = [6.188 + 12.98(a/W) - 41.19(a/W)2 + 54.98(a/W)3 - 22.28(a/W)4]. M is the applied bending moment, B is the specimen thickness (face grooves, when present, may be accounted for by replacing B with n, where Bn is the net thickness at the base of the face grooves; see Fig. 26b), W is the depth of the specimen, a is the depth of the notch plus crack, E is the modulus of elasticity, 2V0 is the total crack mouth opening displacement at the top face of the specimen, and VLL is the total crack mouth opening displacement measured at the point of load application, which will vary depending on the load arm length. Equation 5 is an expression for the stress intensity of a rectangular beam in pure bending and is valid over a wide range of a/W values. It applies to Mode I loading only, however, and the usual tests include a Mode II component from resulting shear stresses. Equations 6 and 7 were determined by fitting experimental compliance data for cantilever bend specimens with a polynomial equation expressing the natural log of the normalized compliance as a function of a/W. These experimental values are in excellent agreement with those determined from Eq 5 for pure bending, even though the stress state at the crack tip will differ for cantilever bending. It has been suggested that analyses using pure bending expressions related to compliance measurement are suitable for testing with the cantilever bend configuration (Ref 29). Crack growth measurements can be made with clip gage readings in conjunction with the crack opening displacement calibrations given above or by any other method that can be verified within ±0.127 mm (±0.005 in.). Examples of various methods are given in Ref 29 and 31. Modified compact specimens (K-decreasing or K-increasing), as shown in Fig. 25(b), are frequently referred to as

1T-WOL (wedge-opening loaded) or modified WOL specimens. Although most frequently used with constantdisplacement (bolt) loading (Ref 28, 32), these specimens have also been used with constant load (Ref 3, 33, 34). The specimen configuration shown in Fig. 25(b) is similar to that adopted by the Navy (Ref 28), except that it does not incorporate face grooves. Equations 8, 9, 10, and 11 can be used to calculate stress intensity levels and normalized crack opening displacements for fatigue precracking, for initiation of stress-corrosion testing, and for subsequent intervals during the test. These equations are based on boundary colocation values determined for this type of specimen configuration with face grooves and bolt loading (threaded bolt against a rigid loading tip) (Ref 29). The polynomial regression equation agrees with experimentally determined colocation values within 1% for 0.2 a/W 0.95:

(Eq 8)

(Eq 9) where x = [1.830 + 4.307(a/W) + 5.871 (a0/W)2 - 17.53(a0/W)3 + 14.57(a0/W)4]

(Eq 10) where y = [1.623 + 3.352(a0/W) + 8.205(a0/W)2 - 19.59(a0/W)3 + 15.23(a0/W)4]

(Eq 11)

where z = [1.623 + 3.352(ai/W) + 8.205(ai/W)2 - 19.59(ai/W)3 + 15.23(ai/W)4]. In Eq 8, 9, 10, and 11, KIo is the desired starting stress intensity, a0 is the starting crack length, P is the load calculated to develop KIo with measured a0, W is the net width of the specimen measured from the load line, KIi is the stress intensity after time interval i, ai is the crack length after time interval i, and 2VLL is the total crack mouth opening displacement at the load line. All other quantities are as defined previously. Double-beam specimens (K-decreasing or K-increasing), which are also referred to as double-cantilever beam

specimens, are similar to modified compact specimens, but because of their greater width or length, they are well suited for studying SCC growth rates over a greater range of KI values. The smaller height of these specimens (Fig. 25(c)) allows more versatility in performing short-transverse tests from moderate thicknesses of material. Like compact specimens, double-beam specimens are generally used with constant-displacement (bolt) loading for convenience, but they can also be used with constant load.

Bolt-loaded specimens used with a test procedure similar to that described in Ref 35 have been extensively employed for short-transverse tests of aluminum alloy products (Ref 34, 36, 37). Equations 12, 13, 14, and 15 are recommended for general use with double-beam specimens:

(Eq 12)

(Eq 13)

(Eq 14)

(Eq 15)

Equation 12 is an expression reported in Ref 38. The simplified Eq 15 provides more versatility with high accuracy for a wider range of specimen configurations and K values (crack growth) than equations previously published (Ref 35). Two early KI calibrations based on stress analysis (Ref 39) and compliance (Ref 40) are illustrated in Fig. 27 and are in excellent agreement. The shape of these curves can also be used as a design guide for preparing specimens. If the test must be completed in the shortest possible time, a0 should be short to capitalize on the fact that the rate of decrease of KI with crack extension is maximum for shallow cracks. However, if maximum accuracy is desired, a deeper crack (effective notch length M, Fig. 25(c) should be chosen so that errors in crack length measurement do not cause significant errors in KI .

Fig. 27 Configuration and KI calibration of a double-beam plate specimen. Normalized stress intensity KI plotted against a/H ratio. (W - a) indifferent, crackline-loaded, single-edge cracked specimen. Source: Ref 26

Although in early work with aluminum alloys (Ref 35, 36) a relatively short effective notch length was used ( a0/H 0.9), deeper notches have been used recently (a0/H 1.2 to 2.2), all with a 2H value of 25.4 mm (1.0 in.) (Ref 3, 34, 36, 37 ). The recommended starting a/H value shown in Fig. 25(c) is about 2 to 2.2, depending on the length of the precrack. Limited tests of a smaller beam height of 2H = 12.7 mm (0.5 in.) have shown little effect on the amount and rate of crack growth in aluminum alloy 7075 plate (Ref 41); however, additional study is needed in this area. An alternative double-cantilever beam specimen has been developed for testing relatively thin sections (typically 6.4 mm, or in., thick) of low-alloy steels (Ref 42). The specimen is stressed by forcing an appropriately dimensioned wedge into the slot. These specimens have been used to determine the effect of hardness of low-alloy steels on their resistance to SCC in environments containing hydrogen sulfide. Constant KI specimens are well suited for studying the mechanisms of SCC, because the stress intensity. KI, is not dependent on crack depth and can be neglected in kinetic studies. Other attractive features are the relatively simple expressions for stress intensity and compliance and the apparent retention of plane-strain conditions in thin plate and sheet specimens. The cost of specimen preparation and instrumentation, however, prohibits its use for extensive SCC characterizations.

Reference 26 provides equations for the analysis of two types of constant KI specimens: the tapered double-beam specimen and the double-torsion loaded single-edge cracked specimen. A recent evaluation of the double-torsion method (Ref 43) used Al-Zn-Mg alloy sheet 3.2 mm (0.125 in.) thick. By using the double-torsion specimen, V-K curves were produced for aluminum alloy 7075-T651 sheet with conventional two-stage growth and plateau velocities that were only slightly higher than those for conventional double-cantilever beam tests of plate.

Other precracked specimen configurations, such as those shown in Fig. 24, can be used for special testing

conditions. Information on the preparation and use of these specimens and the related fracture mechanics equations are given in Ref 26 and 44, 45, 46. Preparation of Precracked Specimens When using precracked SCC test specimens, the investigator must consider the dimensional (size) requirements of the specimen, its crack configuration and orientation, and machining and precracking of the specimen. These considerations are discussed below. Additional guidelines and recommendations on specimen preparation in conjunction with fracture toughness testing are given in Ref 26, 27 and 44, 45, 46. Dimensional Requirements. A basic requirement of all precracked specimen configurations is that the dimensions be

sufficient to maintain predominantly triaxial stress (plane-strain) conditions, in which plastic deformation is limited to a very small region in the vicinity of the crack tip. Experience with fracture toughness testing has shown that for a valid KIc measurement neither the crack depth a nor the thickness B should be less than 2.5(KIc/YS)2, where YS is the yield strength of the material (Ref 28). Because of the uncertainty regarding a minimum thickness for which an invariant value of KISCC can be obtained, guidelines for designing fracture mechanics test specimens should be tentatively followed for SCC test specimens. The threshold stress intensity value should be substituted for KIc in the above expression as a test of its validity. If specimens are to be used for determination of KISCC, the initial specimen size should be based on an estimate of the KISCC of the material. Overestimation of the KISCC value is recommended; therefore, a larger specimen should be used than may eventually be necessary. When determining stress-corrosion crack growth behavior as a function of stress intensity, specimen size should be based on the highest stress intensity at which crack growth rates are to be measured (substitute KIo in the 2.5(KIc/YS)2 expression). Notch Configuration and Orientation. For SCC testing, the depth of the initial crack-starter notch--that is, the

machined slot with a fatigue or mechanical pop-in crack at its apex--can be as short as 0.2W. Guidelines for the depth of the notch depend on the limits of accurate KI calibration with respect to the range of a/W or a/H and the considerations discussed previously for double-beam specimens. Several designs of crack-starter notches are available for most plate specimens. The machined slot is used to simulate a crack, because it is impractical to produce plane cracks of sufficient size and accuracy in plate specimens. ASTM E 399 (Ref 27) recommends that the notch root radius should not be greater than 0.127 mm (0.005 in.), unless the chevron form is used, in which case it may be 0.25 mm (0.01 in.) or less (Fig. 26). This tolerance can be easily achieved with conventional milling and grinding equipment. A significant factor in the SCC testing of thick sections of some metals, such as aluminum and titanium, is the direction of applied stress relative to the grain structure. A standardized plan for identifying the loading direction, the fracture plane, and the direction of crack propagation is shown in Fig. 28.

Fig. 28 Specimen orientation and fracture plane identification. L, length, longitudinal, principal direction of metal working (rolling, extrusion, axis of forging); T, width, long-transverse grain direction; S, thickness, shorttransverse grain direction; C, chord of cylindrical cross section; R, radius of cylindrical cross section. First letter: normal to the fracture plane (loading direction); second letter: direction of crack propagation in fracture plane.

Source: Ref 27

Machining. Specimens of the required orientation should be machined from products in the fully heat-treated and stressrelieved condition to avoid complications due to residual stresses in the finished specimens. Safeguards against the presence of residual stresses are especially important for precracked specimens because these specimens are usually bulky and contain notches that are machined deep into the metal. For specimens of material that cannot easily be completely machined in the fully heat-treated condition, the final thermal treatment can be given before the notching and finishing operations. However, fully machined specimens should be heat treated only when the heat treatment will not result in distortion, residual stress, quench cracking, or detrimental surface conditions. Precracking. Fatigue precracking should be done in accordance with ASTM E 399 (Ref 27). The K level used for

precracking each specimen should not exceed about two thirds of the intended starting K-value for the environmental exposure. This prevents fatigue damage or residual compressive stress at the crack tip, which may alter the SCC behavior, particularly when testing at a K level near the threshold stress intensity for the specimen. Aluminum alloy specimens can also be precracked by pop-in methods (wedge-opening loaded to the point of tensile overload), but steel and titanium alloys are usually too strong and tough to pop in without breaking off one of the specimen arms. Chevron notches are usually used to facilitate starting such mechanical precracks, and face grooves are sometimes necessary to produce straight precracks in tougher alloys (Fig. 26). These modifications may also be necessary to control fatigue precracking of some materials. When a specimen is mechanically precracked by pop in, the load should be maintained and should not be reduced for testing at a lower initial K-value. Reducing the load (crack mouth opening displacement) required for pop in will result in residual compressive stress at the crack tip, which could interfere with SCC initiation. When testing specimens at a relatively low fraction of KIc, fatigue precracking is recommended. Testing Procedure For all methods using precracked specimens, the primary objective is usually to determine KISCC or Kth, threshold stress intensity for SCC for the alloy and environment combination. One procedure, similar to that used with smooth specimens, depends on the initiation of SCC at various levels of applied KIo values. Both constant load (K-increasing) and constantdisplacement (K-decreasing) tests can be used. The latter procedure, which is unique to precracked specimens, involves crack arrest. This technique requires a K-decreasing constant-displacement test. These methods are compared in Fig. 29, which illustrates the shift in the stress intensity factor as SCC growth occurs.

Fig. 29 Schematic comparison of determination of KISCC by crack initiation versus crack arrest. (a) Constantload test. (b) Constant crack opening displacement test. a0 = depth of precrack associated with the initial stress intensity KIo; Vpl = plateau velocity.

K-Increasing Versus K-Decreasing Tests. In constant-load specimens (K-increasing tests), stress parameters can

be quantified with confidence. Because crack growth results in an increasing crack opening, there is less likelihood that corrosion products will block the crack or wedge it open. Crack-length measurements can be made readily with several continuous-monitoring methods.

A wide selection of constant-load specimen geometries are available to suit the test material, experimental facilities, and test objective. Therefore, crack growth can be studied under either bend or tension loading conditions. Specimens can be used to determine KISCC by the initiation of a stress-corrosion crack from a preexisting fatigue crack using a series of specimens or to measure crack growth rates. The principal disadvantages of constant-load specimens are the expense and bulk associated with the need for an external loading system. Bend specimens can be tested in relatively simple cantilever beam equipment, but specimens subjected to tension loading require constant-load creep-rupture equipment or similar testing machines. In this case, expense can be minimized by testing chains of specimens connected by loading links that are designed to prevent unloading upon failure of individual specimens. Because of the size of these loading systems, it is difficult to test constant-load specimens under operating conditions, but they can be tested in environments obtained from operating systems. Constant-displacement specimens (K-decreasing tests) are self-loaded; therefore, external stressing equipment is not required. Their compact dimensions also facilitate exposure to operating service environments. They can be used to determine KISCC by the initiation of stress-corrosion cracks from the fatigue precrack, in which case a series of specimens must be used to bracket the threshold value. This can also be achieved by the arrest of a propagating crack, because under constant-displacement testing conditions stress intensity decreases progressively as crack propagation occurs. In this case, a single specimen suffices in principle; in practice, the use of several replicate specimens is recommended to assess variability in test results. Constant-displacement specimens are subject to several inherent disadvantages. Oxide formation or corrosion products can wedge the crack surfaces open, thus changing the applied displacement and load. Oxide formation or corrosion products can also block the crack mouth, thus preventing the entry of corrodent, and can impair the accuracy of crack length measurements by electrical resistance methods. Applied loads can be measured only indirectly by displacement changes or by other sophisticated instrumentation. Crack arrest must be defined by an arbitrary crack growth rate below which it is impractical to measure cracks accurately (commonly about 10-10 m/s, or 1.5 × 10-5 in./h). Loading Arrangements and Crack Measurement. To monitor crack propagation rate as a function of decreasing

stress intensity when testing constant-displacement loaded specimens, two of the three testing variables must be measured--crack depth (ai) or load (Pi), and crack opening displacement at the load line (VLL). Although crack initiation and growth can be detected from change in either load or crack length, load change is usually more sensitive to these conditions. Therefore, crack advance is easier to detect in specimens loaded in a testing machine, an elastic loading ring, or an instrumented bolt than in specimens loaded with a bolt or wedge. 30(a) and Figure 30(b) illustrates typical loading arrangements for which load changes can be automatically monitored (Ref 3, 33, 47).

Fig. 30(a) Wedge-opening load specimen loaded with instrumented bolt. Source: Ref 47

Fig. 30(b) Ring-loaded wedge-opening load specimen test setup. Box to the left of loading rings contains analog signal conditioning for load and displacement signals. The digital data acquisition system consists of a scanner connected to the analog load and displacement signals, a digital voltmeter, and a portable computer used to read and store data and to control the other instruments. Source: Ref 3

Figure 31 illustrates an ultrasonic method of measuring crack length at the interior (midwidth and quarter widths) of a bolt-loaded double-beam specimen. This method provides a more accurate measure of crack depth than visual measurements made on the specimen surfaces. Various other techniques have been used, such as measurement of beam deflection for cantilever beam specimens (Ref 29) and changes in electrical resistance. Such arrangements, however, require calibration. It is feasible and desirable to obtain crack length measurements with a precision of at least ±0.127 mm (0.005 in.).

Fig. 31 Ultrasonic crack measurement system for double-beam specimens. Bolt-loaded specimen is mounted on translation stage at center. Ultrasonic transducer is located above specimen, and the oscilloscope at left indicates (left to right) the top of the specimen, the crack plane, and the bottom face reflection. Digital readouts of stage position and peak height for the crack front measurement used to make consistent positioning measurements are shown (right). This system has a crack growth resolution of approximately 0.127 mm (0.005 in.). Source: Ref 3

Exposure to Environment. When practical for laboratory accelerated testing, the test environment should be brought into contact with the specimen before it is stressed or immediately afterward; this enhances access of the corrodent to the crack tip to promote earlier initiation of SCC and to decrease variability in test results. Similarly, in certain cases, it may also be beneficial to introduce the corrodent even earlier, that is, during precracking. However, unless facilities are available to begin environmental exposure immediately after precracking, corrodent remaining at the crack tip may promote blunting due to corrosive attack. In addition, corrosion of the specimen surfaces in the small volume of the precrack or the advancing stress-corrosion crack will change the composition of the environment that is in contact with the crack tip and can significantly affect the test results. Therefore, hydrolysis reactions can drastically reduce the pH of the aqueous test environment (Ref 48) and can induce embrittlement of some steels by corrosion product hydrogen. Selection of an appropriate test duration presents problems that vary with the testing system; this includes the

alloy and metallurgical condition, the test environment, and the loading method. Errors in interpretation of the test results can be caused by test durations that are either too short or too long. The optimum length of exposure can be best approached through recognition of meaningful crack propagation rates. What is considered meaningful depends on the available precision of measurement of crack lengths and an acceptably low rate for the criterion of a stress intensity threshold (Fig. 3). A problem also exists with the correlation of SCC crack growth rates in the laboratory test and in an anticipated service environment. The question leads ultimately to the intended application and a determination of what is a tolerable amount of SCC growth for a given length of time. Calculation of Crack Growth Rates. There is no generally accepted procedure for calculating crack growth rate,

da/dt, as a function of stress intensity from crack growth curves. Various approaches exist; the simplest is a graphical a/ t technique that may incorporate smoothing of the a versus t curve (Ref 35, 36, 37). Another widely used approach is smoothing of the crack growth curve by computer techniques for curve fitting the entire a versus t curve by a multipleterm polynomial function (Ref 29). Other techniques include a secant method and an incremental polynomial method, in which derivatives of the smoothed crack growth curve are calculated at various points to determine instantaneous crack growth rates. Instantaneous growth rates are then plotted against the instantaneous stress intensities. KIi, at corresponding time intervals to obtain graphs similar to that shown in Fig. 3. Additional information on the secant and incremental methods, which are often used in fatigue studies, can be found in the article "Fatigue Crack Growth Data Analysis" in Mechanical Testing, Volume 8 of ASM Handbook, formerly 9th Edition Metals Handbook. A limited study of the above four methods of treating crack growth data is presented for a high-strength aluminum alloy in Ref 3. All of the methods used to calculate crack growth rates produced the same general results, which were difficult to interpret because of large amounts of scatter resulting from the use of small crack growth increments. Moreover, the

significance of such graphs is dubious when the corrosivity of the environment and the length of exposure can invalidate the estimate of K by causing gross corrosion product wedging effects and/or crack branching. Reduction of crack length data becomes useless without prior subjective interpretation of crack length versus time curves. Allowances should be made for extraneous effects caused by erratic or apparent initiation of stress-corrosion crack growth, scatter in the measurement data due to excessive crack front curvature, multiple crack planes, crack-tip branching, and gross wedging caused by corrosion products. A simple method of comparing materials by using crack growth curves is based on average growth rates taken from an exposure time of zero to an arbitrary time that is sufficient to achieve significant crack extension in the most SCCsusceptible materials being compared (Ref 41). This method not only rapidly identifies materials with relatively low resistance to SCC, but also provides numerical test results for highly resistant materials that may not develop a KI versus da/dt curve with a definite plateau (see the section "Testing of Aluminum Alloys" in this article).

Dynamic Loading: Slow Strain Rate Testing The most recently developed method for accelerating the SCC process in laboratory testing involves relatively slow strain rate tension testing of a specimen during exposure to appropriate environmental conditions. The application of slow dynamic strain exceeding the elastic limit assists in the SCC initiation. This accelerating technique is consistent with the various proposed general mechanisms of SCC, most of which involve plastic microstrain and film rupture. Slow strain rate tests can be used to test a wide variety of product forms, including parts joined by welding. Tests can be conducted in tension, in bending, or with plain, notched, or precracked specimens. The principal advantage of slow strain rate testing is the rapidity with which the SCC susceptibility of a particular alloy and environment can be assessed. Slow strain rate testing is not terminated after an arbitrary period of time. Testing always ends in specimen fracture, and the mode of fracture is then compared with the criteria of SCC susceptibility for the test material. In addition to its timesaving benefits, less scatter occurs in the test results. Comprehensive discussions on the slow strain rate testing technique can be found in Ref 49, 50, 51, 52. Critical Strain Rate. The most significant variable in slow strain rate testing is the magnitude of strain rate. If the strain rate is too high, ductile fracture will occur before the necessary corrosion reactions can take place. Therefore, relatively low strain rates must be used. However, at too low a strain rate, corrosion may be prevented because of repassivation or film repair so that the necessary reactions of bare metal cannot be sustained, and SCC may not occur. Although typical critical strain rates range from 10-5 to 10-7 s-1 depending on the alloy and environment system, the most severe strain rate must be determined in each case.

The repassivation reaction that is observed at very low strain rates and that prevents the formation of anodic SCC does not occur when cracking is the result of embrittlement by corrosion product hydrogen. This mechanistic difference can be used to distinguish between anodic SCC (active path corrosion) and cathodic SCC (hydrogen embrittlement) as illustrated in Fig. 32.

Fig. 32 Schematic showing the effect of strain rate on SCC and hydrogen-induced cracking. Source: Ref 53

The fastest strain rate that will promote SCC in a given system depends on crack velocity. Generally, the lower the stresscorrosion cracking velocity, the slower the strain rate required. Applied strain rates known to have promoted SCC in metal/environment systems are listed in Table 2. Table 2 Critical strain rate regimes promoting SCC in various metal/environment systems System

Applied strain rate, s-1

Aluminum alloys in chloride solutions

10-4 and 10-7

Copper alloys in ammoniacal and nitrite solutions

10-6

Steels in carbonate, hydroxide, or nitrate solutions and liquefied ammonia

10-6

Magnesium alloys in chromate/chloride solutions

10-5

Stainless steels in chloride solutions

10-6

Stainless steels in high-temperature solutions

10-7

Titanium alloys in chloride solutions

10-5

The most relevant strain rates for various aluminum alloys are illustrated in Fig. 33. These trends illustrate that slow strain rate tests should be performed in a strain rate regime that is appropriate for the given alloy and environment system.

Fig. 33 Strain rate regimes for studying SCC of various aluminum alloys. Corrodent: 3% sodium chloride plus 0.3% hydrogen peroxide. Source: Ref 52

Test Specimen Selection. Standard tension specimens (ASTM E 8) (Ref 54) are generally recommended for use with the specified conditions of gage lengths, radii, and so on, unless specialized studies are being conducted. For initially smooth specimens, the strain rate at the onset of the test is clearly defined: however, once cracks have initiated and grown, straining is likely to concentrate in the vicinity of the crack tip, and the effective strain rate is unknown. Rigorous solutions for determining the strain rate at crack tips or notches are not available, but effective strain rates are likely to be higher than for the same deflection rate applied to plain specimens.

Notched or precracked specimens can be used to restrict cracking to a given location--for example, when testing the heataffected zone associated with a weld. Notched or precracked specimens can also be used to restrict load requirements where bending, as opposed to tensile loading, may offer an added benefit. The section thickness or diameter of such specimens is usually relatively small, so the testing duration is short. Testing Equipment. Constant strain rate apparatus requirements include sufficient stiffness to resist significant

deformation under the loads necessary to fracture the test specimens; a system to provide reproducible, constant strain rates over the range of 10-4 to 10-8 s-1; and a cell to contain the test solution. Auxiliary equipment is used to control environmental conditions and to record test data. The testing equipment can also be instrumented to record loadelongation curves, which is convenient when testing at various strain rates. A typical constant strain rate unit is illustrated schematically in Fig. 34. Various types of corrosion cells may be required to control the test conditions for specific studies.

Fig. 34 Typical slow strain rate test apparatus. Source: Ref 51

In addition to uniaxial tensile units, cantilever constant strain rate apparatus has also been used in which an extension arm attached to a cantilever beam specimen is lowered at a constant rate. This technique has been successfully used to study SCC of low-carbon steel in carbonate-bicarbonate environments to determine crack velocity, critical strain rates, and

inhibitor effectiveness (Ref 55). Additional information on slow strain rate testing equipment and procedures is available in Ref 45 and 49. Assessment of Results. Historically, the principal methods of SCC assessment derived from slow strain rate tension

testing were based on time to failure, maximum gross section stress developed during the tension test, percent elongation, area bounded by the load-elongation curve, and reduction in area. Figure 35 depicts stress-elongation curves that illustrate how stress-corrosion cracks influence the elongation to fracture as well as the maximum load.

Fig. 35 Nominal stress versus elongation curves for carbon-manganese steel in slow-strain rate test in boiling 4 N sodium nitrate and in oil at the same temperature. Source: Ref 50

To eliminate non-SCC effects, parallel tests are conducted in an inert environment, and a ratio of the result obtained in the corrodent divided by the result obtained in the inert environment is commonly used as an index of SCC susceptibility. For example, in Fig. 33, higher SCC resistance is denoted by higher ductility ratios. Figure 36 shows a stress-corroded specimen containing many secondary stress-corrosion cracks and reduced ductility at fracture. Some alloys experience rapid deterioration of mechanical properties on contact with certain corrosive environments; any additional effect of applied straining can best be assessed by comparison with the behavior of unstrained specimens. Therefore, it is essential that the cause of environmental degradation be verified as SCC.

Fig. 36 Photomacrographs of two carbon steel specimens after slow strain rate tests conducted at a strain rate of 2.5 × 10-6 s-1 and 80 °C (180 °F). The ductility ratio in this example was 0.74 (original diameter: 2.54 mm, or 0.100 in.). (left) Ductile fracture in oil. (right) SCC in carbonate solution

Slow strain rate testing is very efficient in comparing environments in terms of their capability to produce SCC, for example, in steels having similar metallurgical characteristics. However, such comparisons are difficult and not very reliable when applied to groups of steels with different characteristics (Ref 53). Slow strain rate testing as generally used does not provide data that can be used for design purposes. Recent work, however, has shown that average SCC velocities, threshold stresses, and threshold strain rates can be obtained with modified techniques combined with microscopy (Ref 50, 55, 56). For example, average SCC crack velocities can be determined from the depth of the largest crack measured on the fracture surfaces of specimens that have failed completely, or in longitudinal sections on the diameter of specimens that have not experienced total failure, divided by the time of testing. With this procedure, SCC is assumed to initiate at the start of the test, which is not always true. With precracked specimens, other methods can be used to monitor crack growth and thus allow determination of crack velocities. The SCC behavior of a pipeline steel (Fig. 37) has been studied by using a precracked cantilever bend specimen in terms of threshold strain rate for crack growth and also in terms of crack growth rates analogous to the Stage II plateau velocity illustrated in Fig. 3. Material properties, such as strength and toughness, that influence SCC performance when measured by tension testing are eliminated as factors; therefore, valid comparisons can be made of alloys with widely different structures and mechanical properties. Additional information on this method of assessment and the effects of strain rate can be found in Ref 57, 58, 59.

Fig. 37 Effects of beam deflection rate on stress-corrosion crack velocity in precracked cantilever bend specimens of a carbon-manganese steel. Tested in a carbonate-bicarbonate solution at 75 °C (165 °F) and at a potential of -650 mV versus SCE. Source: Ref 50

Selection of Test Environments The primary environmental factors in SCC testing are the nature and concentration of anions and cations in aqueous solutions, electrochemical potential, solution pH, the partial pressure and nature of species in gaseous mixtures, and temperature. Separately or in combination, environmental variables can have a profound effect on the thermodynamics and kinetics of the electrochemical processes that control environmentally assisted fracture. Therefore, the choice of environmental conditions provides an important basis for developing accelerated SCC test methods. The environmental requirements for SCC vary with different alloys. Although a mechanical precrack or a critical strain rate provides a worst case for SCC from a mechanical standpoint, there does not appear to be a generally applicable worst

case from an environmental standpoint. However, because the presence of moisture and salt water is universal, the SCC characteristics of alloys in these environments--as well as in any special environment a given engineering structure may experience--are always of interest. Figure 38 illustrates that electrochemical factors can override mechanical factors in determining SCC initiation sites. Three cantilever beam specimens of PH13-8Mo stainless steel were tested in salt water. Specimen A was tested at a high K level. With the participation of the chloride ions, the protective oxide film ruptured at the bottom of the precrack and initiated SCC, which was halted before the beam fractured completely. Specimen B was loaded at a lower K level. After 1300 h, a stress-corrosion crack initiated, but not in the precrack. Crack initiation occurred under the wall of the cell that surrounded the central portion of the specimen and contained the salt water.

Fig. 38 Cantilever beam specimens of PH13-8Mo stainless steel after testing. Experiments demonstrate that electrochemical factors can override mechanical factors in determining initiation sites of SCC. See text for details. Source: Ref 60

Careful examination of this specimen and replicate specimens revealed small crevice corrosion pits under the wall that initiated SCC in an almost smooth surface. Even if these small pits had been as sharp as a fatigue crack, the K level would have been much lower than at the machined and fatigued notch. In the stagnant situation under the cell wall, the stainless steel reacted with the salt water to form hydrochloric acid and other corrosion products from the metal. Therefore, the low pH in a crevice, due to the hydrolysis of chromium corrosion product, overcame the mechanical disadvantage of the lack of a precrack. Specimen C was then tested to verify the effectiveness of electrochemical conditions in crack initiation. Saturated ferric chloride was selected to lower the pH to the range inside an active corrosion pit in the stainless steel; application of the solution to the unnotched beam resulted in the immediate initiation of many cracks in the smooth surface. Hydrochloric acid was found to be equally effective.

Stress-corrosion tests can be divided into two broad environmental classes: those conducted in actual service environments and those conducted under laboratory conditions. Service Environments and Field Testing. The following examples illustrate the value, and in some cases the necessity, of exposure tests performed in actual service environments as an adjunct to laboratory evaluation. The standard 3.5% sodium chloride alternate immersion test data for aluminum alloys 2024 and 7075 proved useless in predicting the serviceability of these aluminum alloys for handling rocket propellant oxidizers such as nitrogen tetroxide and inhibited red fuming nitric acid (Ref 61). The alternate immersion test showed 2024-T351 and 7075-T651 to be susceptible to SCC at low short-transverse stresses, but 2024-T851 and 7075-T7351 were quite resistant even at high stresses. These data were supported by outdoor field tests in seacoast and industrial atmospheres.

However, in proof tests consisting of exposure to the actual service environment--inhibited red fuming nitric acid at 74 °C (165 °F)--SCC occurred in both tempers of 7075 alloy and did not occur in either temper of 2024 alloy (Fig. 39). There were no unexpected failures with the 2219-T87 and 6061-T651 materials, however.

Fig. 39 SCC resistance of various aluminum alloys in inhibited red fuming nitric acid versus alternate immersion in 3.5% sodium chloride solution. Each bar graph represents an individual short-transverse C-ring test specimen machined from rolled plate and stressed at the indicated level. Source: Ref 61

Simulated-service tests should be conducted under conditions duplicating the service environment exactly, as illustrated by the following example with Ti-6Al-4V alloy pressure tanks for propellant-grade nitrogen tetroxide ( Se > Te > S > Bi The role of sulfur as a poison is particularly important because sulfur is commonly encountered and because the chemical form of the sulfur greatly influences its effectiveness as a hydrogen entry promoter. The susceptibility to embrittlement by hydrogen can be demonstrated by the relative resistance to cracking in such environments as wet hydrogen sulfide. In such tests, microstructure has a definite effect on susceptibility. In steels, untempered martensite is the most susceptible phase. Lamellar carbide structures are less desirable than those with spheroidized structures. Quenched-and-tempered microstructures are more resistant than those that have been normalized and tempered (Ref 88). For the same strength level in low-alloy steel, it has been shown that a bainitic structure is more resistant to hydrogen-assisted cracking than a quenched-and-tempered martensitic structure (Ref 89).

Fig. 6 Effect of anion and temperature on hydrogen absorption in a low-carbon steel. All acid concentrations were 2 N. Source: Ref 86

Embrittlement by gaseous hydrogen environments at ambient temperature has been effectively inhibited by the addition of 0.4 to 0.7 vol% oxygen (Ref 82, 90, 91). However, similar additions to a hydrogen sulfide gas environment did not halt the growth of cracks (Ref 91, 92). Because of the higher hydrogen solubility in the high-temperature fcc structure of iron (versus the low-temperature bcc structure), cooling of steel in hydrogen atmospheres from temperatures of the order of 1100 °C (2010 °F) can result in internal damage. Exceeding the solubility limit for hydrogen will result in the embrittlement of hydrogen-sensitive microstructures, such as martensite, formed by rapid cooling of some ferritic alloys. The internal precipitation of hydrogen is believed to be responsible for the generation of fissures, delaminations, or other defects. Such defects have been termed flakes, shatter-cracks, fisheyes, or snowflakes. The defects are generally associated with hydrogen precipitation at voids, laminations, or inclusion/matrix interfaces already present in the steel. A reduced cooling rate, which allows hydrogen to be slowly released from the steel, is a general solution to the problem. Slower cooling will also inhibit the formation of hydrogen-sensitive microstructures. Underbead cracking is an embrittlement phenomenon that is associated with absorption of hydrogen by molten metal during the welding process. Sources of hydrogen include moisture or organic contaminants on the surface of the prepared joint, moisture in low-hydrogen coated electrodes (such as E7018), moisture in fluxcored wire (such as M16), or a highhumidity environment. Upon rapid cooling of the weld, entrapped hydrogen can produce internal fissuring or other damage, as described earlier. In addition, the weld HAZ may contain the martensite phase in quench-hardenable alloys. The HAZ is then embrittled by high levels of entrapped hydrogen. Several steels have exhibited susceptibility to such embrittlement--for example, carbon steels containing 0.25 to 0.35 wt% C, low-alloy steels (such as AISI 4140 to 4340), and martensitic or precipitation-hardening stainless steels. Solutions to the hydrogen damage problems associated with

welding include the use of dry welding electrodes, proper cleaning and degreasing procedures for prepared weld joints, the use of an appropriate preheat before welding, and an adequate postweld heat treatment. Welding electrodes should be kept dry by using a heated rod box. The electrodes should be removed only as needed. If they are moistened or exposed in the ambient atmosphere for prolonged periods, low-hydrogen coated electrodes must be heated at 370 to 425 °C (700 to 800 °F) to remove moisture (Ref 63). Recommended preheat temperatures for steels, as a function of steel composition, section thickness, and electrode type, have been published (Ref 93). Welding procedures for the avoidance of hydrogen cracking in carbon-manganese steels have also been published (Ref 94). Appropriate postweld heat treatments for steels can range from a hydrogen bake-out at 175 °C (350 °F) to a martensite tempering treatment at temperatures as high as 705 °C (1300 °F) (Ref 63). Hydrogen attack is a damage mechanism that is associated with unhardened carbon and low-alloy steels exposed to hydrogen-containing environments at temperatures above 220 °C (430 °F) (Ref 63). Exposure to the environment is known to result in a direct chemical reaction with the carbon in the steel. The reaction occurs between absorbed hydrogen and the iron carbide phase, resulting in the formation of methane:

2H2 + Fe3C → CH4 + 3Fe Unlike nascent hydrogen, the resulting methane gas does not dissolve in the iron lattice. Internal gas pressures develop, leading to the formation of voids, blisters, or cracks. The generated defects lower the strength and ductility of the steel. Because the carbide phase is a reactant in the mechanism, its absence in the vicinity of generated defects serves as direct evidence of the mechanism itself. The recommended service conditions (temperature, hydrogen pressure) for carbon and low-alloy steels are shown by the respective Nelson curves in Fig. 7. Chromium and molybdenum are beneficial alloying elements. This is most likely the result of their high affinity for carbon as well as the stability of their carbides. Hydrogen attack does not occur in austenitic stainless steels (Ref 63). In carbon or low-alloy steels, the extent of hydrogen attack is a function of exposure time.

Fig. 7 Operating limits for three steels in hydrogen service to avoid hydrogen attack. Dashed lines show limits for decarburization, not hydrogen attack. Source: Ref 63

Hydrogen blistering is a mechanism that involves hydrogen damage of unhardened steels near ambient temperature. It is known that the entry of atomic hydrogen into steel can result in its collection, as the molecular species, at internal defects or interfaces. If the entry kinetics are substantial (promoted by an acidic environment, high corrosion rates, and cathodic poisons), the resulting internal pressure will cause internal separation (fissuring or blistering) of the steel. Such damage typically occurs at large, elongated inclusions and results in delaminations known as hydrogen blisters. Field experience indicates that fully killed steels are more susceptible than semikilled steels (Ref 95), but the nature and size of the original inclusions appear to be the key factors with regard to susceptibility. Rimmed steels or free-machining grades with high levels of sulfur or selenium would most likely show a high susceptibility to blistering. Stepwise cracking at the ends of blisters indicates an effect of elongated inclusions in the delamination process (Ref 63, 95). Similar stepwise cracking occurs in the hydrogen-induced failure of low-alloy pipeline steels (Ref 96). Both stepwise cracking and blistering appear to be limited to environments in which acidic corrosion occurs and in which cathodic poisons, such as sulfide, are present to promote hydrogen entry. Solutions to the blistering problem include the use of low-sulfur calcium-treated argon-blown steels. Hot-rolled or annealed (as opposed to cold-rolled) steel is preferred (Ref 63). Silicon-killed steels are preferable to aluminum-killed steels. Also, treatment with synthetic slag or the addition of rare-earth metals can favor the formation of less detrimental globular sulfides (Ref 97). Ultrasonic inspection of the steel (according to ASTM E 114 and A 578) should be done before fabrication to detect laminations and other discontinuities that will promote blister formation. Equipment inspection and blister-venting procedures require unusual care (Ref 63). In services in which blistering can be expected, external support pads should not be continuously welded to the vessel itself. This will prevent hydrogen entrapment at the interface. The permeation of hydrogen through ferritic steels can produce physical separation at mechanical joints. For example, bimetallic tubes, with a carbon steel inner liner, exhibited collapse of the liner due to its exposure to HF. Acid corrosion of the inside surface allowed nascent hydrogen to permeate the steel. Molecular hydrogen gas was formed, and trapped, at the interface with the outer tube section (brass). The accumulation of pressure was found to collapse the inner steel liners (Ref 63). In high-temperature H2/H2S service, weld overlaid 2.25Cr-1Mo steel was found to disbond at the weld interface (Ref 98). In this case, a weld overlay of type 309 stainless steel, followed by type 347 stainless steel, was applied. Hydrogeninduced cracking was found to occur in the transition zone below the weld metal after approximately 3

1 years of service. 2

The disbonding was found to be more severe with higher cooling rates after hydrogen absorption. Out-gassing treatments during the cool down were found to prevent disbonding (Ref 99). Figure 8 shows an example of hydrogen-assisted SCC failure of four AISI 4137 steel bolts having a hardness of 42 HRC. Although the normal service temperature (400 °C, or 750 °F) was too high for hydrogen embrittlement, the bolts were also subjected to extended shutdown periods at ambient temperatures. The corrosive environment contained trace hydrogen chloride and acetic acid vapors as well as calcium chloride if leaks occurred. The exact service life was unknown. The bolt surfaces showed extensive corrosion deposits. Cracks had initiated at both the thread roots and the fillet under the bolt head. Figure 8(b) shows a longitudinal section through the failed end of one bolt. Multiple, branched cracking was present, typical of hydrogen-assisted SCC in hardened steels. Chlorides were detected within the cracks and on the fracture surface. The failed bolts were replaced with 17.4 PH stainless steel bolts (Condition H1150M) having a hardness of 22 HRC (Ref 63).

Fig. 8 4137 steel bolts (hardness: 42 HRC) that failed by hydrogen-assisted SCC caused by acidic chlorides from a leaking polymer solution. (a) Overall view of failed bolts. (b) Longitudinal section through one of the failed bolts in (a) showing multiple, branched hydrogen-assisted stress-corrosion cracks initiating from the thread roots. Source: Ref 63

As an example of hydrogen attack, a section of plain carbon steel (0.22% C and 0.31% Si) had been mistakenly included as a part of a type 304 stainless steel hot-gas bypass line used to handle hydrogen-rich gas at 34 MPa (5000 psi) and 320 °C (610 °F). After 15 months of service, the steel pipe section ruptured, causing a serious fire. Figure 9 shows a section of

3 -in.) OD pipe near the fracture. The pipe had been weakened by hydrogen attack through all but 0.8 mm 4 5 of the 8-mm ( -in.) thick wall. As a result of the hydrogen attack and the internal methane formation, the 16 the 44-mm (1

microstructural damage consisted of holes or voids near the outer surface as well as interconnected grain-boundary fissures in a radial alignment near the inner surface (Fig. 9b). The radially aligned voids preceded both the circumferential crack and pipe rupture (Ref 63).

Fig. 9 Section of ASTM A 106 carbon steel pipe with wall severely damaged by hydrogen attack. The pipe failed after 15 months of service in hydrogen-rich gas at 34.5 MPa (5000 psig) and 320 °C (610 °F). (a) Overall view of failed pipe section. (b) Microstructure of hydrogen-attacked pipe near the midwall. Hydrogen attack produced grain-boundary fissures that are radially aligned. Source: Ref 63

Hydrogen blistering is illustrated in Fig. 10, which shows a cross section of a 152-mm (6-in.) diam blister that formed in the wall of a steel sphere. The sphere had been used to store anhydrous HF for 13.5 years at ambient temperatures. The source of nascent hydrogen gas was the cathodic hydrogen generated by the corrosion reaction between the acid and the steel. The corrosion rate was less than 0.05 mm/yr (2 mils/yr). Figure 10(b) shows the propagation of the blister, with the stepwise cracking (arrow) at its edge caused by the buildup of hydrogen pressure within the blister itself (Ref 63). More information on hydrogen attack is available in the article "Environmentally Induced Cracking" in this Volume.

Fig. 10 Hydrogen blister in 19-mm (

3 -in.) 4

steel plate from a spherical tank used to store anhydrous HF for

13.5 years. (a) Cross section of 152-mm (6-in.) diam blister. (b) Stepwise cracking (arrow) at edge of hydrogen blister shown in (a). Source: Ref 63

Erosion-corrosion is a frequently misinterpreted type of metal deterioration that results from the combined action of

erosion and corrosion. This section will be limited to a discussion of three types--liquid erosion-corrosion, cavitation, and fretting. Abrasive wear, which is erosion without corrosion, will also be discussed for comparison purposes. Liquid erosion-corrosion is the accelerated wastage of a metal or material attributed to the flow of a liquid (Ref 100,

101, 102). Liquid erosion-corrosion damage is characterized by grooves, waves, gullies, rounded holes, and/or horseshoeshaped grooves. Analysis of these marks can help determine the direction of flow. Most metals are susceptible to liquid erosion-corrosion under specific conditions. Carbon steels, for example, can be severely damaged by steam containing entrained water droplets. By contrast, the 300-series stainless steels at about the same hardness and strength level are very resistant to flowing wet steam. Virtually anything that is exposed to a moving liquid is susceptible to liquid erosioncorrosion. Examples include piping systems, particularly at bends, elbows, or wherever there is a change in flow direction or increase in turbulence; pumps; valves, especially flow control and pressure let-down valves; centrifuges; tubular heat exchangers; impellers; and turbine blades. Surface films that form on some metals and alloys are very important in their ability to enhance resistance to liquid erosion-corrosion. Titanium is a reactive metal, but is resistant to liquid erosion-corrosion in many environments because of its very stable titanium dioxide surface film. The 300-series stainless steels, as mentioned above, are also resistant because of their stable passive surface films. Both carbon steel and lead have relatively good resistance to certain concentrations of H2SO4 under low-to-moderate flow conditions. Both depend on a metal sulfate corrosion product film for resistance; however, both will fail fairly rapidly after removal of the sulfate film, even at low velocities. Another example is the carbon steel and some low-alloy steels used to handle petroleum refinery fluids that contain hydrogen sulfide. At low velocities or under stagnant conditions, these materials are normally satisfactory because of formation of a tenacious protective iron sulfide film. However, with increased velocity, the film is eroded away, and very rapid attack occurs. Velocity often increases attack, but it may also decrease attack, depending on the material of construction and the corrosive environment. For example, increasing the velocity causes accelerated attack of carbon steel in steam condensate by increasing the supply of dissolved oxygen and/or carbon dioxide to the steel surface. In cooling water, however, increased velocity often reduces the attack of carbon steel by improving the effectiveness of inhibitors and by reducing deposits and pitting in stagnant areas. Many 300-series stainless steels are subject to pitting and crevice corrosion in seawater. However, they may exhibit good resistance if the seawater is kept flowing at a minimum critical velocity. This prevents the formation of deposits and retards general corrosion, pitting, and crevice attack. Table 4 shows the effects different seawater velocities have on the liquid erosion-corrosion of various metals.

Table 4 Corrosion of metals and alloys in seawater as a function of velocity Material

Typical corrosion rate, mg/dm2/d

0.3 m/s (1 ft/s)(a)

1.2 m/s (4 ft/s)(b)

8.2 m/s (27 ft/s)(c)

Carbon steel

34

72

254

Cast iron

45

...

270

Silicon bronze

1

2

343

Admiralty brass

2

20

170

Hydraulic bronze

4

1

339

G bronze

7

2

280

10% aluminum bronze

5

...

236

Aluminum brass

2

...

105

90Cu-10Ni (0.8% Fe)

5

...

99

70Cu-30Ni (0.5% Fe)

36)

(21)

(21)

(29)

(13)

(13)

(17)

Profuse

Profuse

Sparse

Sparse

Profuse

Sparse

Profuse

Sparse

>1.19

>0.91

0.28

0.1

47)

(>36)

(11)

(4)

(1.2

>0.9

48)

(>36)

(1.19

0.58

0.61

0.46

0.66

0.33

0.61

0.25

0.15

(>47)

(23)

(24)

(18)

(26)

(13)

(24)

(10)

(6)

Single

Absorber, spray area

6.2

60

140

100

Profuse

Profuse

Profuse

Profuse

Profuse

Sparse

Profuse

Sparse

Sparse

0.58

0.10

nil

nil

nil

nil

nil

nil

nil

(23)

(4)

Sparse

Outlet duct

2-4(d)

1.5(e)

55

170

130(d)

335(e)

100(d)

82000(e)

>1.19

>0.91

0.58

0.58

0.48

0.18

0.51

0.53

0.36

(>47)

(>36)

(23)

(23)

(19)

(7)

(20)

(21)

(14)

Profuse

Profuse

Profuse

Profuse

Profuse

Single

Profuse

Profuse

IG etch

Source: Ref 25 (a) Slurry contained 7000 ppm dissolved Cl-. Deposits in the quencher, inlet duct, absorber, and outlet ducting contained 3000-4000 ppm Cl- and 800-1900 ppm F-.

(b) Present as halide gases.

(c) Not tested.

(d) During operation.

(e) During bypass. Bypass condition gas stream contained SO2, SO3, HCl, HF and condensate.

Nuclear Power Applications. Type 304 stainless steel piping has been used in boiling-water nuclear reactor power

plants. The operating temperatures of these reactors are about 290 °C (550 °F), and a wide range of conditions can be present during startup, operation, and shutdown. Because these pipes are joined by welding, there is a possibility of sensitization. This can result in intergranular SCC in chloride-free high-temperature water that contains small amounts of oxygen, for example, 0.2 to 8 ppm. Nondestructive electrochemical tests have been used to evaluate weldments for this service (Ref 26). Type 304 stainless steel with additions of boron (about 1%) has been used to construct spent-fuel storage units, dry storage casks, and transportation casks. The high boron level provides neutron-absorbing properties. More information on nuclear applications is available in the article "Corrosion in the Nuclear Power Industry" in this Volume. Pulp and Paper Industry. In the kraft process, paper is produced by digesting wood chips with a mixture of Na 2S and

NaOH (white liquor). The products is transferred to the brown stock washers to remove the liquor (black liquor) from the brown pulp. After screening, the pulp may go directly to the paper mill to produce unbleached paper or may be directed first to the bleach plant to produce white paper. The digester vapors are condensed, and the condensate is pumped to the brown stock washers. The black liquor from these washers is concentrated and burned with sodium sulfate (Na2SO4) to recover sodium carbonate (Na2CO3) and Na2S. After dissolution in water, this green liquor is treated with calcium hydroxide (Ca(OH)2) to produce NaOH to replenish the white liquor. Pulp bleaching involves treating with various chemicals, including chlorine, chlorine dioxide (ClO2), sodium hypochlorite (NaClO), calcium hypochlorite (Ca(ClO)2), peroxide, caustic soda, quicklime, or oxygen. The sulfite process uses a liquor in the digester that is different from that used in the kraft process. This liquor contains free SO2 dissolved in water, along with SO2 as a bisulfite. The compositions of the specific liquors differ, and the pH can range from 1 for an acid process to 10 for alkaline cooking. Sulfur dioxide for the cooking liquor is produced by burning elemental sulfur, cooling rapidly, absorbing the SO2 in a weak alkaline solution, and fortifying the raw acid. Various alloys are selected for the wide range of corrosion conditions encountered in pulp and paper mills. Paper mill headboxes are typically fabricated from type 316L stainless steel plate with superior surface finish and are sometimes electropolished to prevent scaling, which may affect pulp flow. The blades used to remove paper from the drums have been fabricated from type 410 and 420 stainless steels and from cold-reduced 22Cr-13Ni-5Mn stainless steel. Evaporators and reheaters must deal with corrosive liquors and must minimize scaling to provide optimum heat transfer. Type 304 stainless steel ferrite-free welded tubing has been used in kraft black liquor evaporators. Cleaning is often performed with HCl, which attacks ferrite. In the sulfite process, type 316 ( 2.75% Mo) and type 317 stainless steels have been used in black liquor evaporators. Digester liquor heaters in the kraft and sulfite processes have used 7-Mo stainless for resistance to caustic or chloride SCC. Bleach plants have used type 316 and 317 stainless steels and are upgrading to austenitic grades containing 4.5 and 6% Mo in problem locations. Tightening of environmental regulations has generally increased temperature, chloride level, and acidity in the plant, and this requires grades of stainless steel that are more highly alloyed than those used in the past. Tall oil units have shifted from type 316 and 317 stainless steels to such candidate alloys as 904L or 20Mo-4 stainless steels and most recently to 254SMO and 20Mo-6 stainless steels. Tests including higher-alloyed materials have been coordinated by the Metals Subcommittee of the TAPPI Corrosion and Materials Engineering Committee. Racks of test samples, which included crevices at polytetrafluoroethylene (PTFE) spacers, were submerged in the vat below the washer in the C (chlorination), D (Chlorine dioxide), and H (hypochlorite) stages of several paper mills. The sum of the maximum attack depth on all samples for each alloy--at crevices and remote from crevices--is shown in Fig. 14(a), 14(b), and 14(c). It should be noted that the vertical axes are different in Fig. 14(a), 14(b), and 14(c). Additional information on corrosion in this industry is available in the article "Corrosion in the Pulp and Paper Industry" in this Volume.

Fig. 14(a) Resistance of austenitic stainless steels containing 2.1 to 4.4% Mo to localized corrosion in paper mill bleach plant environment. Total depth of attack has been divided by 4 because there were four crevice sites per specimen. See also Fig. 14(b) and 14(c). Source: Ref 27

Fig. 14(b) Resistance of austenitic stainless steels containing 4.4 to 7.0% Mo to localized corrosion in paper mill bleach plant environment. Total depth of attack has been divided by 4 because there were four crevice sites per specimen. See also Fig. 14(a) and 14(c). Source: Ref 27

Fig. 14(c) Resistance of ferritic and duplex stainless steels to localized corrosion in paper mill bleach plant environment. Total depth of attack has been divided by 4 because there were four crevice sites per specimen. See also Fig. 14(a) and 14(b). Source: Ref 27

Transportation Industry. Stainless steels are used in a wide range of components in transportation that are both functional and decorative. Bright automobile parts, such as trim, fasteners, wheel covers, mirror mounts, and windshield wiper arms, have generally been fabricated from 17Cr or 18Cr-8Ni stainless steel of similar grades. Example alloys include type 430, 304, 304, and 305 stainless steels. Type 302HQ-FM remains a candidate for such applications as wheel nuts, and Custom 455 stainless has been used as wheel lock nuts. Use of type 301 stainless steel for wheel covers has diminished with the weight reduction programs of the automotive industry.

Stainless steels also serve many nondecorative functions in automotive design. Small-diameter shafts of type 416 and, occasionally, type 303 stainless steels have been used in connection with power equipment, such as windows, door locks, and antennas. Solenoid grades, such as type 430FR stainless steels, have also found application. Type 409 stainless steel has been used for mufflers and catalytic converters for many years, but it is now being employed throughout the exhaust system. The article "Corrosion in the Automotive Industry" in this Volume contains detailed information on corrosion in the automotive environment. In railroad cars, external and structural stainless steels provide durability, low-cost maintenance, and superior safety through crashworthiness. The fire resistance of stainless steel is a significant safety advantage. Modified type 409 stainless steel is used as structural component in buses. Types 430 and 304 are used for exposed functional parts on buses. Type 304 stainless steel has provided economical performance in truck trailers. For tank trucks, type 304 has been the most frequently used stainless steel, but type 316 and higher-alloyed grades have been used where appropriate to carry more corrosive chemicals safely over the highways.

Stainless steels are used for seagoing chemical tankers, with types 304, 316, 317, and alloy 2205 being selected according to the corrosivity of the cargoes being carried. Conscientious adherence to cleaning procedures between cargo changeovers has allowed these grades to give many years of service with a great variety of corrosive cargoes. In aerospace, quench-hardenable and precipitation-hardenable stainless steels have been used in varying applications. Heat treatments are chosen to optimize fracture toughness and resistance to SCC. Stainless steel grades 15-5PH and PH13-8Mo have been used in structural parts, and A286 and PH3-8Mo stainless steels have served as fasteners. Parts in cooler sections of the engine have been fabricated from type 410 or A286 stainless steel. Custom 455, 17-4PH, 17-7PH, and 15-5PH stainless steels have been used in the space shuttle program (see the articles "Corrosion in the Aircraft Industry" and "Corrosion in the Aerospace Industry" in this Volume). Architectural Applications. Typically, type 430 or 304 has been used in architectural applications. In bold exposure these grades are generally satisfactory; however, in marine and industrially contaminated atmospheres, type 316 is often suggested and has performed well (see the article "Corrosion in Structures" in this Volume).

In all applications, but particularly in these cases where appearance is important, it is essential that any chemical cleaning solutions be thoroughly rinsed from the metal.

Corrosion Testing The physical and financial risks involved in selecting stainless steels for particular applications can be reduced through consideration of corrosion tests. However, care must be taken when selecting a corrosion test. The test must relate to the type of corrosion possible in the application. The steel should be tested in the condition in which it will be applied. The test conditions should be representative of the operating conditions and all reasonably anticipated excursions of operating conditions. Corrosion tests vary in their degree of simulation of operation in terms of the design of the specimen and the selection of medium and test conditions. Standard test use specimens of defined nature and geometry exposed in precisely defined media and conditions. Standard tests can confirm that a particular lot of steel conforms to the level of performance expected of a standard grade. Standard tests can also rank the performance of standard and proprietary grades. The relevance of test results to performance in particular applications increases as the specimen is made to resemble more closely the final fabricated structure--for example, bent, welded, stressed, or creviced. Relevance also increases as the test medium and conditions more closely approach the most severe operating conditions. However, many types of failures occur only after extended exposures to operating cycles. Therefore, there is often an effort to accelerate testing by increasing the severity of one or more environmental factors, such as temperature, concentration, aeration, and pH. Care must be taken that the altered conditions do not give spurious results. For example, an excessive temperature may either introduce a new failure mode or prevent a failure mode relevant to the actual application. The effects of minor constituents or impurities on corrosion are of special concern in simulated testing. Pitting and crevice corrosion are readily tested in the laboratory by using small coupons and controlled-temperature

conditions. A procedure for such tests using 6% FeCl3 (10% FeCl3·6H2O) is described in ASTM G 48 (Ref 28). This procedure is performed in 3 days. The coupon may be evaluated in terms of weight loss, pit depth, pit density, and appearance. Several suggestions for methods of pitting evaluation are given in ASTM G 46 (Ref 29). Reference 28 also describes the construction of a crevice corrosion coupon (Fig. 15). It is possible to determine a temperature below which crevice corrosion is not initiated for a particular material and test environment. This critical crevice temperature (CCT) provides a useful ranking of stainless steels. For the CCT to be directly applicable in design, it is necessary to determine that the test medium and conditions relate to the most severe conditions to be encountered in service.

Fig. 15 Assembled crevice corrosion test specimen. Source: Ref 30

Figure 16 shows one of several frequently used specimens with a multiple crevice assembly. The presence of many separate crevices helps to deal with the statistical nature of corrosion initiation. The severity of the crevices can be regulated by means of a standard crevice design and the use of a selected torque in its application.

Fig. 16 Multiple-crevice cylinders for use in crevice corrosion testing. Source: Ref 30

Laboratory media do not necessarily have the same response of corrosivity as a function of temperature as do engineering environments. For example, the ASTM G 48 solution is thought to be roughly comparable to seawater at ambient temperatures. However, the corrosivity of FeCl3 increases steadily with temperature. The response of seawater to increasing temperature is quite complex, relating to such factors as concentration of oxygen and biological activity. Also, the various families of stainless steels will be internally consistent, but will differ from one another in response to a particular medium. Pitting and crevice corrosion may also be evaluated by electrochemical techniques. When immersed in a particular medium a metal coupon will assume a potential that can be measured relative to a standard reference electrode. It is then possible to impress a potential on the coupon and observe the corrosion as measured by the resulting current. Various techniques of scanning the potential range provide extremely useful data on corrosion resistance. Figure 17 demonstrates a simplified view of how these tests may indicate the corrosion resistance for various materials and media.

Fig. 17 Schematics showing how electrochemical tests can indicate the susceptibility to pitting of a material in a given environment. (a) Specimen has good resistance to pitting. (b) Specimen has poor resistance to pitting. In both cases, attack occurs at the highest potentials. Source: Ref 30

The nature of intergranular sensitization has been discussed earlier in this article. There are many corrosion tests for detecting susceptibility to preferential attack at the grain boundaries. The appropriate media and test conditions vary widely for the different families of stainless steels. Table 11 summarizes the ASTM tests for intergranular sensitization. Figure 18 shows that electrochemical techniques may also be used, as in the electrochemical potentiostatic reactivation (EPR) test. Table 11 ASTM standard tests for susceptibility to intergranular corrosion in stainless alloys ASTM test method

Test medium and duration

Alloys

Phases detected

A 262, practice A

Oxalic acid etch; etch test

AISI types 304, 304L, 316, 316L, 317L, 321, 347 casting alloys

Chromium carbide

A 262, practice B

Fe2(SO4)3-H2SO4; 120 h

Same as above

Chromium carbide,

phase(a)

A 262, practice C

HNO3 (Huey test); 240 h

Same as above

Chromium carbide,

phase(b)

A 262, practice D

HNO3-HF; 4 h

AISI types 316, 316L, 317, 317L,

Chromium carbide

A 262, practice E

CuSO4-16%H2SO4, with copper contact; 24 h

Austenitic stainless steels

Chromium carbide

A 708 (formerly A 393)

CuSO4-16%H2SO4, without copper contact; 72 h

Austenitic stainless steels

Chromium carbide

G 28

Fe2(SO4)3-H2SO4; 24-120 h

Hastelloy alloys C-276 and G; 20Cb-3; Inconel alloys 600, 625, 800, and 825

Carbide and/or intermetallic phases(c)

A 763, practice X

Fe2(SO4)3-H2SO4; 24-120 h

AISI types 403 and 446; E-Brite, 29-4, 29-4-2

Chromium carbide and nitride intermetallic phases(d)

A 763, practice Y

CuSO4-50% H2SO4; 96-120 h

AISI types 446, XM27, XM33, 29-4, 29-4-2

Chromium carbide and nitride

A 763, practice Z

CuSO4-16% H2SO4; 24 h

AISI types 430, 434, 436, 439, 444

Chromium carbide and nitride

Source: Ref 30 (a) There is some effect of

(b) Detects

phase in type 321 stainless steel.

phase in AISI types 316, 316L, 317, 317L, and 321.

(c) Carbides and perhaps other phases detected, depending on the alloy system.

(d) Detects

and

phases, which do not cause intergranular attack in unstabilized iron-chromium-molybdenum alloys.

Fig. 18 Schematic showing the use of the EPR test to evaluate sensitization. The specimen is first polarized up to a passive potential at which the metal resists corrosion. Potential is then swept back through the active region, where corrosion may occur. Source: Ref 30

Stress-corrosion cracking covers all types of corrosion involving the combined action of tensile stress and corrodent. Important variables include the level of stress, the presence of oxygen, the concentration of corrodent, temperature, and the conditions of heat transfer. It is important to recognize the type of corrodent likely to produce cracking in a particular family of steel. For example, austenitic stainless steels are susceptible to chloride SCC (Table 12). Martensitic and ferritic grades are susceptible to cracking related to hydrogen embrittlement.

Table 12 Stress-corrosion cracking resistance of stainless steels Grade

Stress-corrosion cracking test(a)

Boiling 42% MgCl2

Wick test

Boiling 25% NaCl

AISI type 304

F(b)

F

F

AISI type 316

F

F

F

AISI type 317

F

[P(c) or F](d)

(P or F)

Type 317LM

F

(P or F)

(P or F)

Alloy 904L

F

(P or F)

(P or F)

AL-6XN

F

P

P

254SMO

F

P

P

20Mo-6

F

P

P

AISI type 409

P

P

P

Type 439

P

P

P

AISI type 444

P

P

P

E-Brite

P

P

P

Sea-Cure

F

P

P

MONIT

F

P

P

AL 29-4

P

P

P

AL 29-4-2

F

P

P

AL 29-4C

P

P

P

3RE60

F

NT

NT

2205

F

NT

(P or F)(e)

Ferralium

F

NT

(P or F)(e)

Source: Ref 6 (a) U-bend tests, stressed beyond yielding.

(b) Fails, cracking observed.

(c) Passes, no cracking observed.

(d) Susceptibility of grade to SCC determined by variation of composition within specified range.

(e) Susceptibility of grade to SCC determined by variation of thermal history.

It is important to realize that corrosion tests are designed to single out one particular corrosion mechanism. Therefore, determining the suitability of a stainless steel for a particular application will usually require consideration of more than one type of test. No single chemical or electrochemical test has been shown to be an all-purpose measure of corrosion resistance. More information on corrosion testing is available in the Section "Corrosion Testing and Evaluation" in this Volume.

References 1.

"Standard Practices for Detecting Susceptibility to Intergranular Corrosion Attack in Austenitic Stainless Steels," A 262, Annual Book of ASTM Standards, American Society for Testing and Materials 2. "Standard Practices for Detecting Susceptibility to Intergranular Attack in Ferritic Stainless Steels," A 763, Annual Book of ASTM Standards, American Society for Testing and Materials 3. "Standard Recommended Practice for Cleaning and Descaling Stainless Steel Parts, Equipment, and Systems," A 380, Annual Book of ASTM Standards, American Society for Testing and Materials 4. T. DeBold, Passivating Stainless Steel Parts, Mach. Tool Blue Book, Nov 1986 5. "Passivation Treatments for Corrosion-Resisting Steels," Federal Specification QQ-P-35B, United States Government Printing Office, April 1973 6. R.M. Davison et al., A Review of Worldwide Developments in Stainless Steels in Specialty Steels and Hard Materials, Pergamon Press, 1983, p 67-85 7. Corrosion Resistance of the Austenitic Chromium-Nickel Stainless Steels in Atmospheric Environments, The International Nickel Company, Inc. 1963 8. K.L. Money and W.W. Kirk, Stress Corrosion Cracking Behavior of Wrought Fe-Cr-Ni Alloys in Marine Atmosphere, Mater. Perform., Vol 17, July 1978, p 28-36 9. M. Henthorne. T.A. DeBold, and R.J. Yinger, "Custom 450--A new High Strength Stainless Steel," Paper 53, presented at Corrosion/72, National Association of Corrosion Engineers, 1972 10. The Role of Stainless Steels in Desalination, American Iron and Steel Institute, 1974 11. M.A. Streicher, Analysis of Crevice Corrosion Data From Two Sea Water Exposure Tests on Stainless Alloys, Mater. Perform., Vol 22, May 1983, p 37-50

12. A.H. Tuthill and C.M. Schillmoller, Guidelines for Selection of Marine Materials, The International Nickel Company, Inc. 1971 13. R.M. Kain, "Crevice Corrosion Resistance of Austenitic Stainless Steels in Ambient and Elevated Temperature Seawater," Paper 230, presented at Corrosion/79, National Association of Corrosion Engineers, 1979 14. F.L. LaQue and H.R. Copson, Ed., Corrosion Resistance of Metals and Alloys, Reinhold, 1963, p 375-445 15. J.E. Truman, in Corrosion: Metal/Environment Reactions, Vol 1, L.L. Shreir, Ed., Newness-Butterworths, 1976, p 352 16. M.A. Streicher, Development of Pitting Resistant Fe-Cr-Mo Alloys, Corrosion, Vol 30, 1974, p 77-91 17. H.O. Teeple, Corrosion by Some Organic Acids and Related Compounds, Corrosion, Vol 8, Jan 1952, p 14-28 18. T.A. DeBold, J.W. Martin, and J.C. Tverberg, Duplex Stainless Offers Strength and Corrosion Resistance, in Duplex Stainless Steels, R.A. Lula, Ed., American Society for Metals, 1983, p 169-189 19. L.A. Morris, in Handbook of Stainless Steels, D. Peckner and I.M. Bernstein, Ed., McGraw-Hill, 1977, p 17-1 20. "Material Requirements: Sulfide Stress Cracking Resistant Metallic Materials for Oil Field Equipment," MR-01-84, National Association of Corrosion Engineers 21. J.R. Kearns, M.J. Johnson, and J.F. Grubb, "Accelerated Corrosion in Dissimilar Metal Crevices," Paper 228, presented at Corrosion/86, National Association of Corrosion Engineers, 1986 22. L.S. Redmerski, J.J. Eckenrod, and K.E. Pinnow, "Cathodic Protection of Seawater-Cooled Power Plant Condensers Operating With High Performance Ferritic Stainless Steel Tubing," Paper 208, presented at Corrosion/85, National Association of Corrosion Engineers, 1985 23. E.C. Hoxie and G.W. Tuffnell, A Summary of INCO Corrosion Tests in Power Plant Flue Gas Scrubbing Processes, in Resolving Corrosion Problems in Air Pollution Control Equipment, National Association of Corrosion Engineers, 1976 24. Effective Use of Stainless Steel in FGD Scrubber Systems, American Iron and Steel Institute, 1978 25. G.T. Paul and R.W. Ross, Jr., "Corrosion Performance in FGD Systems at Laramie River and Dallman Stations," Paper 194, presented at Corrosion/83, National Association of Corrosion Engineers, 1983 26. A.P. Majidi and M.A. Streicher, "Four Non-Destructive Electrochemical Tests for Detecting Sensitization in Type 304 and 304L Stainless Steels," Paper 62, presented at Corrosion/85, National Association of Corrosion Engineers, 1985 27. A.H. Tuthill, Resistance, of Highly Alloyed Materials and Titanium to Localized Corrosion in Bleach Plant Environments, Mater. Perform., Vol 24, Sept 1985, p 43-49 28. "Standard Test Methods for Pitting and Crevice Corrosion Resistance of Stainless Steels and Related Alloys by the Use of Ferric Chloride Solution," G 48, Annual Book of ASTM Standards, American Society for Testing and Materials 29. "Standard Recommended Practice for Examination and Evaluation of Pitting Corrosion," G 46, Annual Book of ASTM Standards, American Society for Testing and Materials 30. T.A. DeBold, Which Corrosion Test for Stainless Steels, Mater. Eng., Vol 2 (No. 1), July 1980

Corrosion of Cast Irons Donald R. Stickle, The Duriron Company, Inc.

Introduction CAST IRON is a generic term that identifies a large family of ferrous alloys. Cast irons are primarily alloys of iron that contain more than 2% carbon and 1% or more silicon. Low raw material costs and relative ease of manufacture make cast irons the least expensive of the engineering metals. Cast irons can be cast into intricate shapes because of their excellent fluidity and relatively low melting points and can be alloyed for improvement of corrosion resistance and strength. With proper alloying, the corrosion resistance of cast irons can equal of exceed that of stainless steels and nickel-base alloys. Because of the excellent properties obtainable with these low-cost engineering materials, cast irons fluid wide application in environments that demand good corrosion resistance. Services in which cast irons are used for their excellent corrosion resistance include water, soils, acids, alkalies, saline solutions, organic compounds, sulfur compounds, and liquid metals. In some services, alloyed cast irons offer the only economical alternative for constructing equipment.

Basic Metallurgy of Cast Irons The metallurgy of cast irons is similar to that of steels except that sufficient silicon is present to necessitate use of the iron-silicon-carbon ternary phase diagram rather than the simple iron-carbon binary diagram. Figure 1 shows a section of the iron-iron carbide-silicon ternary diagram at 2% Si. The eutectic and eutectoid points in the iron-silicon-carbon diagram are both affected by the introduction of silicon into the system. In the 1 to 3% Si levels normally found in cast irons, eutectic carbon levels are related to silicon levels as follows:

%C +

(%Si) = 4.3

(Eq 1)

where %C is the eutectic carbon level, and %Si is the silicon level in the cast iron. The metallurgy of cast iron can occur in the metastable iron-iron carbide system, the stable iron-graphite system, or both. This causes structures of cast irons to be more complex than those of steel and more susceptible to processing conditions.

Fig. 1 Section of the iron-iron carbide-silicon ternary phase diagram at 2% Si

An appreciable portion of carbon in cast irons separates during solidification and appears as a separate constituent in the microstructure. The level of silicon in the cast iron has a strong effect on the manner in which carbon segregates in the microstructure. Higher silicon levels favor the formation of graphite, but lower silicon levels favor the formation of iron carbides. The form and shape in which the carbon occurs determine the type of cast iron (Table 1). Table 1 Summary of cast iron classification based on carbon form and shape Type of cast iron

Carbon form and shape

White cast iron

Iron carbide compound

Malleable cast iron

Irregularly shaped nodules of graphite

Gray cast iron

Graphite flakes

Ductile cast iron

Spherical graphite nodules

Compacted graphite cast iron

Short, fat interconnected flakes (intermediate between ductile and gray cast iron)

The structure of the metal matrix around the carbon-rich constituent establishes the class of iron within each type of iron. Four basic matrix structures occur in cast iron: ferrite, pearlite, bainite, and martensite.

Ferrite is generally a soft constituent, but it can be solid solution hardened by silicon. When silicon levels are below 3%,

the ferrite matrix is readily machined, but exhibits poor wear resistance. Above 14% Si, the ferritic matrix becomes very hard and wear resistant, but is essentially nonmachinable. The low carbon content of the ferrite phase makes hardening difficult. Ferrite can be observed in cast irons upon solidification, but is generally present as the result of special annealing heat treatments. High silicon levels promote the formation of ferritic matrices in the as-cast condition. Pearlite consists of alternate layers of ferrite and iron carbide (Fe3C, or cementite). It is very strong and tough. The hardness, strength, machinability, and wear resistance of pearlitic matrices vary with the fineness of its laminations. The carbon content of pearlite is variable and depends on the analysis of the iron and its cooling rate. Bainite is an acicular structure in cast ions that can be obtained by heat treating, alloying, or combinations of these.

Bainitic structures provide very high strength at a machinable hardness. Martensitic structures also occur in cast irons. These structures can be obtained by alloying, heat treating, or a

combination of these practices. Martensitic microstructures are the hardest, most wear-resistant structures obtainable in cast ions. Molybdenum, nickel, manganese, and chromium can be used to produce martensitic or bainitic structures. Silicon has a negative effect on martensite formation, because it promotes the formation of pearlite or ferrite.

Influence of Alloying Alloying elements can play a dominant role in the susceptibility of cast ions to corrosion attack. The alloying elements generally used to enhance the corrosion resistance of cast ions include silicon, nickel, chromium, copper, and molybdenum. Other alloying elements, such as vanadium and titanium, are sometimes used, but not to the extent of the five elements mentioned. Silicon is the most important alloying element used to improve the corrosion resistance of cast irons. Silicon is generally

not considered an alloying element in cast ions until levels exceed 3%. Silicon levels between 3 and 14% offer some increase in corrosion resistance to the alloy, but above about 14% Si, the corrosion resistance of the cast iron increases dramatically. Silicon levels up to 17% have been used to enhance the corrosion resistance of the alloy further, but silicon levels over 16% make the alloy extremely brittle and difficult to manufacture. Even at 14% Si, the strength and ductility of the material is low, and special design and manufacturing parameters are required to produce and use these alloys. Alloying with silicon promotes the formation of strongly adherent surface films in cast irons. Considerable time may be required to establish these films fully on the castings. Consequently, in some services, corrosion rates may be relatively high for the first few hours or even days of exposure, then may decline to extremely low steady-state rates for the rest of the time the parts are exposed to the corrosive environment (Fig. 2).

Fig. 2 Corrosion rates of high-silicon cast irons as a function of time and corrosive media

Nickel is used to enhance the corrosion resistance of cast irons in a number of applications. Nickel increases corrosion

resistance by the formation of protective oxide films on the surfaces of the castings. Up to 4% Ni is added in combination with chromium to improve both strength and corrosion resistance in cast iron alloys. The enhanced hardness and corrosion resistance obtained is particularly important for improving the erosion-corrosion resistance of the material. Nickel additions enhance the corrosion resistance of cast irons to reducing acids and alkalies. Nickel additions of 12% or greater are necessary to optimize the corrosion resistance of cast irons. Nickel is not as common an alloying addition as either silicon or chromium for enhancing the corrosion resistance in cast irons. It is much more important as a strengthening and hardening addition. Chromium is frequently added alone and in combination with nickel and/or silicon to increase the corrosion resistance of cast irons. As with nickel, small additions of chromium are used to refine graphite and matrix microstructures. These refinements enhance the corrosion resistance of cast irons in seawater and weak acids. Chromium additions of 15 to 30% improve the corrosion resistance of cast irons to oxidizing acids, such as nitric acid (HNO3).

Chromium increases the corrosion resistance of cast iron by the formation of protective oxides on the surfaces of castings. The oxides formed will resist oxidizing acids, but will be of little benefit under reducing conditions. High chromium additions, like higher silicon additions, reduce the ductility of cast irons. Copper is added to cast irons in special cases. Copper additions of 0.25 to 1% increase the resistance of cast iron to

dilute acetic (CH3COOH), sulfuric (H2SO4), and hydrochloric (HCl) acids as well as acid mine water. Small additions of copper are also made to cast irons to enhance atmospheric-corrosion resistance. Additions of up to 10% are made to some high nickel-chromium cast irons to increase corrosion resistance. The exact mechanism by which copper improves the corrosion resistance of cast irons is not known. Molybdenum. Although an important use of molybdenum in cast irons is to increase strength structural uniformity, it is

also used to enhance corrosion resistance, particularly in high-silicon cast irons. Molybdenum is particularly useful in HCl. As little as 1% Mo is helpful in some high-silicon irons, but for optimum resistance, 3 to 4% Mo is added. Other Alloying Additions. In general, other alloying additions to cast irons have a minimal effect on corrosion

resistance. Vanadium and titanium enhance the graphite morphology and matrix structure and impart slightly increased corrosion resistance to cast irons. Few other additions are made to cast irons that have any significant effect on corrosion resistance.

Influence of Microstructure Although the graphite shape and the amount of massive carbides present are critical to mechanical properties, these structural variables do not have a strong effect on corrosion resistance. Flake graphite structures may trap corrosion products and retard corrosion slightly in some applications. Under unusual circumstances, graphite may act cathodically with regard to the metal matrix and may accelerate attack. The structure of the matrix has a slight influence on corrosion resistance, but the effect is small compared to that of composition. In gray irons, ferrite structure are generally the least resistant, and graphite flakes exhibit the greatest corrosion resistance. Pearlite and cementite show intermediate corrosion resistance. Shrinkage or porosity can degrade the corrosion resistance of cast iron parts. The presence of porosity permits the corrosive medium to enter the body of the casting and can provide continuous leakage paths for corrosives in pressurecontaining components.

Commercially Available Cast Irons Based on corrosion resistance, cast irons can be grouped into five basic categories. Each will be discussed. Unalloyed gray, ductile, malleable, and white cast irons represent the first and largest category. All of these

materials contain carbon and silicon of 3% or less and no deliberate additions of nickel, chromium, copper, or molybdenum. As a group, these materials exhibit corrosion resistance that equals or slightly exceeds that of unalloyed steels, but they show the highest rates of attack for cast irons. These materials are available in a wide variety of configurations and alloys. Major ASTM standards that cover these materials are listed in Table 2. Table 2 ASTM standards that include unalloyed cast irons Standard

Materials/products covered

A 47

Malleable iron castings

A 48

Gray iron castings

A 74

Cast iron soil pipe and fittings

A 126

Gray iron castings for valves, flanges, and pipe fittings

A 159

Automotive gray iron castings

A 197

Cupola malleable iron

A 220

Pearlitic malleable iron castings

A 278

Gray iron castings for pressure-containing parts for temperatures up to 345 °C (650 °F)

A 319

Gray iron castings for elevated temperatures for nonpressure-containing parts

A 395

Ferritic ductile iron pressure-retaining castings for use at elevated temperatures

A 476

Ductile iron castings for paper mill dryer rolls

A 536

Ductile iron castings

A 602

Automotive melleable iron castings

A 716

Ductile iron culvert pipe

A 746

Ductile iron gravity sewer pipe

Low and moderately alloyed irons constitute the second major class. These irons contain the iron and silicon of unalloyed cast irons plus up to several percent of nickel, copper, chromium, or molybdenum. As a group, these materials exhibit two to three times the service life of unalloyed cast irons. Major ASTM standards that include these materials are listed in Table 3.

Table 3 ASTM standards that include low alloyed cast iron materials Standard

Materials/products covered

A 159

Automotive gray iron castings

A 319

Gray iron castings for elevated temperatures for nonpressure-containing parts

A 532

Abrasion-resistant cast irons

Note: Because most cast iron standards make chemical composition subordinate to mechanical properties, many of the standards listed in Table 2 may also be used to purchase low alloyed cast iron materials. High-nickel austenitic cast irons represent a third major class of cast irons for corrosion service. These materials contain large percentages of nickel and copper and are fairly resistant to such acids as concentrated H 2SO4 and phosphoric (H3PO4) acid at slightly elevated temperatures, HCl at room temperature, and such organic acids as CH3COOH, oleic, and stearic. When nickel levels exceed 18%, austenitic cast irons are nearly immune to alkali or caustics, although stress corrosion can occur. High-nickel cast irons can be nodularized to yield ductile irons. They can be purchased to the ASTM standards listed in Table 4.

Table 4 ASTM standards that include high-nickel austenitic cast iron materials Standard

Materials/products covered

A 436

Austenitic gray iron castings

A 439

Austenitic ductile iron castings

A 571

Austenitic ductile iron castings for pressure-containing parts suitable for low-temperature service

High-chromium cast irons are the fourth class of corrosion-resistant cast irons. These materials are basically white

cast irons alloyed with 12 to 30% Cr. Other alloying elements may also be added to improve resistance to specific environments. When chromium levels exceed 20%, high-chromium cast irons exhibit good resistance to oxidizing acids, particularly HNO3. High-chromium irons are not resistant to reducing acids. They are used in saline solutions, organic acids, and marine and industrial atmospheres. These materials display excellent resistance to abrasion, and with proper alloying additions, they can also resist combinations of abrasives liquids, including some dilute acid solutions. Highchromium cast irons are covered in ASTM standard A 532. In addition, some proprietary alloys not covered by national standards are produced for special applications. High-silicon cast irons represent the fifth class of corrosion-resistant cast irons. The principal alloying element is 12

to 18% Si, with more than 14.2% Si needed to develop excellent corrosion resistance. Chromium and molybdenum are also used in combination with silicon to develop corrosion resistance to specific environments. High-silicon cast irons represent the most universally corrosion-resistant alloys available at moderate cost. When silicon levels exceed 14.2% high-silicon cast irons exhibit excellent resistance to H2SO4, HNO3, HCl, CH3COOH, and most other mineral and organic acids and corrosives. These materials display good resistance in oxidizing and reducing environments and are not appreciably affected by concentration or temperature. Exceptions to universal resistance are hydrofluoric acid (HF), fluoride salts, sulfurous acid (H2SO3), sulfite compounds, strong alkalies, and alternating acid-alkali conditions. Highsilicon cast irons are defined in ASTM standards A 518 and A 861. In addition, some proprietary compositions not included in these standards, such as alloy SD77 (Fe-4Cr-3Mo-16Si-1Mn-1C), are manufactured for high-temperature HCl service.

Forms of Corrosion Cast irons exhibit the same general forms of corrosion as other metals and alloys. Examples of the forms of corrosion observed in cast irons include: • • • • • • • • • •

Uniform or general attack Galvanic or two-metal corrosion Crevice corrosion Pitting Intergranular corrosion Selective leaching Erosion-corrosion Stress corrosion Corrosion fatigue Fretting corrosion

Graphite Corrosion. A form of corrosion unique to cast irons is a selective leaching attack commonly referred to as

graphitic corrosion. Graphitic corrosion is observed in gray cast irons in relatively mild environments in which selective leaching of iron leaves a graphite network. Selective leaching of the iron takes place because the graphite is cathodic to iron and the gray iron structure establishes an excellent galvanic cell. This form of corrosion generally occurs only when corrosion rates are low. If the metal corrodes more rapidly, the entire surface, including the graphite, is removed, and more or less uniform corrosion occurs. Graphitic corrosion can cause significant problems because, although no dimensional changes occur, the cast iron loses its strength and metallic properties. Thus, without detection, potentially dangerous situations may develop in pressure-containing applications. Graphitic corrosion is observed only in gray cast irons. In both nodular and malleable iron, the lack of graphite flakes provides no network to hold the corrosion products together. Fretting corrosion is commonly observed when vibration or slight relative motion occurs between parts under load.

The relative resistance of cast iron to this form of attack is influenced by such variables as lubrication, hardness variations between materials, the presence of gaskets, and coatings. Table 5 compares the relative fretting resistance of cast iron under different combinations of these variables.

Table 5 Relative fretting resistance of cast iron Poor

Average

Good

Aluminum on cast iron Magnesium on cast iron Cast iron on chrome plate Laminated plastic on cast iron Bakelite on cast iron Cast iron on tin plate Cast iron on cast iron with coating of shellac

Cast iron on cast iron Copper on cast iron Brass on cast iron Zinc on cast iron Cast iron on silver plate Cast iron on copper plate Cast iron on amalgamated copper plate Cast iron on cast iron with rough surface

Cast iron on cast iron with phosphate coating Cast iron on cast iron with coating or rubber cement Cast iron on cast iron with coating of tungsten sulfide Cast iron on cast iron with rubber gasket Cast iron on cast iron with Molykote lubricant Cast iron on stainless with Molykote lubricant

Source: Ref 1 Pitting and Crevice Corrosion. The presence of chlorides and/or crevices or other shielded areas presents conditions that are favorable to the pitting and/or crevice corrosion of cast iron. Pitting has been reported in such environments as dilute alkylaryl sulfonates, antimony trichloride (SbCl3), and calm seawater. Alloying can influence the resistance of cast irons to pitting and crevice corrosion. For example, in calm seawater, nickel additions reduce the susceptibility of cast irons to pitting attack. High-silicon cast irons with chromium and/or molybdenum offer enhanced resistance to pitting and crevice corrosion. Although microstructural variations probably exert some influence on susceptibility to crevice corrosion and pitting, there are few reports of this relationship. Intergranular attack is relatively rare in cast irons. In stainless steels, in which this type of attack is most commonly

observed, intergranular attack is related to chromium depletion adjacent to grain boundaries. Because only the highchromium cast irons depend on chromium to form passive films for resistance to corrosion attack, few instances of intergranular attack related to chromium depletion have been reported. The only reference to intergranular attack in cast irons involves ammonium nitrate (NH4NO3), in which unalloyed cast irons are reported to be intergranularly attacked. Because this form of selective attack is relatively rare in cast irons, no significant references to the influence of either structure or chemistry on intergranular attack have been reported. Erosion-Corrosion. Fluid flow by itself or in combination with solid particles can cause erosion-corrosion attack in

cast irons. Two methods are known to enhance the erosion-corrosion resistance of cast irons. First, the hardness of the cast irons can be increased through solid-solution hardening or phase transformation induced hardness increases. For example, 14.5% Si additions to cast irons cause substantial solid-solution hardening of the ferritic matrix. In such environments as the sulfate liquors encountered in the pulp and paper industry, this hardness increase enables high-silicon iron equipment to be successfully used, while lower-hardness unalloyed cast irons fail rapidly by severe erosioncorrosion. Use of martensitic or white cast irons can also improve the erosion-corrosion resistance of cast irons as a result of hardness increases. Second, better inherent corrosion resistance can also be used to increase the erosion-corrosion resistance of cast irons. Austenitic nickel cast irons can have hardnesses similar to unalloyed cast irons, but may exhibit better erosion resistance because of the improved inherent resistance of nickel-alloyed irons compared to unalloyed irons. Microstructure can also affect erosion-corrosion resistance slightly. Gray cast irons generally show better resistance than steels under erosioncorrosion conditions. This improvement is related to the presence of the graphite network in the gray cast iron. Iron is corroded from the gray iron matrix as in steel, but the graphite network that is not corroded traps corrosion products; this layer of corrosion products and graphite offers additional protection against erosion-corrosion attack. Stress-corrosion cracking (SCC) is observed in cast irons under certain combinations of environment and stress.

Because stress is necessary to initiate SCC and because design factors often limit stresses in castings to relatively low levels, SCC is not observed as often in cast irons as in other more highly stressed components. However, under certain conditions, SCC can be a serious problem. Because unalloyed cast irons are generally similar to ordinary steels in resistance to corrosion, the same environments that cause SCC in steels will likely cause problems in cast irons. Environments that may cause SCC in unalloyed cast irons include (Ref 2):

• • • • • • • • • • • • •

Sodium hydroxide (NaOH) solutions NaOH-Na2SiO2 solutions Calcium nitrate (Ca(NO3)2) solutions NH4NO3 solutions Sodium nitrate (NaNO3) solutions Mercuric nitrate (Hg(NO3)2) solutions Mixed acids (H2SO4-HNO3) Hydrogen cyanide (HCN) solutions Seawater Acidic hydrogen sulfide (H2S) solutions Molten sodium-lead alloys Acid chloride solutions Oleum

Graphite morphology can play an important role in SCC resistance in certain environments. In oleum (fuming H2SO4), flake graphite structures present special problems. Acid tends to penetrate along graphite flakes and corrodes the iron matrix. The corrosion products formed build up pressure and eventually crack the iron. This problem is found in both gray irons and high-silicon irons, which have flake graphite morphologies, but is not seen in ductile irons that have nodular graphite shapes.

Resistance to Corrosive Environments No single grade of cast iron will resist all corrosive environments. However, a cast iron can be identified that will resist most of the corrosives commonly used in industrial environments. Cast irons suitable for the more common corrosive environments are discussed below. Sulfuric Acid. Unalloyed, low-alloyed, and high-nickel austenitic as well as high-silicon cast irons are used in H2SO4 applications. Use of unalloyed and low-alloyed cast iron is limited to low-velocity low-temperature concentrated (>70%) H2SO4 service. Unalloyed cast iron is rarely used in dilute or intermediate concentrations, because corrosion rates are substantial. In concentrated H2SO4 as well as other acids, ductile iron is generally considered superior to gray iron, and ferritic matrix irons are superior to pearlitic matrix irons. In hot, concentrated acids, graphitization of the gray iron can occur. In oleum, unalloyed gray iron will corrode at very low rates. However, acid will penetrate along the graphite flakes, and the corrosion product will form and build up sufficient pressure to split the iron. Interconnecting graphite is believed to be necessary to cause this form of cracking; therefore, ductile and malleable irons are generally acceptable for this service. Some potential galvanic corrosion between cast iron and steel has been reported in 100% H2SO4.

High-nickel austenitic cast irons exhibit acceptable corrosion resistance in room temperature and slightly elevated temperature services. As shown in Fig. 3, they are adequate over the entire range of H2SO4 concentrations, but are a second choice compared to high-silicon cast irons.

Fig. 3 Corrosion of high-nickel austenitic cast iron in H2SO4 as a function of acid concentration and temperature. Source: Ref 2

High-silicon cast irons are the best choice among the cast irons and perhaps among the commonly available engineering material for resistance to H2SO4. The material resists the entire H2SO4 concentration range at temperatures to boiling (Fig. 4). Rapid attack occurs at concentrations over 100% and in services containing free sulfur trioxide (SO3). High-silicon cast irons are relatively slow to passivate in H2SO4 services. Corrosion rates are relatively high for the first 24 to 48 h of exposure and then decrease to very low steady-state rates (Fig. 2).

Fig. 4 Corrosion of high-silicon cast iron in H2SO4 as a function of acid concentration and temperature

Nitric Acid. All types of cast iron, except high-nickel austenitic iron, find some applications in HNO3. The use of

unalloyed cast iron in HNO3 is limited to low-temperature low-velocity concentrated acid service. Even in this service, caution must be exercised to avoid dilution of acid because the unalloyed and low-alloyed cast irons both corrode very

rapidly in dilute or intermediate concentrations at any temperature. High-nickel austenitic cast irons exhibit essentially the same resistance as unalloyed cast iron to HNO3 and cannot be economically justified for this service. High-chromium cast irons with chromium contents over 20% give excellent resistance to HNO3, particularly in dilute concentrations (Fig. 5). High-temperature boiling solutions attack these grades of cast iron.

Fig. 5 Corrosion of high-chromium cast iron in HNO3 as a function of acid concentration and temperature. Source: Ref 2

High-silicon cast irons also offer excellent resistance to HNO3. Resistance is exhibited over essentially all concentration and temperature ranges with the exception of dilute, hot acids (Fig. 6). High-silicon cast iron equipment has been used for many years in the manufacture and handling of HNO3 mixed with other chemicals, such as H2SO4, sulfates, and nitrates. Contamination of the HNO3 with HF, such as might be experienced in pickling solutions, may accelerate attack of the high-silicon iron to unacceptable levels.

Fig. 6 Corrosion of high-silicon cast iron in HNO3 as a function of concentration and temperature

Hydrochloric Acid. Use of cast irons is relatively limited in HCl. Unalloyed cast iron is unsuitable for any HCl service.

Rapid corrosion occurs at a pH of 5 or lower, particularly if appreciable velocity is involved. Aeration or oxidizing conditions, such as the presence of metallic salts, result in rapid destructive attack of unalloyed cast irons even in very dilute HCl solutions. High-nickel austenitic cast irons offer some resistance to all HCl concentrations at room temperature or below. Highchromium cast irons are not suitable for HCl services. High-silicon cast irons offer the best resistance to HCl of any cast iron. When alloyed with 4 to 5% Cr, high-silicon cast iron is suitable for all concentrations of HCl to 28 °C (80 °F). When high-silicon cast iron is alloyed with chromium, molybdenum, and higher silicon levels, the temperature for use can be increased (Fig. 7). In concentrations up to 20%, ferric ions (Fe3+) or other oxidizing agents inhibit corrosion attack on high-silicon iron alloyed with chromium. At over 20% acid concentration, oxidizers accelerate attack on the alloy. As in H2SO4, corrosion rates in high-silicon cast iron are initially high in the first 24 to 48 h of exposure then decrease to very low steady-state rates (Fig. 2).

Fig. 7 Isocorrosion diagram for two high-silicon cast irons in HCl. A, Fe-14.3Si-4Cr-0.5Mo; B, Fe-16Si-4Cr-3Mo

Phosphoric Acid. All cast irons find some application in H3PO4, but the presence of contaminants must be carefully

evaluated before selecting a material for this service. Unalloyed cast iron finds little use in H3PO4, with the exception of concentrated acids. Even in concentrated acids, use may be severely limited by the presence of fluorides, chlorides, or H2SO4. High-nickel cast irons find some application in H3PO4 at and slightly above room temperature. These cast irons can be used over the entire H3PO4 concentration range. Impurities in the acid may greatly restrict the applicability of this grade of cast iron. High-chromium cast irons exhibit generally low rates of attack in H3PO4 up to 60% concentration. High-silicon cast irons show good-to-excellent resistance at all concentrations and temperatures or pure acid. The presence of fluoride ions (F-) in H3PO4 makes the high-silicon irons unacceptable for use. Organic acids and compounds are generally not as corrosive as mineral acids; consequently, cast irons find many

applications in handling these materials. Unalloyed cast iron can be used to handle concentrated CH3COOH and fatty acids, but will be attacked by more dilute solutions. Unalloyed cast irons are used to handle methyl, ethyl, butyl, and amyl alcohols. If the alcohols are contaminated with water an air, discoloration of the alcohols may occur. Unalloyed cast irons can also be used to handle glycerine, although slight discoloration of the glycerine may result.

Austenitic nickel cast irons exhibit adequate resistance to CH3COOH, oleic acid, and stearic acid. High-chromium cast irons are adequate for CH3COOH, but will be more severely corroded by formic acid (HCOOH). High-chromium cast irons are excellent for lactic and citric solutions. High-silicon cast irons show excellent resistance to most organic acids, including HCOOH and oxalic acid, in all temperature and concentration ranges. High-silicon cast irons also exhibit excellent resistance to alcohols and glycerine. Alkali solutions require material selections that are distinctly different from those of acid solutions. Alkalies include

NaOH, potassium hydroxide (KOH), sodium silicate (Na2SiO3), and similar chemicals that contain sodium, potassium, or lithium. Unalloyed cast irons exhibit generally good resistance to alkalies--approximately equivalent to that of steel. These unalloyed cast irons are not attacked by dilute alkalies at any temperature. Hot Alkalies at concentrations exceeding 30% attack unalloyed iron. Temperatures should not exceed 80 °C (175 °F) for concentrations up to 70% if corrosion rates of less than 0.25 mm/yr (10 mils/yr) are desired. Ductile and gray iron exhibit about equal resistance to alkalies; however, ductile iron is susceptible to cracking in highly alkaline solutions, but gray iron is not. Alloying with 3 to 5% Ni substantially improves the resistance of cast irons to alkalies. High-nickel austenitic cast irons offer even better resistance to alkalies than unalloyed or low-nickel cast irons. High-silicon cast irons show good resistance to relatively dilute solutions of NaOH at moderate temperatures, but should not be applied for more concentrated conditions at elevated temperatures. High-silicon cast irons are usually economical over unalloyed and nickel cast irons in alkali solutions only when other corrosives are involved for which the lesser alloys are unsuitable. High-chromium cast irons have inferior resistance to alkali solutions and are generally not recommended for alkali services. Atmospheric corrosion is basically of interest only for unalloyed and low-alloy cast irons. Atmospheric corrosion

rates are determined by the relative humidity and the presence of various gases and solid particles in the air. The high humidity, sulfur dioxide (SO2) or similar compounds found in many industrialized areas, and chlorides found in marine atmospheres increase the rate of attack on cast irons. Cast irons typically exhibit very low corrosion rates in industrial environments--generally under 0.13 mm/yr (5 mils/yr)-and the cast irons are usually found to corrode at lower rates than steel structures in the same environment. White cast irons show the lowest rate of corrosion of the unalloyed materials. Pearlitic irons are generally more resistant that ferritic irons to atmospheric corrosion. In marine atmospheres, unalloyed cast irons also exhibit relatively low rates of corrosion. Low alloy additions are sometimes made to improve corrosion resistance further. Higher alloy additions are even more beneficial, but are rarely warranted. Gray iron offers some added resistance over ductile iron in marine atmospheres. Corrosion in Soils. Cast iron use in soils, as in atmospheric corrosion, is basically limited to unalloyed and low-alloyed

cast irons. Corrosion in soils is a function of soil porosity, drainage, and dissolved constituents in the soil. Irregular soil contact can cause pitting, and poor drainage increases corrosion rates substantially above the rates in well-drained soils. Neither metal structure nor graphite morphology has an important influence on the corrosion of cast irons in soils. Some alloying additions are made to improve the resistance of cast irons to attack in soils. For example, 3% Ni additions to cast iron are made to reduce initial attack in cast irons in poorly drained soils. Alloyed cast irons would exhibit better resistance than unalloyed or low-alloyed cast irons, but are rarely needed for soil applications, because unalloyed cast irons generally have long service lives. Anodes placed in soils are frequently constructed from high-silicon cast iron. The high-silicon cast iron is not needed to resist the basic soil environment but rather to extend service life when subjected to the high electrical current discharge rates commonly used in protective anodes. Corrosion in Water. Unalloyed and low-alloyed cast irons are the primary cast irons used in water. The corrosion resistance of unalloyed cast iron in water is determined by its ability to form protective scales. In hard water, corrosion rates are generally low because of the formation of calcium carbonate (CaCO3) scales on the surface of the iron. In softened or deionized water, the protective scales cannot be fully developed, and some corrosion will occur.

In industrial waste waters, corrosion rates are primarily a function of the contaminants present. Acid pH waters increase corrosion, but alkaline pH waters lower rates. Chlorides increase the corrosion rates of unalloyed cast irons, although the influence of chlorides is small at a neutral pH. Seawater presents some special problems for cast irons. Gray iron may experience graphitic corrosion in calm seawater. It will also be galvanically active in contact with most stainless steels, copper-nickel alloys, titanium, and Hastelloy C. Because these materials are frequently used in seawater structures, this potential for galvanic corrosion must be considered. In calm seawater, the corrosion resistance of cast iron is not greatly affected by the presence of crevices. However, intermittent exposure to seawater is very corrosive to unalloyed cast irons. Use of high-alloy cast irons in water is relatively limited. High-nickel austenitic cast irons are used to increase the resistance of cast iron components to pitting in calm seawater. High-silicon cast iron is used to produce anodes for the anodic protection systems used in seawater and brackish water. Corrosion in Saline Solutions. The presence of salts in water can have dramatic effects on the selection of suitable grades of cast iron. Unalloyed cast irons exhibit very low corrosion in such salts as cyanides, silicates, carbonates, and sulfides, which hydrolyze to form alkaline solutions. However, in salts such as ferric chloride (FeCl3), cupric chloride (CuCl2), stannic salts, and mercuric salts, which hydrolyze to form acid solutions, unalloyed cast irons experience much higher rates. In salts that form dilute acid solutions, high-nickel cast irons are acceptable. More acidic and oxidizing salts, such as FeCl3, usually necessitate the use of high-silicon cast irons.

Chlorides and sulfates of alkali metals yield neutral solutions, and unalloyed cast iron experiences very low corrosion rates in these solutions. More highly alloyed cast irons also exhibit low rates, but cannot be economically justified for this application. Unalloyed cast irons are suitable for oxidizing salts, such as chromates, nitrates, nitrites, and permanganates, when the pH is neutral or alkaline. However, if the pH is less than 7, corrosion rates can increase substantially. At the lower pH with oxidizing salts, high-silicon cast iron is an excellent material selection. Ammonium salts are generally corrosive to unalloyed iron. High-nickel, high-chromium, and high-silicon cast irons provide good resistance to these salts. Other Environments. Unalloyed cast iron is used as a melting crucible for such low-melting metals as lead, zinc,

cadmium, magnesium, and aluminum. Resistance to molten metals is summarized in Table 6. Ceramic coatings and washes are sometimes used to inhibit metal attack on cast irons.

Table 6 Resistance of gray cast iron to liquid metals at 300 and 600 °C (570 and 1110 °F) Liquid metal

Liquid metal melting point, °C

Resistance of gray cast iron(a)

300 °C (570 °F)

600 °C (1110 °F)

Mercury

-38.8

Unknown

Unknown

Sodium, potassium, and mixtures

-12.3 to 97.9

Limited

Poor

Gallium

29.8

Unknown

Unknown

Bismuth-lead-tin

97

Good

Unknown

Bismuth-lead

125

Unknown

Unknown

Tin

321.9

Limited

Poor

Bismuth

271.3

Unknown

Unknown

Lead

327

Good at 327 °C (621 °F)

Unknown

Indium

156.4

Unknown

Unknown

Lithium

186

Unknown

Unknown

Thallium

303

Unknown

Unknown

Cadmium

321

Good at 321 °C (610 °F)

Good

Zinc

419.5

...

Poor

Antimony

630.5

...

Poor at 630.5 °C (1167 °F)

Magnesium

651

...

Good at 651 °C (1204 °F)

Aluminum

660

...

Poor at 660 °C (1220 °F)

Source: Ref 3 (a) Good, considered for long-time use, 10 mils/yr); Unknown, no data for these temperatures.

Cast iron can also be used in hydrogen chloride and chloride gases. In dry hydrogen chloride, unalloyed cast iron is suitable to 205 °C (400 °F), while in dry chlorine, unalloyed cast iron is suitable to 175 °C (350 °F). If moisture is present, unalloyed cast iron is unacceptable at any temperature.

Coatings Four general categories of coatings are used on cast irons to enhance corrosion resistance: metallic, organic, conversion, and enamel coatings. Coatings on cast irons are generally used to enhance the corrosion resistance of unalloyed and lowalloy cast irons. High-alloy cast irons are rarely coated. Metallic coatings are used to enhance the corrosion resistance of cast irons. These coatings may either be sacrificial

metal coatings, such as zinc, or barrier metal coatings, such as nickel-phosphorus. From a corrosion standpoint, these two classes of coatings have important differences. Sacrificial coatings are anodic when compared to iron, and the coatings corrode preferentially and protect the cast iron substrate. Small cracks and porosity in the coatings have a minimal overall effect on the performance of the coatings. Barrier coatings are cathodic compared to iron, and the coatings can protect the cast iron substrate only when porosity or cracks are not present. If there are defects in the coatings, the service environment will attack the cast iron substrate at these imperfections, and the galvanic couple set up between the relatively inert coating and the casting may accelerate attack on the cast iron. Several methods are used to apply metallic coatings to cast iron. Cast irons may be electroplated, hot dipped, flame sprayed, diffusion coated, or hard faced. Table 7 lists the metals that can be applied by these various techniques. Table 7 Summary of metallic coating techniques to enhance corrosion resistance of cast irons Coating technique

Metals/alloys applied

Electroplating

Cadmium, chromium, copper, lead, nickel, zinc, tin, tin-nickel, brass, bronze

Hot dipped

Zinc, tin, lead, lead-tin, aluminum

Hard facing

Cobalt-base alloys, nickel-base alloys, metal carbides, high-chromium ferrous alloys, high-manganese ferrous alloys, high chromium and nickel ferrous alloys

Flame spraying

Zinc, aluminum, lead, iron, bronze, copper, nickel, ceramics, cermets

Diffusion coating

Aluminum, chromium, nickel-phosphorus, zinc, nitrogen, carbon

Zinc is one of the most widely used coatings on cast irons. Although zinc is anodic to iron, its corrosion rate is very low, and it provides relatively long-term protection for the cast iron substrate. A small amount of zinc will protect a large area of cast iron. Zinc coatings provide optimum protection in rural or arid areas. Other metal coatings are also commonly used on cast irons. Cadmium provides atmospheric protection similar to that of zinc. Tin coatings are frequently used to improve the corrosion resistance of equipment intended for food handling, and aluminum coatings protect against corrosive environments containing sulfur fumes, organic acids, salts, and compounds of nitrate-phosphate chemicals. Lead and lead-tin coating are primarily applied to enhance the corrosion resistance of iron castings to H2SO3 and H2SO4. Nickel-phosphorus diffusion coatings offer corrosion resistance approaching that obtainable with stainless steel. Organic coatings can be applied to cast irons to provide short-term or long-term corrosion resistance. Short-term rust

preventatives include oil, solvent-petroleum-base inhibitors and film formers dissolved in petroleum solvents, emulsifiedpetroleum-base coatings modified to form a stable emulsion in water, and wax.

For longer-term protection and resistance to more corrosive environments, rubber-base coatings, bituminous paints, asphaltic compounds, or thermoset and thermoplastic coatings can be applied. Rubber-case coatings include chlorinated rubber neoprene, and Hypalon. These coatings are noted for their mechanical properties and corrosion resistance but not for their decorative appearance. Bituminous paints have very low water permeability and provide high resistance in cast iron castings exposed to water. Use of bituminous paints is limited to applications that require good resistance to water, weak acids, alkalies, and salts. Asphaltic compounds are used to increase the resistance to cast irons to alkalies, sewage, acids, and continued exposure to tap water. Their application range is similar to that of bituminous paints. Cast irons are also lined with thermoset and thermoplastics, such as epoxy and polyethylene, to resist attack by fluids. Fluorocarbon coatings offer superior corrosion resistance except in abrasion services. Fluorocarbon coatings applied to cast irons include such materials as polytetrafluoroethylene (PTFE), perfluoroalkoxy resins (PFA), and fluorinated ethylene polypropylene (FEP). Fluorocarbon coatings resist most common industrial services and can be used to 205 °C (400 °F). Cast iron lined with fluorocarbons is very competitive with stainless, nickel-base, and even titanium and zirconium materials in terms of range of services covered and product cost. Conversion coatings are produced when the metal on the surface of the cast iron reacts with another element or

compound to produce an iron-containing compound. Common conversion coatings include phosphate coatings, oxide coatings, and chromate coatings. Phosphate coatings enhance the resistance of cast iron to corrosion in sheltered atmospheric exposure. If the surface of the casting is oxidized and black iron oxide or magnetic is formed, the corrosion resistance of the iron can be enhanced, particularly if the oxide layer is impregnated with oil or wax. Chromate coatings are formed by immersing the iron castings in an aqueous solution of chromic acid (H2CrO4) or chromium salts. Chromate coatings are sometimes used as a supplement to cadmium plating in order to prevent the formation of powdery corrosion products. The overall benefits of conversion coatings are small with regard to atmospheric corrosion. Enamel Coatings. In the enamel coating of cast irons, glass frits are melted on the surface and form a hard, tenacious

bond to the cast iron substrate. Good resistance to all acids except HF can be obtained with the proper selection and application of the enamel coating. Alkaline-resistant coatings can also be applied, but they do not exhibit the same general resistance to alkalies as acids do. Proper design and application are essential for developing enhanced corrosion resistance on cast irons with enamel coatings. Any cracks, spalling, or other coating imperfections may permit rapid attack of the underlying cast iron.

Selection of Cast Irons Cast irons can provide excellent resistance to a wide range of corrosion environments when properly matched with the service environment for which they are intended. The basic parameters to consider before selecting cast irons for corrosion services include: • • • • • • • • • •

Concentration of solution components in weight percent Contaminants, even at parts per million levels pH of solution Temperatures and its potential range and rate of change Degree of aeration Percent and type of solids Continuous or intermittent operation Upset potential: maximum temperature and concentration Unusual conditions, such as high velocity and vacuum Materials currently used in the system and potential for galvanic corrosion

Although it is advisable to consider each of the parameters before ultimate selection of a cast iron, the information needed to assess all variables of importance properly is often lacking. In such cases, introduction of test coupons of the candidate materials into the process stream should be considered before extensive purchases of equipment are made. If neither test coupons nor complete service data are viable alternatives, consultation with a reputable manufacturer of the equipment or the cast iron with a history of applications in the area of interest should be considered.

References 1. J.R. McDowell, in Symposium on Fretting Corrosion, STP 144, American Society for Testing and Materials, 1952, p 24 2. E.C. Miller, Liquid Metals Handbook, 2nd ed., Government Printing Office, 1952, p 144 3. R.I. Higgins, Corrosion of Cast Iron, J. Res., Feb 1956, p 165-177 Selected References • • • • • •

Corrosion Data Survey, 6th ed., National Association of Corrosion Engineers, 1985 M.G. Fontana, Corrosion Engineering, 3rd ed., McGraw-Hill, 1986 "High Silicon Iron Alloys for Corrosion Services," Bulletin A12e, The Duriron Company, April 1972 Properties and Selection: Irons, Steels, and High-Performance Alloys Vol 1, ASM Handbook, formerly 10th ed., Metals Handbook, ASM International, 1990 C.F. Walton, Ed., The Gray Iron Castings Handbook, A.L. Garber, 1957 C.F. Walton, Ed., Gray and Ductile Iron Castings Handbook, R.R. Donnelley & Sons, 1971

Corrosion of Cast Steels Raymond W. Monroe, Maynard Steel Casting Company; Steven J. Pawel, University of Tennessee

Introduction STEEL CASTING COMPOSITIONS are generally divided into the categories of carbon, low-alloy, corrosion-resistant, or heat-resistant, depending on alloy content and intended service. Castings are classified as corrosion resistant if they are capable of sustained operation when exposed to attack by corrosive agents at service temperatures normally below 315 °C (600 °F). The high-alloy ferrous-base compositions are usually given the name stainless steel, although this name has been questioned. Actually, they are widely referred to as cast stainless steels. Some of the high alloys, such as 12% Cr steel, exhibit many of the familiar physical characteristics of carbon and low-alloy steels, and some of their mechanical properties, such as hardness and tensile strength, can be altered by suitable heat treatment. The alloys of higher chromium content (20 to 30%)--chromium-nickel and nickel-chromium--do not show the changes in phase observed in ordinary steel when heated or cooled in the range from room temperature to the melting point. Consequently, these materials are nonhardenable, and their mechanical properties depend on their composition rather than heat treatment. The high-alloy steels differ from carbon and low-alloy steels in other respects, such as their production and properties. Special consideration must be given to each grade with regard to casting design and foundry practice. For example, such elements as chromium, nickel, carbon, nitrogen, silicon, molybdenum, and niobium may exert a profound influence on the ultimate structure of these complex alloys; therefore, balancing of the alloy compositions is frequently required to obtain a satisfactory product. The chemical ranges used in the manufacture of wrought stainless alloys are not used to produce castings, because a different balance of alloying elements may be required to provide castability, desired mechanical properties, and optimum corrosion resistance. Corrosion resistance is a relative term that depends on the particular environment to which a specific alloy is exposed. Carbon and low-alloy steels are considered resistant only to very mild corrosives, but the various high-alloy grades are applicable for varying situations from mild to severe services, depending on the particular conditions involved. It is often misleading to list the comparative corrosion rates of different alloys exposed to the same corroding medium. In this article, no attempt will be made to recommend alloys for specific applications, and the data supplied should be used only as a general guideline. Alloy casting users will find it helpful to consult materials and corrosion specialists when selecting alloys for a particular application. The factors that must be considered in material selection include:

• • • • • •

The principal corrosive agents and their concentrations Known or suspected impurities, including abrasive materials and their concentration Average operating temperature, including variations even if encountered only for short periods Presence (or absence) of oxygen or other gases in solution Continuous or intermittent operation Fluid velocity

Each of these can have a vital effect on the service life of both cast and wrought equipment, and such detailed information usually must be provided. Many rapid failures are traceable to these details being overlooked--often when the information was available. Selection of the most economical alloy is often made by the judicious use of corrosion data. However, discretion is suggested in evaluating the relative corrosion rates of various steels because of the uncertainties of the actual test or service conditions. Corrosion rates determined in controlled laboratory tests should be applied cautiously when considering actual service. The best information is obtained from equipment used under similar operating conditions. However, exposing samples to service conditions will also provide valuable information.

Corrosion of Carbon and Low-Alloy Cast Steels Unless shielded by a protective coating, iron and steel will corrode in the presence of water and oxygen; therefore, steel will corrode when it is exposed to moist air. The rate at which corrosion proceeds in the atmosphere depends on the corroding medium, the conditions of the particular location in which the material is in use, and the steps that have been taken to prevent corrosion. The rate of corrosion also depends on the character of the steel as determined by its chemical composition and heat treatment. The probable rate of corrosion of a material in an environment can generally be estimated only from long-term tests. Cast steel and wrought steel of similar analysis and heat treatment exhibit about the same corrosion resistance in the same environments. Plain carbon steel and some of the low-alloy steels do not ordinarily resist drastic corrosive conditions, although there are some exceptions, such as strong sulfuric acid (H2SO4). To increase the corrosion resistance of steel significantly, it is necessary to resort to extensive alloying. Small amounts of copper and nickel slightly improve the resistance of steel to atmospheric attack, but appreciably larger amounts of other elements, such as chromium or nickel, improve resistance significantly. Atmospheric Corrosion. A 15-year research program compared the corrosion resistance of nine cast steels in marine and industrial atmospheres. Table 1 shows the compositions of the cast steels tested. The cast steel specimens exposed

were 13-mm ( -in.) thick, 100- × 150-mm (4- × 6-in.) panels with beveled edges. The surfaces of half the specimens were machined. Specimens of each composition and surface condition were divided into three groups. One group was exposed to an industrial atmosphere at East Chicago, IN, and the other two groups were exposed to marine atmospheres 24 and 240 m (80 and 800 ft) from the ocean at Kure Beach, NC. The weight losses of the specimens during exposure were converted to corrosion rates in terms of millimeters (mils) per year. The results of this research are shown in Fig. 1, 2, 3, and 4.

Table 1 Compositions of cast steels tested in atmospheric corrosion Cast steel

Composition, %(a)

Ni

Cu

Mn

Cr

V

C

Mo

P

S

Si

Other

Carbon, grade A

0.10

0.13

0.61

0.21

0.03

0.14

trace

0.016

0.026

0.41

...

Nickel-chromium-molybdenum

0.56

0.13

0.80

0.60

0.04

0.26

0.15

...

...

0.44

...

1Ni-1.7Mn

1.08

0.08

1.70

0.08

0.04

0.27

...

0.02

0.023

0.42

...

2% Ni

2.26

0.12

0.77

0.19

0.03

0.17

trace

0.017

0.021

0.65

...

Carbon, grade B

0.03

0.03

0.65

0.10

0.04

0.25

...

0.011

0.021

0.51

...

1% Cu

0.04

0.94

0.87

0.11

0.07

0.28

...

...

...

0.42

...

1.36Mn-0.09V

0.01

0.15

1.36

0.08

0.09

0.37

...

0.031

0.038

0.34

...

1.42% Mn

0.01

0.13

1.42

0.16

0.04

0.37

...

0.027

0.022

0.38

...

1.5Mn-0.05Ti

0.01

0.11

1.48

0.04

0.03

0.33

...

0.016

0.025

0.40

0.05 Ti

Source: Ref 1 (a) All compositions contain balance of iron.

Fig. 1 Corrosion rates of various cast steels in a marine atmosphere. Nonmachined specimens were exposed 24 m (80 ft) from the ocean at Kure Beach, NC. Source: Ref 1

Fig. 2 Corrosion rates of various cast steels exposed at the 240-m (800-ft) site at Kure Beach, NC. Specimens were not machined. Source: Ref 1

Fig. 3 Corrosion rates for cast steels in an industrial atmosphere. Nonmachined specimens were exposed at East Chicago, IN. Source: Ref 1

Fig. 4 Corrosion rates of machined and non-machined specimens of cast steels after 7 years in three environments. The effect of surface finish on corrosion rates is negligible. Source: Ref 1

Figure 5 shows the results of another portion of this project. Corrosion rates for a 3-year exposure of various cast steels, wrought steels, and malleable iron in both atmospheres are compared. The following conclusions can be drawn from these tests:

•

•

•

•

• •

The condition of the specimen surface has no significant effect on the corrosion resistance of cast steels. Unmachined surfaces with the casting skin intact have corrosion rates similar to those of machined surfaces regardless of the atmospheric environment The highest corrosion rate occurs in the marine atmosphere 24 m (80 ft) from the ocean, with lower but similar corrosion rates occurring in the industrial atmosphere and the marine atmosphere 240 m (800 ft) from the ocean The corrosion rate of cast steel decreases as a function of time, because corrosion products (scale and rust coating) build up and act as a protective coating on the cast steel surface. However, the corrosion rate of the most resistant cast steel (2% Ni) is always less than that of lesser corrosion-resistant cast steels Cast steels containing small percentages of nickel, copper, or chromium as alloying elements have corrosion resistance superior to that of cast carbon steels and those containing manganese when exposed to atmospheric environments Increasing the nickel and the chromium contents of cast steel increases the corrosion resistance in all three of the atmospheric environments All cast steels have greater corrosion resistance than malleable iron in industrial atmospheres and are superior or equivalent to the wrought steels in this environment. The corrosion rate in the marine atmosphere depends primarily on the alloy content. The cast carbon steel is much superior to the AISI 1020 wrought steel, but is slightly inferior to malleable iron (Ref 1)

Fig. 5 Comparison of corrosion rates of cast steels, malleable cast iron, and wrought steel after 3 years of exposure in two atmospheres. Source: Ref 1

Other Environments. Several low- and high-alloy cast steels have been studied regarding their corrosion resistance to

high-temperature steam. Test specimens 150 mm (6 in.) in length and 13 mm ( -in.) in diameter were machined from

test coupons and then exposed to steam at 650 °C (1200 °F) for 570 h. The steel compositions and test results are given in Table 2. Table 3 shows the resistance of cast steels to petroleum corrosion, and Tables 4 and 5 supply similar data relating to water and acid attack. These data show the value of higher chromium content for improved corrosion resistance. Table 2 Corrosion of cast carbon and alloy steels in steam at 650 °C (1200 °F) for 570 h Type of steel

Composition, %

Average penetration rate

C

Cr

Ni

Mo

mm/yr

mils/yr

Carbon

0.24

...

...

...

0.3

12

Carbon

0.25

...

...

...

0.28

11

Carbon-molybdenum

0.21

...

...

0.49

0.3

12

Carbon-molybdenum

0.20

...

...

0.49

0.25

10

Nickel-chromium-molybdenum

0.35

0.64

2.13

0.26

0.25

10

Nickel-chromium-molybdenum

0.28

0.73

2.25

0.26

0.25

10

5Cr-molybdenum

0.22

5.07

...

0.47

0.1

4

5Cr-molybdenum

0.27

5.49

...

0.43

0.1

4

7Cr-molybdenum(a)

0.11

7.33

...

0.59

0.05

2

9Cr-1.5Mo

0.23

9.09

...

1.56

0.025

1

Source: Ref 1 (a) Not a cast steel.

Table 3 Petroleum corrosion resistance of cast steels 1000-h test in petroleum vapor under 780 N (175 lb) of pressure at 345 °C (650 °F) Type of material

Cast carbon steel

Weight loss

mg/cm2

mg/in.2

3040

196

Cast steel, 2Ni-0.75Cr

2370

153

Seamless tubing, 5% Cr

1540

99.2

Cast steel, 5Cr-1W

950

61.5

Cast steel, 5Cr-0.5Mo

730

47

Cast steel, 12% Cr

6.4

100

Stainless steel, 18Cr-8Ni

2.1

30

Source: Ref 1

Table 4 Corrosion of cast steels in waters Corrosive medium

Exposure time, months

Corrosion factor(a)

Fe-0.29C-0.69Mn-0.44Si

Fe-0.32C-0.66Mn-1.12Cr

Fe-0.11C-0.41Mn-3.58Cr

2

100

85

58

6

100

73

61

2

100

60

26

6

100

80

40

2

100

93

30

6

100

109

25

Hot water

1

100

100

64

0.05% H2SO4

2

100

71

68

6

100

89

102

2

100

223

61

Tap water

Seawater

Alternate immersion and drying

0.50% H2SO4

Source: Ref 1

(a) Corrosion factor is the ratio of average penetration rate of the alloy in question to Fe-0.29C-0.69Mn-0.44Si steel.

Table 5 Corrosion of cast chromium and carbon steels in mineral acids Steel

Weight loss in 5 h

5% H2SO4

5% HCl

5% HNO3

mg/cm2

mg/in.2

mg/cm2

mg/in.2

mg/cm2

mg/in.2

Carbon steel, 0.31% C

2.7

17.42

2.1

13.55

80.79

521.1

Chromium steel, 0.30C-2.42Cr

4.9

31.6

5.41

34.9

47.36

305.5

Corrosion of Cast Stainless Steels Cast stainless steels are usually specified on the basis of composition by using the alloy designation system established by the Alloy Casting Institute (ACI). The ACI designations, such as CF-8M, have been adopted by the American Society for Testing and Materials (ASTM) and are preferred for cast alloys over the designations used by the American Iron and Steel Institute (AISI) for similar wrought steels. The first letter of the ACI designation indicates whether the alloy is intended primarily for liquid corrosion service (C) or heat-resistant service (H). The second letter denotes the nominal chromium-nickel type, as shown in Fig. 6. As the nickel content increases, the second letter in the ACI designation increases from A to Z. The numerals following the two letters refer to the maximum carbon content (percent × 100) of the alloy. If additional alloying elements are included, they can be denoted by the addition of one or more letters after the maximum carbon content. Thus, the designation CF-8M refers to an alloy for corrosion-resistant service (C) of the 19Cr-9Ni (F) type, with a maximum carbon content of 0.08% and containing molybdenum (M).

Fig. 6 Chromium and nickel contents in ACl standard grades of heat- and corrosion-resistant castings. See text for details. Source: Ref 2

Corrosion-resistant cast steels are also often classified on the basis of microstructure. The classifications are not completely independent, and a classification by composition often involves microstructural distinctions. Cast corrosionresistant alloy compositions are listed in Table 6. Table 6 Compositions of ACI heat- and corrosion-resistant casting alloys ACI designation

Wrought alloy type(a)

Composition, % (balance iron)(b)

C

Mn

Si

P

S

Cr

Ni

Other elements

CA-15

410

0.15

1.00

1.50

0.04

0.04

11.5-14

1

0.5Mo(c)

CA-15M

...

0.15

1.00

0.65

0.04

0.04

11.5014.0

1.00

0.15-1.00Mo

CA-40

420

0.200.40

1.00

1.50

0.04

0.04

11.5-14

1

0.5Mo(c)

CA-6NM

...

0.06

1.00

1.00

0.04

0.03

11.5-14.0

3.5-4.5

0.4-1.0Mo

CA-6N

...

0.06

0.50

1.00

0.02

0.02

10.5-12.0

6.0-8.0

CB-30

431

0.30

1.00

1.50

0.04

0.04

18-21

2

...

CB-7Cu-1

...

0.07

0.70

1.00

0.035

0.03

14.0-15.5

4.5-5.5

0.15-0.35Nb, 0.05N, 2.5-3.2Cu

CB-7Cu-2

...

0.07

0.70

1.00

0.035

0.03

14.0-15.5

4.5-5.5

0.15-0.35Nb, 0.05N, 2.5-3.2Cu

CC-50

446

0.50

1.00

1.50

0.04

0.04

26-30

4

CD-4MCu

...

0.04

1.00

1.00

0.04

0.04

24.5-26.5

4.756.00

1.75-2.25Mo, 2.75-3.25Cu

CE-30

...

0.30

1.50

2.00

0.04

0.04

26-30

8-11

...

CF-3

304L

0.03

1.50

2.00

0.04

0.04

17-21

8-21

...

CF-8

304

0.08

1.50

2.00

0.04

0.04

18-21

8-11

...

CF-20

302

0.20

1.50

2.00

0.04

0.04

18-21

8-11

...

CF-3M

316L

0.03

1.50

1.50

0.04

0.04

17-21

9-13

2.0-3.0Mo

CF-8M

316

0.08

1.50

2.00

0.04

0.04

18-21

9-12

2.0-3.0Mo

CF-8C

347

0.08

1.50

2.00

0.04

0.04

18-21

9-12

3 × C min, 1.0 max Nb

CF-16F

303

0.16

1.50

2.00

0.17

0.04

18-21

9-12

1.5Mo, 0.2-0.35Se

CG-12

...

0.12

1.50

2.00

0.04

0.04

20-23

10-13

CG-8M

317

0.08

1.50

1.50

0.04

0.04

18-21

9-13

3.0-4.0Mo

CH-20

309

0.20

1.50

2.00

0.04

0.04

22-26

12-15

...

CK-20

310

0.20

2.00

2.00

0.04

0.04

23-27

19-22

...

CN-7M

...

0.07

1.50

1.50

0.04

0.04

19-22

27.530.5

2.0-3.0Mo, 3.0-4.0Cu

CN-7MS

...

0.07

1.00

2.503.50

0.04

0.03

18-20

22-25

2.0-3.0Mo, 1.5-2.0Cu

CW-12M

...

0.12

1.00

1.50

0.04

0.03

15.5-20

bal

7.5Fe

CY-40

...

0.40

1.50

3.00

0.03

0.03

14-17

bal

11.0Fe

CZ-100

...

1.00

1.50

2.00

0.03

0.03

...

bal

3.0Fe, 1.25Cu

N-12M

...

0.12

1.00

1.00

0.04

0.03

1.0

bal

0.26-0.33Mo, 6.0Fe

M-35

...

0.35

1.50

2.00

0.03

0.03

...

bal

28-33Cu, 3.5Fe

HA

...

0.20

0.350.65

1.00

0.04

0.04

8-10

...

0.90-1.20Mo

HC

446

0.50

1.00

2.00

0.04

0.04

26-30

4

0.5Mo(c)

HD

327

0.50

1.50

2.00

0.04

0.04

26-30

4-7

0.5Mo(c)

HE

...

0.200.50

2.00

2.00

0.04

0.04

26-30

8-11

0.5Mo(c)

HF

302B

0.200.40

2.00

2.00

0.04

0.04

18-23

8-12

0.5Mo(c)

HH

309

0.20-

2.00

2.00

0.04

0.04

24-28

11-14

0.5Mo(c), 0.2N

0.60V,

2.50Co,

0.50

HI

...

0.200.50

2.00

2.00

0.04

0.04

26-30

14-18

0.5Mo(c)

HK

310

0.200.60

2.00

2.00

0.04

0.04

24-28

18-22

0.5Mo(c)

HL

...

0.200.60

2.00

2.00

0.04

0.04

28-32

18-22

0.5Mo(c)

HN

...

0.200.50

2.00

2.00

0.04

0.04

19-23

23-27

0.5Mo(c)

HP

...

0.350.75

2.00

2.50

0.04

0.04

24-28

33-37

0.5Mo(c)

HP-50WZ

...

0.450.55

2.00

2.00

0.04

0.04

24-28

33-37

4.0-6.0W, 0.2-1.0Zr

HT

330

0.350.75

2.00

2.50

0.04

0.04

15-19

33-37

0.5Mo(c)

HU

...

0.350.75

2.00

2.50

0.04

0.04

17-21

37-41

0.5Mo(c)

HW

...

0.350.75

2.00

2.50

0.04

0.04

10-14

58-62

0.5Mo(c)

HX

...

0.350.75

2.00

2.50

0.04

0.04

15-19

64-68

0.5Mo(c)

Source: Ref 3 (a) Cast alloy chemical composition ranges are not the same as the wrought composition ranges; buyers should use cast alloy designations for proper identification of castings.

(b) Maximum, unless range is given.

(c) Molybdenum not intentionally added.

Composition and Microstructure The principal alloying element in the high-alloy family is usually chromium, which, through the formation of protective oxide films, is the first step for these alloys in achieving stainless quality. For all practical purposes, stainless behavior requires at least 12% Cr. As will be discussed later, corrosion resistance further improves with additions of chromium to at least the 30% level. As shown in Table 5, nickel and lesser amounts of molybdenum and other elements are often added to the iron-chromium matrix.

Although chromium is the ferrite and martensite promoter, nickel is an austenite promoter. By varying the amounts and ratios of these two elements (or their equivalents), almost any desired combination of microstructure, strength, or other property can be achieved. Equally important is heat treatment. Temperature, time at temperature, and cooling rate must be controlled to obtain the desired results. It is useful to think of the compositions of high-alloy steels in terms of the balance between austenite promoters and ferrite promoters. This is done on the widely used Schaeffler-type diagrams (Fig. 7). The phases shown are those that persist after cooling to room temperature at rates normally used in fabrication (Ref 2, 3).

Fig. 7 Schaeffler diagram showing the amount of ferrite and austenite present in weldments as a function of chromium and nickel equivalents. Source: Ref 2

The empirical correlations shown in Fig. 7 can be understood from the following. The field designated as martensite encompasses such alloys as CA-15, CA-6NM, and even CB-7Cu. These alloys contain 12 to 17% Cr, with adequate nickel, molybdenum, and carbon to promote high hardenability, that is, the ability to transform completely to martensite when cooled at even the moderate rates associated with the air cooling of heavy sections. High alloys have low thermal conductivities and cool slowly. To obtain the desired properties, a full heat treatment is required after casting; that is, the casting is austenitized by heating to 870 to 980 °C (1600 to 1800 °F), cooled to room temperature to produce the hard martensite, and then tempered at 595 to 760 °C (1100 to 1400 °F) until the desired combination of strength, toughness, ductility, and resistance to corrosion or stress corrosion is obtained (Ref 2, 3). Increasing the nickel equivalent (moving vertically in Fig. 7) eventually results in an alloy that is fully austenitic, such a CC-20, CH-20, CK-20, or CN-7M. These alloys are extremely ductile, tough, and corrosion resistant. On the other hand, the yield and tensile strength may be relatively low for the fully austenitic alloys. Because these high-nickel alloys are fully austenitic, they are nonmagnetic. Heat treatment consists of a single step: water quenching from a relatively high temperature at which carbides have been taken into solution. Solution treatment may also homogenize the structure, but because no transformation occurs, there can be no grain refinement. The solutionizing step and rapid cooling ensure maximum resistance to corrosion. Temperatures between 1040 and 1205 °C (1900 and 2200 °F) are usually required (Ref 2, 3). Adding chromium to the lean alloys (proceeding horizontally in Fig. 7) stabilizes the -ferrite that forms when the casting solidifies. Examples are CB-30 and CC-50. With high chromium content, these alloys have relatively good resistance to corrosion, particularly in sulfur-bearing atmospheres. However, being single-phase, they are nonhardenable, have

moderate-to-low strength, and are often used as-cast or after only a simple solutioning treatment. Ferritic alloys also have poor toughness (Ref 2, 3). Between the fields designated M, A, and F in Fig. 7 are regions indicating the possibility of two or more phases in the alloys. Commercially, the most important of these alloys are the ones in which austenite and ferrite coexist, such as CF-3, CF-8, CF-3M, CF-8M, CG-8M, and CE-30. These alloys usually contain 3 to 30% ferrite in a matrix of austenite. Predicting and controlling ferrite content is vital to the successful application of these materials. Duplex alloys offer superior strength, weldability, and corrosion resistance. Strength, for example, increases directly with ferrite content. Achieving specified minimums may necessitate controlling the ferrite within narrow bands. Figure 8 and Schoefer's equations are used for this purpose. These duplex alloys should be solution treated and rapidly cooled before use to ensure maximum resistance to corrosion (Ref 2, 3).

Fig. 8 Schoefer diagram for estimating the average ferrite content in austenitic iron-chromium-nickel alloy castings. Source: Ref 2

The presence of ferrite is not entirely beneficial. Ferrite tends to reduce toughness, although this is not of great concern given the extremely high toughness of the austenite matrix. However, in applications that require exposure to elevated temperatures, usually 315 °C (600 °F) and higher, the metallurgical changes associated with the ferrite can be severe and detrimental. In the low end of this temperature range, the reductions in toughness observed have been attributed to carbide precipitation or reactions associated with 475-°C embrittlement. The 475-°C embrittlement is caused by precipitation of an intermetallic phase with a composition of approximately 80Cr-20Fe. The name derives from the fact that this embrittlement is most severe and rapid when it occurs at approximately 475 °C (885 °F). At 540 °C (1000 °F) and above, the ferrite phase may transform to a complex iron-chromium-nickel-molybdenum intermetallic compound known as phase, which reduces toughness, corrosion resistance, and creep ductility. The extent of the reduction increases with time and temperature to about 815 °C (1500 °F) and may persist to 925 °C (1700 °F). In extreme cases, Charpy V-notch energy at room temperature may be reduced 95% from its initial value (Ref 2, 3). More information on the metallography and microstructures of these alloys is available in the article "Stainless Steel Casting Alloys" in Metallography and Microstructures, Volume 9 of ASM Handbook, formerly 9th Edition Metals Handbook.

Corrosion Behavior of H-Type Alloys The ACI heat-resistant (H-type) alloys must be able to withstand temperatures exceeding 1095 °C (2000 °F) in the most severe high-temperature service. An important factor pertaining to the corrosion behavior of these alloys is chromium content. Chromium imparts resistance to oxidation and sulfidation at high temperatures by forming a passive oxide film. Heat-resistant casting alloys must also have good resistance to carburization. More information on the corrosion of metals and alloys in high-temperature gases is available in the article "Fundamentals of Corrosion in Gases" in this Volume. Oxidation. Resistance to oxidation increases directly with chromium content (Fig. 9). For the most severe service at

temperatures above 1095 °C (2000 °F), 25% or more chromium is required. Additions of nickel, silicon, manganese, and aluminum promote the formation of relatively impermeable oxide films that retard further scaling. Thermal cycling is extremely damaging to oxidation resistance because it leads to breaking, cracking, or spalling of the protective oxide film. The best performance is obtained with austenitic alloys containing 40 to 50% combined nickel and chromium. Figure 10 shows the behavior of the H-type grades.

Fig. 9 Effect of chromium on oxidation resistance of cast steels. Specimens (13-mm, or 0.5-in., cubes) were exposed for 48 h at 1000 °C (1830 °F). Source: Ref 3

Fig. 10 Corrosion behavior of ACI H-type (heat-resistant) alloy castings in air (a) and in oxidizing flue gases containing 5 grains of sulfur per 2.8 m3 (100 ft3) of gas (b). Source: Ref 3

Sulfidation environments are becoming increasingly important. Petroleum processing, coal conversion, utility and

chemical applications, and waste incineration have heightened the need for alloys resistant to sulfidation attack in relatively weak oxidizing or reducing environments. Fortunately, high chromium and silicon contents increase resistance to sulfur-bearing environments. On the other hand, nickel has been found to be detrimental to the most aggressive gases. The problem is attributable to the formation of low-melting nickel-sulfur eutectics. These produce highly destructive liquid phases at temperatures even below 815 °C (1500 °F). Once formed, the liquid may run onto adjacent surfaces and rapidly corrode other metals. The behavior of H-type grades in sulfidizing environments is represented in Fig. 11.

Fig. 11 Corrosion behavior of ACI H-type alloys in 100-h tests at 980 °C (1800 °F) in reducing sulfur-bearing gases. (a) Gas contained 5 grains of sulfur per 2.8 m3 (100 ft3) of gas. (b) Gas contained 300 grains of sulfur per 2.8 m3 (100 ft3) of gas. (c) Gas contained 100 grains of sulfur per 2.8 m3 (100 ft3) of gas; test at constant temperature. (d) Some sulfur content as gas in (c), but cooled to 150 °C (300 °F) each 12 h

Carburization. High alloys are often used in nonoxidizing atmospheres in which carbon diffusion into metal surfaces is

possible. Depending on chromium content, temperature, and carburizing potential, the surface may become extremely rich in chromium carbides, rendering it hard and possibly susceptible to cracking. Silicon and nickel are thought to be beneficial and enhance resistance to carburization. Corrosion Behavior of C-Type Alloys The ACI C-type (for liquid corrosion service) stainless steels must resist corrosion in the various environments in which they regularly serve. In this section, the general principles and important highlights of corrosion behavior will be discussed as influenced by the metallurgy of these materials. Topics include general corrosion, intergranular corrosion, localized corrosion, corrosion fatigue, and stress corrosion. General Corrosion of Martensitic Alloys. The martensitic grades include CA-15, CA-15M, CA-6NM, CA-6NM-B,

CA-40, CB-7Cu-1 and CB-7Cu-2. These alloys are generally used in applications requiring high strength and some corrosion resistance. Alloy CA-15 typically exhibits a microstructure of martensite and ferrite. This alloy contains the minimum amount of chromium to be considered a stainless steel (11 to 14% Cr) and as such may not be used in aggressive environments. It

does, however, exhibit good atmospheric-corrosion resistance, and it resists staining by many organic environments. Alloy CA-15M may contain slightly more molybdenum than CA-15 (up to 1% Mo) and therefore may have improved general corrosion resistance in relatively mild environments. Alloy CA-6NM is similar to CA-15M except that it contains more nickel and molybdenum, which improves its general corrosion resistance. Alloy CA-6NM-B is a lower-carbon version of this alloy. The lower strength level promotes resistance to sulfide stress cracking. Alloy CA-40 is a higherstrength version of CA-15, and it also exhibits excellent atmospheric-corrosion resistance after a normalize and temper heat treatment. Microstructurally, the CB-7Cu alloys usually consist of mixed martensite and ferrite, and because of the increased chromium and nickel levels compared to the other martensitic alloys, they offer improved corrosion resistance to seawater and some mild acids. These alloys also have good atmospheric-corrosion resistance. The CB-7Cu alloys are hardenable and offer the possibility of increased strength and improved corrosion resistance among the martensitic alloys. General Corrosion of Ferritic Alloys. Alloys CB-30 and CC-50 are higher-carbon and higher-chromium alloys than the CA alloys previously mentioned. Each alloy is predominantly ferritic, although a small amount of martensite may be found in CB-30. Alloy CB-30 contains 18 to 21% Cr and is used in chemical-processing and oil-refining applications. The chromium content is sufficient to have good corrosion resistance to many acids, including nitric acid (HNO3). Figure 12 shows an isocorrosion diagram for CB-30 in HNO3. Alloy CC-50 contains substantially more chromium (26 to 30%) and offers relatively high resistance to localized corrosion and high resistance to many acids, including dilute H2SO4 and such oxidizing acids as HNO3.

Fig. 12 Isocorrosion diagram for ACI CB-30 in HNO3. Castings were annealed at 790 °C (1450 °F), furnace cooled to 540 °C (1000 °F), and then air cooled to room temperature.

General Corrosion of Austenitic and Duplex Alloys. Alloy CF-8 typically contains approximately 19% Cr and

9% Ni and is essentially the cast equivalent of AISI 304-type wrought alloys. Alloy CF-8 may be fully austenitic, but it more commonly contains some residual ferrite (3 to 30%) in an austenite matrix. In the solution-treated condition, this alloy has excellent resistance to a wide variety of acids. It is particularly resistant to highly oxidizing acids, such as boiling HNO3. Figure 13 shows isocorrosion diagrams for CF-8 in HNO3, phosphoric acid (H3PO4), and sodium hydroxide (NaOH). The duplex nature of the microstructure of this alloy imparts additional resistance to stress-corrosion cracking (SCC) compared to its wholly austenitic counterparts. Alloy CF-3 is a reduced-carbon version of CF-8 with essentially identical corrosion resistance except that CF-3 is much less susceptible to sensitization (Fig. 14). For applications in which the corrosion resistance of the weld heat-affected zone (HAZ) may be critical, CF-3 is a common material selection.

Fig. 13 Isocorrosion diagrams for ACI CF-8 in HNO3 (a), H3PO4 (b and c), and NaOH solutions (d and e). (b) and (d) Tests performed in a closed container at equilibrium pressure. (c) and (e) Tested at atmospheric pressure

Fig. 14 Isocorrosion diagram for solution-treated quenched and sensitized ACI CF-3 in HNO3

Alloys CF-8A and CF-3A contain more ferrite than their CF-8 and CF-3 counterparts. Because the higher ferrite content is achieved by increasing the chromium/nickel equivalent ratio, the CF-8A and CF-3A alloys may have slightly higher chromium or slightly lower nickel contents than the low-ferrite equivalents. In general, the corrosion resistance is very similar, but the strength increases with ferrite content. Because of the high ferrite content, service should be restricted to temperature below 400 °C (750 °F) due to the possibility of severe embrittlement. Alloy CF-8C is the niobium-stabilized grade of the CF-8 alloy class. This alloy contains small amounts of niobium, which tend to form carbides preferentially over chromium carbides and improve intergranular corrosion resistance in applications involving relatively high service temperatures. Alloys CF-8M, CF-3M, CF-8MA, and CF-3MA are molybdenum-bearing (2 to 3%) versions of the CF-8 and CF-3 alloys. The addition of 2 to 3% Mo increases resistance to corrosion by seawater and improves resistance to many chloride-bearing environments. The presence of 2 to 3% Mo also improves crevice corrosion and pitting resistance compared to the CF-8 and CF-3 alloys. Molybdenum-bearing alloys are generally not as resistant to highly oxidizing environments (this is particularly true for boiling HNO3), but for weakly oxidizing environments and reducing environments, Mo-bearing alloys are generally superior. Alloy CF-16F is a selenium-bearing free-machining grade of cast stainless steel. Because CF-16F nominally contains 19% Cr and 10% Ni, it has adequate corrosion resistance to a wide range of corrodents, but the large number of selenide inclusions makes surface deterioration and pitting definite possibilities. Alloy CF-20 is a fully austenitic, relatively high-strength corrosion-resistant alloy. The 19% Cr content provides resistance to many types of oxidizing acids, but the high carbon content makes it imperative that this alloy be utilized in the solution-treated condition for environments known to cause intergranular corrosion. Alloy CE-30 is a nominally 27Cr-9Ni alloy that normally contains 10 to 20% ferrite in an austenite matrix. The high carbon and ferrite contents provide relatively high strength. The high chromium content and duplex structure act to minimize corrosion because of the formation of chromium carbides in the microstructure. This particular alloy is known for good resistance to sulfurous acid and sulfuric acid, and it is extensively used in the pulp and paper industry (see the article "Corrosion in the Pulp and Paper Industry" in this Volume). Alloy CG-8M is slightly more highly alloyed than the CF-8M alloys, with the primary addition being increased molybdenum (3 to 4%). The increased amount of molybdenum provides superior corrosion resistance to halide-bearing media and reducing acids, particularly H2SO3 and H2SO4 solutions. The high molybdenum content, however, renders CG8M generally unsuitable in highly oxidizing environments. Alloy CD-4MCu is the most highly alloyed material in this group of alloys; it has a nominal composition of Fe-26Cr-5Ni2Mo-3Cu. The chromium/nickel equivalent ratio for this alloy is quite high, and a microstructure containing

approximately equal amounts of ferrite and austenite is common. The low carbon content and high chromium content render the alloy relatively immune to intergranular corrosion. High chromium and molybdenum provide a high degree of localized corrosion resistance (crevices and pitting), and the duplex microstructure provides SCC resistance in many environments. This alloy can be precipitation hardened to provide strength and is also relatively resistant to abrasion and erosion-corrosion. Figures 15 and 16 show isocorrosion diagrams for CD-4MCu in HNO3 and H2SO4, respectively.

Fig. 15 Isocorrosion diagram for ACI CD-4MCu in HNO3. The material was solution treated at 1120 °C (2050 °F) and water quenched.

Fig. 16 Isocorrosion diagram for ACI CD-4MCu in H2SO4. The material was solution annealed at 1120 °C (2050 °F) and water quenched.

Fully Austenitic Alloys. Alloys CH-10 and CH-20 are fully austenitic and contain 22 to 26% Cr and 12 to 15% Ni.

The high chromium content minimizes the tendency toward the formation of chromium-depleted zones due to sensitization. These alloys are used for handling paper pulp solutions and are known for good resistance to dilute H2SO4 and HNO3. Alloy CK-20 contains 23 to 27% Cr and 19 to 22% Ni and is less susceptible than CH-20 to intergranular corrosion attack in many acids after brief exposures to the chromium carbide formation temperature range. Maximum corrosion resistance is achieved by solution treatment. Alloy CK-20 possesses good corrosion resistance to many acids and, because of its fully austenitic structure, can be used at relatively high temperature.

Alloy CN-7M, with a nominal composition of Fe-29Ni-20Cr-2.5Mo-3.5Cu, exhibits excellent corrosion resistance in a wide variety of environments and is often used for H2SO4 service. Figure 17 shows isocorrosion diagrams for CN-7M in H2SO4, HNO3, H3PO4, and NaOH. Relatively high resistance to intergranular corrosion and SCC make this alloy attractive for very many applications. Although relatively highly alloyed, the fully austenitic structure of CN-7M may lead to SCC susceptibility for some environments and stress states.

Fig. 17 Isocorrosion diagrams for solution-annealed and quenched ACl CN-7M in H2SO4, HNO3, NaOH, and H3PO4. (a), (b), (d), and (f) Tested at atmospheric pressure. (c) and (e) Tested at equilibrium pressure in a closed container. See Fig. 13 for legend.

Intergranular Corrosion of Austenitic and Duplex Alloys. The optimum corrosion resistance for these alloys is

developed by solution treatment. Depending on the specific alloy in question, temperatures between 1040 and 1205 °C (1900 and 2200 °F) are required to ensure complete solution of all carbides and phases, such as and , the sometimes form in highly alloyed stainless steels. Alloys containing relatively high total alloy content, particularly high molybdenum content, often require the higher solution treatment temperature. Water quenching from the temperature range of 1040 to 1205 °C (1900 to 2200 °F) normally completes the solution treatment. Failure to solution treat a particular alloy or an improper solution treatment may seriously compromise the observed corrosion resistance in service. Inadvertent or unavoidable heat treatment in the temperature range of 480 to 820 °C (900 to 1500 °F)--for example, welding--may destroy the intergranular corrosion resistance of the alloy. When austenitic or duplex (ferrite in austenite matrix) stainless steels are heated in or cooled slowly through this temperature range, chromium-rich carbides form at grain boundaries is austenitic alloys and at ferrite/austenite interfaces in duplex alloys. These carbides deplete the surrounding matrix of chromium, thus diminishing the corrosion resistance of the alloy. An alloy in this condition of reduced corrosion resistance due to the formation of chromium carbides is said to be sensitized. In small amounts, these carbides may lead to localized pitting in the alloy, but if the chromium-depleted zones are extensive throughout the alloy or HAZ of a weld, the alloy may disintegrate intergranularly in some environments. If solution treatment of the alloy after casting and/or welding is impractical or impossible, the metallurgist has several tools from which to choose to minimize potential intergranular corrosion problems. The low carbon grades CF-3 and CF3M are commonly used as a solution to the sensitization incurred during welding. The low carbon content (0.03% C maximum) of these alloys precludes the formation of an extensive number of chromium carbides. In addition, these alloys

normally contain 3 to 30% ferrite in an austenitic matrix. By virtue of rapid carbide precipitation kinetics at ferrite/austenite interfaces compared to austenite/austenite interfaces, carbide precipitation is confined to ferrite/austenite boundaries in alloys containing a minimum of about 3 to 5% ferrite (Ref 4, 5). If the ferrite network is discontinuous in the austenite matrix (depending on the amount, size, and distribution of ferrite pools), then extensive intergranular corrosion will not be a problem in most of the environments to which these alloys would be subjected. An example of attack at the ferrite/austenite boundaries is shown in Fig. 18. These low-carbon alloys need not sacrifice significant strength compared to their high-carbon counterparts, because nitrogen may be added to increase strength. However, a large amount of nitrogen will begin to reduce the ferrite content, which will cancel some of the strength gained by interstitial hardening. Appropriate adjustment of the chromium/nickel equivalent ratio is beneficial in such cases. Fortunately, nitrogen is also beneficial to the corrosion resistance of austenitic and duplex stainless steels (Ref 6). Nitrogen seems to retard sensitization and improve the resistance to pitting and crevice corrosion of many stainless steels.

Fig. 18 Ferrite/austenite grain-boundary ditching in as-cast ACI CF-8. The specimen, which contained 3% ferrite, was EPR tested. SEM micrograph. 4550×. Source: Ref 5

The standard practices of ASTM A 262 (Ref 7) are commonly implemented to predict and measure the susceptibility of austenitic and duplex stainless steels to intergranular corrosion. Table 7 indicates some representative results for CF-type alloys as tested according to practices A, B, and C of Ref 7 as well as two electrochemical tests described in Ref 10 and 11. Table 8 lists the compositions of the alloys investigated. The data indicate the superior resistance of the low-carbon alloys to intergranular corrosion. Table 7 also indicates that for highly oxidizing environments (represented here by A 262C-boiling HNO3) the CF-3 and CF-3M alloys are equivalent in the solution-treated condition but that subsequent heat treatment causes the corrosion resistance of the CF-3M alloys to deteriorate rapidly for service in oxidizing environments (Ref 9). In addition, the degree of chromium depletion necessary to cause susceptibility to intergranular corrosion appears to increase in the presence of molybdenum (Ref 5). The passive film stability imparted by molybdenum may offset the loss of solid-solution chromium for mild degrees of sensitization.

Table 7 Intergranular corrosion test results for ACI casting alloys Metallurgical condition

Solution treated

Simulated weld repair

Solution treated, held 1 h at 650 °C (1200 °F)

Test(a)

Alloy(b)/Test results(c)

CF8 (4)

CF8 (11)

CF8 (20)

CF8M (5)

CF8M (11)

CF8M (20)

CF3 (2)

CF3 (5)

CF3 (8)

CF3M (5)

CF3M (9)

CF3M (16)

A 262A

P

P

P

P

P

P

P

P

P

P

P

P

A 262B

P

P

P

P

P

P

P

P

P

P

P

P

A 262C

P

P

P

P

P

P

P

P

P

P

P

P

EPR

P

P

P

P

P

P

P*

P*

P*

P

P

P

JEPR

P

P

P

P

P

P

P

P

P

P

P

P

A 262A

X

X

X

X

X

X

P

P

P

P

P

P

A 262B

X

X

X

X

X

X

P

P

P

P

P

P

A 262C

X

X

X

X

X

X

P

P

P

P

P

P

EPR

X

X

X

P

P

P

P*

P*

P*

P

P

P

JEPR

X

X

X

P

P

P

P

P

P

P

P

P

A 262A

X

X

X

X

X

X

X

X

X

X

X

X

A 262B

X

X

X

X

X

X

P

P

P

P

P

P

A 262C

X

X

X

X

X

X

P

P

P

X

X

X

EPR

X

X

X

X

X

X

X/P*

X/P*

X/P*

X/P

P

P

JEPR

X

X

X

P

X

X

P

P

P

P

P

P

As-cast

A 262A

X

X

X

X

X

X

X

X

X

X

X

X

A 262B

X

X

X

X

X

X

P

P

P

P

X

P

A 262C

X

X

X

X

X

X

P**

P**

P**

X

X

X

EPR

X

X

X

X

X

X

X/P*

X/P*

X/P*

X/P

X/P

P

JEPR

X

X

X

X

X

X

X/P

P

P

P

P

P

Source: Ref 5, 8, 9 (a) See Ref 7 for details of ASTM A 262 practices. EPR, electrochemical potentiokinetic reactivation test; see Ref 10 for details. JEPR, Japanese electrochemical potentiokinetic reactivation test: see Ref 11 for details.

(b) Parenthetical value is the percentage of ferrite. See Table 8 for alloy compositions.

(c) P, pass; X, fail, based on the following criteria: A 262A ditching, 99% Cu

High-copper alloys

C16200-C19600

>96% Cu

Brasses

C205-C28580

Cu-Zn

Leaded brasses

C31200-C38590

Cu-Zn-Pb

Tin brasses

C40400-C49080

Cu-Zn-Sn-Pb

Phosphor bronzes

C50100-C52400

Cu-Sn-P

Leaded phosphor bronzes

C53200-C54800

Cu-Sn-Pb-P

Copper-phosphorus and copper-silver-phosphorus alloys

C55180-C55284

Cu-P-Ag

Aluminum bronzes

C60600-C64400

Cu-Al-Ni-Fe-Si-Sn

Silicon bronzes

C64700-C66100

Cu-Si-Sn

Other copper-zinc alloys

C66400-C69900

...

Wrought alloys

Copper-nickels

C70000-C79900

Cu-Ni-Fe

Nickel silvers

C73200-C79900

Cu-Ni-Zn

Coppers

C80100-C81100

>99% Cu

High-copper alloys

C81300-C82800

>94% Cu

Red and leaded red brasses

C83300-C85800

Cu-Zn-Sn-Pb (75-89% Cu)

Yellow and leaded yellow brasses

C85200-C85800

Cu-Zn-Sn-Pb (57-74% Cu)

Manganese and leaded manganese bronzes

C86100-C86800

Cu-Zn-Mn-Fe-Pb

Silicon bronzes, silicon brasses

C87300--C87900

Cu-Zn-Si

Tin bronzes and leaded tin bronzes

C90200-C94500

Cu-Sn-Zn-Pb

Nickel-tin bronzes

C94700-C94900

Cu-Ni-Sn-Sn-Zn-Pb

Aluminum bronzes

C95200-C95810

Cu-Al-Fe-Ni

Copper-nickels

C96200-C96800

Cu-Ni-Fe

Nickel silvers

C97300-C97800

Cu-Ni-Zn-Pb-Sn

Leaded coppers

C98200-C98800

Cu-Pb

Miscellaneous alloys

C99300-C99750

...

Cast alloys

Coppers and high-copper alloys have similar corrosion resistance. They have excellent resistance to seawater corrosion and biofouling, but are susceptible to erosion-corrosion at high water velocities. The high-copper alloys are primarily used in applications that require enhanced mechanical performance, often at slightly elevated temperature, with good thermal or electrical conductivity. Processing for increased strength in the high-copper alloys generally improves their resistance to erosion-corrosion. A number of alloys in this category have been developed for electronic applications-such as contact clips, springs, and lead frames--that require specific mechanical properties, relatively high electrical conductivity, and atmospheric-corrosion resistance. Brasses are basically copper-zinc alloys and are the most widely used group of copper alloys. The resistance of brasses

to corrosion by aqueous solutions does not change markedly as long as the zinc content does not exceed about 15%;

above 15% Zn, dezincification may occur. Quiescent or slowly moving saline solutions, brackish waters, and mildly acidic solutions are environments that often lead to the dezincification of unmodified brasses. Susceptibility to stress-corrosion cracking (SCC) is significantly affected by zinc content; alloys that contain more zinc are more susceptible. Resistance increases substantially as zinc content decreases from 15 to 0%. Stress-corrosion cracking is practically unknown in commercial copper. Elements such as lead, tellurium, beryllium, chromium, phosphorus, and manganese have little or no effect on the corrosion resistance of coppers and binary copper-zinc alloys. These elements are added to enhance such mechanical properties as machinability, strength, and hardness. Tin Brasses. Tin additions significantly increase the corrosion resistance of some brasses, especially resistance to dezincification. Examples of this effect are two tin-bearing brasses: uninhibited admiralty metal (no active UNS number) and naval brass (C46400). Uninhibited admiralty metal was once widely used to make heat-exchanger tubes; it has largely been replaced by inhibited grades of admiralty metal (C44300, C44400, and C44500), which have even greater resistance to dealloying. Admiralty metal is a variation of cartridge brass (C26000) that is produced by adding about 1% Sn to the basic 70Cu-30Zn composition. Similarly, naval brass is the alloy resulting from the addition of 0.75% Sn to the basic 60Cu-40Zn composition of Muntz metal (C28000).

Cast brasses for marine use are also modified by the addition of tin, lead, and, sometimes, nickel. This group of alloys is known by various names, including composition bronze, ounce metal, and valve metal. These older designations are used less frequently, because they have been supplanted by alloy numbers under the UNS or Copper Development Association (CDA) system. The cast marine brasses are used for plumbing goods in moderate-performance seawater piping systems or in deck hardware, for which they are subsequently chrome plated. Aluminum Brasses. An important constituent of the corrosion film on a brass that contains a few percent aluminum in

addition to copper and zinc is aluminum oxide (Al2O3), which markedly increases resistance to impingement attack in turbulent high-velocity saline water. For example, the arsenical aluminum brass C68700 (76Cu-22Zn-2Al) is frequently used for marine condensers and heat exchangers in which impingement attack is likely to pose a serious problem. Aluminum brasses are susceptible to dezincification unless they are inhibited, which is usually done by adding 0.02 to 0.10% As. Inhibited Alloys. Addition of phosphorus, arsenic, or antimony (typically 0.02 to 0.10%) to admiralty metal, naval

brass, or aluminum brass effectively produces high resistance to dezincification. Inhibited alloys have been extensively used for such components as condenser tubes, which must accumulate years of continuous service between shutdowns for repair or replacement. Phosphor Bronzes. Addition of tin and phosphorus to copper produces good resistance to flowing seawater and to most nonoxidizing acids except hydrochloric (HCI). Alloys containing 8 to 10% Sn have high resistance to impingement attack. Phosphor bronzes are much less susceptible to SCC than brasses and are similar to copper in resistance to sulfur attack. Tin bronzes--alloys of copper and tin--tend to be used primarily in the cast form, in which they are modified by further alloy additions of lead, zinc, and nickel. Like the cast brasses, the cast tin bronzes are occasionally identified by older, more colorful names that reflect their historic uses, such as G Bronze, Gun Metal, Navy M Bronze, and steam bronze. Contemporary uses include pumps, valves, gears, and bushings. Wrought tin bronzes are known as phosphor bronzes and find use in high strength wire applications, such as wire rope. This group of alloys has fair resistance to impingement and good resistance to biofouling. Copper Nickels. Alloy C71500 (Cu-30Ni) has the best general resistance to aqueous corrosion of all the commercially

important copper alloys, but C70600 (Cu-10Ni) is often selected because it offers good resistance at lower cost. Both of these alloys, although well suited to applications in the chemical industry, have been most extensively used for condenser tubes and heat-exchanger tubes in recirculating steam systems. They are superior to coppers and to other copper alloys in resisting acid solutions and are highly resistant to SCC and impingement corrosion. Nickel Silvers. The two most common nickel silvers are C75200 (65Cu-18Ni-17Zn) and C77000 (55Cu-18Ni-27Zn).

They have good resistance to corrosion in both fresh and salt waters. Primarily because their relatively high nickel contents inhibit dezincification, C75200 and C77000 are usually much more resistant to corrosion in saline solutions than brasses of similar copper content.

Copper-silicon alloys generally have the same corrosion resistance as copper, but they have higher mechanical

properties and superior weldability. These alloys appear to be much more resistant to SCC than the common brasses. Silicon bronzes are susceptible to embrittlement by high-pressure steam and should be tested for suitability in the service environment before being specified for components to be used at elevated temperature. Aluminum bronzes containing 5 to 12% Al have excellent resistance to impingement corrosion and high-temperature

oxidation. Aluminum bronzes are used for beater bars and for blades in wood pulp machines because of their ability to withstand mechanical abrasion and chemical attack by sulfite solutions. In most practical commercial applications, the corrosion characteristics of aluminum bronzes are primarily related to aluminum content. Alloys with up to 8% Al normally have completely face-centered cubic (fcc) structures and good resistance to corrosion attack. As aluminum content increases above 8%, - duplex structures appear. The phase is a high-temperature phase retained at room temperature upon fast cooling from 565 °C (1050 °F) or above. Slow cooling for long exposure at temperatures from 320 to 565 °C (610 to 1050 °F) tends to decompose the phase into a brittle + 2 eutectoid having either a lamellar or a nodular structure. The phase is less resistant to corrosion than the phase, and eutectoid structures are even more susceptible to attack. Depending on specific environmental conditions, phase or eutectoid structure in aluminum bronze can be selectively attacked by a mechanism similar to the dezincification of brasses. Proper quench-and-temper treatment of duplex alloys, such as C62400 and C95400, produces a tempered structure with reprecipitated acicular crystals, a combination that is often superior in corrosion resistance to the normal annealed structures. Iron-rich particles are distributed as small round or rosette particles throughout the structures of aluminum bronzes containing more than about 0.5% Fe. These particles sometimes impart a rusty tinge to the surface, but have no known effect on corrosion rates. Nickel-aluminum bronzes are more complex in structure with the introduction of the phase. Nickel appears to alter the corrosion characteristics of the phase to provide greater resistance to dealloying and cavitation-erosion in most liquids. For C63200 and perhaps C95800, quench-and-temper treatments may yield even greater resistance to dealloying. Alloy C95700, a high-manganese cast aluminum bronze, is somewhat inferior in corrosion resistance to C95500 and C95800, which are low in manganese and slightly higher in aluminum. Aluminum bronzes are generally suitable for service in nonoxidizing mineral acids, such as phosphoric (H3PO4) sulfuric (H2SO4), and HCl; organic acids, such as lactic, acetic (CH3COOH), or oxalic; neutral saline solutions, such as sodium chloride (NaCI) or potassium chloride (KCl); alkalies, such as sodium hydroxide (NaOH), potassium hydroxide (KOH), and anhydrous ammonium hydroxide (NH4OH); and various natural waters including sea, brackish, and potable waters. Environments to be avoided include nitric acid (HNO3); some metallic salts, such as ferric chloride (FeCl3) and chromic acid (H2CrO4); moist chlorinated hydrocarbons; and moist HN3. Aeration can result in accelerated corrosion in many media that appear to be compatible. Exposure under high tensile stress to moist NH3 can result in SCC. In certain environments, corrosion can lower the fatigue limit to 25 to 50% of the normal atmospheric value.

Types of Attack Coppers and copper alloys, like most other metals and alloys, are susceptible to several forms of corrosion, depending primarily on environmental conditions. Table 2 lists the identifying characteristics of the forms of corrosion that commonly attack copper metals as well as the most effective means of combating each.

Table 2 Guide to corrosion of copper alloys Form of attack

Characteristics

Preventive measures

General thinning

Uniform metal removal

Select proper alloy for environmental conditions based on weight loss data.

Galvanic corrosion

Corrosion preferentially near a more cathodic metal

Avoid electrically coupling dissimilar metals; maintain optimum ratio of anode to cathode area; maintain optimum concentration of oxidizing constituent in corroding medium.

Pitting

Localized pits, tubercles; water line pitting; crevice corrosion; pitting under foreign objects or dirt

Alloy selection; design to avoid crevices; keep metal clean.

Impingement, Erosion-corrosion cavitation

Erosion attack from turbulent flow plus dissolved gases, generally as lines of pits in direction of fluid flow

Design for streamlined flow; keep velocity low; remove gases from liquid phase; use erosion-resistant alloy.

Fretting

Chafing or galling, often occurring during shipment

Lubricate contacting surfaces; interleave sheets of paper between sheets of metal; decrease load on bearing surfaces.

Intergranular corrosion

Corrosion along grain boundaries without visible signs of cracking

Select proper alloy for environmental conditions based on metallographic examination of corrosion specimens.

Dealloying

Preferential dissolution of zinc or nickel, resulting in a layer of sponge copper

Select proper alloy for environmental conditions based on metallographic examination of corrosion specimens.

Corrosion fatigue

Several transgranular cracks

Select proper alloy based on fatigue tests in service environment; reduce mean or alternating stress.

SCC

Cracking, usually intergranular but sometimes transgranular, that is often fairly rapid

Select proper alloy based on stress-corrosion tests; reduce applied or residual stress; remove mercury compounds or NH3 from environment.

General Corrosion General corrosion is the well-distributed attack of an entire surface with little or no localized penetration. It is the least damaging of all forms of attack. General corrosion is the only form of corrosion for which weight loss data can be used to estimate penetration rates accurately. General corrosion of copper alloys results from prolonged contact with environments in which the corrosion rate is very low, such as fresh, brackish, and salt waters; many types of soil; neutral, alkaline, and acid salt solutions; organic acids; and sugar juices. Other substances that cause uniform thinning at a faster rate include oxidizing acids, sulfur-bearing compounds, NH3, and cyanides. Additional information on this form of attack is available in the article "General Corrosion" in this Volume. Galvanic Corrosion An electrochemical potential almost always exists between two dissimilar metals when they are immersed in a conductive solution. If two dissimilar metals are in electrical contact with each other and immersed in a conductive solution, a potential results that enhances the corrosion of the more electronegative member of the couple (the anode) and partly or completely protects the more electropositive member (the cathode). Copper metals are almost always cathodic to other

common structural metals, such as steel and aluminum. When steel or aluminum is put in contact with a copper metal, the corrosion rate of the steel or aluminum increases, but that of the copper metal decreases. The common grades of stainless steel exhibit variable behavior; that is, copper metals may be anodic or cathodic to the stainless steel, depending on conditions of exposure. Copper metals usually corrode preferentially when coupled with high-nickel alloys, titanium, or graphite. Additional information on this subject is available in the section "Galvanic Corrosion." of the article "General Corrosion" in this Volume. Corrosion potentials of copper metals generally range from -0.2 to -0.4 V when measured against a saturated calomel electrode (SCE); the potential of pure copper is about -0.3 V. Alloying additions of zinc or aluminum move the potential toward the anodic (more electronegative) end of the range; additions of tin or nickel move the potential toward the cathodic (less electronegative) end. Galvanic corrosion between two copper metals is seldom a significant problem, because the potential difference is so small. Table 3 lists a galvanic series of metals and alloys valid for dilute aqueous solutions, such as seawater and weak acids. The metals that are grouped together can be coupled to each other without significant galvanic damage. However, the connecting of metals from different groups leads to damage of the more anodic metal; the larger the difference in galvanic potential between groups, the greater the corrosion. Accelerated damage due to galvanic effects is usually greatest near the junction, where the electrochemical current density is the highest. Table 3 Galvanic series in seawater Anodic End

Magnesium

Magnesium alloys

Zinc

Galvanized steel

Aluminum alloy 5052H

Aluminum alloy 3004

Aluminum alloy 3003

Aluminum alloy 1100

Aluminum alloy 6053

Alclad aluminum alloys

Cadmium

Aluminum alloy 2017

Aluminum alloy 2024

Low-carbon steel

Wrought iron

Cast iron

Ni-resist cast iron

AISI type 410 stainless steel (active)

50Pb-50Sn solder

AISI type 304 stainless steel (active)

AISI type 316 stainless steel (active)

Lead

Tin

Muntz metal (C28000)

Manganese bronze (C67500)

Naval brass (C46400)

Nickel (active)

Inconel (active)

Cartridge brass (C26000)

Admiralty metal (C44300)

Aluminum bronze (C61400)

Red brass (C23000)

Copper (C11000)

Silicon bronze (C65100)

Copper-nickel, 30% (C71500)

Nickel (passive)

Inconel (passive)

Monel

AISI type 304 stainless steel (passive)

AISI type 316 stainless steel (passive)

Silver

Gold

Platinum

Cathodic End

Another factor that affects galvanic corrosion is area ratio. An unfavorable area ratio exists when the cathodic area is large and the anodic area in small. The corrosion rate of the small anodic area may be several hundred times greater than if the anodic and cathodic areas were equal in size. Conversely, when a large anodic area is coupled to a small cathodic area, current density and damage due to galvanic corrosion are much less. For example, copper rivets (cathodic) used to fasten steel plates together lasted longer than 1.5 years in seawater, but steel rivets used to fasten copper plates were completely destroyed during the same period. Five principal methods are available for eliminating or significantly reducing galvanic corrosion: • • • • •

Select dissimilar metals that are as close as possible to each other in the galvanic series Avoid coupling small anodes to large cathodes Insulate dissimilar metals completely wherever practicable Apply coatings and keep them in good repair, particularly on the cathodic member Use a sacrificial anode; that is, couple the system to a third metal that is anodic to both structural metals

Pitting As with most commercial metals, corrosion of copper metals results in pitting under certain conditions. Pitting is sometimes general over the entire surface, giving the metal an irregular and roughened appearance. In other cases, pits are concentrated in specific areas and are of various sizes and shapes. Detailed information on this form of attack is available in the section "Pitting" in the article "Localized Corrosion" in this Volume. Localized pitting is the most damaging form of corrosive attack because it reduces load-carrying capacity and increases stress concentration by creating depressions or holes in the metal. Pitting is the usual form of corrosive attack at

surfaces on which there are incomplete protective films, nonprotective deposits of scale, or extraneous deposits of dirt or other foreign substances. Copper alloys do not corrode primarily by pitting, but because of metallurgical and environmental factors that are not completely understood, the corroded surface does show a tendency toward nonuniformity. In seawater, pitting tends to occur more often under conditions of relatively low water velocity, typically less than 0.6 to 0.9 m/s (2 to 3 ft/s). The occurrence of pitting is somewhat random regarding the specific location of a pit on the surface as well as whether it will even occur on a particular metal sample. Long-term tests of copper alloys show that the average pit depth does not continually increase with extended times of exposure. Instead, pits tend to reach a certain limit beyond which little apparent increase in depth occurs. Of the copper alloys, the most pit resistant are the aluminum bronzes with less than 8% Al and the low-zinc brasses. Copper nickels and tin bronzes tend to have intermediate pitting resistance, but the highcopper alloys and silicon bronzes are somewhat more prone to pitting. Crevice corrosion is a form of localized corrosion that occurs near a crevice formed either by two metal surfaces or a

metal and a nonmetal surface. Like pitting, crevice attack is a random occurrence, the precise location of which cannot always be predicted. Also, like pitting, the depth of attack appears to level off rather than to increase continually with time. This depth is usually less than that from pitting, and for most copper alloys, it will be less than 400 m (15.8 mils). For most copper alloys, the location of the attack will be outside but immediately adjacent to the crevice due to the formation of metal ion concentration cells. Classic crevice corrosion resulting from oxygen depletion and attack within crevices is less common in copper alloys. Aluminum- and chromium-bearing copper alloys, which form more passive surface films, are susceptible to differential oxygen cell attack, as are aluminum alloys and stainless steels. The occurrence of crevice attack is somewhat statistical in nature, with the odds of it occurring and its severity increasing if the area within a crevice is small compared to the area outside the crevice. Other conditions that will increase the odds of crevice attack are higher water temperatures or a flow condition on the surface outside the crevice. Local cell action similar to crevice attack may also result from the presence of foreign objects or debris, such as dirt, pieces of shell, or vegetation, or it may result from rust, permeable scales, or uneven accumulation of corrosion product on the metallic surface. This type of attack can sometimes be controlled by cleaning the surfaces. For example, condensers and heat exchangers are cleaned periodically to prevent deposit attack. Water line attack is a term used to describe pitting due to a differential oxygen cell functioning between the well-

aerated surface layer of a liquid and the oxygen-starved layer immediately beneath it. The pitting occurs immediately below the water line. Impingement Various forms of impingement attack occur where gases, vapors, or liquids impinge on metal surfaces at high velocities, such as in condensers or heat exchangers. Rapidly moving turbulent water can strip away the protective films from copper alloys. When this occurs, the metal corrodes at a more rapid rate in an attempt to reestablish this film, but because the films are being swept away as rapidly as they are being formed, the corrosion rate remains constant and high. The conditions under which the corrosion product film is removed are different for each alloy and are discussed in the section "Corrosion of Copper Alloys in Specific Environments" in this article. Additional information on various types of impingement attack is available in the article "Mechanically Assisted Degradation" in this Volume. Erosion-corrosion is characterized by undercut grooves, waves, ruts, gullies, and rounded holes; it usually exhibits a

directional pattern. Pits are elongated in the direction of flow and are undercut on the downstream side. When the condition becomes severe, it may result in a pattern of horseshoe-shaped grooves or pits with their open ends pointing downstream. As attack progresses, the pits may join, forming fairly large patches of undercut pits. When this form of corrosion occurs in a condenser tube, it is usually confined to a region near the inlet end of the tube where fluid flow is rapid and turbulent. If some of the tubes in a bundle become plugged, the velocity is increased in the remaining tubes; therefore, the unit should be kept as clean as possible. Erosion-corrosion is most often found with waters containing low levels of sulfur compounds and with polluted, contaminated, or silty salt water or brackish water. The erosive action locally removes protective films, thus contributing to the formation of concentration cells and to localized pitting of anodic sites. Cavitation is a phenomenon that occurs in moving water when the flow is disturbed so as to create a local pressure drop.

Under these conditions, a vapor bubble will form and then collapse, applying a momentary stress of up to 1379 MPa (200

ksi) to the surface. The current theories of cavitation state that this repeated mechanical working of the surface creates a local fatigue situation that aids the removal of metal. This is in agreement with the observations that the harder alloys tend to have greater resistance to cavitation and that there is often an incubation period before the onset of cavitation attack. Of the copper alloys, aluminum bronze has the best cavitation resistance. Cavitation damage will be confined to the area where the bubbles collapse, usually immediately downstream of the low-pressure zone. Impingement attack can be reduced, and the life of the unit extended, by decreasing fluid velocity, streamlining the flow, and removing entrained air. This is usually accomplished by redesigning water boxes, injector nozzles, and piping to reduce or eliminate low-pressure pockets, obstructions to smooth flow, abrupt changes in flow direction, and other features that cause local regions of high-velocity or turbulent flow. Condensers and heat exchangers are less susceptible to impingement attack if they are made of one of the aluminum brasses or copper nickels, which are more erosion resistant than the brasses or tin brasses. Erosion-resistant inserts at tube inlets and epoxy-type coatings are often effective repair methods in existing shell and tube heat exchangers. When contaminated waters are involved, filtering or screening the liquids and cleaning the surfaces can be very effective in minimizing impingement attack. The use of cathodic protection can lessen all forms of localized attack except cavitation. Fretting Another form of attack, called fretting or fretting corrosion, appears as pits or grooves in the metal surface that are surrounded or filled with corrosion product. Fretting is sometimes referred to as chafing, road burn, friction oxidation, wear oxidation, or galling. The basic requirements for fretting are as follows: • • • •

Repeated relative (sliding) motion between two surfaces must occur. The relative amplitude of the motion may be very small--motion of only a few tenths of a millimeter is typical The interface must be under load Both load and relative motion must be sufficient to produce deformation of the interface Oxygen and/or moisture must be present

Fretting does not occur on lubricated surfaces in continuous motion, such as axle bearings, but instead on dry interfaces subject to repeated, small relative displacements. A classic type of fretting occurs during shipment of bundles of mill products having flat faces. Fretting is not confined to coppers and copper alloys, but has been recognized on almost every kind of surface--steel, aluminum, noble metals, mica, and glass. Fretting can be controlled, and sometimes eliminated, by: • • • •

Lubricating with low-viscosity high-tenacity oils to reduce friction at the interface between the two metals and to exclude oxygen from the interface Separating the faying surfaces by interleaving an insulating material Increasing the load to reduce motion between faying surfaces; this may be difficult in practice, because only a minute amount of relative motion is necessary to produce fretting Decreasing the load at bearing surfaces to increase the relative motion between parts

Detailed information is available in the section "Fretting" of the article "Mechanically Assisted Degradation" in this Volume. Intergranular Corrosion Intergranular corrosion is an infrequently encountered form of attack that occurs most often in applications involving high-pressure steam. This type of corrosion penetrates the metal along grain boundaries--often to a depth of several grains--which distinguishes it from surface roughening. Mechanical stress is apparently not a factor in intergranular corrosion. The alloys that appear to be the most susceptible to this form of attack are Muntz metal, admiralty metal,

aluminum brasses, and silicon bronzes. Additional information is provided in the section "Intergranular Corrosion" of the article "Metallurgically Influenced Corrosion" in this Volume. Dealloying Dealloying is a corrosion process in which the more active metal is selectively removed from an alloy, leaving behind a weak deposit of the more noble metal. Copper-zinc alloys containing more than 15% Zn are susceptible to a dealloying process called dezincification. In the dezincification of brass, selective removal of zinc leaves a relatively porous and weak layer of copper and copper oxide. Corrosion of a similar nature continues beneath the primary corrosion layer, resulting in gradual replacement of sound brass by weak, porous copper. Unless arrested, dealloying eventually penetrates the metal, weakening it structurally and allowing liquids or gases to leak through the porous mass in the remaining structure. The term plug-type dealloying refers to the dealloying that occurs in local areas; surrounding areas are usually unaffected or only slightly corroded. In uniform-layer dealloying, the active component of the alloy is leached out over a broad area of the surface. Dezincification is the usual form of corrosion for uninhibited brasses in prolonged contact with waters high in oxygen and carbon dioxide (CO2). It is frequently encountered with quiescent or slowly moving solutions. Slightly acidic water, low in salt content and at room temperature, is likely to produce uniform attack, but neutral or alkaline water, high in salt content and above room temperature, often produces plug-type attack. Brasses with copper contents of 85% or more resist dezincification. Dezincification of brasses with two-phase structures is generally more severe, particularly if the second phase is continuous; it usually occurs in two stages: the high-zinc phase, followed by the lower-zinc phase. Tin tends to inhibit dealloying, especially in cast alloys. Alloys C46400 (naval brass) and C67500 (manganese bronze), which are - brasses containing about 1% Sn, are widely used for naval equipment and have reasonably good resistance to dezincification. Addition of a small amount of phosphorous, arsenic, or antimony to admiralty metal (an all71Cu-28Zn-1Sn brass) inhibits dezincification. Inhibitors are not entirely effective in preventing dezincification of the - brasses, because they do not prevent dezincification of the phase. Where dezincification is a problem, red brass, commercial bronze, inhibited admiralty metal, and inhibited aluminum brass can be successfully used. In some cases, the economic penalty of avoiding dealloying by selecting a low-zinc alloy may be unacceptable. Low-zinc alloy tubing requires fittings that are available only as sand castings, but fittings for higher-zinc tube can be die cast or forged much more economically. Where selection of a low-zinc alloy is unacceptable, inhibited yellow brasses are generally preferred. Dealloying has been observed in other alloys. Dealloying of aluminum occurs in some copper-aluminum alloys, particularly with those having more than 8% Al. It is especially severe in alloys with continuous phase and usually occurs as plug-type dealloying. Nickel additions exceeding 3.5% or heat treatment to produce an + microstructure prevents dealloying. Dealloying of nickel in C71500 is rare, have been observed at temperatures over 100 °C (212 °F), low flow conditions, and high local heat flux. Dealloying of tin in cast tin bronzes has been observed as a rare occurrence in hot brine or steam. Cathodic protection generally protects all but the two-phase copper-zinc alloys from dealloying. Additional information on this form of attack is available in the section "Dealloying Corrosion" of the article "Metallurgically Influenced Corrosion" in this Volume. Corrosion Fatigue The combined action of corrosion (usually pitting corrosion) and cyclic stress may result in corrosion fatigue cracking. Like ordinary fatigue cracks, corrosion fatigue cracks generally propagate at right angles to the maximum tensile stress in the affected region. However, cracks resulting from simultaneous fluctuating stress and corrosion propagate much more rapidly than cracks caused solely by fluctuating stress. Also, corrosion fatigue failure usually involves several parallel cracks, but it is rare for more than one crack to be found in a part that has failed by simple fatigue. The cracks shown in Fig. 1 are characteristic of service failures resulting from corrosion fatigue.

Fig. 1 Typical corrosion fatigue cracking of a copper alloy. Transgranular cracks originate at the base of corrosion pits on the roughened inner surface of a tube. Etched. About 150×

Ordinarily, corrosion fatigue can be readily identified by the presence of several cracks emanating from corrosion pits. Cracks not visible to the unaided eye or at low magnification can be made visible by deep etching or plastic deformation or can be detected by eddy-current inspection. Corrosion fatigue cracking is often transgranular, but there is evidence that certain environments induce intergranular cracking in copper metals. In addition to effective resistance to corrosion, copper and copper alloys also resist corrosion fatigue in many applications involving repeated stress and corrosion. These applications include such parts as springs, switches, diaphragms, bellows, aircraft and automotive gasoline and oil lines, tubes for condensers and heat exchangers, and fourdrinier wire for the paper industry. Copper alloys that are high in fatigue limit and resistance to corrosion in the service environment are more likely to have good resistance to corrosion fatigue. Alloys frequently used in applications involving both cyclic stress and corrosion include beryllium coppers, phosphor bronzes, aluminum bronzes, and copper nickels. More information on corrosion fatigue is available in the section "Corrosion Fatigue" of the article "Mechanically Assisted Degradation" in this Volume. Stress-Corrosion Cracking Stress-corrosion cracking and season cracking describe the same phenomenon--the apparently spontaneous cracking of stressed metal. Stress-corrosion cracking is often intergranular (Fig. 2), but transgranular cracking may occur in some alloys in certain environments. Stress-corrosion cracking occurs only if a susceptible alloy is subjected to the combined effects of sustained stress and certain chemical substances.

Fig. 2 Typical SCC in a copper alloy. Intergranular cracking in an etched specimen. About 60×

Mechanism. Copper alloys crack in a wide variety of electrolytes. In some cases, the crack surfaces have the distinctive

brittle appearance that is associated with SCC. In other cases, the threshold stress for cracking may be close to that observed in air, and the fracture surfaces resemble those of samples fractured in air. It is also clear in many systems that cracking occurs at low threshold stresses only when certain environmental conditions exist. Variables that control this threshold stress in a specific environment include pH, potential of the metal, temperature, extent of cold work before the test, and minor alloying elements in the copper alloy. The best nonquantitative interpretation of SCC is the following. Stress-corrosion cracking occurs in those environmental/metal systems in which the rate of corrosion is low; the corrosion that does occur proceeds in a highly localized manner. Intergranular attack, selective removal of an alloy component, pitting, attack at a metal/precipitate interface, or surface flaws, when they occur in the presence of a surface tensile stress, may lead to a surface defect at the base of which the stress intensity factor, KI, exceeds the threshold stress intensity for SCC KIscc, for that specific environment/alloy system under the conditions selected for the test or encountered in service. Whether or not a crack propagates depends on the specimen geometry and how the magnitude of the stress field at the crack tip changes as the crack develops. The critical factor is how the metal reacts at the crack tip. If the metallurgical structure or the kinetics of chemical corrosion at the crack tip is such that a small radius of curvature (sharp crack tip) is maintained at the crack tip, the crack will continue to propagate because the local stress at the crack tip is high. High rates of corrosion at the crack tip, which lead to a large radius of curvature (blunt), will favor pitting rather than crack growth. A sharp crack tip is favored by: • • • • •

Selective removal of one component of an alloy with the resulting development of local voids that provide a brittle crack path Brittle fracture of a corrosion product coating at the base of a crack that continually reforms Attack along the interface of two discrete phases Intergranular attack that does not spread laterally Surface energy considerations that encourage intrusion of the environment (a liquid metal in particular) into minute flaws

Since the discovery by E. Mattsson that a medium containing ammonium sulfate [(NH4)2SO4], NH4OH, and copper sulfate (CuSO4) is an excellent one for studying the fundamentals of the SCC process caused by NH3, many researchers have used this electrolyte, and the name Mattsson's solution has been given to this solution (Ref 1). Much of the knowledge of the specifics of SCC by NH3 solutions has been obtained from brass exposed to this solution while under a tensile stress. The chemistry and the electrochemistry of the brass-NH3 system was recently reviewed and analyzed (Ref 2). Cupric (Cu2+) ammonium complex was concluded to be necessary for the occurrence of SCC under open-circuit conditions in oxygenated NH3 solutions. This complex becomes a component in the predominant cathodic reaction:

+ e-

+ 2NH3

(Eq 3)

Equation 3 permits cracking by cyclic rupture of a Cu2O film generated at the crack tip (Ref 3) or by a mechanism involving dezincification (Ref 4). Cracking can also occur in deoxygenated solutions in the absence of significant concentrations of the Cu2+ ions provided the cuprous (Cu+) complexes are available. It was suggested that the role of the Cu+ complex is to provide a cathodic reaction, in this case allowing dezincification to occur. These findings are consistent with the recognition that SCC failures of brass are not limited to environments containing NH3. The most damaging evidence against the film rupture model is given in Ref 5. In this study, the tarnish film that formed on unstressed 70Cu-30Zn brass during exposure for 48 h to an NH4OH-(NH4)2SO4-CuSO4 electrolyte at pH 7.2 was shown to fracture transgranularly when fractured in air. The reported film rupture mechanism predicts that these films should fracture intergranularly. The transgranular cracks do not propagate when a stressed specimen is immersed in the electrolyte; instead, very rapid intergranular SCC is observed. These facts are also difficult to reconcile with the repeated film rupture model. It was first shown in 1972 that dezincification of 70Cu-30Zn brass occurs in the crack during SCC in an ammonium salt environment (Ref 4). More recently, mechanical strain was found to lead to dezincification of both 85Cu-15Zn and 70Cu30Zn alloys in an NH4OH-(NH4)SO4-CuSO4 electrolyte (Ref 6). Unstressed samples of the same alloys did not show dezincification. Strain-induced dealloying was further shown to occur in both intergranular (copper-zinc) and transgranular (copper-zinc-nickel) (Ref 7). These observations indicated that stress corrosion of copper alloys is integrally related to strain-induced dealloying. Conditions Leading to SCC. Ammonia and ammonium compounds are the corrosive substances most often associated

with SCC of copper alloys. These compounds are sometimes present in the atmosphere; in other cases, they are in cleaning compounds or in chemicals used to treat boiler water. Both oxygen and moisture must be present for NH3 to be corrosive to copper alloys; other compounds, such as CO2, are thought to accelerate SCC in NH3 atmospheres. Moisture films on metal surfaces will dissolve significant quantities of NH3, even from atmospheres with low NH3 concentrations. A specific corrosive environment and sustained stress are the primary causes of SCC; microstructure and alloy composition may affect the rate of crack propagation in susceptible alloys. Microstructure and composition can be most effectively controlled by selecting the correct combination of alloy, forming process, thermal treatment, and metalfinishing process. Although test results may indicate that a finished part is not susceptible to SCC, such an indication does not ensure complete freedom from cracking, particularly where service stresses are high. Applied and residual stresses can both lead to failure by SCC. Susceptibility is largely a function of stress magnitude. Stresses near the yield strength are usually required, but parts have failed under much lower stresses. In general, the higher the stress, the weaker the corroding medium must be to cause SCC. The reverse is also true: the stronger the corroding medium, the lower the required stress. Sources of Stress. Applied stresses result from ordinary service loading or from fabricating techniques, such as

riveting, bolting, shrink fitting, brazing, and welding. Residual stresses are of two types: differential-strain stresses, which result from nonuniform plastic strain during cold forming, and differential-thermal-contraction stresses, which result from nonuniform heating and/or cooling. Residual stresses induced by nonuniform straining are primarily influenced by the method of fabrication. In some fabricating processes, it is possible to cold work a metal extensively and yet produce only a low level of residual stress. For example, residual stress in a drawn tube is influenced by die angle and amount of reduction. Wide-angle dies (about 32°) produce higher residual stresses than narrow-angle dies (about 8°). Light reductions yield high residual stresses because only the surface of the alloy is stressed; heavy reductions yield low residual stresses because the region of cold working extends deeper into the metal. Most drawing operations can be planned so that residual stresses are low and susceptibility to SCC is negligible. Residual stresses resulting from upsetting, stretching, or spinning are more difficult to evaluate and to control by varying tooling and process conditions. For these operations, SCC can be prevented more effectively by selecting a resistant alloy or by treating the metal after fabrication.

Alloy Composition. Brasses containing less than 15% Zn are highly resistant to SCC. Phosphorus-deoxidized copper

and tough pitch copper rarely exhibit SCC, even under severe conditions. On the other hand, brasses containing 20 to 40% Zn are highly susceptible. Susceptibility increases only slightly as zinc content is increased from 20 to 40%. There is no indication that the other elements commonly added to brasses increase the probability of SCC. Phosphorus, arsenic, magnesium, tellurium, tin, beryllium, and manganese are thought to decrease susceptibility under some conditions. Addition of 1.5% Si is known to decrease the probability of cracking. Altering the microstructure cannot make a susceptible alloy totally resistant to SCC. However, the rapidity with which susceptible alloys crack appears to be affected by grain size and structure. All other factors being equal, the rate of cracking increases with grain size. The effects of structure on SCC are not sharply defined, primarily because they are interrelated with effects of both composition and stress. Control Measures. Stress-corrosion cracking can be controlled, and sometimes prevented, by selecting copper alloys

that have high resistance to cracking (notably those with less than 15% Zn); by reducing residual stress to a safe level by thermal stress relief, which can usually be applied without significantly decreasing strength; or by altering the environment, such as by changing the predominant chemical species present or introducing a corrosion inhibitor. Residual and assembly stresses can be eliminated by recrystallization annealing after forming or assembly. Recrystallization annealing cannot be used when the integrity of the structure depends on the higher strength of strainhardened metal, which always contains a certain amount of residual stress. Thermal stress relief (sometimes called relief annealing) can be specified when the higher strength of a cold-worked temper must be retained. Thermal stress relief consists of heating the part for a relatively short time at low temperature. Specific times and temperatures depend on alloy composition, severity of deformation, prevailing stresses, and the size of the load being heated. Usually, time is from 30 min to 1 h and temperature is from 150 to 425 °C (300 to 795 °F). Table 4 lists typical stress-relieving times and temperatures for some of the more common copper alloys. Table 4 Typical stress-relieving parameters for some common copper alloys Common name

UNS number

Temperature

°C

°F

Time, h

Commercial bronze

C22000

205

400

1

Cartridge brass

C26000

260

500

1

Muntz metal

C28000

190

375

Admiralty metal

C44300, C44400, C44500

300

575

1

Phosphor bronze, 5 or 10%

C51000, C52400

190

375

1

Silicon bronze

C65500

370

700

1

Aluminum bronze

C61300, C61400

400

750

1

The exact thermal treatment should be established by examining specific parts for residual stress. If such examination indicates that a thermal treatment is insufficient, temperature and/or time should be adjusted until satisfactory results are obtained. Parts in the center of a furnace load may not reach the desired temperature as soon as parts around the periphery. Therefore, it may be necessary to compensate for furnace loading when setting process controls or to limit the number of parts that can be stress relieved together. Mechanical methods, such as stretching, flexing, bending, straightening between rollers, peening, and shot blasting, can also be used to reduce residual stresses to a safe level. These methods depend on plastic deformation to decrease dangerous tensile stresses or to convert them to less objectionable compressive stresses. Additional information on SCC is available in the section "Stress-Corrosion Cracking" of the article "Environmentally Induced Cracking" in this Volume.

Corrosion of Copper Alloys in Specific Environments Selection of a suitably resistant material requires consideration of the many factors that influence corrosion. Operating records are the most reliable guidelines as long as the data are accurately interpreted. Some of the information in this article has been collected over a period of 20 years or more. Results of short-term laboratory and field testing are also described, but these data may not be as reliable for solving certain problems. Laboratory corrosion tests often do not duplicate such operating factors as stress, velocity, galvanic coupling, concentration cells, initial surface conditions, and contamination of the surrounding medium. If damage occurs by pitting, intergranular corrosion, or dealloying (as in dezincification) or if a thick adherent scale forms, corrosion rates calculated from a change in weight may be misleading. From these forms of corrosion, estimates of reduction in mechanical strength are often more meaningful. Corrosion fatigue and SCC are also potential sources of failure that cannot be predicted from routine measurements of weight loss or dimensional change. Over the years, experience has been the best criterion for selecting the most suitable alloy for a given environment. The CDA has compiled much field experience in the form of the ratings shown in Table 5. Similar data for cast alloys are given in Table 6. These tables should be used only as a guide; small changes in the environmental conditions sometimes degrade the performance of a given alloy from "suitable" to "not suitable." Table 5 Corrosion ratings of wrought copper alloys in various corrosive media This table is intended to serve only as a general guide to the behavior of copper and copper alloys in corrosive environments. It is impossible to cover in a simple tabulation the performance of a material for all possible variations of temperature, concentration, velocity, impurity content, degree of aeration, and stress. The ratings are based on general performance; they should be used with caution, and then only for the purpose of screening candidate alloys. The letters E, G, F, and P have the following significance: E, excellent: resists corrosion under almost all conditions of service G, good: some corrosion will take place, but satisfactory service can be expected under all but the most severe conditions. F, fair: corrosion rates are higher than for the G classification, but the metal can be used if needed for a property other than corrosion resistance and if either the amount of corrosion does not cause excessive maintenance expense or the effects of corrosion can be lessened, such as by use of coatings or inhibitors. P, poor: corrosion rates are high, and service is generally unsatisfactory. Corrosive medium

Coppers

Low-zinc brasses

High-zinc brasses

Special brasses

Phosphor bronzes

Aluminum bronzes

Silicon bronzes

Copper nickels

Nickel silvers

Acetate solvents

E

E

G

E

E

E

E

E

E

Acetic acid(a)

E

E

P

P

E

E

E

E

G

Acetone

E

E

E

E

E

E

E

E

E

Acetylene(b)

P

P

(b)

P

P

P

P

P

P

Alcohols(a)

E

E

E

E

E

E

E

E

E

Aldehydes

E

E

F

F

E

E

E

E

E

Alkylamines

G

G

G

G

G

G

G

G

G

Alumina

E

E

E

E

E

E

E

E

E

Aluminum chloride

G

G

P

P

G

G

G

G

G

Aluminum hydroxide

E

E

E

E

E

E

E

E

E

G

G

P

G

G

G

G

E

G

Ammonia, dry

E

E

E

E

E

E

E

E

E

Ammonia moist(c)

P

P

P

P

P

P

P

F

P

Ammonium chloride(c)

P

P

P

P

P

P

P

F

P

Ammonium hydroxide(c)

P

P

P

P

P

P

P

F

P

Ammonium nitrate(c)

P

P

P

P

P

P

P

F

P

Ammonium sulfate(c)

F

F

P

P

F

F

F

G

F

Aniline and aniline dyes

F

F

F

F

F

F

F

F

F

Asphalt

E

E

E

E

E

E

E

E

E

E

E

E

E

E

E

E

E

E

Aluminum and alum

sulfate

Atmosphere:

Industrial(c)

Marine

E

E

E

E

E

E

E

E

E

Rural

E

E

E

E

E

E

E

E

E

Barium carbonate

E

E

E

E

E

E

E

E

E

Barium chloride

G

G

F

F

G

G

G

G

G

Barium hydroxide

E

E

G

E

E

E

E

E

E

Barium sulfate

E

E

E

E

E

E

E

E

E

Beer(a)

E

E

G

E

E

E

E

E

E

Beet-sugar syrup(a)

E

E

G

E

E

E

E

E

E

Benzene, benzol

E

E

E

E

E

E

E

E

E

Benzoic acid

E

E

E

E

E

E

E

E

E

Black liquor, sulfate process

P

P

P

P

P

P

P

G

P

Bleaching (wet)

G

G

P

G

G

G

G

G

G

Borax

E

E

E

E

E

E

E

E

E

Bordeaux mixture

E

E

G

E

E

E

E

E

E

Boric acid

E

E

G

E

E

E

E

E

E

Brines

G

G

P

G

G

G

G

E

E

Bromine, dry

E

E

E

E

E

E

E

E

E

Bromine, moist

G

G

P

F

G

G

G

G

G

Butane(d)

E

E

E

E

E

E

E

E

E

Calcium bisulfate

G

G

P

G

G

G

G

G

G

benzine,

powder

Calcium chloride

G

G

F

G

G

G

G

G

G

Calcium hydroxide

E

E

G

E

E

E

E

E

E

Calcium hypochlorite

G

G

P

G

G

G

G

G

G

Cane-sugar syrup(a)

E

E

E

E

E

E

E

E

E

Carbolic (phenol)

F

G

P

G

G

G

G

G

G

Carbonated beverages(a)(e)

E

E

E

E

E

E

E

E

E

Carbon dioxide, dry

E

E

E

E

E

E

E

E

E

Carbon moist(a)(e)

dioxide,

E

E

E

E

E

E

E

E

E

Carbon tetrachloride (dry)

E

E

E

E

E

E

E

E

E

Carbon tetrachloride (moist)

G

G

F

G

E

E

E

E

E

Castor oil

E

E

E

E

E

E

E

E

E

Chlorine, dry(f)

E

E

E

E

E

E

E

E

E

Chlorine, moist

F

F

P

F

F

F

F

G

F

Chloracetic acid

G

F

P

F

G

G

G

G

G

Chloroform, dry

E

E

E

E

E

E

E

E

E

Chromic acid

P

P

P

P

P

P

P

P

P

Citric acid(a)

E

E

F

E

E

E

E

E

E

Copper chloride

F

F

P

F

F

F

F

F

F

Copper nitrate

F

F

P

F

F

F

F

F

F

acid

Copper sulfate

G

G

P

G

G

G

G

E

G

Corn oil(a)

E

E

G

E

E

E

E

E

E

Cottonseed oil(a)

E

E

G

E

E

E

E

E

E

Creosote

E

E

G

E

E

E

E

E

E

Downtherm "A"

E

E

E

E

E

E

E

E

E

Ethanol amine

G

G

G

G

G

G

G

G

G

Ethers

E

E

E

E

E

E

E

E

E

E

E

G

E

E

E

E

E

E

Ethylene glycol

E

E

G

E

E

E

E

E

E

Ferric chloride

P

P

P

P

P

P

P

P

P

Ferric sulfate

P

P

P

P

P

P

P

P

P

Ferrous chloride

G

G

P

G

G

G

G

G

G

Ferrous sulfate

G

G

P

G

G

G

G

G

G

Formaldehyde (aldehydes)

E

E

G

E

E

E

E

E

E

Formic acid

G

G

P

F

G

G

G

G

G

Freon, dry

E

E

E

E

E

E

E

E

E

Freon, moist

E

E

E

E

E

E

E

E

E

Fuel oil, light

E

E

E

E

E

E

E

E

E

Fuel oil, heavy

E

E

G

E

E

E

E

E

E

Furfural

E

E

F

E

E

E

E

E

E

Ethyl (esters)

acetate

Gasoline

E

E

E

E

E

E

E

E

E

Gelatin(a)

E

E

E

E

E

E

E

E

E

Glucose(a)

E

E

E

E

E

E

E

E

E

Glue

E

E

G

E

E

E

E

E

E

Glycerin

E

E

G

E

E

E

E

E

E

Hydrobromic acid

F

F

P

F

F

F

F

F

F

Hydrocarbons

E

E

E

E

E

E

E

E

E

Hydrochloric (muriatic)

acid

F

F

P

F

F

F

F

F

F

Hydrocyanic dry

acid,

E

E

E

E

E

E

E

E

E

Hydrocyanic moist

acid,

P

P

P

P

P

P

P

P

P

Hydrofluoric anhydrous

acid,

G

G

P

G

G

G

G

G

G

Hydrofluoric hydrated

acid,

F

F

P

F

F

F

F

F

F

Hydrofluosilicic acid

G

G

P

G

G

G

G

G

G

Hydrogen(d)

E

E

E

E

E

E

E

E

E

Hydrogen peroxide up to 10%

G

G

F

G

G

G

G

G

G

Hydrogen peroxide over 10%

P

P

P

P

P

P

P

P

P

Hydrogen dry

sulfide,

E

E

E

E

E

E

E

E

E

Hydrogen moist

sulfide,

P

P

F

F

P

P

P

F

F

Kerosine

E

E

E

E

E

E

E

E

E

Ketones

E

E

E

E

E

E

E

E

E

Lacquers

E

E

E

E

E

E

E

E

E

E

E

E

E

E

E

E

E

E

Lactic acid(a)

E

E

F

E

E

E

E

E

E

Lime

E

E

E

E

E

E

E

E

E

Lime sulfur

P

P

F

F

P

P

P

F

F

Linseed oil

G

G

G

G

G

G

G

G

G

Lithium compounds

G

G

P

F

G

G

G

E

E

Magnesium chloride

G

G

F

F

G

G

G

G

G

Magnesium hydroxide

E

E

G

E

E

E

E

E

E

Magnesium sulfate

E

E

G

E

E

E

E

E

E

Mercury mercury salts

P

P

P

P

P

P

P

P

P

Milk(a)

E

E

G

E

E

E

E

E

E

Molasses

E

E

G

E

E

E

E

E

E

Natural gas(d)

E

E

E

E

E

E

E

E

E

Nickel chloride

F

F

P

F

F

F

F

F

F

Nickel sulfate

F

F

P

F

F

F

F

F

F

Nitric acid

P

P

P

P

P

P

P

P

P

Oleic acid

G

G

F

G

G

G

G

G

G

Lacquer (solvents)

thinners

or

Oxalic acid(g)

E

E

P

P

E

E

E

E

E

Oxygen(h)

E

E

E

E

E

E

E

E

E

Palmitic acid

G

G

F

G

G

G

G

G

G

Paraffin

E

E

E

E

E

E

E

E

E

Phosphoric acid

G

G

P

F

G

G

G

G

G

Picric acid

P

P

P

P

P

P

P

P

P

Potassium carbonate

E

G

E

E

E

E

E

E

E

Potassium chloride

G

G

P

F

G

G

G

E

E

Potassium cyanide

P

P

P

P

P

P

P

P

P

Potassium dichromate (acid)

P

P

P

P

P

P

P

P

P

Potassium hydroxide

G

G

F

G

G

G

G

E

E

Potassium sulfate

E

E

G

E

E

E

E

E

E

Propane(d)

E

E

E

E

E

E

E

E

E

Rosin

E

E

E

E

E

E

E

E

E

Seawater

G

G

F

E

G

E

G

E

E

Sewage

E

E

F

E

E

E

E

E

E

Silver salts

P

P

P

P

P

P

P

P

P

Soap solution

E

E

E

E

E

E

E

E

E

Sodium bicarbonate

E

E

G

E

E

E

E

E

E

Sodium bisulfate

G

G

F

G

G

G

G

E

E

Sodium carbonate

E

E

G

E

E

E

E

E

E

Sodium chloride

G

G

P

F

G

G

G

E

E

Sodium chromate

E

E

E

E

E

E

E

E

E

Sodium cyanide

P

P

P

P

P

P

P

P

P

Sodium dichromate (acid)

P

P

P

P

P

P

P

P

P

Sodium hydroxide

G

G

F

G

G

G

G

E

E

Sodium hypochlorite

G

G

P

G

G

G

G

G

G

Sodium nitrate

G

G

P

F

G

G

G

E

E

Sodium peroxide

F

F

P

F

F

F

F

G

G

Sodium phosphate

E

E

G

E

E

E

E

E

E

Sodium silicate

E

E

G

E

E

E

E

E

E

Sodium sulfate

E

E

G

E

E

E

E

E

E

Sodium sulfide

P

P

F

F

P

P

P

F

F

Sodium thiosulfate

P

P

F

F

P

P

P

F

F

Steam

E

E

F

E

E

E

F

E

E

Stearic acid

E

E

F

E

E

E

E

E

E

Sugar solutions

E

E

G

E

E

E

E

E

E

Sulfur, solid

G

G

E

G

G

G

G

E

G

Sulfur, molten

P

P

P

P

P

P

P

P

P

Sulfur (dry)

E

E

E

E

E

E

E

E

E

chloride

chloride

P

P

P

P

P

P

P

P

P

Sulfur dioxide (dry)

E

E

E

E

E

E

E

E

E

Sulfur (moist)

dioxide

G

G

P

G

G

G

G

F

F

Sulfur trioxide (dry)

E

E

E

E

E

E

E

E

E

Sulfuric 95%(i)

acid

80-

G

G

P

F

G

G

G

G

G

Sulfuric 80%(i)

acid

40-

F

F

F

P

F

F

F

F

F

Sulfuric acid 40%(i)

G

G

P

F

G

G

G

G

G

Sulfurous acid

G

G

P

G

G

G

G

F

F

Tannic acid

E

E

E

E

E

E

E

E

E

Tartaric acid(a)

E

E

G

E

E

E

E

E

E

Toluene

E

E

E

E

E

E

E

E

E

Trichloracetic acid

G

G

P

F

G

G

G

G

G

Trichlorethylene (dry)

E

E

E

E

E

E

E

E

E

Trichlorethylene (moist)

G

G

F

G

E

E

E

E

E

Turpentine

E

E

E

E

E

E

E

E

E

Varnish

E

E

E

E

E

E

E

E

E

Vinegar(a)

E

E

P

F

E

E

E

E

G

Water, acidic mine

F

F

P

F

G

F

F

P

F

Water, potable

E

E

G

E

E

E

E

E

E

Sulfur (moist)

Water condensate(c)

E

E

E

E

E

E

E

E

E

Wetting agents(j)

E

E

E

E

E

E

E

E

E

Whiskey(a)

E

E

E

E

E

E

E

E

E

White water

G

G

G

E

E

E

E

E

E

Zinc chloride

G

G

P

G

G

G

G

G

G

Zinc sulfate

E

E

P

E

E

E

E

E

E

(a) Copper and copper alloys are resistant to corrosion by most food products. Traces of copper may be dissolved and affect taste or color of the products. In such cases, copper alloys are often tin coated.

(b) Acetylene forms an explosive compound with copper when moisture or certain impurities are present and the gas is under pressure. Alloys containing less than 65% Cu are satisfactory; when the gas is not under pressure, other copper alloys are satisfactory.

(c) Precautions should be taken to avoid SCC.

(d) At elevated temperatures, hydrogen will react with tough pitch copper, causing failure by embrittlement.

(e) Where air is present, corrosion rate may be increased.

(f) Below 150 °C (300 °F), corrosion rate is very low; above this temperature, corrosion is appreciable and increases rapidly with temperature.

(g) Aeration and elevated temperature may increase corrosion rate substantially.

(h) Excessive oxidation may begin above 120 °C (250 °F). If moisture is present, oxidation may begin at lower temperatures.

(i) Use of high-zinc brasses should be avoided in acids because of the likelihood of rapid corrosion by dezincification. Copper, low-zinc brasses, phosphor bronzes, silicon bronzes, aluminum bronzes, and copper nickels offer good resistance to corrosion by hot and cold dilute H2SO4 and to corrosion by cold concentrated H2SO4. Intermediate concentrations of H2SO4 are sometimes more corrosive to copper alloys than either concentrated or dilute acid. Concentrated H2SO4 may be corrosive at elevated temperatures due to breakdown of acid and formation of metallic sulfides and sulfur dioxide, which cause localized pitting. Tests indicate that copper alloys may undergo pitting in 90 to 95% H2SO4 at about 50 °C (122 °F), in 80% acid at about 70 °C (160 °F), and in 60% acid at about 100 °C (212 °F).

(j) Wetting agents may increase corrosion rates of copper and copper alloys slightly to substantially when carbon dioxide or oxygen is present by preventing formation of a film on the metal surface and by combining (in some instances) with the dissolved copper to produce a green, insoluble compound.

Table 6 Corrosion ratings of cast copper alloys in various media The letters A, B, and C have the following significance: A, recommended; B, acceptable; C, not recommended Corrosive medium

Copper

Tin bronze

Leaded tin bronze

Highleaded tin bronze

Leaded red brass

Leaded semi-red brass

Leaded yellow brass

Leaded highstrength yellow brass

Highstrength yellow brass

Aluminum bronze

Leaded nickel brass

Leaded nickel bronze

Silicon bronze

Silicon brass

Acetate solvents

B

A

A

A

A

A

B

A

A

A

A

A

A

B

20%

A

C

B

C

B

C

C

C

C

A

C

A

A

B

50%

A

C

B

C

B

C

C

C

C

A

C

B

A

B

Glacial

A

A

A

C

A

C

C

C

C

A

B

B

A

A

Acetone

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Acetylene(a)

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Alcohols(b)

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Aluminum chloride

C

C

C

C

C

C

C

C

C

B

C

C

C

C

Aluminum sulfate

B

B

B

B

B

C

C

C

C

A

C

C

A

A

Acetic acid

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Ammonia, moisture-free

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Ammonium chloride

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Ammonium hydroxide

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Ammonium nitrate

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Ammonium sulfate

B

B

B

B

B

C

C

C

C

A

C

C

A

A

Aniline and aniline dyes

C

C

C

C

C

C

C

C

C

B

C

C

C

C

Asphalt

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Barium chloride

A

A

A

A

A

C

C

C

C

A

A

A

A

C

Barium sulfide

C

C

C

C

C

C

C

C

B

C

C

C

C

C

Beer(b)

A

A

B

B

B

C

C

C

A

A

C

A

A

B

Beet-sugar syrup

A

A

B

B

B

A

A

A

B

A

A

A

B

B

Benzine

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Ammonia, gas

moist

Benzol

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Boric acid

A

A

A

A

A

A

A

B

A

A

A

A

A

A

Butane

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Calcium bisulfite

A

A

B

B

B

C

C

C

C

A

B

A

A

B

Calcium (acid)

chloride

B

B

B

B

B

B

C

C

C

A

C

C

A

C

Calcium (alkaline)

chloride

C

C

C

C

C

C

C

C

C

A

C

A

C

B

Calcium hydroxide

C

C

C

C

C

C

C

C

C

B

C

C

C

C

Calcium hypochlorite

C

C

B

B

B

C

C

C

C

B

C

C

C

C

Cane-sugar syrups

A

A

B

A

B

A

A

A

A

A

A

A

A

B

Carbonated beverages(b)

A

C

C

C

C

C

C

C

C

A

C

C

A

C

Carbon dry

dioxide,

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Carbon moist(b)

dioxide,

B

B

B

C

B

C

C

C

C

A

C

A

A

B

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Carbon

tetrachloride, dry

Carbon tetrachloride, moist

B

B

B

B

B

B

B

B

B

B

B

A

A

A

Chlorine, dry

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Chlorine, moist

C

C

B

B

B

C

C

C

C

C

C

C

C

C

Chromic acid

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Citric acid

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Copper sulfate

B

A

A

A

A

C

C

C

C

B

B

B

A

A

Cottonseed oil(b)

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Creosote

B

B

B

B

B

C

C

C

C

A

B

B

B

B

Ethers

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Ethylene glycol

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Ferric sulfate

chloride,

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Ferrous sulfate

chloride,

C

C

C

C

C

C

C

C

C

C

C

C

C

C

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Formaldehyde

Formic acid

A

A

A

A

A

B

B

B

B

A

B

B

B

C

Freon

A

A

A

A

A

A

A

A

A

A

A

A

A

B

Fuel oil

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Furfural

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Gasoline

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Gelatin(b)

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Glucose

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Glue

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Glycerin

A

A

A

A

A

A

A

A

A

A

A

A

A

A

C

C

C

C

C

C

C

C

C

B

C

C

C

C

Hydrofluoric acid

B

B

B

B

B

B

B

B

B

A

B

B

B

B

Hydrofluosilicic acid

B

B

B

B

B

C

C

C

C

B

C

C

B

C

Hydrogen

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Hydrogen peroxide

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Hydrochloric muriatic acid

or

Hydrogen dry

sulfide,

C

C

C

C

C

C

C

C

C

B

C

C

B

C

Hydrogen moist

sulfide,

C

C

C

C

C

C

C

C

C

B

C

C

C

C

Lacquers

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Lacquer thinners

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Lactic acid

A

A

A

A

A

C

C

C

C

A

C

C

A

C

Linseed oil

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Black liquor

B

B

B

B

B

C

C

C

C

B

C

C

B

B

Green liquor

C

C

C

C

C

C

C

C

C

B

C

C

C

B

White liquor

C

C

C

C

C

C

C

C

C

A

C

C

C

B

Magnesium chloride

A

A

A

A

A

C

C

C

C

A

C

C

A

B

Magnesium hydroxide

B

B

B

B

B

B

B

B

B

A

B

B

B

B

Magnesium sulfate

A

A

A

A

B

C

C

C

C

A

C

B

A

B

Liquors

Mercury, mercury salts

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Milk(b)

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Molasses(b)

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Natural gas

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Nickel chloride

A

A

A

A

A

C

C

C

C

B

C

C

A

C

Nickel sulfate

A

A

A

A

A

C

C

C

C

A

C

C

A

C

Nitric acid

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Oleic acid

A

A

B

B

B

C

C

C

C

A

C

A

A

B

Oxalic acid

A

A

B

B

B

C

C

C

C

A

C

A

A

B

Phosphoric acid

A

A

A

A

A

C

C

C

C

A

C

A

A

A

Picric acid

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Potassium chloride

A

A

A

A

A

C

C

C

C

A

C

C

A

C

Potassium cyanide

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Potassium hydroxide

C

C

C

C

C

C

C

C

C

A

C

C

C

C

Potassium sulfate

A

A

A

A

A

C

C

C

C

A

C

C

A

C

Propane gas

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Seawater

A

A

A

A

A

C

C

C

C

A

C

C

B

B

Soap solutions

A

A

A

A

B

C

C

C

C

A

C

C

A

C

Sodium bicarbonate

A

A

A

A

A

A

A

A

A

A

A

A

A

B

Sodium bisulfate

C

C

C

C

C

C

C

C

C

A

C

C

C

C

Sodium carbonate

C

A

A

A

A

C

C

C

C

A

C

C

C

A

Sodium chloride

A

A

A

A

A

B

C

C

C

A

C

C

A

C

Sodium cyanide

C

C

C

C

C

C

C

C

C

B

C

C

C

C

Sodium hydroxide

C

C

C

C

C

C

C

C

C

A

C

C

C

C

Sodium hypochlorite

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Sodium nitrate

B

B

B

B

B

B

B

B

B

A

B

B

A

A

Sodium peroxide

B

B

B

B

B

B

B

B

B

B

B

B

B

B

Sodium phosphate

A

A

A

A

A

A

A

A

A

A

A

A

A

A

sulfate,

A

A

B

B

B

B

C

C

C

A

C

C

A

B

Sodium sulfide, thiosulfate

C

C

C

C

C

C

C

C

C

B

C

C

C

C

Stearic acid

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Sulfur, solid

C

C

C

C

C

C

C

C

C

A

C

C

C

C

Sulfur chloride

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Sulfur dioxide, dry

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Sulfur moist

dioxide,

A

A

A

B

B

C

C

C

C

A

C

C

A

B

Sulfur trioxide, dry

A

A

A

A

A

A

A

A

A

A

A

A

A

A

78% or less

B

B

B

B

B

C

C

C

C

A

C

C

B

B

78% to 90%

C

C

C

C

C

C

C

C

C

B

C

C

C

C

90% to 95%

C

C

C

C

C

C

C

C

C

B

C

C

C

C

Fuming

C

C

C

C

C

C

C

C

C

A

C

C

C

C

Sodium silicate

Sulfuric acid

Tannic acid

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Tartaric acid

B

A

A

A

A

A

A

A

A

A

A

A

A

A

Toluene

B

B

A

A

A

B

B

B

B

B

B

B

B

A

Trichlorethylene, dry

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Trichlorethylene, moist

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Turpentine

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Varnish

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Vinegar

A

A

B

B

B

C

C

C

C

B

C

C

A

B

Water, acid mine

C

C

C

C

C

C

C

C

C

C

C

C

C

C

Water, condensate

A

A

A

A

A

A

A

A

A

A

A

A

A

A

Water, potable

A

A

A

A

A

A

B

B

B

A

A

A

A

A

Whiskey(b)

A

A

C

C

C

C

C

C

C

A

C

C

A

C

Zinc chloride

C

C

C

C

C

C

C

C

C

B

C

C

B

C

Zinc sulfate

A

A

A

A

A

C

C

C

C

B

C

A

A

C

(a) Acetylene forms an explosive compound with copper when moist or when certain impurities are present and the gas is under pressure. Alloys containing less than 65% Cu are satisfactory for this use. When gas is not under pressure, other copper alloys are satisfactory.

(b) Copper and copper alloys resist corrosion by most food products. Traces of copper may be dissolved and affect taste or color. In such cases, copper metals are often tin coated.

Whenever there is a lack of operating experience, whenever reported test conditions do not closely match the conditions for which alloy selection is being made, and whenever there is doubt as to the applicability of published data, it is always best to conduct an independent test. Field tests are the most reliable. Laboratory tests can be equally valuable, but only if operating conditions are precisely defined and then accurately simulated in the laboratory. Long-term tests are generally preferred because the reaction that dominates the initial stages of corrosion may differ significantly from the reaction that dominates later on. If short-term tests must be used as the basis for alloy selection, the test program should be supplemented with field tests so that the laboratory results can be reevaluated in light of true operating experience. Erroneous conclusions based on laboratory results can also be reached by measuring corrosion damage inaccurately, especially when corrosion is slight. It is common practice to express test results in terms of penetration or average reduction in metal thickness, even when corrosion was actually measured by weight loss. Weight loss or averagepenetration data are valid only when corrosion is uniform. When corrosion occurs predominantly by pitting or some other localized form or when corrosion is intergranular or involves the formation of a thick, adherent scale, direct measurement of the extent of corrosion provides the most reliable information. A common technique is to measure the maximum depth of penetration observed on a metallographic cross section through the region of interest. Statistical averaging of repeated measurements on one or more specimens may or may not be warranted. Despite the deficiencies in laboratory testing, information gained in this manner serves as a useful starting point for alloy selection. Operating experience may later indicate the need for a more discriminating selection. Atmospheric Exposure Comprehensive tests conducted over a 20-year period under the supervision of the American Society for Testing and Materials (ASTM), as well as many service records, have confirmed the suitability of copper and copper alloys for atmospheric exposure (Table 7). Copper and copper alloys resist corrosion by industrial, marine, and rural atmospheres except atmospheres containing NH3 or certain other agents where SCC has been observed in high-zinc alloys (>20% Zn). The copper metals most widely used in atmospheric exposure are C11000, C22000, C23000, C38500, and C75200. Alloy C11000 is an effective material for roofing, flashings, gutters, and down-spouts. Table 7 Atmospheric corrosion of selected copper alloys Alloy

Corrosion rates at indicated locations(a)

Altoona, PA

New York, NY

mils/yr m/yr

Key West, FL

mils/yr m/yr

La Jolla, CA

mils/yr m/yr

mils/yr

State PA

m/yr

College,

Phoenix, AZ

mils/yr

m/yr

mils/yr m/yr

C11000

1.40

0.055

1.38

0.054

0.56

0.022

1.27

0.050

0.43

0.017

0.13

0.005

C12000

1.32

0.052

1.22

0.048

0.51

0.020

1.42

0.056

0.36

0.014

0.08

0.003

C23000

1.88

0.074

1.88

0.074

0.56

0.022

0.33

0.013

0.46

0.018

0.10

0.004

C26000

3.05

0.120

2.41

0.095

0.20

0.008

0.15

0.006

0.46

0.018

0.10

0.004

C52100

2.24

0.088

2.54

0.100

0.71

0.028

2.31

0.091

0.33

0.013

0.13

0.005

C65500

1.65

0.065

1.73

0.068

...

...

1.38

0.054

0.51

0.020

0.15

0.006

C44200

2.13

0.084

2.51

0.099

...

...

0.33

0.013

0.53

0.021

0.10

0.004

70Cu-29Ni1Sn(b)

2.64

0.104

2.13

0.084

0.28

0.011

0.36

0.014

0.48

0.019

0.10

0.004

(a) Derived from 20-year exposure tests. Types of atmospheres: Altoona, industrial; New York City, industrial marine; Key West, tropical rural marine; La Jolla, humid marine; State College, northern rural; Phoenix, dry rural.

(b) Although obsolete, this alloy indicates the corrosion resistance expected of C71500.

The colors of different copper alloys are often important in architectural applications, and color may be the primary criterion for selecting a specific alloy. After surface preparation, such as sanding or polishing, different copper alloys vary in color from silver to yellow to gold to reddish shades. Different alloys having the same initial color may show differences in color after weathering under similar conditions. Therefore, alloys having the same or nearly the same composition are usually used together for consistency of appearance in a specific structure. Copper alloys are often specified for marine atmosphere exposures because of the attractive and protective patina they form during the exposure. In marine atmospheric exposures, this patina consists of a film of basic copper chloride or carbonate, sometimes with an inner layer of Cu2O. The severity of the corrosion attack in marine atmospheres is somewhat less than that in industrial atmospheres but greater than that in rural atmospheres. However, these rates decrease with time. Individual differences in corrosion rates do exist between alloys, but these differences are frequently less than the differences caused by environmental factors. Thus, it becomes possible to classify the corrosion behavior of copper alloys in a marine atmosphere into two general categories: those alloys that corrode at a moderate rate and include high-copper alloys, silicon bronze, and tin bronze and those alloys that corrode at a slower rate and include brass, aluminum bronze, nickel silver, and copper nickel. The average metal loss, d, of the former group can be approximated by d = 0.1 t ; the latter group can be approximated by d = 0.1 t . In both equations, t is exposure time. These relationships are shown as solid lines in Fig. 3.

Fig. 3 Typical corrosion rates of representative copper alloys in a marine atmosphere. (a) Average data for copper, silicon bronze, and phosphor bronze. (b) Average data for brass, aluminum bronze, nickel silver, and copper-nickel

Environmental factors can cause this median thickness loss to vary by as much as 50% or more in a few extreme cases. Figure 3 shows the extent of this variation as a pair of dashed lines forming an envelope around the median. Those environmental factors that tend to accelerate metal loss include high humidity, high temperatures (either ambient or due to solar radiation), proximity to the ocean, long times of wetness, and the presence of pollutants in the atmosphere. The converse of these conditions would tend to retard metal loss. Metallurgical factors can also affect metal loss. Within a given alloy family, those with a higher alloy content tend to corrode at a lower rate. Surface finish also plays a role in that a highly polished metal will corrode slower than one with a rougher surface. Finally, design details can affect corrosion behavior. For example, designs that allow the collection and stagnation of rainwater will often exhibit wastage rates in the puddle areas that are more typical of those encountered in seawater immersions.

Certain copper alloys are susceptible to various types of localized corrosion that can greatly affect their utility in a marine atmosphere. Brasses and nickel silvers containing more than 15% Zn can suffer from dealloying. The extent of this attack is greater on alloys that contain higher proportions of zinc. In addition, these same alloys are subject to SCC in the presence of small quantities of NH3 or other gaseous pollutants. Inhibited grades of these alloys are available that resist dealloying but are susceptible to SCC. Alloys containing large amounts of manganese tend to be somewhat prone to pitting in marine atmospheres, as are the cobalt-containing beryllium-coppers. A tendency toward intergranular corrosion has been observed in silicon bronzes and aluminum brass, but its occurrence is somewhat sporadic. On the whole, however, even under somewhat adverse conditions, the average thickness losses for copper alloys in a marine atmosphere tend to be very slight, typically under 50 m (Fig. 3). Thus, copper alloys can be safely specified for applications requiring long-term durability in a marine atmosphere. Design considerations for the atmospheric use of copper alloys include allowance for free drainage of structures, the possibility of staining from runoff water, and the use of smooth or polished surfaces. Soils and Groundwater Copper, zinc, lead, and iron are the metals most commonly used in underground construction. Data compiled by the National Bureau of Standards (NBS) compare the behavior of these materials in soils of the following four types: wellaerated acid soils low in soluble salts (Cecil clay loam), poorly aerated soils (Lake Charles clay), alkaline soils high in soluble salts (Docas clay), and soils high in sulfides (Rifle peat). Corrosion data as a function of time for copper, iron, lead, and zinc exposed to these four types of soil are given in Fig. 4. Copper exhibits high resistance to corrosion by these soils, which are representative of most soils found in the United States. Where local soil conditions are unusually corrosive, it may be necessary to use some means of protection, such as cathodic protection, neutralizing backfill (limestone, for example), protective coating, or wrapping.

Fig. 4 Corrosion of copper, iron, lead and zinc in four different soils

For many years, NBS has conducted studies on the corrosion of underground structures to determine the specific behavior of metals and alloys when exposed for long periods in a wide range of soils. Results indicate that tough pitch coppers,

deoxidized coppers, silicon bronzes, and low-zinc brasses behave essentially alike. Soils containing cinders with high concentrations of sulfides, chlorides, or hydrogen ions (H+) corrode these materials. In this type of contaminated soil, the corrosion rates of copper-zinc alloys containing more than about 22% Zn increase with zinc content. Corrosion generally results from dezincification. In soils that contain only sulfides, corrosion rates of the copper-zinc alloys decrease with increasing zinc content, and no dezincification occurs. Although not included in these tests, inhibited admiralty metals would offer significant resistance to dezincification. Electric cables that contain copper are often buried underground. A recent study investigated the corrosion behavior of phosphorus-deoxidized copper (C12200) in four soil types: gravel, salt marsh, swamp, and clay (Ref 8). After 3 years of exposure, uniform corrosion rates were found to vary between 1.3 and 8.8 m/yr (0.05 to 0.35 mil/yr). No pitting attack was observed. In general, the corrosion rate was highest for soils of lowest resistivity. The possibility of disposing of nuclear waste in copper containers buried deep underground is currently under investigation. Except for the mining and oil industries, underground construction is usually limited to the first few tens of meters from the surface; an underground waste disposal vault would probably be located at a depth of 500 to 1000 m (1640 to 3280 ft) in stable bedrock. At these depths, the environment differs in several respects from that nearer the surface. With increasing depth, the natural groundwaters tend to become more saline and less oxidizing. In addition, the pressures exerted by hydrostatic and lithostatic forces become greater. These aspects affect the design and corrosion behavior of any metallic structure buried at such great depths. A copper nuclear waste disposal container would be surrounded by a compacted claylike material. This serves a dual purpose: first it acts as a physical barrier, reducing the rate of transport of species to and from the container, and second, it provides some chemical buffering effects and effectively increases the pH of the environment. Both of these properties are beneficial in terms of the corrosion resistance of copper. The clay most likely to be used is a montmorillonite clay, such as sodium bentonite. In the compacted form, this clay swells when wet and would effectively seal all cracks in the surrounding rock. The low permeability of the clay ensures that there would be no mass flow of groundwater and that transport of dissolved species would occur by diffusion only. The rate of diffusion in the clay is perhaps 100 times slower than in free solution. This slow rate of diffusion applies not only to the transport of oxidants, such as dissolved oxygen (O2) or sulfide ions (S2-), to the copper surface but also to the diffusion of soluble corrosion products away from the surface. The net effect is reduction in the corrosion rate of copper compared with that in free solution. One study suggests that under such conditions uniform corrosion of oxygen-free electronic copper (C10100) would only amount to 1.1 mm (43.4 mils) in 106 years (Ref 9). Experimental results indicate that the clay may reduce the corrosion rate by about a factor of ten over that in bulk solution, although these results suggest a corrosion rate of about 1 m/yr (0.04 mils/yr) (Ref 10). Naturally occurring saline waters are also found deep underground. Although the composition and concentration of these groundwaters vary from site to site, the concentration of dissolved species generally increases with depth (Ref 11). Such groundwaters are encountered in mines, during oil drilling, and in deep boreholes. The waters have a complex composition, often being mixtures of sodium (Na+), calcium (Ca2+), magnesium (Mg2+), chloride (Cl-) sulfate (

)

and bicarbonate ( ) ions as well as trace amounts of other ions. Iron minerals in the bedrock react with dissolved oxygen in the groundwater and produce less oxidizing conditions than are found in waters nearer the surface. Additional information on the corrosion of nuclear waste containment materials is available in the section "Corrosion of Containment Materials for Radioactive Waste" of the article "Corrosion in the Nuclear Power Industry" in this Volume. The corrosion rate of copper in quiescent groundwaters tends to decrease with time. This is due to the formation of a protective film, an example of which is shown in Fig. 5. The underlying layer consists of species from the groundwater as well as copper. This layer is brittle and is extensively cracked, permitting continued dissolution of copper ions into solution. In Fig. 5, some of these copper ions have precipitated on the underlying layer in the form of cupric hydroxychloride [CuCl2·3(Cu(OH)2)] and copper oxide crystals. The corrosion layer is not truly passivating, and corrosion will continue, although at a reduced rate.

Fig. 5 Scanning electron micrograph of the corrosion product formed on C10100 in complex groundwater at 150 °C (300 °F). A, underlying film containing copper, silicon, calcium, chlorine, and magnesium; B, crystals of CuCl2-3(Cu(OH)2); C, crystals of CuO or Cu2O. Courtesy of F. King and C.D. Litke

For both copper and copper alloys, corrosion rate depends strongly on the amount of dissolved oxygen present. The data in Table 8 illustrate this point for both pure copper and Cu-10Ni in various synthetic groundwaters. These data are derived from experiments lasting from 2 to 4 weeks; therefore, they include the high initial rates of corrosion and do not represent long-term corrosion rates. However, they do not serve to show that deoxygenation of the solution results in at least an order of magnitude decrease in the short-term corrosion rate. It is also apparent from these data that, in aerated solutions at least, the addition of nickel decreases the uniform corrosion rate of copper. This is due to the formation of a more highly protective surface film. Table 8 Short-term corrosion rates of copper alloys in saline groundwaters Alloy

C10100

Copper

Type of groundwater

Synthetic 55 g/L TDS(a)

Brine A 306 g/L TDS

Seawater 35 g/L TDS

Oxygen concentration, g/g

Temperature

Corrosion rate

°C

°F

1300(a)

B

>33(a)

>1300(a)

(a) Specimen was destroyed.

Table 5 Corrosion of various nickel-base alloys in pure H3PO4 Acid was nitrogen purged before the test, which was conducted at 280 °C (535 °F). Alloy

Molybdenum content, %

Corrosion rate

108% H3PO4

112% H3PO4

mm/yr

mils/yr

mm/yr

mils/yr

G-30

5.5

9.4

372

32.5

1280

G-3

7

7.9

310

24.4

961

625

9

4.6

180

14.9

590

C-22

13

2.6

102

3.9

152

C-276

16

1.4

54

1.8

72

B-2

28

0.23

9

0.05

2

Hydrofluoric acid is commercially available in concentrations ranging from 30 to 70% and as anhydrous HF. Plain carbon steel can withstand the anhydrous acid and is extensively used and rail cars carrying anhydrous HF. Of the nickelbase alloys, alloy 400 has been the most extensively examined (Ref 37). The alloy possesses good corrosion resistance in all concentrations of HF up to a temperature of about 120 °C (250 °F). However, the presence of oxygen in the solution is detrimental (Table 6).

Table 6 Effect of oxygen on corrosion of alloy 400 in HF solutions The acids were purged with nitrogen containing various amounts of oxygen. Concentration of oxygen in purge gas, ppm

Corrosion rate

Boiling (112 °C, or 234 °F) 38% HF

Boiling (108 °C, or 226 °F) 48% HF

Liquid phase

Vapor phase

Liquid phase

Vapor phase

mm/yr

mils/yr

mm/yr

mils/yr

mm/yr

mils/yr

mm/yr

mils/yr

0.003

0.003

...

6

(a) Iron tolerance equals manganese content of alloy times 0.032.

(b) Iron tolerance equals manganese content of alloy times 0.021.

(c) Magnesium content of AZ63 reported as 0.2%

Fig. 2 Effect of nickel and copper contamination on the salt spray corrosion performance of die-cast AZ91 alloy. Source: Ref 3

It should also be noted that the nickel tolerance depends strongly on the cast form, which influences grain size, with the low-pressure cast alloys showing just a 10-ppm tolerance for nickel in the as-cast (F) temper. Therefore, alloys intended for low-pressure cast applications should be of the lowest possible nickel level (Ref 4). The low tolerance limits for the contaminants in AM60 alloy when compared to AZ91 alloy can be related to the absence of zinc. Zinc is thought to improve the tolerance of magnesium-aluminum alloys for all three contaminants, but it is limited to 1 to 3% because of its detrimental effects on microshrinkage porosity and its accelerating effect on corrosion. For the rare-earth, thorium, and zinc alloys containing zirconium, the normal saltwater corrosion resistance is only moderately reduced when compared to high-purity magnesium-aluminum alloys--0.5 to 0.76 mm/yr (20 to 30 mils/yr) as opposed to less than 0.25 mm/yr (10 mils/yr) in 5% salt spray--but contaminants must again be controlled. This zirconium alloying element is effective in this case because it serves as a strong grain refiner for magnesium alloys and it precipitates the iron contaminant from the alloys before casting. However, if alloys containing more than 0.5 to 0.7% Ag or more than 2.7 to 3% Zn are used, a sacrifice in corrosion resistance should beexpected (Fig. 1). Nevertheless, when properly finished, these alloys provide excellent service in harsh environments. Heat-Treating, Grain Size, and Cold-Work Effects. Using controlled-purity AZ91 alloy cast in both high- and

low-pressure forms, the contaminant tolerance limits have been defined as summarized in Table 4 for the as-cast (F), the solution-treated (T4, held 16 h at 410 °C, or 775 °F and quenched), and the solution-treated and aged (T6, held 16 h at 410 °C, or 775 °F, quenched, and aged 4 h at 215 °C, or 420 °F) (Ref 4, 5).

Table 4 Contaminant tolerance limits versus temper and cast form for AZ91 alloy High-pressure die cast, 5-10 m average grain size; low-pressure cast, 100-200 m average grain size Contaminant, %

Critical contaminant limit(a)

High pressure F

Low pressure

F

T4

T6

Iron

0.032 Mn

0.032 Mn

0.035 Mn

0.046 Mn

Nickel

0.0050

0.0010

0.001

0.001

Copper

0.040

0.040

450 °C (840 °F). The rate of oxidation accelerates with time up to 450 °C (840 °F), is constant at T = 450 °C (840 °F), and decreases with time at temperatures above 450 °C (840 °F). The data in Ref 2, along with those of several other investigators, are summarized in Fig. 1. The deviation in the curve from linearity at 450 °C (840 °F) was attributed to a change in the character of the oxide from a highly porous and nonprotective layer to a more compact scale, although all of these oxides were identified as nominally UO2.

Fig. 1 Oxidation rates of uranium in air and in oxygen

Water Vapor. Uranium reacts with water vapor according to:

U + (2 + x)H2O UO2+x + (2 + x)H2

(Eq 6)

where x is between 0 and 0.1 in the absence of oxygen. In the most recent investigation of the uranium-water vapor reaction, Equation 6 was found to the linear between 30 and 80 °C (85 and 175 °F) (Ref 9). The reaction rate is proportional to the square root of the humidity and obeys Eq 7:

(Eq 7)

where r is the fractional relative humidity. These data and Eq 7 agree very well with the results in Ref 10 and 11 (Fig. 2). Both investigators found a water vapor pressure dependency of 1:2.1. The crystal structure of the UO2 is fcc. The reaction product forms loose flakes that readily spall from the base metal. The quantity of hydrogen produced is typically from 1 to 12% less than the 2 mol predicted for every mole of oxide produced.

Fig. 2 Oxidation rate of uranium at 100 °C (212 °F) versus water vapor pressure

Water Vapor-Oxygen Mixtures. Until the work done in the early 1980s, which is discussed in Ref 11, it was thought that uranium in a water vapor-oxygen mixture reacted with water vapor, forming UO2 and releasing hydrogen that combined with the free oxygen to form more water vapor. This work, combining both thermogravimetric and gas spectrometric techniques, showed that uranium reacts with oxygen directly in a water vapor-oxygen mixture. The water vapor affects the uranium oxide structure formed, producing a more defective, or nonstoichiometric, oxide than that formed in the absence of water vapor. Thus, uranium reacts with oxygen in the water vapor-oxygen mixture according to:

4U + 9/2 O2

U4O9

(Eq 8)

The most recent and extensive investigation of th uranium-oxygen-water vapor system measured the kinetics between 40 and 100 °C (105 and 212 °F), and between 11 and 75% relative humidity (Ref 12). The reaction rate is given by:

(Eq 9)

Equation 9 is linear and not proportional to either the water vapor pressure (relative humidity) or the oxygen pressure. However, at oxygen concentrations from 100 to 10 ppm, the reaction reverts to the uranium-water vapor reaction shown in Fig. 3. The reaction product, identified by x-ray diffraction analysis (Ref 11) as U4O9 or UO2.25, is fcc and is compact and adherent. No hydrogen is produced.

Fig. 3 Oxidation rate of uranium versus oxygen pressure in water vapor at 100 °C (212 °F). Source: Ref 10

Water. Although numerous papers and review articles have been written concerning the reactivity of uranium in water,

that is, the corrosion of uranium, most of this work is decades old. The consensus from these investigations is that the rate of corrosion of uranium in water at 25 °C (75 °F) equals the rate of oxidation of uranium in water vapor at 100 °C (212 °F) as the relative humidity approaches 100%, that rate being 0.57 mg/cm2/h (Ref 13). The reaction product has been identified as UO3·0.8H2O (Ref 13). Aqueous Solutions. The corrosion properties of uranium in aqueous solutions are highly dependent on the pH of the medium, owing to the participation of hydrogen ion (H+) in the redox reactions involved in corrosion. General corrosion in caustic aqueous solutions in the presence of oxygen normally proceeds by coupling the anodic dissolution reaction to the oxygen to hydroxl ion (OH-) cathodic reaction. The H+ ion to hydrogen gas reaction does not seem to become significant until a low pH is reached. Factors in the specific corrosion mechanism that lead to the accumulation of hydrogen in the region of the surface film and film/metal interface cause increased general corrosion rates.

Effect of Alloying Elements on Oxidation Response Most of the uranium alloys under study and in use have come from binary and higher combinations of the -miscible elements titanium, molybdenum, niobium, and zirconium. Addition of these elements stabilizes the phase and increases corrosion resistance. Titanium. The most recent and extensive study involving the oxidation of a uranium-titanium alloy investigated the

oxidation of a U-0.75Ti alloy in environments containing oxygen and/or water vapor at 140 and 100 °C (285 and 212 °F) (Ref 14). The reaction rate of U-0.75Ti with water vapor at 140 °C (285 °F) was found to be linear and proportional to the water vapor pressure (K ) for water vapor pressures between 267 and 2670 Pa (2 and 20 torr) (Fig. 4). Hydrogen was produced by the reaction at a rate of approximately 2 mol for every 1 mol of UO2 formed. A pressure dependency of one-half suggests that a dissociative/adsorption equilibrium exists at the reaction interface.

Fig. 4 Oxidation rate of U-0.75Ti versus water vapor pressure. Source Ref 14

In contrast, the behavior exhibited by U-0.75Ti when tested at 100 °C (212 °F) in water vapor showed no pressure dependency on water vapor after approximately 50 h of exposure (Fig. 4). The rate measured was 1.4 × 10-3 mg/cm2/h. This occurrence was believed to be due to the reaction rate being controlled by a solid-state diffusion process rather than a dissociative adsorption process. The reaction rate of U-0.75Ti with oxygen was also linear and independent of the oxygen pressure for oxygen pressures between 67 Pa and 133 kPa (0.5 and 1000 torr) at both temperatures. The reaction rates at 140 and 100 °C (285 and 212 °F) were approximately 3.2 × 10-3 and 1.2 × 10-4 mg/cm2/h, respectively. The reaction rate was thought to be controlled by the reaction of the metal with oxygen ions (O2-) at the metal/oxide interface. The reaction rate of U-0.75Ti in environments containing both water vapor and oxygen was found to be linear and not dependent on either water vapor or oxygen pressure for the ranges investigated at both temperatures. The oxidation rates were 3.7 × 10-3 and 1.5 × 10-4 mg/cm2/h at 140 and 100 °C (285 and 212 °F), respectively. Thus, the addition of water vapor to a pure oxygen environment had only a small effect on the oxidation kinetics. Also, the mechanism for oxidation in the mixed environment was thought to be the same as that in oxygen, that is, the reaction of metal with O2- ions at the metal/oxide interface. The reaction products were identified as oxygen-rich variants of the form UO2 + x, where x is greater than 0 but less than 0.25. X-ray diffraction analyses of samples from each of the three environments showed that the crystal structure of the oxides was fcc. The oxidation rates for this alloy were one to two orders of magnitude lower than the rates for uranium in equivalent environments. Otherwise, the alloy behaved similarly to unalloyed uranium in all three environments. In general, the oxidation resistance of uranium increases with increasing additions of titanium up to an addition level of 2 wt% (Ref 15). Molybdenum. The published literature on the atmospheric corrosion of uranium-molybdenum alloys is very limited.

The available literature on the corrosion of all uranium-molybdenum alloys in both moist air and moist nitrogen has been reviewed. The only conclusions that can be made are that these alloys have a corrosion resistance comparable to uraniumtitanium alloys of the same atomic percentage and that the corrosion resistance increases with molybdenum content.

Niobium. Although a number of investigations of the oxidation of uranium-niobium binary alloys have been reported in

recent years, the two studies that included the most alloys and controlled conditions are discussed in Ref 16 and 17. In the first study, the alloys were tested in water-saturated nitrogen or oxygen at 75 °C (165 °F), and in the second, specimens were exposed to pure water, water vapor plus oxygen, or water vapor plus air at 125 °C (255 °F). In both experiments, the alloys showed considerably reduced reaction rates compared to unalloyed uranium, with the resistance of the alloy to reaction with the moisture increasing with niobium content, and the presence of oxygen in the test environment reduced the rate of hydrogen evolution. In all cases, more hydrogen was absorbed by the alloy than was released into the atmosphere. The amount of hydrogen released into the atmosphere decreased significantly with increasing niobium content, but all of the alloys absorbed roughly the same amount of hydrogen. The most recent investigation of a uranium-niobium alloy involved the atmospheric corrosion of the U-6Nb alloy (Ref 18). The temperature range investigated was between 50 and 110 °C (120 and 230 °F), and the corresponding relative humidity ranged from 75 to 17%. The oxidation rate of U-6Nb in water vapor was so slow that it was not possible to derive an equation for the corrosion rate, even at 110 °C (230 °F). In moist air, the reaction was slightly faster than had been measured in moist nitrogen, but was still about 1/500 of the rate for unalloyed uranium. Niobium-Zirconium. The uranium-niobium-zirconium alloy of the most interest is a U-7.5Nb-2.5Zr alloy designated

as mulberry, which was given the label of a "stainless" uranium alloy. In humid environments, mulberry reacts similarly to the binary U-6Nb alloy. Oxidation data for mulberry in humid environments are provided in Ref 19.

Corrosion Behavior of Uranium Alloys The factors that control metallic corrosion can be categorized as thermodynamic or kinetic. The electrochemical principles of thermodynamics are those of reversible cells and standard potential; kinetics is governed by polarization techniques. Humid Air. Figure 5 illustrates the relationship between the percent of alloying additions and corrosion response in hot,

humid air. The data show the rate of hydrogen generation of selected uranium alloys exposed to 100% relative humidity in oxygen at 75 °C (165 °F). This relationship is shown as approximately linear. Thus, increasing the total amount of alloying additions, irrespective of the alloying element, increases the resistance to corrosion.

Fig. 5 Rate of hydrogen generation versus alloy addition of selected uranium alloys.

quad, U-0.5Nb-0.5Mo-

0.5Zr-0.5Ti; quad, U-0.75Nb-0.75Mo-0.75Zr-0.5Ti; 1 quad, U-1Nb-1Mo-1Zr-0.5Ti; 1 quint, U-1Nb-1Mo-1Zr0.5Ti-0.5V. Source: Ref 16

Water. In addition to work at Argonne National Laboratory, Battelle Columbus Laboratories, and the Army Materials

and Mechanics Research Center (AMMRC), a series of Russian articles (Ref 20) discusses the relative resistance of numerous uranium alloys to boiling water at 100 °C (212 °F). It has been shown that the corrosion resistance of uranium alloys to boiling water is inversely proportional to the percentage of -uranium present in the particular compositions. An increase in the corrosion resistance with alloying is due to the formation of the phase. The most promising alloying additives--molybdenum, niobium, and zirconium--have been extensively studied. Dilute Salt Solutions. Because the corrosion resistance of uranium alloys in hot, moist air and in boiling water has

been found to be proportional to the total alloying content, it is expected that this proportionality also holds true for corrosion in dilute salt solutions. Electrochemical measurements have indicated that this should be the case. In one study, standard or corrosion potentials of various uranium alloys were measured in a 0.001 M potassium chloride (KCl) solution against a standard calomel electrode (Ref 21). Figure 6 shows these measured potentials versus the total alloying content of the alloys. Although the relationship is not strictly linear between potential and alloy content, it is directly related. Kinetic data on the corrosion rates of uranium alloys in dilute salt solutions are scarce, but those that have been published confirm this relationship.

Fig. 6 Rest potential versus total alloy content for uranium alloys in 0.001 M KCl at 25 °C (75 °F). 0.5Nb-0.5Mo-0.5Zr-0.5Ti; 1Nb-1Mo-1Zr-0.5Ti-0.5V

quad, U-

quad, U-0.75Nb-0.75Mo-0.75Zr-0.5Ti; 1 quad, U-1Nb-1Mo-1Zr-0.5Ti; 1 quint, U-

Ocean Water. As with the preceding corrosion environments, the corrosion susceptibility of uranium alloys in ocean

water is expected to decrease with increasing alloy content. In one investigation, the corrosion potentials of a number of uranium alloys were measured in artificial seawater (Fig. 7). The observed trend of decreasing corrosion potential with increasing alloying content is very similar to that discussed previously for the dilute salt solution test. Thermodynamics again predicts a decreasing corrosion rate with increased alloy content. The kinetics of the uranium alloy-ocean water reaction were measured gravimetrically, the results showed a logarithmic relationship between the corrosion rate and the total alloy content (Fig. 8).

Fig. 7 Rest potential versus total alloy content for uranium alloys in ocean water at 20 °C (70 °F). Source: Ref 22

Fig. 8 Corrosion rate versus total alloy content for uranium alloys in ocean water at 20 °C (70 °F). Source: Ref 23

Acids and Bases. An indication of the possible anodic reactions can be obtained by examining the Pourbaix diagram for uranium shown in Fig. 9. In the potential range from -1.8 to 1.2 V at low pHs (0 to 2.0), uranium forms primarily

soluble species. The uranium ion U3+ forms in the active region near the corrosion potential, and the uranyl ion ( )forms in the transpassive region. In the passive region, UO2 undoubtedly forms. Anodic polarization techniques can be used to study the ease of transition from the active to the passive state as well as the dissolution behavior of the metal and its alloys. The transition from the active to the passive state is accompanied by a decrease in corrosion rate of the order of 104 to 106, which is extremely significant for many applications.

Fig. 9 Pourbaix diagram for uranium

Anodic polarization techniques have been used to study the effects of alloying constituent, temperature, solution chemistry, solution concentration, pH, and presence of chloride on the corrosion response of uranium alloys (Ref 24). An example of the effect of alloying on anodic polarization is shown in Fig. 10; the passive current densities vary inversely with alloy content. Figure 11 shows an example of the effect of solution chemistry , particularly the addition of chloride, on the anodic polarization behavior. The uranium-molybdenum alloy passivates more easily in sodium sulfate (Na2SO4) than sulfuric acid (H2SO4), and the addition of chloride prevents passivation entirely.

Fig. 10 Effect of alloying on the anodic polarization of uranium in H2SO4 at 25 °C (75 °F)

Fig. 11 Anodic polarization curves for uranium in differing aqueous solutions at 25 °C (75 °F)

Additional conclusions have been reached based on this work, as follows. Uranium binary alloys exhibit active-passive behavior in sodium hydroxide (NaOH), ammonium hydroxide (NH4OH), sodium nitrate (NaNO3), sodium chromate (Na2CrO4), and ammonium chromate ([NH4]2CrO4). The critical current densities for passivity were inversely proportional to the H2SO4 concentration. Unlike those of most metals, the dissolution rates of uranium alloys decrease with increasing acid concentration. Chloride additions as small as 0.005 M affect the anodic polarization curve, but chromates, sulfates, and nitrates inhibit pitting at this low concentration of chloride. The uranium-titanium alloys were found to be more resistant to basic solutions than the uranium-molybdenum alloys.

Effect of Microstructure on Corrosion Properties The influence of microstructure on the corrosion properties of uranium alloys is very much dependent on the particular uranium alloy. For example, the corrosion response of the U-0.75Ti alloy changes very little with process variables, but the U-6Nb alloy changes significantly. U-0.75Ti. Research conducted for the United States Air Force measured the corrosion response of U-0.75Ti penetrators manufactured by various processes. The processing parameters included those in extrusion (temperature and rate), solutionizing (temperature, time, and media), aging (temperature, time, and atmosphere), and quenching (rate and media). In addition, U-0.75Ti penetrators made by such forming techniques as casting, swaging, forging, or grinding were tested. In all, more than 50 different combinations of processing techniques were investigated. The test environments chosen were a salt fog (to satisfy a military specification) and hot, moist nitrogen. The results showed that the largest difference in corrosion response to the salt fog environment between any two lots of material was a factor of two (Ref 25, 26). For hot, moist nitrogen testing, the difference in final weight loss between the most and least corroded material was a factor of ten. In that the weight loss in moist nitrogen for even the most susceptible material was very small, these differences were indeed small. Thus, the metal processing of the U-0.75Ti alloy did not significantly change the corrosion response. U-6Nb. An investigation of the effect of cooling rate on the microstructure, mechanical behavior, corrosion resistance,

and subsequent age hardenability of U-6Nb revealed that cooling rates exceeding 20 °C/s (36 °F/s) cause the parent phase to transform martensitically to a niobium-supersaturated variant of the phase (Ref 27). This martensitic phase

exhibits low hardness and strength, high ductility, good corrosion resistance, and substantial age hardenability. As cooling rates decrease from 10 °C/s (18 °F/s) to 0.2 °C/s (0.4 °F/s), fine-scale microstructural changes (consistent with spinodal decomposition) occur to an increasing extent. These changes were found to produce large increases in hardness and strength as well as large decreases in ductility, slight decreases in corrosion resistance, and slight changes in age hardenability. At cooling rates less than 0.2 °C/s (0.4 °F/s), the parent phase undergoes cellular decomposition to a course two-phase lamellar microstructure. This lamellar microstructure was seen to exhibit intermediate strength and ductility, substantially reduced corrosion resistance, and no age hardenability. Figure 12 shows the corrosion behavior of U-6Nb as influenced by cooling rate. Very rapid cooling results in a low rest potential, as would be expected for a corrosion-resistant material. At low cooling rates, the rest potential is substantially higher than that for any of the intermediate or high cooling rates, which suggests that a substantial decrease in corrosion resistance occurs at low cooling rates. Corrosion tests done at 75 °C (165 °F) in 95% relative humidity air or nitrogen confirmed the interferences drawn from the rest potential measurements. Samples cooled at intermediate and high rates exhibited no measurable weight change or visual tarnishing, but the samples cooled at low rates formed a black surface film and had an average weight gain of 0.01 mg/cm2/d (equivalent to a penetration rate for UO2 of 17 m/yr, or (0.67 mils/yr). More information on the microstructures and metallography or uranium and uranium alloys is available in the article "Uranium and Uranium Alloys" in Metallography and Microstructures, Volume 9 of ASM Handbook, formerly 9th Edition Metals Handbook.

Fig. 12 Effects of cooling rate on rest potential of U-6Nb in 0.001 M KCl. The error bar indicates the uncertainty in the potential measurement. Source: Ref 27

Galvanic Behavior Some uses of uranium-base alloys require that they be in contact with other metals. When two dissimilar metals are electrically connected and immersed in a suitable electrolyte, a galvanic current flows between them. The driving force for the current flow is the potential difference between the two dissimilar metals. An electromotive force (emf) series is a list of metals and alloys arranged in order of the potentials generated when electrodes of each material are compared with one another in a specific environment. An emf series of a chloride-containing aqueous solution is often used for purposes of comparison for many applications. Uranium-Niobium Alloys. The galvanic-corrosion behavior of uranium and uranium alloys, particularly uraniumniobium alloys, was investigated in ocean water and hydrochloric acid (HCl), and the emf series of selected materials generated in ocean water and in 0.1 N HCl are given in Tables 1 and 2. In addition, galvanic couples were immersed in 0.1 N HCl for 4 to 96 h, and the corrosion rates were determined for each couple. The emf series, based on electrode potentials in 0.1 N HCl at 25 and 70 °C (75 and 160 °F) and in oxygen-saturated ocean water, was found to be essentially the same as the series based on the gravimetric galvanic-corrosion tests. Therefore, the emf series based only on electrode potentials is useful for indicating potential galvanic-corrosion effects.

Table 1 Electromotive force series of selected materials in 0.1 N HCl Alloy

Electrode potentials, mV

Oxygensaturated, 25 °C (75 °F)

Air-equilibrated

25 °C (75 °F)

70 °C (160 °F)

Aluminum alloy 7178

-740

-760

-805

Uranium

-740

-755

-790

U-4.5Nb

-465

-475

-600

4340 steel

-445

-460

-470

U-6Nb

-395

-420

-465

U-8Nb

-375

-400

-445

Ti-6Al-4V

-310

-385

-375

Mulberry

-305

-340

-410

U-10Mo

-170

-190

-240

Type 304 stainless steel

230

75

-230

Source: Ref 22

Table 2 Electromotive force series of selected materials in ocean water at 25 °C (75 °F) Ocean water prepared according to ASTM D 1141 Alloy(a)

Electrode potential, mV

Oxygen-saturated

Air-equilibrated

Aluminum alloy 7178

-800

-800

Tuballoy (depleted uranium)

-770

-795

U-4.5Nb

-525

-530

4340 steel

-480

-540

U-6Nb

-470

-460

U-8Nb

-430

-415

Ti-6Al-4V

-390

-350

Mulberry

-370

...

U-10Mo

-300

...

Type 304 stainless steel(b)

-225

-250

Source: Ref 22 (a) Alloys are listed in order of increasing mobility from top to bottom as determined by electrode potentials in oxygen-saturated ocean water.

(b) Passive.

Polynary Uranium Alloys. The galvanic-corrosion behavior of polynary uranium alloys, particularly quaternary alloys, was studied in humid air and 4% salt spray; the results are summarized in Table 3. Galvanic corrosion became significant in water-saturated and condensing conditions, the different quad alloys behaved similarly, and aluminum alloy 6061-T6 and AISI type 304 stainless steel showed the best compatibility with the uranium alloys in resistance to humid atmospheres.

Table 3 Galvanic corrosion of uranium alloy couples Couple

Test duration, days

Environment

Surface observation

Appearance of interface at 100 ×

Uranium

Other material

U-593/stainless steel(a)

90

4% NaCl

Heavy oxidation of U

Slight pitting

Normal

U-581/aluminum(b)

90

4% NaCl

Heavy oxidation of U

Corrosive attack

Slight pitting

U-581/stainless steel(b)

180

30% relative humidity

No apparent change

Normal

Normal

U-583/aluminum(c)

180

30% relative humidity

No apparent change

Normal

Normal

U-584/1042 steel(d)

76

30% relative humidity

No apparent change

Normal

Slight pitting

U-584/4340 steel(d)

76

30% relative humidity

No apparent change

Slight pitting

Slight pitting

U-584/1042 steel(d)

76

100% humidity

relative

Oxidation

Normal

Corrosive attack

U-584/4340 steel(d)

76

100% humidity

relative

Oxidation

Corrosive attack

Slight attack

corrosive

U-584/1042 steel(d)

76

4% NaCl

Corrosion oxidation

Deep pitting

Slight attack

corrosive

U-584/4340 steel(d)

76

4% NaCl

Corrosion and oxidation

Deep pitting

Slight attack

corrosive

U-581/stainless steel(b)

180

100% humidity

relative

Corrosion and oxidation

Slight attack

corrosive

Slight attack

corrosive

U-581/aluminum(b)

180

100% humidity

relative

Corrosion and oxidation

Slight attack

corrosive

Slight attack

corrosive

and

heavy

Uranium alloy compositions: Source: Ref 28 (a) U-0.99Mo-1.02Nb-1.27Zr-0.49Ti;

(b) U-1.40Mo-1.50Nb-1.47Zr-0.52Ti;

(c) U-1.80Mo-2.01Nb-1.90Zr-0.54Ti;

(d) U-1.03Mo-1.04Nb-0.98Zr-0.62Ti.

Galvanic Behavior Some uses of uranium-base alloys require that they be in contact with other metals. When two dissimilar metals are electrically connected and immersed in a suitable electrolyte, a galvanic current flows between them. The driving force for the current flow is the potential difference between the two dissimilar metals. An electromotive force (emf) series is a list of metals and alloys arranged in order of the potentials generated when electrodes of each material are compared with one another in a specific environment. An emf series of a chloride-containing aqueous solution is often used for purposes of comparison for many applications. Uranium-Niobium Alloys. The galvanic-corrosion behavior of uranium and uranium alloys, particularly uraniumniobium alloys, was investigated in ocean water and hydrochloric acid (HCl), and the emf series of selected materials generated in ocean water and in 0.1 N HCl are given in Tables 1 and 2. In addition, galvanic couples were immersed in 0.1 N HCl for 4 to 96 h, and the corrosion rates were determined for each couple. The emf series, based on electrode potentials in 0.1 N HCl at 25 and 70 °C (75 and 160 °F) and in oxygen-saturated ocean water, was found to be essentially

the same as the series based on the gravimetric galvanic-corrosion tests. Therefore, the emf series based only on electrode potentials is useful for indicating potential galvanic-corrosion effects. Table 1 Electromotive force series of selected materials in 0.1 N HCl Alloy

Electrode potentials, mV

Oxygensaturated, 25 °C (75 °F)

Air-equilibrated

25 °C (75 °F)

70 °C (160 °F)

Aluminum alloy 7178

-740

-760

-805

Uranium

-740

-755

-790

U-4.5Nb

-465

-475

-600

4340 steel

-445

-460

-470

U-6Nb

-395

-420

-465

U-8Nb

-375

-400

-445

Ti-6Al-4V

-310

-385

-375

Mulberry

-305

-340

-410

U-10Mo

-170

-190

-240

Type 304 stainless steel

230

75

-230

Source: Ref 22

Table 2 Electromotive force series of selected materials in ocean water at 25 °C (75 °F) Ocean water prepared according to ASTM D 1141 Alloy(a)

Electrode potential, mV

Oxygen-saturated

Air-equilibrated

Aluminum alloy 7178

-800

-800

Tuballoy (depleted uranium)

-770

-795

U-4.5Nb

-525

-530

4340 steel

-480

-540

U-6Nb

-470

-460

U-8Nb

-430

-415

Ti-6Al-4V

-390

-350

Mulberry

-370

...

U-10Mo

-300

...

Type 304 stainless steel(b)

-225

-250

Source: Ref 22 (a) Alloys are listed in order of increasing mobility from top to bottom as determined by electrode potentials in oxygen-saturated ocean water.

(b) Passive.

Polynary Uranium Alloys. The galvanic-corrosion behavior of polynary uranium alloys, particularly quaternary alloys, was studied in humid air and 4% salt spray; the results are summarized in Table 3. Galvanic corrosion became significant in water-saturated and condensing conditions, the different quad alloys behaved similarly, and aluminum alloy 6061-T6 and AISI type 304 stainless steel showed the best compatibility with the uranium alloys in resistance to humid atmospheres.

Table 3 Galvanic corrosion of uranium alloy couples Couple

Test duration, days

Environment

Surface observation

Appearance of interface at 100 ×

Uranium

Other material

U-593/stainless steel(a)

90

4% NaCl

Heavy oxidation of U

Slight pitting

Normal

U-581/aluminum(b)

90

4% NaCl

Heavy oxidation of U

Corrosive attack

Slight pitting

U-581/stainless steel(b)

180

30% relative humidity

No apparent change

Normal

Normal

U-583/aluminum(c)

180

30% relative humidity

No apparent change

Normal

Normal

U-584/1042 steel(d)

76

30% relative humidity

No apparent change

Normal

Slight pitting

U-584/4340 steel(d)

76

30% relative humidity

No apparent change

Slight pitting

Slight pitting

U-584/1042 steel(d)

76

100% humidity

relative

Oxidation

Normal

Corrosive attack

U-584/4340 steel(d)

76

100% humidity

relative

Oxidation

Corrosive attack

Slight attack

corrosive

U-584/1042 steel(d)

76

4% NaCl

Corrosion oxidation

Deep pitting

Slight attack

corrosive

U-584/4340 steel(d)

76

4% NaCl

Corrosion and oxidation

Deep pitting

Slight attack

corrosive

U-581/stainless steel(b)

180

100% humidity

relative

Corrosion and oxidation

Slight attack

corrosive

Slight attack

corrosive

U-581/aluminum(b)

180

100% humidity

relative

Corrosion and oxidation

Slight attack

corrosive

Slight attack

corrosive

Uranium alloy compositions: Source: Ref 28 (a) U-0.99Mo-1.02Nb-1.27Zr-0.49Ti;

and

heavy

(b) U-1.40Mo-1.50Nb-1.47Zr-0.52Ti;

(c) U-1.80Mo-2.01Nb-1.90Zr-0.54Ti;

(d) U-1.03Mo-1.04Nb-0.98Zr-0.62Ti.

Stress-Corrosion Cracking of Uranium Alloys One of the problems that has been encountered in the use of uranium alloys is SCC. Stress-corrosion cracking is defined as a cracking process that requires the simultaneous action of a corrosive environment and a sustained stress. In general, as the corrosion resistance and strength of an alloy increase, its susceptibility to SCC also increases. The conventional test methods use time to failure measurements on smooth specimens, incorporating both initiation and propagation processes, or precracked specimens in a fracture mechanics test, which tends to eliminate the initiation stages. Uranium-Titanium Alloys. The U-0.75Ti alloy has been the only composition in the U-Ti system that has received a significant amount of stress-corrosion testing. Crack propagation tests using precracked specimens conducted in oxygen, hydrogen, water vapor, water, dry air, wet air, and chloride solutions have shown that water is the species responsible for cracking (Table 4). Oxygen apparently inhibits crack propagation in this alloy. The effect of the strength level on the susceptibility of U-0.75Ti to SCC has also been investigated (Ref 30). Figure 13 shows the relationship between ultimate tensile strength and KIscc, the threshold stress intensity for SCC. The data show that KIscc decreased linearly by a factor of two in the strength range investigated. All of the tests conducted in this alloy system found that transgranular cracking was the SCC mode, and there was no evidence of cracking along prior -phase grain boundaries.

Table 4 SCC thresholds for U-0.75Ti in air and in aqueous environments Heat treatment

Yield strength

MPa

ksi

380 °C (715 °F), 6 h

986

143

380 °C (715 °F), 6 h

986

380 °C (715 °F), 6 h

Environment

KIscc

MPa

ksi

Dry air

42

38

143

100% relative humidity air

27.5

25

986

143

50 ppm Cl-

17.6

16

380 °C (715 °F), 6 h

986

143

3.5% NaCl

17.6

16

Air cooled

607

88

Water

23

21

Air cooled

607

88

3.5% NaCl

16.5

15

450 °C (840 °F)

...

...

Air

28.6

26

450 °C (840 °F)

...

...

Water

15.4

14

450 °C (840 °F)

...

...

3.5% NaCl

11

10

Source: Ref 29

Fig. 13 Plane-strain threshold for SCC propagation versus ultimate tensile strength for U-0.75Ti in 50 ppm Clsolution at 25 °C (75 °F). Source: Ref 30

Uranium-Molybdenum Alloys. Uranium alloys with molybdenum concentrations from 0.6 to 12% have been found to be susceptible to SCC (Ref 29). Below 5% Mo, a series of metastable -phase alloys is formed that has been found to be susceptible to cracking. Tests done in both wet and dry conditions on -phase alloys containing more than 5% Mo have shown that, contrary to the behavior observed with uranium-titanium alloys, oxygen is the species primarily responsible for SCC. Carbon content is important to the cracking behavior, alloys with higher carbon contents being more susceptible. The heat treating of quenched alloys can lead to significant improvements in SCC resistance in that alloys with equilibrium microstructures are much more resistant to cracking than metastable alloys. The predominant fracture mode observed in uranium-molybdenum alloys has been transgranular, particularly in nonaqueous environments. An intergranular mode has been observed in precracked tests conducted in aqueous environments (Ref 31). Uranium-Niobium Alloys. The principal interest in the U-Nb system has centered on alloys with 2.3, 4.5, 6, and 8%

Nb by weight, respectively. The SCC response is very much dependent on the particular alloy. The 2.3 and 4.5% Nb alloys are subject to SCC in water vapor, but the 6% Nb alloy is not. However, 6 and 8% Nb alloys will crack in environments containing oxygen, and water vapor will accelerate the SCC in these environments (Ref 32). The U-4.5Nb alloy was shown to crack in either water vapor or oxygen and was more susceptible when both were present (Ref 32). Figure 14 shows that for the U-4.5Nb alloy the crack velocity decreases and KIscc increases as the oxygen pressure is decreased (Ref 32). The chloride ion (Cl-) was also found to be very deleterious to the U-4.5Nb alloy in aqueous solutions (Ref 33). Specimens loaded at 33MPa (30 ksi ) failed in 270 min in distilled water and in as little as 1 min in chloride solutions. However, another study revealed that the U-6Nb alloy was not susceptible to SCC initiation in chloride solutions when aged at temperatures below 200 °C (390 °F), but material aged between 250 and 400 °C (480 and 750 °F) was susceptible (Ref 34). When overaged (aged at temperatures above 600 °C, or 1110 °F), the alloy exhibited an increased resistance to cracking. In general, the data have shown for all of the uranium-niobium alloys that overaged specimens are more resistant to SCC than underaged specimens.

Fig. 14 Crack velocity versus SCC threshold as a function of oxygen pressure for U-4.5Nb. Source: Ref 32

Both transgranular and intergranular cracking was observed in the U-Nb system, with the mode observed being dependent on the alloy composition, the environment, and the type of specimen. In chloride solutions, smooth-specimen tests showed that the U-2.3Nb alloys cracked transgranularly and that the U-6Nb and U-8Nb alloys cracked intergranularly. The U-4.5Nb alloy exhibited intergranular cracking in smooth specimens and transgranular cracking in precracked specimens. Uranium-Niobium-Zirconium Alloys. The only uranium-niobium-zirconium alloy that has been studied to any significant extent is mulberry. The SCC behavior of mulberry is similar to that of U-6Nb. The SCC behavior of this alloy was extensively studied in the 1970s and was later reviewed in Ref 29 and 32. Two types of SCC are observed in the alloy, each with its own fracture mode. The intergranular cracking initiates easily, but requires oxygen, water, and chloride to do so. The transgranular SCC does not initiate easily, but can propagate in pure oxygen or in solutions containing strong oxidizers.

An investigation of the effect of heat treatment on the SCC susceptibility of mulberry showed that the higher the aging temperature the greater the resistance to SCC initiation (Table 5). It was also found that crack propagation was fastest for material aged at intermediate temperature (350 to 450 °C, or 660 to 840 °F) and was slowest in the material aged at 150 °C (300 °F), the standard aging temperature for this alloy.

Table 5 Threshold stress for SCC initiation of mulberry in dilute chloride solutions as a function of heat treatment Heat treatment

KIscc

MPa

ksi

150 °C (300 °F), 1 h

65

Source: Ref 35 Polynary Uranium Alloys. The SCC behavior of a series of polynary uranium alloys containing small quantities of all

the potent stabilizers (titanium, niobium, molybdenum, and zirconium) has been investigated, primarily by AMMRC (Ref 36). All of the SCC studies have been conducted on as-extruded material that has a partial microstructure of equilibrium phases in addition to the phase from quenching. Precracked specimen tests have yielded KIscc values for polynary alloys tested in distilled water and in chloride solutions (Table 6). The data show that chloride increases the susceptibility to SCC. Tests done on smooth specimens have shown that these alloys are very resistant to crack initiation in humid air, but water promotes crack propagation in precracked specimens. All of the SCC failures have resulted in a transgranular stress-corrosion fracture. Table 6 SCC thresholds for polynary uranium alloys in aqueous solutions Alloy

U-0.75Nb-0.75Mo-0.75Zr-0.5Ti

U-1Nb-1Mo-1Zr-0.5Ti

Yield strength

MPa

ksi

772

112

1172

170

Environment

KIscc

MPa

ksi

Water

44

40

3.5% NaCl

13

12

Water

31

28

50 ppm Cl-

10

9

3.5% NaCl

8

7

1627

U-1Nb-1Mo-1Zr-0.5Ti-0.5V

236

Water

10

9

50 ppm Cl-

5.5

5

3.5% NaCl

5.5

5

Stress-Corrosion Cracking of Uranium Alloys One of the problems that has been encountered in the use of uranium alloys is SCC. Stress-corrosion cracking is defined as a cracking process that requires the simultaneous action of a corrosive environment and a sustained stress. In general, as the corrosion resistance and strength of an alloy increase, its susceptibility to SCC also increases. The conventional test methods use time to failure measurements on smooth specimens, incorporating both initiation and propagation processes, or precracked specimens in a fracture mechanics test, which tends to eliminate the initiation stages. Uranium-Titanium Alloys. The U-0.75Ti alloy has been the only composition in the U-Ti system that has received a significant amount of stress-corrosion testing. Crack propagation tests using precracked specimens conducted in oxygen, hydrogen, water vapor, water, dry air, wet air, and chloride solutions have shown that water is the species responsible for cracking (Table 4). Oxygen apparently inhibits crack propagation in this alloy. The effect of the strength level on the susceptibility of U-0.75Ti to SCC has also been investigated (Ref 30). Figure 13 shows the relationship between ultimate tensile strength and KIscc, the threshold stress intensity for SCC. The data show that KIscc decreased linearly by a factor of two in the strength range investigated. All of the tests conducted in this alloy system found that transgranular cracking was the SCC mode, and there was no evidence of cracking along prior -phase grain boundaries.

Table 4 SCC thresholds for U-0.75Ti in air and in aqueous environments Heat treatment

Yield strength

MPa

ksi

380 °C (715 °F), 6 h

986

143

380 °C (715 °F), 6 h

986

380 °C (715 °F), 6 h

Environment

KIscc

MPa

ksi

Dry air

42

38

143

100% relative humidity air

27.5

25

986

143

50 ppm Cl-

17.6

16

380 °C (715 °F), 6 h

986

143

3.5% NaCl

17.6

16

Air cooled

607

88

Water

23

21

Air cooled

607

88

3.5% NaCl

16.5

15

450 °C (840 °F)

...

...

Air

28.6

26

450 °C (840 °F)

...

...

Water

15.4

14

450 °C (840 °F)

...

...

3.5% NaCl

11

10

Source: Ref 29

Fig. 13 Plane-strain threshold for SCC propagation versus ultimate tensile strength for U-0.75Ti in 50 ppm Clsolution at 25 °C (75 °F). Source: Ref 30

Uranium-Molybdenum Alloys. Uranium alloys with molybdenum concentrations from 0.6 to 12% have been found to be susceptible to SCC (Ref 29). Below 5% Mo, a series of metastable -phase alloys is formed that has been found to be susceptible to cracking. Tests done in both wet and dry conditions on -phase alloys containing more than 5% Mo have shown that, contrary to the behavior observed with uranium-titanium alloys, oxygen is the species primarily responsible for SCC. Carbon content is important to the cracking behavior, alloys with higher carbon contents being more susceptible. The heat treating of quenched alloys can lead to significant improvements in SCC resistance in that alloys with equilibrium microstructures are much more resistant to cracking than metastable alloys. The predominant fracture mode observed in uranium-molybdenum alloys has been transgranular, particularly in nonaqueous environments. An intergranular mode has been observed in precracked tests conducted in aqueous environments (Ref 31). Uranium-Niobium Alloys. The principal interest in the U-Nb system has centered on alloys with 2.3, 4.5, 6, and 8%

Nb by weight, respectively. The SCC response is very much dependent on the particular alloy. The 2.3 and 4.5% Nb alloys are subject to SCC in water vapor, but the 6% Nb alloy is not. However, 6 and 8% Nb alloys will crack in environments containing oxygen, and water vapor will accelerate the SCC in these environments (Ref 32). The U-4.5Nb alloy was shown to crack in either water vapor or oxygen and was more susceptible when both were present (Ref 32). Figure 14 shows that for the U-4.5Nb alloy the crack velocity decreases and KIscc increases as the oxygen pressure is decreased (Ref 32). The chloride ion (Cl-) was also found to be very deleterious to the U-4.5Nb alloy in aqueous solutions (Ref 33). Specimens loaded at 33MPa (30 ksi ) failed in 270 min in distilled water and in as little as 1 min in chloride solutions. However, another study revealed that the U-6Nb alloy was not susceptible to SCC initiation in chloride solutions when aged at temperatures below 200 °C (390 °F), but material aged between 250 and 400 °C (480 and 750 °F) was susceptible (Ref 34). When overaged (aged at temperatures above 600 °C, or 1110 °F), the alloy exhibited an increased resistance to cracking. In general, the data have shown for all of the uranium-niobium alloys that overaged specimens are more resistant to SCC than underaged specimens.

Fig. 14 Crack velocity versus SCC threshold as a function of oxygen pressure for U-4.5Nb. Source: Ref 32

Both transgranular and intergranular cracking was observed in the U-Nb system, with the mode observed being dependent on the alloy composition, the environment, and the type of specimen. In chloride solutions, smooth-specimen tests showed that the U-2.3Nb alloys cracked transgranularly and that the U-6Nb and U-8Nb alloys cracked intergranularly. The U-4.5Nb alloy exhibited intergranular cracking in smooth specimens and transgranular cracking in precracked specimens. Uranium-Niobium-Zirconium Alloys. The only uranium-niobium-zirconium alloy that has been studied to any significant extent is mulberry. The SCC behavior of mulberry is similar to that of U-6Nb. The SCC behavior of this alloy was extensively studied in the 1970s and was later reviewed in Ref 29 and 32. Two types of SCC are observed in the alloy, each with its own fracture mode. The intergranular cracking initiates easily, but requires oxygen, water, and chloride to do so. The transgranular SCC does not initiate easily, but can propagate in pure oxygen or in solutions containing strong oxidizers.

An investigation of the effect of heat treatment on the SCC susceptibility of mulberry showed that the higher the aging temperature the greater the resistance to SCC initiation (Table 5). It was also found that crack propagation was fastest for material aged at intermediate temperature (350 to 450 °C, or 660 to 840 °F) and was slowest in the material aged at 150 °C (300 °F), the standard aging temperature for this alloy.

Table 5 Threshold stress for SCC initiation of mulberry in dilute chloride solutions as a function of heat treatment Heat treatment

KIscc

MPa

ksi

150 °C (300 °F), 1 h

65

Source: Ref 35 Polynary Uranium Alloys. The SCC behavior of a series of polynary uranium alloys containing small quantities of all

the potent stabilizers (titanium, niobium, molybdenum, and zirconium) has been investigated, primarily by AMMRC (Ref 36). All of the SCC studies have been conducted on as-extruded material that has a partial microstructure of equilibrium phases in addition to the phase from quenching. Precracked specimen tests have yielded KIscc values for polynary alloys tested in distilled water and in chloride solutions (Table 6). The data show that chloride increases the susceptibility to SCC. Tests done on smooth specimens have shown that these alloys are very resistant to crack initiation in humid air, but water promotes crack propagation in precracked specimens. All of the SCC failures have resulted in a transgranular stress-corrosion fracture. Table 6 SCC thresholds for polynary uranium alloys in aqueous solutions Alloy

U-0.75Nb-0.75Mo-0.75Zr-0.5Ti

U-1Nb-1Mo-1Zr-0.5Ti

Yield strength

MPa

ksi

772

112

1172

170

Environment

KIscc

MPa

ksi

Water

44

40

3.5% NaCl

13

12

Water

31

28

50 ppm Cl-

10

9

3.5% NaCl

8

7

U-1Nb-1Mo-1Zr-0.5Ti-0.5V

1627

236

Water

10

9

50 ppm Cl-

5.5

5

3.5% NaCl

5.5

5

References 1. 2. 3. 4. 5. 6. 7. 8.

9.

10. 11. 12. 13. 14.

15. 16. 17. 18. 19.

A.G. Ritchie, A Review of the Rates of Reaction of Uranium with Oxygen and Water Vapor at Temperatures up to 300 C. J. Nucl. Mater., Vol 102, 1981, p 170 L. Baker, Jr. and J.D. Bingle, The Kinetics of Oxidation of Uranium Between 300 and 625 C, J. Nucl. Mater., Vol 20, 1966, p 11 L. Leibowitz, J.G. Schnizlein, J.D. Bingle, and R.C. Vogel, The Kinetics of Oxidation of Uranium Between 125 and 250 °C, J. Electrochem. Soc., Vol 108, 1961, p 1155-1160 J.W. Isaacs and J.N. Wanklyn, "The Reaction of Uranium with Air at High Temperatures," Report No. AERE-R-3559, Atomic Energy Research Establishment, Harwell, England, Dec 1960 W.J. Megaw, R.C. Chadwick, A.C. Wells, and J.E. Bridges, The Oxidation and Release of Iodine-131 from Uranium Slugs Oxidizing in Air and CO2, J. Nucl. Energy, Parts A/B, Vol 15, 1961, p 176-184 R.K. Hilliard, "Oxidation of Uranium in Air at High Temperatures," U.S. Atomic Energy Commission Contract Report No. HW-58022, General Electric Co., Hanford Atomic Products Operation, 1958 A.F. Bessonov and V.G. Vlasov, Oxidation Mechanism of Metallic Uranium, Phys. Metals Metallogr., Vol 12, 1961, p 89-94 (translation of Fix. Metal. Metalloved., Vol 12, 1961, p 403-408) J.E. Antill and P. Murray, Reactions Between Fuel Elements and Gaseous Coolants, in Progress in Nuclear Energy, Ser. IV Technology, Engineering and Safety, Vol 3, C.M. Nicholls, Ed., Pergamon Press, 1960, p 65-86 A.G. Ritchie, R.C. Greenwood, S.J. Randles, D.R. Netherton, and J. Whitehorn, "Measurement of the Rate of the Uranium-Water Vapor Reaction," Report O 4/86. Atomic Weapons Research Establishment, June 1986 C.C. Colmenares et al., "Oxidation of Uranium Studied by Gravimetric and Positron Annihilation Techniques," UCRL-85549, Lawrence Livermore National Laboratories, April 1981 L.J. Weirick, "The Oxidation of Uranium in Low Partial Pressures of Oxygen and Water Vapor at 100 C," SAND83-0618, Sandia National Laboratories, June 1984 A.G. Ritchie, R.C. Greenwood, and S.J. Randles, "The Kinetics of the Uranium Oxygen-Water Vapor Reaction Between 40 and 100 C," Report O 10/85, Atomic Weapons Research Establishment, Oct 1985 S. Orman, "Oxidation of Uranium in Water and Water Vapor," Report O 25/64, Atomic Weapons Research Establishment, 1964 C.J. Greenholt and L.J. Weirick, "The Oxidation of Uranium-0.75 Weight Percent Titanium in Environments Containing Oxygen and/or Water Vapor at 140 C," SAND85-1688, Sandia National Laboratories, Sept 1985, to be published in J. Nucl. Mater. A.M. Ammons, Oak Ridge Y-12 Plant, private communication, March 1975 N.J. Magnani, "The Reaction of Uranium and Its Alloys with Water Vapor at Low Temperatures," SAND74-0145, Sandia National Laboratories , Aug 1974 S. Orman, Oxidation of Uranium and Uranium Alloys, in Physical Metallurgy of Uranium Alloys, J.J. Burke, D.A. Colling, A.E. Gorum, and J. Greenspan, Ed., Brook Hill, 1974 A.G. Ritchie, R.C. Greenwood, S.J. Randles, D.R. Netherton, and J. Whitehorn, "The Atmospheric Corrosion of Uranium-6% Niobium Alloy," Preprint, Atomic Weapons Research Establishment, Nov 1985 N.J. Magnani and H. Romero, "The Reaction of Water Vapor with U-7 wt%Nb-2 wt%Zr and U-4 wt%Nb," SC-RR-72-0635, Sandia National Laboratories, Sept 1972

20. V.B. Kishinevski, L.I. Gomozov, and O.S. Ivanov, Corrosion Resistance in Water of Some Alloys of Uranium with Zirconium, Niobium and Molybdenum, in Physical Chemistry of Alloys and Refractory Compounds of Thorium and Uranium, Report AEC-tr-7212, O.S. Ivanov, Ed., U.S. Atomic Energy Commission, 1972, translated from Russian 21. L.J. Weirick, Protective Coatings for Uranium Alloys, in Physical Metallurgy of Uranium Alloys, J.J. Burke, D.A. Colling, A.E. Gorum, and J. Greenspan, Ed., Brook Hill, 1974 22. J.M. Macki and R.L. Kochen, "The Galvanic Corrosion Behavior of Uranium Alloys in Hydrochloric Acid and Ocean Water," RFP-1592, Rocky Flats Plant, Feb 1971 23. J.M. Macki and R.L. Kochen, "The Corrosion Behavior of Uranium-Base U-Nb, U-Nb-Zr, and U-Mo Alloys in Hydrochloric Acid and Ocean Water," RFP-1586, Rocky Flats Plant, Feb 1971 24. M. Levy and C.V. Zabielski, Electrochemical Behavior of Some Binary and Polynary Uranium Alloys, in Physical Metallurgy of Uranium Alloys, J.J. Burke, D.A. Colling, A.E. Gorum, and J. Greenspan, Ed., Brook Hill, 1974 25. L.J. Weirick, "Corrosion Testing of the General Electric Mantech GAU 8/A Penetrator," SAND76-8055, Sandia National Laboratories, Feb 1977 26. H.R. Johnson and L. J. Weirick, "Corrosion Testing of the General Electric Mantech II GAU 8/A Penetrator," SAND78-8009, Sandia National Laboratories, May 1978 27. K.H. Eckelmeyer, A.D. Romig, Jr., and L.J. Weirick, The Effect of Quench Rate on the Microstructure, Mechanical Properties, and Corrosion Behavior of U-6 wt pct Nb, Metall. Trans. Vol 15A, July 1984, p 1319 28. C. Levy, "Uranium Alloys for XM-673 Projectile, Engineering Program," AMMRC SP 72-17, Army Materials and Mechanics Research Center, Oct 1972 29. N.J. Magnani, Stress Corrosion Cracking of Uranium Alloys, in Physical Metallurgy of Uranium Alloys, J.J. Burke, D.A. Colling, A.E. Gorum, and J. Greenspan, Ed., Brook Hill, 1974 30. N.J. Magnani, The Effect of Environment, Orientation and Strength Level on the Stress Corrosion Behavior of U-0.75wt.%Ti, J. Nucl. Mater., Vol 54, 1974, p 108 31. N.J. Magnani, The Effect of Environment on the Cracking Behavior of Selected Uranium Alloys, "SCR72-2661, Sandia National Laboratories, March 1972 32. N.J. Magnani, Hydrogen Embrittlement and Stress-Corrosion Cracking of Uranium and Uranium Alloys, in Advances in Corrosion Science and Technology, Vol 6, M.G. Fontana and R.W. Staehle, Plenum Press, 1976 33. N.J. Magnani, The Effects of Chloride Ions on the Cracking Behavior of U-7.5wt%Nb-2.5wt%Zr and U4.5wt%Nb, J. Nucl. Mater., Vol 42, 1972, p 271 34. J.W. Koger, "Stress-Corrosion Cracking of Uranium Alloys," Y-DA-5624, Y-12 Plant, 1973 35. L.J. Weirick, The Effect of Heat Treatment Upon the Stress-Corrosion Cracking of Mulberry (U-7.5Nb2.5Zr), Corrosion, Vol 31, 1975, p 5 36. W.F. Czyrklis and M. Levy, Stress Corrosion Cracking Behavior of Uranium Alloys, Corrosion, Vol 30, 1974, p 181 37. G.S. Petit, R.R. Wright, C.A. Keinberger, and C.W. Weber, "Formation of Corrosion-Resistant Oxide Film on Uranium," K-1778, Oak Ridge Gaseous Diffusion Plant, 1969 38. O. Flint, J.J. Polling, and A. Charlesby, The Anodic Oxidation of Uranium Acta Metall., Vol 2, 1954, p 696 39. T.S. Prevender, Sandia National Laboratories, private communication 40. S. Orman and P. Walker, The Corrosion of Uranium and Its Prevention by Organic Coatings, J. Oil Colour Chem. Assoc., Vol 48, 1965, p 233 41. C.A. Colmenares, "Aluminum and Polymeric Coatings for Protection of Uranium," UCID-19970, Lawrence Livermore National Laboratory, Dec 1983 42. C.E. Miller, "Producibility Study, Cartridge 20MM, DS Mk 149," Final Report, Olin Energy Systems Operations, Olin Corporation, Dec 1974

43. J.M. Macki and R.L. Kochen, "The Corrosion and Stress-Corrosion Cracking of Painted U-4.2wt.%Nb," RFP-1891, Rocky Flats Plant, Aug 1972 44. L.J. Weirick, Electroplating and Corrosion Protection, in Surface Metallurgy of Uranium and Uranium Alloys, Vol 3, Metallurgical Technology of Uranium and Uranium Alloys, American Society for Metals, 1981 45. G.W. Vest, Ion Plating in Corrosion Protection, in Surface Metallurgy of Uranium and Uranium Alloys, Vol 3, Metallurgical Technology of Uranium and Uranium Alloys, American Society for Metals, 1981 46. J.K. Gore and R. Seegmiller, Surface Treatments and Electro-plated Coatings on Uranium, Plating, Vol 50, 1963, p 215 47. J.G. Beach and C.L. Faust, "Electroplates on Thorium and Uranium for Corrosion Protection and to Aid Joining," BMI-1537, Battelle Memorial Institute, Aug 1961 48. P.D. Anderson, P.R. Coronado, and L.M. Berry, U.S. Patent 3,341,350, granted Sept 1967, Method of Preparing a Uranium Artide for a Protective Coating 49. J.W. Dini and P.R. Coronado, Preparation of Uranium for Electro-plating with Nickel, Trans. IMF, Vol 47, 1969, p 1 50. L.J. Weirick, "Evaluation of Metallic Coatings for the Corrosion Protection of a Uranium-3/4 Weight Percent Titanium Alloy," SLL-73-5024, Sandia National Laboratories, Feb 1974 51. L.J. Weirick, Electrochemical Determination of Porosity in Nickel Electroplates on a Uranium Alloy, J. Electrochem. Soc., Vol 47 (No. 1), 1975, p 937 52. L.J. Weirick and D.L. Douglass, Effect of Thin Electrodeposited Nickel Coatings on the Corrosion Behavior of U-0.75Ti, Corrosion, Vol 36 (No. 6), 1976, p 209 53. D.M. Mattox, Film Deposition Using Accelerated Ions, Electrochem. Technol., Vol 2, 1964, p 295 54. R.D. Bland, A Parametric Study of Ion-Plated Aluminum Coatings on Uranium, Electrochem. Technol., Vol 6, 1968, p 272

Corrosion Materials

of

Powder

Metallurgy

Erhard Klar, SCM Metal Products

Introduction POWDER METALLURGY (P/M) is a branch of metallurgy that may be defined as the technology and art of producing metal powders and using them to construct massive materials and shaped objects. Although the commercial use of P/M dates back to the beginning of the 20th century, the major classes of P/M materials developed during the first four decades (cemented carbides, composite electrical contacts, self-lubricating bronze bearings, and composite friction materials) did not pose any major problems with regard to corrosion resistance in that the service lives of these materials were normally not limited by their corrosion properties. With the appearance of iron-base sintered structural parts in the 1940s, the P/M method began to compete with other metal-forming techniques, such as casting, machining, and forging. The rapid increase in the utilization of iron and copper powders from 1940 to 1970 is linked to the advent of mass production in the automotive industry, and the successful use of P/M technology was based mainly on labor and material savings. With the growth of the industry and the development of stronger sintered parts came the demand for improved corrosion resistance, which led to the development of several surface treatments for sintered iron and steel parts (steam treatment; impregnation with oils, plastics, and waxes; metallization; tumbling with fillers; and electroplating) and to the development of sintered stainless steels in the 1950s. With the development of fully dense P/M materials in the 1950s and 1960s, and particularly with the appearance of rapid solidification technology in recent years, it became increasingly clear that P/M processing was capable not only of cost savings from reduced labor and reduced material usage (net shape and near net shape processing) but also of producing superior materials. The superior microstructures possible with P/M technology were responsible for significant improvements in mechanical and magnetic properties, workability, and corrosion resistance. Two classes of materials in which such improved properties are of utmost importance are fully dense P/M superalloys and P/M aluminum alloys. Powder metallurgy superalloys are currently used in advanced turbine engines. Although wrought P/M aluminum alloys are still in the evaluation stage, they have been used in aerospace structures on a limited trial basis in non-critical applications. The corrosion properties of the above-mentioned major classes of P/M materials will be described in the following sections. Corrosion resistance will be compared with that of similar conventional alloys where possible. Brief accounts will be given for the manufacture and use of these materials. More detailed descriptions of all aspects of the P/M process, such as powder production, compaction, sintering, and other state-of-the art consolidation methods, are available in Powder Metal Technologies and Applications, Volume 7 of the ASM Handbook. In addition to these major classes of P/M materials, there are many P/M specialty materials in which corrosion resistance may also play an important role (see, for example, the article "Corrosion of Cemented Carbides," in this Volume). Information on such materials is also provided in Powder Metal Technologies and Applications, Volume 7 of the ASM Handbook.

Sintered Iron Base P/M Parts Powder metallurgy iron and carbon steel parts exhibit little resistance to corrosion. Therefore, steam treatment and/or impregnation with oils is widely used to prevent or minimize rusting in mild environments. Steam treatment will be described in this section. Oxidation of small sintered iron and steel parts (Fig. 1) with superheated steam has been in use for over 40 years as an economical process to improve their hardness, compressive strength, wear, and corrosion resistance. This treatment can be combined with tempering and is then referred to as steam tempering. It leaves a tightly adhering bluish to black oxide that is usually acceptable as a final finish. The pores become coated or filled with the oxide as a result of the lower density of the oxide. This provides a certain pressure tightness and minimizes the entry of contaminants.

Fig. 1 P/M parts steam treated using the cycle shown in Fig. 2

The Process. A typical procedure for the steam treatment (Fig. 2) of clean (degreased) parts involves heating to a

temperature above 100 °C (212 °F) but below the critical point of air oxidation and discoloration (425 °C, or 800 °F). A temperature of 315 °C (600 °F) is typically used for the first stage. Parts are soaked for 15 min or until the center of the load is above 100 °C (212 °F). Steam is then introduced into the chamber at a high flow rate (4.5 to 165 kg/h, or 10 to 360 lbs/h, depending on furnace size) to purge air from the furnace through the relief valve.

Fig. 2 Typical steam-treating cycle for iron P/M parts

Once the purge has been completed, steam flow is halved and temperature is increased to between 425 and 620 °C (800 and 1150 °F). The load is then soaked for 30 min to 3 h, depending on the amount of oxide required. When the soak has been completed and the load removed, the furnace is cooled to below 425 °C (800 °F) for the next load (Ref 1). With correct processing, only two types of iron oxide form: FeO and Fe3O4. The furnace is purged of air before admitting the steam in order to avoid the formation of ferric oxide (Fe2O3) (red rust). Furthermore, the coldest part of the furnace charge must be above 100 °C (212 °F) before admission of steam to prevent the formation of iron hydroxide (Fe(OH)3) and Fe2O3. Dark blue or light gray oxides form, depending on the temperature of the second soak. The disadvantages of the treatment include some losses in tensile strength, impact resistance, and ductility (Ref 2). Details on the kinetics of oxidation, thermochemical equilibria, permeability and dimensional changes, oxide adherence, and hardness increases can be found in Ref 3, 4, 5, 6. Corrosion Resistance. Phase composition and other structural characteristics of oxidation products depend on the

oxidation technique and the cooling conditions (Ref 4). For example, the corrosion resistance of a low alloy sintered steel [0.12% C, 0.01% Si (max), 0.24% Mn, 0.010% P, 0.025% S, 0.02% Cr, 0.01% Ni (max), 0.019% Al, 1.86% Cu, 0.01% Ti (max)] with a density of 6.4 g/cm3 in 3% aqueous sodium chloride (NaCl) (by immersion) was inferior at low (580 °C, or 1075 °F) and high (630 °C, or 1165 °F) oxidation temperatures. The latter was attributed to the lower adherence of the oxides to the metallic phase.

In one study, potentiostatic anodic polarization measurements of steam-treated carbon steel in dilute (0.5%) sulfuric acid (H2SO4) and aqueous NaCl solutions (400 ppm chloride ion Cl-), revealed significant reduction of corrosion rates for the Cl- environment, while H2SO4 testing showed an increase in corrosion rate (Ref 7). The latter was attributed to the solubility of the oxides in dilute H2SO4 and their increased attack compared to the unoxidized steel.

Sintered (Porous) P/M Stainless Steels In its first decade of rapid commercialization, sintered stainless steel use increased at a compound annual rate of nearly 20% to reach a consumption of 2000 tons per year in 1973 for the North American market. During this period, the major developmental efforts were aimed at improving the compacting properties of water-atomized stainless steel powder. Interest in corrosion resistance increased after the market had peaked at a consumption of about 3000 tons per year. Although powder and processing requirements for improving corrosion resistance were better defined and improved stainless steel powders became available, the basic understanding of corrosion as well as corrosion data development for sintered stainless steels are both still in their infancy. There are many factors that distinguish sintered stainless steel from cast and wrought stainless steel. Complicating the issue is the fact that the corrosion resistance of sintered stainless steels depends as much on the sintering process as it does on the properties of the powder. Neither guidelines nor standards relating to corrosion behavior exist for most of the critical composition and process parameters. Furthermore, much of the published literature on specific corrosion data of sintered stainless steels is nearly obsolete because of the lack of information on process conditions. Application and Selection of Sintered P/M Stainless Steels In the absence of detailed and reliable corrosion data, tentative selection of a P/M stainless steel for a specific application is made by following the same principles developed for cast and wrought stainless steels. Thus, for better corrosion resistance, the austenitic grades are preferred. However, type 410L stainless steel is often used for its good abrasion resistance. Because the corrosion resistance of sintered stainless steels depends so much on powder quality and partsprocessing details, appropriate field testing is advisable to ensure compliance with specifications. Table 1 provides on overview of market segments and applications for sintered stainless steels. The 300-series austenitic grades account for about two-thirds of total usage, and among the austenitic grades, type 316L is the most important. In terms of market distribution, automotive applications constitute the largest volume, followed by hardware and tools, filters, appliances, office machines, and a large segment of miscellaneous uses. Table 1 Applications for P/M stainless steels Part

Alloy

Aerospace

Seatback tray slides

316L

Galley latches

316L

Jet fuel refueling impellers

316L

Foam generators

316L

Agriculture

Fungicide spray equipment

316L

Appliances

Automatic dishwasher components

304L

Automatic washer components

304L

Garbage disposal components

410L

Pot handles

316L

Coffee filters

316L-Si

Electric knives

316L

Blenders

303L

Can opener gears

410L

Automotive

Rearview mirror mounts

316L, 434L

Brake components

434L

Seat belt locks

304L

Windshield wiper pinions

410L

Windshield wiper arms

316L

Manifold heat control valves

304L

Building and construction

Plumbing fixtures

303L

Spacers and washers

316L

Sprinkler system nozzles

316L

Shower heads

316L

Window hardware

304L, 316L

Thermostats

410L

Chemical

Filters

304L-Si, 316L

High corrosion resistance filters

830

Cartridge assemblies

316L-Si

Electrical and electronic

Limit switches

410L

G-frame motor sleeves

303L

Rotary switches

316L

Magnetic clutches

410L, 440A

Battery nuts

830

Electrical testing probe jaws

316L

Hardware

Lock components

304L, 316L

Threaded fasteners

303L

Fasteners

316L

Quick-disconnect levers

303L, 316L

Industrial

Water and gas meter parts

316L

Filters, liquid and gas

316L-Si

Recording fuel meters

303L

Fuel flow meter devices

410L

Pipe flange clamps

316L

High polymer filtering

316L-Si

Jewelry

Coins, medals, medallions

316L

Watch cases

316L

Watch band parts

316L

Marine

Propeller thrust hubs

316L

Cam cleats

304L

Medical

Centrifugal drive couplings

316L

Dental equipment

304L

Hearing aids

316L

Anesthetic vaporizers

316L

Office equipment

Nonmagnetic card stops

316L

Dictating machine switches

316L

Computer knobs

316L

Recreation and leisure

Fishing rod guides

304L, 316L

Fishing rod gear ratchets

316L

Photographic equipment

316L

Soft drink vending machines

830, 316L

Travel trailer water pumps

316L

Powder Production and Compositions Stainless steel powders suitable for cold compaction are produced by water atomization. Table 2 lists the typical powder compositions of the important commercial grades of stainless steel powders. Table 2 Compositions of commercial P/M stainless steels Alloy

Composition, %

Cr

Ni

Si

Mo

Cu

Su

Mn

C

S

P

Fe

O(ppm)

Austenitic grades

303

17-18

12-13

0.6-0.8

...

...

...

0.3(a)

0.03(a)

0.1-0.3

0.03(a)

rem

...

304L

18-19

10-12

0.7-0.9

...

...

...

0.3(a)

0.03(a)

0.03(a)

0.03(a)

rem

1000-2000

304LSC

18-20

10-12

0.8-1.0

...

2(b)

1(b)

0.3(a)

0.03(a)

0.03(a)

0.03(a)

rem

...

316L

16.5-17.5

13-14

0.7-0.9

2-2.5

...

...

0.3(a)

0.03(a)

0.03(a)

0.03(a)

rem

1000-2000

...

0.7-0.9

...

...

...

0.1-0.5

0.05(a)

0.03(a)

0.03(a)

rem

1500-2500

Martensitic grade

410L

12-13

Ferritic grades

430L

16-17

...

0.7-0.9

...

...

...

0.3(a)

0.03(a)

0.03(a)

0.03(a)

rem

...

434L

16-18

...

0.7-0.9

0.5-1.5

...

...

0.3(a)

0.03(a)

0.03(a)

0.03(a)

rem

...

(a) Maximum.

(b) Typical

Properties of Sintered Parts Sintered properties depend not only on powder characteristics but also on processing and sintering conditions. The effects of the most important processing parameters are shown in Fig. 3, 4, and 5. Sintering in a nitrogen-containing atmosphere results in the absorption of considerable amounts of nitrogen, along with an increase in strength and a decrease in ductility. Additional mechanical properties are summarized in Table 3. Table 3 Typical mechanical properties of medium-density P/M stainless All materials sintered in dissociated NH3 MPIF designation

SS-303

SS-316

SS-410

Composition, %

Tensile strength

0.2% yield strength

Elongation in 25 mm (1 in.), %

Density, g/cm3

Cr

Ni

Mo

Si

Fe

MPa

ksi

MPa

ksi

17

12

...

0.7

rem

241

35

220

32

1

6.2

17

12

...

0.7

rem

358

52

324

47

2

6.5

16

13

2

0.7

rem

262

38

220

32

2

6.2

16

13

2

0.7

rem

372

54

275

40

4

6.6

12

...

...

0.8

rem

289

42

283

41

150 to 200 °C/min (270 to 360 °F) through critical temperature range (700 to 1000 °C, or 1290 to 1830 °F).

Use higher sintering temperature.

Use intermediate dew points (-37 to -45 °C, or 35 to -50 °F) in cooling zone of furnace.

Use tin-modified stainless steel powders.

Oxygen

Excessive oxygen in powder; excessive dew point of sintering atmosphere; slow cooling after sintering

Inferior resistance to general corrosion

Use low oxygen content powder, preferably 200 °C/min (360 °F/min)

For nitrogen-containing atmospheres, used dew point of -37 to -45 °C (-35 to -50 °F) in cooling zone.

For sintering in H2, ensure that water vapor

9, 11, 14

content of atmosphere is below 50 ppm.

Density of sintered part

High sintered density

Inferior resistance to crevice corrosion

Use lower density to increase pore size and circulation of corrodent. In acidic environments, corrosion resistance improves with increasing density due to a decrease of specific surface area.

9, 11, 15, 16, 17

Effect of Iron Contamination. Contamination of stainless steel powder with iron or iron-base powder may originate at the powder producer or the part manufacturer. Even extremely small amounts of iron contamination have a disastrous effect on the corrosion resistance of sintered parts in a saline environment. Utmost cleanliness, for example, through the use of separate production facilities and dedicated equipment, is mandatory. In saline solutions, active iron or iron-base particles form galvanic couples with the passive stainless steel and corrode anodically in preference to the stainless steel. Figure 7 shows this type of corrosion for iron particles embedded in the surface of a pressed and sintered type 316L part. Rusting occurs within minutes after exposure. The buildup of the initial corrosion product forms a crevice in which oxygen depletion causes acidification of the solution inside the part and further corrosion.

Fig. 7 Small circles of rust around iron particles embedded in the surface of sintered type 316L stainless steel after testing in 5% aqueous NaCl. 35× Source: Ref 9

Because of its severity, iron contamination overshadows the other factors that affect corrosion resistance (Fig. 8). Active iron or iron alloy particles present in stainless steel powder or on the surface of a sintered stainless steel part will be revealed by placing the powder or part in a concentrated aqueous solution of copper sulfate (CuSO4). The dissolved copper plates out on the iron particles within minutes, making them easy to identify with a low-magnification microscope. The powder must be tested in the unlubricated condition because lubricant will prevent the solution from wetting the powder. Experiments with very fine iron powder particles combined with high-temperature (>1260 °C, or 2300 °F), sintering have shown that this type of corrosion can be avoided if the sintering conditions result in complete alloying of the iron particles with the stainless steel matrix (Ref 9).

Fig. 8 Typical corrosion behavior of regular and copper-tin modified (type 316LSC) sintered type 316L stainless steel sintered in dissociated NH3 under various conditions of cooling and contamination. B rating indicates that 758 MPa (110 ksi) for an experimental mechanically alloyed aluminum-lithium binary alloy. Preliminary baseline corrosion studies on binary aluminum-lithium alloys have indicated that susceptibility to SCC is possible when the lithium concentration exceeds the room-temperature solid solubility limit (Ref 83). Overaged tempers have been found to be nonsusceptible, but peak-aged tempers have given the highest susceptibility. In one investigation, the SCC behavior of two aluminum-lithium-copper P/M alloys (2.6% Li) was studied with and without magnesium additions by using three test methods: threshold stress test on tuning fork specimens, slow crack growth tests on fracture mechanics specimens, and slow strain rate tests on electrically isolated tensile coupons (Ref 84). Each of these test methods, together with other experimental parameters was found to yield important information on the SCC behavior of the alloys. Another study compared the SCC susceptibility of the aluminum-lithium alloys (Ref 85). It was found that alloys exposed to aqueous 3.5% NaCl exhibited time-dependent fracture and that SCC susceptibility was composition dependent. Alloy X2020 and mechanically alloyed Al-3Li-2.1Mg appeared very resistant to SCC. The lower limit of threshold stress necessary to cause fracture in the most susceptible P/M alloy (Al-2.6Li-1.4Cu-1.6Mg) was about 355 MPa (51.5 ksi). The P/M aluminum-lithium-copper alloys were susceptible to intergranular crack initiation in aqueous NaCl, but were resistant to sustained subcritical crack growth for stress intensity factors up to 15 MPa (13.7 ksi ). Intergranular corrosion began at active pitting sites correlated to oxide particles strung along the extrusion direction of the alloys. Elevated-Temperature Alloys. Table 22 shows the compositions and Table 23 shows the elevated-temperature

properties of P/M aluminum alloys exhibiting superior elevated-temperature strength. These alloys are currently under development, but preliminary results indicate that they have good corrosion and SCC resistance. In these ternary and

quaternary alloys, iron functions as the major dispersoid-forming element. The solute levels of iron and other intermetallic compound forming transition metals are considerably higher than is acceptable in conventional alloys (Ref 86). Table 22 Nominal compositions of elevated-temperature P/M aluminum alloys Alloy

Composition, %

Fe

Ce

V

Mo

Zr

Al

CU-78

8.0

4.0

...

...

...

bal

Allied 1

9.3

...

3.5

...

...

bal

Allied 2

10.1

...

2.3

...

3.2

bal

Pratt & Whitney

8.0

...

...

2.0

...

bal

Table 23 Mechanical properties of elevated-temperature P/M aluminum alloys Alloy

CU 78

Al-8Fe-2Mo

Allied 1

Temperature

Ultimate tensile strength

Yield strength (0.2% offset)

Elongation, in 50 mm (2 in.), %

°C

°F

MPa

ksi

MPa

ksi

20

70

565

82

448

65

5

230

450

424

61

391

57

5

345

650

165

24

124

18

7

20

70

519

75

414

60

3

230

450

370

54

331

48

NA

345

650

207

30

172

25

NA

20

70

621

90

596

86

6.3

230

450

433

63

425

62

6.8

345

650

270

39

253

37

11.4

Allied 2

20

70

645

94

632

91

8.7

230

450

420

61

414

60

7.3

345

650

252

37

240

35

10.2

NA, not available. Source: Ref 59

Figure 47 shows the superior high-temperature (230 to 340 °C, or 450 to 650 °F) strength of some of these alloys as well as of aluminum-iron-vanadium-silicon alloys made by the planar flow cast process in comparison to conventional aluminum alloys. Figure 48 shows the much improved saline environment (ASTM B 117) weight loss data for several high-temperature aluminum-iron-zirconium-vanadium alloys compared to some conventional alloys (Ref 87).

Fig. 47 Elevated-temperature strength of advanced aluminum alloys. Source: Ref 86

Fig. 48 Saline environment weight loss data for several aluminum alloys. Source: Ref 86

References 1. 2. 3.

4. 5. 6.

7. 8. 9.

D.J. Orzeske, Steam Treatment Improves P/M Part Quality, Ind. Heat., Vol 53, May 1986, p 34-37 D. Feinberg, Surface Finishing of Powder Metallurgy Parts, Metal Powder Industries Federation, 1968 K. Razavizadeh and B.L. Davies, Influence of Powder Type and Density on Pore Closure and Surface Hardness Changes Resulting from Steam Treatment of Sintered Iron, Powder Metall., Vol 22 (No. 4), 1979, p 187-192 K. Volenik, H. Volrabova, J. Neid, M. Sebenji, and J. Cirak, Structure of Oxidation Products of Sintered Steel in Superheated Steam, Powder Metal., Vol 21 (No. 3), 1978, p 149-154 P. Franklin and B.C. Davies, The Effects of Steam Oxidation on Porosity in Sintered Iron, Powder Metall., Vol 20 (No. 1), 1977, p 11-16 G.F. Bocchini, A. Gallo, and I. Montevechi, Treatment of Sintered Steel Parts in Steam Atmosphere, Proceedings of the 2nd International Congress on Heat Treatment of Materials: 1st National Congress on Metallurgical Coatings, Florence, Italy, Associazone Italiana di Metallurgia, 1982, p 1091-1102 B.J. Santer and B.D. Cash, Effect of Surface Treatment on the Corrosion Properties of Ferrous Powder Metallurgy Parts, Powder Metall., Vol 17 (No. 34), 1974, p 319-330 "P/M Standards and Specifications," MPIF 35, Metal Powder Industries Federation M.A. Pao and E. Klar, Corrosion Phenomena in Regular and Tin-modified P/M Stainless Steels, in Progress in Powder Metallurgy, H.S. Nayar, S.M. Kaufman, and K.E. Meiners, Ed., Metal Powder Industries Federation, 1984, p 431-444

10. D. Ro and E. Klar, Corrosive Behavior of P/M Austenitic Stainless Steels, in Modern Developments in Powder Metallurgy, Vol 13, H.H. Hausner and P.W. Taubenblat, Ed., Metal Powder Industries Federation, 1980, p 247-287 11. R.L. Sands, G.F. Bidmead, and D.A. Oliver, The Corrosion Resistance of Sintered Stainless Steels, in Modern Developments in Powder Metallurgy, Vol 2, H.H. Hausner, Ed., Plenum Press, 1966, p 73-85 12. H.S. Nayar, R.M. German, and W.R. Johnson, The Effect of Sintering on the Corrosion Resistance of 316L Stainless Steel, in Progress in Powder Metallurgy, Vol 37, 1981, p 255-265 13. G. Lei, R.M. German, and H.S. Nayar, Influence of Sintering Variables on the Corrosion Resistance of 316L Stainless Steel, Powder Metall. Int., Vol 15 (No. 2), 1983, p 70-76 14. R.H. Shay, T.L. Ellison, and K.R. Berger, Control of Nitrogen Absorption and Surface Oxidation of Austenitic Stainless Steels in H-N Atmospheres, in Progress in Powder Metallurgy, Vol 39, H.S. Nayar, S.M. Kaufman, and K.E. Meiners, Ed., Metal Powder Industries Federation, 1983, p 411-430 15. K.H. Moyer, The Burn-Off Characteristics of Common Lubricants in 316L Powder Compacts, Int. J. Powder Metall., Vol 7 (No. 3), 1971, p 33-43 16. S.K. Chatterjee, M.E. Warwick, and D.J. Maykuth, The Effect of Tin, Copper, Nickel, and Molybdenum on the Mechanical Properties and Corrosion Resistance of Sintered Stainless Steel (AISI 304L), in Modern Developments in Powder Metallurgy, Vol 16, E.N. Aqua and C.I. Whitman, Ed., Metal Powder Industries Federation, 1984, p 277-293 17. T. Kato, K. Kusaka, and T. Hisada, Influence of Cu Addition on Some Properties of SUS 304L Stainless Steel Powders, Denki Seiko (Electro. Furn. Steel), Vol 51 (No. 4), Nov 1980, p 252-263 18. A.J. Sedriks, Corrosion of Stainless Steels, John Wiley & Sons, 1979, p 15 19. M.A. Pao and E. Klar, On the Corrosion Resistance of P/M Austenitic Stainless Steels, Proceedings of the International Powder Metallurgy Conference, Florence, Italy, Associazone Italiano di Metallurgia, 1982 20. G. Lei, R.M. German, and H.S. Nayar, Corrosion Control in Sintered Austenitic Stainless Steels, in Progress in Powder Metallurgy, Vol 39, H.S. Nayar, S.M. Kaufman, and K.E. Meiners, Ed., Metal Powder Industries Federation, 1984, p 391-410 21. K. Kusaka, T. Kato, and T. Hisada, Influence of S, Cu, and Sn Additions on the Properties of AISI 304L Type Sintered Stainless Steel, in Modern Developments in Powder Metallurgy, Vol 16, E.N. Aqua and C.I. Whitman, Ed., Metal Powder Industries Federation, 1984, p 247-259 22. D. Itzhak and S. Harush, The Effect of Sn Addition on the Corrosion Behavior of Sintered Stainless Steel in H So, Corros. Sci., Vol 25 (No. 10), 1985, p 883-888 23. F.M.F. Jones, The Effect of Processing Variables on the Properties of Type 316L Powder Compacts, in Progress in Powder Metallurgy, Vol 30, Metal Powder Industries Federation, 1970, p 25-50 24. T. Kato and K. Kusaka, On Some Properties of Sintered Stainless Steels at Elevated Temperatures, Powder Metall., Vol 27 (No. 5). July 1980, p 2-8 25. O.W. Reen and G.O. Hughes, Evaluating Stainless Steel Powder Metal Parts, Precis. Met., Vol 35 (No 8), Aug 1977, p 53-54 26. M.H. Tikkanen, Corrosion Resistance of Sintered P/M Stainless Steels and Possibilities for Increasing It, Scand. J. Metall., Vol 11, 1982, p 211-215 27. T. Takeda and K. Tamura, Compacting and Sintering of Chrome-Nickel Austenitic Stainless Stress Powders. H.B. Transl 8311 from Powder and Powder Metallurgy (Japan), Vol 17 (No. 2), 1970, p 70-76 28. T.J. Treharne, U.S. Patent 4,536,228, granted 20 Aug 1985, Corrosion Inhibition in Sintered Stainless Steel 29. A. Kempster, J.R. Smith, and C.C. Hanson, Chromium Diffusion Coatings on Sintered Stainless Steel, Metal Powder Report, MPR Publishing Services Ltd., England, June 1986, p 455-460 30. C. Tornberg, "The Manufacture of Seamless Stainless Steel Tubes from Powder," Paper 8410-013, presented at the 1984 ASME International Conference on New Developments in Stainless Steel Technology, Detroit, MI, American Society of Mechanical Engineers, 1984, p 1-6 31. "Standard Practices for Detecting Susceptibility to Intergranular Attack in Austenitic Stainless Steels," A 262, Annual Book of ASTM Standards, American Society for Testing and Materials

32. G.S. Yurek, D. Eisen, and A.J. Garrat-Reed, Metall. Trans. A, Vol 13A, 1982, p 473 33. P.C. Searson and R.M. Latanision, The Corrosion and Oxidation Resistance of Iron-and Aluminum-Based Powder Metallurgical Alloys, Corros. Sci., Vol 25 (No. 10). 1985, p 947-968 34. G.H. Gessinger, Recent Developments in Powder Metallurgy of Superalloys. Powder Metall. Int., Vol 13 (No. 2), 1981, p 93-101 35. R.F. Singer, Recent Developments and Trends in High Strength P/M Materials, Powder Metall. Int., Vol 17 (No 6), 1985, p 284-288 36. L.R. Curwick, The Mechanical Alloying Process: Powder to Mill Product, in Frontiers of High Temperature Materials, J.S. Benjamin, Ed., of the International Conference on Oxide Dispersion Strengthened Superalloys by Mechanical Alloying, Inco Alloy Products Company, 1981, p 3-10 37. J.S. Benjamin and T.E. Volin, The Mechanism of Mechanical Alloying, Metall. Trans., Vol 5, 1974, p 1929-1934 38. F.S. Pettit and G.H. Meier, Oxidation and Hot Corrosion of Superalloys, in Supperalloys, 1984, M. Gell et al., Ed., Proceedings of the Fifth International Symposium on Superalloys, The Metallurgical Society, 1984, p 651-687 39. C.T. Sims, A History of Superalloy Metallurgy for Superalloy Metallurgists, in Superalloys, 1984, M. Gell et al., Ed., Proceedings of the Fifth International Symposium on Superalloys, The Metallurgical Society, 1984, p 309-419 40. G.H. Gessinger, Powder Metallurgy of Superalloys, Butterworths, 1984 41. J. Stringer, B.A. Wilcox, and P.I. Jaffee, Oxid. Met., Vol 5, 1972, p 11 42. C.S. Giggins and F.S. Pettit, The Oxidation of TD Ni Cr (Ni-20 Cr-2 Vol. pct ThO2) Between 900 and 1200 C, Metall. Trans., Vol 2, 1971, p 1071-1078 43. F. Perry, Oxide-Dispersion-Strengthened P/M Alloys Produced for Severe Service Applications, Ind. Heat., Vol 49 (No. 5), May 1982, p 22-25 44. J.H. Weber and P.S. Gilman, Environmentally Induced Porosity in Ni-Cr Oxide Dispersion Strengthened Alloys, Scr. Metall., Vol 18, 1984, p 479-482 45. J.H. Weber, High Temperature Oxide Dispersion Strengthened Alloys, Proceedings of the 25th National SAMPE Symposium and Exhibition, Society for the Advancement of Material and Process Engineering, 1980, p 752-763 46. G.H. Gessinger, High Temperature Alloys for Gas Turbines, D. Coutsouradis et al., Ed., Applied Science, 1978, p 817 47. P.L. Antona, A. Bennani, P. Cavalloti, and O. Ducati, Heat Treatments and Oxidation Behavior of Some P/M Ni-Base Superalloys, in European Symposium on Powder Metallurgy, Vol 1, Jernkontoret Activity Group B, 1978, p 137-142 48. W.K. Uhl, M.R. Pendley, and S. McEvoy, "Evaluation of HIP Nickel-Base Alloys for Extreme Sour Service," Paper 219, presented at Corrosion/84, New Orleans, LA, National Association of Corrosion Engineers, April 1984 49. S. Floreen, Effects of Environment on Intermediate Temperature Crack Growth in Superalloys, in Proceedings of the AIME Symposium (Louisville, KY), American Society of Mining, Metallurgical, and Petroleum Engineers, 1981 50. G.A.J. Hack, Inconel Alloy MA 6000--A New Material for High Temperature Turbine Blades, Met. Powder Rep., Vol 36 (No. 9), Sept 1981, p 425-429 51. D.H. Boone, D.A. Crane, and D.P. Whittle, Thin Solid Films, Vol 84, 1981, p 39 52. F.R. Wermuth and A.R. Stetson, Report NASA CR-120852, National Aeronautics and Space Administration, 1971 53. M.A. Gedwill, T.K. Glasgow, and R.S. Levine, "A New Diffusion Inhibited Oxidation Resistant Coating for Superalloys," NASA TM 82687, National Aeronautics and Space Administration, 1981 54. T.K. Glasgow and G.J. Santoro, Oxidation and Hot Corrosion of Coated and Bare Oxide Dispersion Strengthened Superalloy MA 755E, Oxid. Met., Vol 15 (No. 314), April 1986, p 251-276

55. S.G. Roberts, An Exploratory Investigation of Prealloyed Powders of Aluminum, in Powder Metallurgy, W. Leszynski, Ed., 1961, p 799-817 56. A.P. Haarr, Final Report, Section III, Contract No DA-36-04-ORD-3559RD, DDC, Alexandria, VA, 31 May 1966 57. H. Jones, Rapid Solidification of Metals and Alloys, Monograph 8, Institution of Metallurgists, 1982 58. J.P. Lyle and W.S. Cebulak, Powder Metallurgy Approach for Microstructure and Properties in High Strength Aluminum Alloys, Metall. Trans. A, Vol 6A, 1975, p 685-699 59. J.R. Pickens, Review: Aluminum Powder Metallurgy Technology for High-Strength Applications, J. Mater. Sci., Vol 16, 1981, p 1437-1457 60. J.R. Pickens, High Strength Aluminum Alloys Made by Powder Metallurgy: A Brief Overview, In Rapidly Quenched Metals, S. Steeb and H. Warlimont, Ed., Elsevier, 1985, p 1711-1718 61. T.E. Tietz and I.G. Palmer, Advanced P/M Aluminum Alloys in Proc. 1981 ASM Mater. Sci. Sem., Advances in Powder Metallurgy, 1981 62. J.R. Pickens and E.A. Starke, Jr., The Effect of Rapid Solidification on the Microstructures and Properties of Aluminum Powder Metallurgy Alloys, in Rapid Solidification Processing: Principle and Technologies, III, R. Mehrabian, Ed., National Bureau of Standards, p 150-170 63. F.H. Froes and J.R. Pickens, Powder Metallurgy of Light Metal Alloys for Demanding Applications, J. Metals, Vol 36 (No. 1), 1984, p 14-28 64. J.R. Pickens, K.S. Kumar, and T.J. Langan, High Strength Aluminum Alloy Development, to be published in The Proceedings of the 33rd Sagamore Army Materials Research Conference, Corrosion Prevention and Control, Vol 33, AMMRC 65. E. Lavernia, B. Poggiali, I. Servi, J. Clark, F. Katrak, and N. Grant, Rapidly Solidified Aluminum Alloys: A Market Assessment, J. Met., Vol 37 (No. 11), Nov 1985, p 35-38 66. N.J. Grant, in Proceedings of the International Conference on the Rapid Solidification Process, Claitor's Publishing, p 230 67. A. Hafeez, "Corrosion Resistance of Wrought P/M Alloys," Tech Brief 2 3/85, Aluminum Company of America, Pittsburgh, PA 68. G.J. Hildeman, L.C. Labarre, A. Hafeez, and L.M. Angers, "Microstructural, Mechanical Property and Corrosion Evaluations of 7xxx P/M Alloy CW 67," Internal Report, Aluminum Company of America 69. L. Christodoulou, J.R. Gordon, and J.R. Pickens, Effect of Co Content on the Stress-Corrosion Cracking Behavior of 7091-Type Aluminum Powder Alloys, Metall. Trans. A, Vol 16A, 1985, p 945-951 70. L. Christodoulou, T.J. Langan, D. Venables, J.A.S. Green, and J.R. Pickens, Crack-Tip Chemistry of Stress-Corrosion Cracks in 7xxx Aluminum Powder Alloys, accepted for publication in Corrosion NACE 71. J.R. Pickens and L. Christodoulou, The Stress-Corrosion Cracking Behavior of High-Strength Aluminum Powder Metallurgy Alloys, accepted for publication in Metall. Trans. A 72. J.R. Pickens, L. Christodoulou, and T.J. Langan, Report DAA 929-81-C-0031, Part I, Martin Marietta, 1983 73. A. Hafeez, Aluminum Corporation of America, private communication, Sept 1986 74. J.S. Santner and M. Kumar, Corrosion-Fatigue Crack Propagation Rates in Conventional 7075 and P/M X7091 Aluminum Alloys, in Corrosion Fatigue: Mechanics, Metallurgy, Electrochemistry and Engineering, STP 801, T.W. Crooker and B.N. Leis, Ed., American Society for Testing and Materials, 1983, p 229-255 75. "Standard Test Method for Exfoliation Susceptibility in 2xxx and 7xxx Series Aluminum Alloys (EXCO Test)," G 34, Annual Book of ASTM Standards, American Society for Testing and Materials 76. D.L. Erich and S.J. Donachie, Benefits of Mechanically Alloyed Aluminum, Met. Prog., Vol 121 (No. 2), Feb 1982 77. "Standard Recommended Practice for Alternate Immersion Stress Corrosion Testing in 3.5% Sodium Chloride Solution," G 44, Annual Book of ASTM Standards, American Society for Testing and Materials 78. J.R. Pickens, R.D. Schelleng, S.J. Donachie, and T.J. Nichol, High Strength Aluminum Alloy and Process,

79.

80. 81. 82.

83.

84.

85. 86.

87.

U.S. Patent 4,292,079, 1981 J.R. Pickens, Techniques for Assessing the Corrosion Properties of Aluminum Powder Metallurgy Alloys, in Proceedings of Rapidly Solidified Powder Aluminum Alloys Conference, American Society for Testing and Materials, 1984 "Standard Test Method for Visual Assessment of Exfoliation Corrosion Susceptibility of 5xxx Series Aluminum Alloys (ASSET Test)," G 66, Annual Book of ASTM Standards, American Society for Metals W.J.D. Shaw, Surface Corrosion Comparisons of Some Aluminum Alloys in 3.5% NaCl Solution, in Microstructural Science, Vol 12, Elsevier, p 243-261 J.R. Pickens, Mechanically Alloyed Dispersion Strengthened Aluminum-Lithium Alloy, U.S. Patent 4,532,106, 1985; European Examiner Application No. 45622A, 1982; Japanese Examiner Application No. 57857/82, 1982 L. Christodoulou, L. Struble, and J.R. Pickens, Stress-Corrosion Cracking in Al-Li Binary Alloys, in Proceedings of the Second International Aluminum-Lithium Conference (Monterey, CA), The Metallurgical Society, 1983, p 561-579 P.P. Pizzo, R. Galvin, and H.G. Nelson, Utilizing Various Test Methods to Study the Stress Corrosion Behavior of Al-Li-Cu Alloys, in Environment-Sensitive Fracture: Evaluation and Comparison of Test Methods, STP 821, S.W. Dean, E.N. Pugh, and G.M. Ugiansky, Ed., American Society for Testing and Materials, 1984, p 173-201 P.P. Pizzo, R. Galvin, and H.G. Nelson, Stress-Corrosion Behavior of Aluminum-Lithium Alloys in Aqueous Salt Environments, in Aluminum-Lithium Alloys II, The Metallurgical Society, 1984, p 627-656 C.M. Adam and R.E. Lewis, High Performance Aluminum Alloys, in Rapidly Solidified Crystalline Alloys, S.K. Das, B.H. Kear, and C.M. Adam, Ed., Proceedings of TMS-AIME Northeast Regional Meeting, The Metallurgical Society, 1985 "Standard Method of Salt Spray (Fog) Testing," B 117, Annual Book of ASTM Standards, American Society for Testing and Materials

Corrosion of Cemented Carbides Herbert S. Kalish, Adamas Carbide Corporation

Introduction CEMENTED CARBIDES consist of hard refractory metal compounds that have a lower-melting ductile metal binder or cement (internationally, the term hardmetal is used in preference to the term cemented carbides, which is used almost exclusively in the United States). Figure 1 shows microstructures of both the basic tungsten carbide-cobalt (WC-Co) materials and materials containing titanium carbide (TiC) and tantalum carbide (TaC). Table 1 shows the physical properties of the commonly available refractory metal or hard metal carbides used to make cemented carbides. Only two-WC and TiC--are used as true base compound materials that comprise over 50% of the composition. Tungsten carbide base materials are by far the most predominant and have been in widespread use for more than 50 years. They were originally used as early as 1916 (Ref 3, 4, 5). During this time, it was found that WC could be combined with cobalt to make a high-hardness, wear-resistant, strong material. This material was initially used for wire drawing dies instead of diamond dies.

Table 1 Physical properties of carbides used in the manufacture of cemented carbides Carbide

Microhardness, kg/mm2

Melting point

°C

°F

Density, g/cm3

TiC

3200

3200

5790

4.94

VC

2950

2830

5125

5.71

HfC

2700

3890

7030

12.76

ZrC

2560

3530

6385

6.56

NbC

2400

3500

6330

7.80

Cr2C3

2280

1895

3440

6.66

WC

2080

2600

4710

15.67

Mo2C

1950

2675

4850

9.18

TaC

1790

3780

6835

14.50

Fig. 1 Microstructures of WC-Co (a, c, and e) and WC-TaC-TiC-Co (b, d, and f) cemented carbides. In a, c, and e, the white areas are cobalt binder phase. In b, d, and f, the darker, more rounded grains are the WxTayTizC cubic solid-solution phase. (a) and (b) Fine grain structures. (c) and (d) Medium grain structures. (e) and (f) Coarse grain structures. All 1500×. Source: Ref 1 and 2

The first key to the successful development of cemented carbides was that these refractory metal compounds, particularly WC, are best produced as powders. In fact, the only logical way to produce tungsten is the hydrogen reduction of WO3 or ammonium paratungstate powder into tungsten metal powder. The carburization of tungsten to WC also results in a fine powder. The second key was the discovery of the eutectic system WC-Co (Fig. 2). Liquid-phase sintering is possible well below the melting point of the WC and even below the melting point of cobalt.

Fig. 2 Quasi-binary phase diagram for the WC-Co system

Cemented WC is produced by mixing from 3 wt% or less up to as much as 30 wt% of cobalt metal powder with a balance of WC powder. The mixed powders are ball milled, generally in volatile solvents, for times ranging from a few hours to as long as 7 days. Alternatively, the powders are milled in an attritor for 1 to 10 h. A suitable transient binder is added to the powder, which is then pelletized and pressed to form the shape. Finally, the part is sintered at temperatures between 1300 and 1600 °C (2370 and 2910 °F), most often in vacuum. Because a liquid phase is formed during sintering, virtually 100% density is achieved. More information on the production of cemented carbides is available in the articles "Cermets and Cemented Carbides" and "Production Sintering Practices" in Powder Metal Technologies and Applications, Volume 7 of the ASM Handbook.

Effect of Composition on Properties The two most common variables in cemented carbides are the cobalt or binder content and the grain size. As shown in Fig. 3, increased grain size decreases hardness, and increased cobalt content also decreases hardness (Ref 6). Increased contents of cobalt or other binders, however, are necessary to increase strength. As shown in Fig. 4, strength increases with increased cobalt content; although a maximum appears to occur at about 15 to 18% Co, this is true only for transverse rupture strength (Ref 6). Very high impact strength requires very high cobalt contents (up to 25 or 30 wt%) and coarse-grain carbide. In corrosion applications, however, the binder content ranges from virtually nil (there are some socalled "binderless" compositions that actually contain 1 to 2% binder) up to about 10%, with exceptions running up to 15% binder.

Fig. 3 Effect of cobalt content and grain size on the hardness of WC-Co cemented carbides

Fig. 4 Effect of cobalt content and grain size on the transverse rupture strength of WC-Co cemented carbides

Cemented carbides are not selected for corrosion applications per se. They are extremely important in corrosion conditions in which high hardness, wear resistance, or abrasion resistance is required. When this is the case and the selection of a cemented carbide is logical, the corrosion-resistant properties are examined. For ordinary corrosion resistance, many metals and ceramics are better choices, but when wear resistance is also a requirement, the cemented carbide is needed. Binder Composition and Content. The corrosion resistance of cemented carbides is based on the two very different

components. The cobalt binder has very poor corrosion and oxidation resistance, and the WC has excellent corrosion resistance and good oxidation resistance. Alternate binders, such as nickel, have better corrosion resistance than cobalt and are used in spite of their lower hardness and strength. Nickel is a superior binder for cemented TiC and therefore is

used in all cemented TiC materials regardless of the need for corrosion resistance. In some applications, cemented TiC shows repair corrosion resistance, and in other applications, cemented WC is better. The addition of nickel to the usual cobalt binder used for WC, or the substitution of it entirely for cobalt, always improves corrosion resistance. There is, however, a sacrifice in strength, hardness, and wear resistance. A chromium addition also enhances corrosion resistance. The most important variable in the corrosion of cemented carbides is the binder content. Because the binder corrodes more than the carbide, the smaller the amount of binder the better. On the other hand, decreasing the binder decreases the strength. Carbides. Additions of TaC and TiC to the WC-Co materials are common for the compositions used for machining steel. These additives give the carbide crater resistance. Cratering on the top of a metal-cutting insert is the result of a physicochemical reaction. The addition of TaC and/or TiC will slow this reaction; indeed, it has been found that TaC also enhances the outright chemical corrosion resistance of these materials.

Other additives, such as chromium carbide (Cr2C3), molybdenum carbide (Mo2C), niobium carbide (NbC), and vanadium carbide (VC), are often added in small quantities as grain growth inhibitors. Little has been published about their effect on corrosion, but chromium has been shown to be a beneficial binder additive to WC-Ni binder compositions (Ref 7). Vanadium carbide and Mo2C will probably have a weakening effect on the strength of a WC-base hardmetal. For TiC-base hardmetals, Mo2C is invariably added to the composition, but there are no known studies of the effect of molybdenum on corrosion resistance. The molybdenum has always been added to enhance the liquid-phase sintering of the TiC-base compositions. In general, these compositions have been made for their hardness and strength characteristics, with corrosion resistance being a secondary consideration. Most rescent TiC-base compositions have titanium nitride (TiN) added, and this has been shown to improve the corrosion resistance (Ref 8). Perhaps it is not surprising that compositions developed primarily for machining should show improved corrosion resistance. In machining, there is heat with resultant oxidation and often corrosionlike mechanisms. Thus, some of the improved machining compositions also show better corrosion behavior. On the other hand, optimum corrosion resistance is obtained by tailoring the composition and amount of the binder phase. This can result in lower-strength materials with limited usefulness in machining applications. Because carbon is the basis of cemented carbides, its variation within a given composition is very important to properties and corrosion resistance. Figure 5 shows the range of carbon content allowable in the simple WC-Co compositions as cobalt content is varied (Ref 9, 10, 11). Corrosion-resistant compositions have three problems: • • •

The lower the cobalt or binder content, the better the resistance to corrosion, but this limits the safe zone, in which neither carbon porosity nor phase (hard, brittle M6C or M12C intermetallics) exist The lower the carbon content, the better the corrosion resistance, but falling into the -phase zone results in embrittlement of the material The addition of alternate binders, such as nickel, decreases the safe zone

In making corrosion-resistant cemented carbides, manufacturers must be aware of these problems and limitations. Information on the metallography and microstructures of these materials is available in the article "Cemented Carbides" in Metallography and Microstructure, Volume 9 of ASM Handbook, formerly 9th Edition Metals Handbook.

Fig. 5 Effect of cobalt content and carbon content on the phases present in WC-Co cemented carbides

Applications of Cemented Carbides The major applications of cemented carbides actually involve environments that are inherently corrosive. For example, the major use of cemented carbides is for metal-cutting (machining) applications. In these applications, extreme heat is generated whether or not coolants are used, and in those cases in which coolants are used, the corrosive attack of the coolant is a factor in the performance of the cutting tool. In general, however, very little heed is paid to this factor; cemented carbides are more often chosen for their wear resistance in such applications as mining and oil well drilling. In actuality, there is a corrosive environment to be contended with in mining (Ref 12) and oil well drilling; the natural waters and other fluids involved are often very corrosive. Other well-known examples in which cemented carbide is performing in a corrosive environment include balls for ball point pens and dental drills. In both of these examples, the corrosion resistance of the most frequently used WC-6Co composition was serendipitous. The material was selected for its wear resistance. It just happens to have good corrosion resistance in the saline and ink solutions. The dulling of cemented carbide saw tips used for sawing green or unseasoned wood is a corrosive as well as a wear phenomenon (see the section "Saw Tips and Corrosion" in this article). Examples of the use of cemented carbide in true corrosion applications include the following: • • • • • • • • • • •

Ball point pen balls Dental drills and burrs Surgical and orthodontic tweezers, pliers, and clamps Valve seats Valve balls and valve stems Valve and shaft seals (seal rings) Spray nozzles Pulverizing hammers Compressor plungers Bearings Cage mills

• •

Ball mill linings and balls Internal parts in industrial meters

The article "Cermets and Cemented Carbides" in Powder Metal Technologies and Applications, Volume 7 of the ASM Handbook contains more information on applications for cemented carbides.

Selection of Cemented Carbides for Corrosion Applications The selection of cemented carbides is a very difficult problem for the user. There has been a lack of standardization on the part of the producers, and this lack has not been answered by any national or international standards organization. Some attempts have been made to standardize with regard to metal-cutting applications. There is International Organization for Standardization (ISO) standard 513 for metal-cutting applications for carbide (Ref 13). It is widely used in Europe and most other parts of the industrial world, but it is not recognized in the United States (Ref 11). In addition, there is no ISO standard for cemented carbides used for wear, mining, or corrosion applications, and if any exist in other industrialized countries, the producers choose to ignore them, or they may be so broad that a given producer can have three or more grades falling into one category (Ref 14, 15). The producers also tend to disregard attempts at standardization in the hope of having a unique product. Even in the established WC-6Co grades, the producers offer several different varieties based on different grain size or different minor element additions. For example, the company that developed the WC-6Co composition about 70 years ago offers five different grades of this composition, and two of them have identical published properties. They are not alone. In some cases, three grades are shown with the same composition and properties. A good example is the nine differently designated 6% Co compositions of one company. Five of the nine are indeed different because of small TaC additions or grain size, but one of the compositions has three designations, and two of them have two designations. More often, the reason for the multiple designations of the same composition is that one designation is for cutting tools, another for wear parts or dies, and another for mining. Another problem area is the selection of composition by the manufacturer. For example, if one producer establishes a WC-25Co composition, another producer will make and market a grade with 24% Co, and another a product with 26% Co. Despite these problems, Table 2 lists the properties of various representative grades for corrosion applications; Table 3 lists approximate compositions and proprietary designations for a number of corrosion-resistant grades.

Table 2 Some physical properties of corrosion-resistant cemented carbide grades Properties of a carbon steel, a tool steel, and a cast cobalt alloy are included for comparison. Special attributes

Proprietary designation

Nominal composition, wt%

Hardness, HRA

Density, g/cm3

Transverse rupture strength

Abrasion resistance factor(a)

Coefficient of thermal expansion, m/m · K

Thermal conductivity

W/m · K

cal/cm·s·°C

4.9

125.5

0.30

2.1

5.2

108.8

0.26

320

3.4

5.5

104.6

0.25

2758

400

6.8

5.5

96.2

0.23

14.28

3103

450

17

6.2

87.9

0.21

88.4

13.84

3448

500

32

6.8

83.7

0.20

...

89.6

14.29

2241

325

16.5

5.8

87.9

0.21

...

...

89.0

14.09

2069

300

18

7.1

83.7

0.20

12.5

...

13.5

93.0

5.63

1379

200

22

7.8

16.7

0.04

17.5

1.0

11.0

91.7

5.71

1724

250

28

8.4

16.7

0.04

WC

Co

TaC

TiC

Ni

Cr

Mo2C

MPa

ksi

GU-2(b)

96.5

3

0.5

...

...

...

...

93.3

15.30

1655

240

1.8

PWX(b)

94.0

5.5

0.5

...

...

...

...

92.5

15.05

2137

310

A(b)

94.0

6.0

...

...

...

...

...

91.8

15.00

2206

B(b)

91.0

9.0

...

...

...

...

...

90.8

14.70

BB(b)

87.0

13.0

...

...

...

...

...

89.5

Toughness

GU-1(b)

81.5

18.0

0.5

...

...

...

...

Gall resistance

474(b)

79.0

12

9

...

...

...

GG(b)

60.0

12

28

...

...

Titan 80(b)

...

...

...

74

Titan 60(b)

...

...

...

70.5

Abrasion-resistant, wear, and structural grades

Maximum abrasion resistance

Oxidation resistance

Special corrosion resistance

Titan 50(b)

...

...

...

66.5

22.5

1.0

10.0

K602(c)

88.2

1.8

10.0

...

...

...

...

94.3

15.6

759

110

...

4.9

...

...

K701(c)

85.8

10.1

...

...

...

4.1

...

92.0

14.0

1138

165

...

6.5

62.8

0.15 (d)

K703(c)

93.3

5.8

...

...

...

0.9

...

91.5

14.7

1931

280

...

4.5

...

...

K714(c)

88.4

6.1

4.5

1.0

...

...

...

92.5

13.1

1827

265

1.8(d)

4.0

...

...

K801(c)

93.7

...

0.3

...

6.0

...

...

90.0

14.8

2103

305

17(d)

5.6

96.2

0.23(d)

K803(c)

89.0

...

...

1.0

10.0

...

...

91.0

14.4

2000

290

...

5.6

...

...

HD-15(b)

85.0

15

...

...

...

...

...

87.4

14.10

3172

460

30

6.5

83.7

0.20

HD-20(b)

80.0

20

...

...

...

...

...

85.3

13.60

3103

450

45

6.8

83.7

0.20

HD-25(b)

75.0

25

...

...

...

...

...

83.5

13.15

2965

430

65

7.5

83.7

0.20

HD-20T(b)

75.0

20

5

...

...

...

...

85.3

13.55

2896

420

46

7.1

83.7

0.20

HD-25T(b)

70.0

25

5

...

...

...

...

83.5

13.15

2827

410

67

7.8

83.7

0.20

Grades for heading and forming dies

Impact resistance

Gall resistance

Mining grades

Strength and impact resistance

575(b)

94.0

6

...

...

...

...

...

90.8

15.00

2413

350

8.1

4.9

104.6

0.25

569(b)

90.0

10

...

...

...

...

...

88.6

14.51

2930

425

13

5.8

104.6

0.25

783(b)

89.0

11

...

...

...

...

...

88.1

14.41

3103

450

19

5.8

104.6

0.25

502(b)

88.0

12

...

...

...

...

...

87.6

14.31

2965

430

21

6.2

104.6

0.25

...

...

...

...

...

...

...

...

To 79

7.8

To 1379

200

>140

14.8

50.2

0.12

Noncarbide metals

Carbon steel

(tensile strength)

T1 tool steel

...

...

...

...

...

...

...

...

To 87

8.7

3448

500

70

12.6

...

...

Cast Co-Cr-W alloy

...

...

...

...

...

...

...

...

To 83

8.6

2069

300

110

13-16

...

...

Source: Ref 16, 17 (a) Determined in accordance with ASTM B 657 (Ref 2). The lower the number, the better the resistance to abrasion.

(b) Adamas designation.

(c) Kennametal designation.

(d) Values estimated from available data.

Table 3 Representative compositions and proprietary designations of corrosion-resistant cemented carbide grades oprietary designations

Adamas Carbide

Anderson Strathclyde

Carbidie

Carmet

Danit

General Carbide

General Electric Carboloy

GTE Valenite

GU2

CA

CD20

CA8

K04

GC003

999

VC3

PWX

CF

CD24

CA306

K10

GC005

895

VC32

A

CG

CD30

CA4

K20

GC106

883

...

B

...

CD35F

CA12

K30

GC009

...

VC152

BB

...

CD40

CA10

...

GC313

258

VC11

GU1

...

CD650

...

...

...

...

...

474

...

...

...

...

...

...

VC047

Titan 8

...

...

CA100

...

...

...

VC83

Titan 60

...

...

...

...

...

...

...

Titan 50

...

...

...

...

...

...

...

...

...

...

R10

...

...

...

VC099

HD15

CPM

CD50

CA11

DG30

GC315

268

VC12

HD20

...

...

CA20

DG40

GC320

...

VC13

HD25

...

...

CA225

DG50

GC325

190

VC14

HD20T

...

CD60

...

...

...

...

...

HD25T

...

CD70

...

...

...

...

...

575

CR

...

CA3

B030

GC206

44A

...

569

CM

...

2102

B050

GC410

90

...

783

...

CD337

CA411

B055

GC411

115

...

502

CT

...

CA412

B060

GC412

120

...

Representative composition, %

Proprietary designations

WC

Co

TaC

TiC

Ni

Cr

Mo2C

Kennametal

Krupp Widia

Mefasa

Metallwerk Plansee

Mitsubishi

Sandvik

Sumitomo

Teledyne Firth Sterling

96.5

3.0

0.5

...

...

...

...

K11

THF

...

H03T

...

CS05

...

HF

94.0

5.5

0.5

...

...

...

...

K68

GT05

K1

H10T

GTi05

CS10

HI

HA

94.0

6.0

...

...

...

...

...

K6

GT10

K2

H16T

GTi10

HML

G10E

H6

91.0

9.0

...

...

...

...

...

K9

GT15

MK30

H30T

GTi15

H10F

G3

H8

87.0

13.0

...

...

...

...

...

...

GT3H

...

H40T

GTi20

R4

G5

H81

81.5

18.0

0.5

...

...

...

...

...

...

...

...

GTi40

...

...

...

...

...

...

74.0

12.5

...

13.5

K165

...

...

FO5T

NX33

CN02

...

...

...

...

...

70.5

17.5

1.0

11.0

...

...

...

F10T

NX55

...

T12A

...

...

...

...

66.5

22.5

1.0

10.0

...

TTF

...

...

...

...

T12B

...

88.2

1.8

10.0

...

...

...

...

K602

...

...

...

...

...

...

...

85.8

10.1

...

...

...

4.1

...

K701

...

...

...

...

...

...

...

93.3

5.8

...

...

...

0.9

...

K703

...

...

...

...

...

...

...

88.4

6.1

4.5

1.0

...

...

...

K714

...

...

...

...

...

...

...

93.7

...

0.3

...

6.0

...

...

K801

...

...

WC6Ni

...

...

...

...

89.0

...

...

1.0

10.0

...

...

K803

...

...

TCR30

...

...

...

...

85.0

15.0

...

...

...

...

...

SP212

BT40

G3

B50T

...

CT60

G6

MPD160

80.0

20.0

...

...

...

...

...

...

...

G4

H60T

GTi40

CT75

G7

...

75.0

25.0

...

...

...

...

...

...

GT55

G5

H70T

...

CT85

G8

...

75.0

20.0

5.0

...

...

...

...

K91

...

...

...

...

...

...

ND20

70.0

25.0

5.0

...

...

...

...

K90

...

...

...

...

...

...

ND25

94.0

6.0

...

...

...

...

...

K3404

BT10

K3

B10T

...

CT30

...

HAN6

90.0

10.0

...

...

...

...

...

K3070

BT25

MK35

B30T

...

CT45

G3

MPD10

89.0

11.0

...

...

...

...

...

K3047

...

MK40

B36T

...

CT50

...

MPD11

88.0

12.0

...

...

...

...

...

K3030

BT30

...

B40T

...

...

G5

...

Table 3 includes grades from 16 manufacturers worldwide. These are meant to be representative only in a general sense. There are well over 100 manufacturers throughout the world (over 25 in the United States alone); therefore, it is not feasible to include all. In addition, cross comparisons are not precisely possible. For example, a grade listed with an approximate composition of WC-25Co may be cross referenced with a comparable grade that contains only 24% Co. Reference 15 contains more complete data on any grade, and manufacturers can be consulted for more information. In addition to small differences in cobalt content from one manufacturer to another, there are small differences in minor additives and in grain size. For example, with the 6% Co grades, there are two basic grain size classes--fine and coarse-but these two are not precisely the same from one manufacturer to another. Some have a slightly finer or coarser size within the defined category of fine and coarse. Again, precise standards are lacking. References 1, 2, and 18 are the attempts at standardization, but they are useful only in a general sense; moreover, no producer ever refers to the specifics of these standards in designating the cemented carbide it produces. Another factor is the intentional addition of minor elements such as tantalum, titanium, vanadium, chromium, and molybdenum as grain growth inhibitors or the inadvertent introduction of minor amounts of these and other elements in the raw materials or through recycling. These elements affect hardness and strength and cannot be discounted in the selection of a cemented carbide for corrosion applications. Other processing variables also affect properties and performance. Among the important results of processing variables is the amount of porosity in the final cemented carbide product. In some cases, the porosity is negligible, and theoretical density is achieved. In other cases, porosity is present. This can be rated in accordance with ISO 4505 (Ref 19) or ASTM B 276 (Ref 20), both of which are based on the same standard photomicrographs. The ultimate in freedom from porosity is achieved by hot isostatic pressing. This operation, when carried out properly at about 138 MPa (20,000 psi) and at temperatures of 1200 to 1400 °C (2190 to 2550 °F), has no detrimental or beneficial effect on the cemented carbide except for the removal of the last vestiges of porosity. Table 4 shows some typical values of selected mechanical properties of cemented carbide grades that may also be helpful in selecting a proper composition for a particular application. Table 4 Selected mechanical properties of corrosion-resistant cemented carbide grades Proprietary designation

Poisson's ratio

Charpy V-notch impact resistance(a)

Tensile strength

Compressive strength

Modulus of elasticity

J

in.·lb

MPa

ksi

MPa

ksi

GPa

106 psi

GU-2(b)

0.21

1.24

11

1034

150

6068

880

662

96

PWX(b)

0.21

1.36

12

1241

180

5929

860

652

94.5

A(b)

0.23

1.47

13

1310

190

5516

800

648

94

B(b)

0.26

1.69

15

1586

230

4482

650

607

88

BB(b)

0.28

2.71

24

1793

260

4137

600

545

79

GU-1(b)

0.27

3.38

30

1862

270

4068

590

510

74

474(b)

0.26

1.58

14

1586

230

4206

610

538

78

GG(b)

0.26

1.47

13

1517

220

4137

600

524

76

Titan 80(b)

0.20

0.90

8

1103

160

3448

500

448

65

Titan 60(b)

0.22

1.02

9

1172

170

3275

475

414

60

K602(c)

0.21

0.23

2

...

...

5653

820

586

85

K701(c)

0.24

0.28

2.5

...

...

...

...

531

77

K703(c)

...

0.45

4

...

...

6033

875

627

91

K714(c)

0.20

0.79

7

...

...

5998

870

552

80

K801(c)

0.25

0.90

8

...

...

5275

765

621

90

K803(c)

0.23

1.36

12

...

...

5447

790

552

80

HD15(b)

0.30

2.82

25

1862

270

3965

575

531

77

HD20(b)

0.30

3.05

27

1793

260

3723

540

496

72

HD25(b)

0.30

3.28

29

1724

250

3516

510

462

67

HD20T(b)

0.27

2.94

26

1724

250

3792

550

483

70

HD25T(b)

0.27

3.16

28

1655

240

3620

525

455

66

575(b)

0.27

1.36

12

1517

220

5171

750

641

93

569(b)

0.29

1.92

17

1793

260

4309

625

579

84

783(b)

0.29

2.15

19

1862

270

4240

615

572

83

502(b)

0.29

2.37

21

1862

270

4137

600

565

82

Source: Ref 16, 17 (a) Values extrapolated from available data. Not necessarily based on actual Charpy V-notch impact tests.

(b) Adamas designation.

(c) Kennametal designation.

Corrosion in Aqueous Media The corrosion of cemented carbides is based on the solubility of the key ingredients used in the various compositions. Although some alloying occurs, the solubility of the WC or TiC in cobalt or nickel is very limited. The main alloying in the WC-Co compositions is primarily based on the addition of TiC, TaC, and NbC, which form cubic-phase solid solutions with WC. Table 5 shows the relative solubilities of the chief constituents of cemented carbides in various media. Tungsten carbide is insoluble in most acids as well as in basic and salt solutions. It is soluble only in very strong mixtures of nitric acid plus hydrochloric acid (HNO3 + HCl) and HNO3 plus hydrofluoric acid (HF). Cobalt and nickel show the same significant solubility in all acids. Even so, the nickel binder compositions show somewhat less attack in some acid solutions than the cobalt binder alloys. From this elementary information, it is obvious that the lower the binder content, the less the corrosion. Table 5 Relative solubilities in acids and bases of the basic constituents of cemented carbides Constituent

Medium and solubility(a)

Dilute HNO3

HCl

H2SO4

20HNO3-60HCl20H2O

25HNO3-25HF50H2O

Alkali solutions

Salt solutions

Cobalt

V

SI

SI

V

V

I

I

Nickel

V

SI

SI

V

V

I

I

WC

I

I

I

S

S

I

I

TaC

I

I

Sl?(b)

S

S

I

I

TiC

S

I

I

S

S

I

I

Source: Ref 21, 22 (a) Solubility: V, very soluble; Sl, slightly soluble; I, insoluble; S, soluble.

(b) Data from Ref 21 and Ref 22 are contradictory.

Corrosion of cemented carbides, therefore, is generally based on the surface depletion of the binder phase such that at the surface region only a carbide skeleton remains; because the applications are invariably for wear or abrasion, this skeleton is rapidly worn away. At low binder phase contents, the rate of attack is diminished, and in conditions in which the corrosion is not too severe, the reduced binder content will be beneficial. In more severe corrosion, however, the use of a cobalt binder is prohibited, and the WC-Co grade is simply not resistant enough. In these cases, certain corrosion-resistant grades should be used. The most common of these are WC with nickel alloy binders and TiC-Ni-Mo2C-base cemented carbide. Figure 6 shows the corrosion rate as a function of pH for these different types of cemented carbides tested in buffered solutions. These tests included a final surface wear treatment by tumbling in order to obtain a true value of the depth of the corroded surface.

Fig. 6 Corrosion rate of various cemented carbide grades as a function of pH. Source: Ref 23

As can be seen in Fig. 6, straight WC-Co grades are resistant down to pH 7. This is also valid for WC-Co grades containing cubic carbides such as TiC, TaC, and NbC. The highest corrosion resistance is obtained for certain alloyed TiC-Ni grades, which are resistant down to about pH 1, but compared to the straight WC-Co grades, they are less tough and have lower thermal conductivity. They also have the disadvantages of being difficult to grind and braze; therefore, they are used only in specific applications. In many corrosion-wear situations, the proper choice is specially alloyed WC-Ni grades, which are resistant down to pH 2 to 3. Even in certain solutions with pH values less than 2, they have proved to be resistant to corrosion. Because WC is the hard principal constituent and because nickel and cobalt are similar metals in many respects, their mechanical and thermal properties are comparable to those of the straight WC-Co grades. The pH value is one of the most important parameters when determining the corrosivity of a medium, but other factors such as temperature and electrical conductivity also have a great influence. The latter is dependent on the ion concentration, that is, the amount of dissolved salts in the solution. Thus, one cannot define the corrosivity of a certain medium in a simple way, and accordingly, no general rules that are valid in all situations can be given. However, Table 6 gives general guidelines for the corrosion resistance of WC-Co and TiC-Ni cemented carbides in various roomtemperature media. Table 7 gives compatibility data for several types of cemented carbides in aqueous media at various temperatures, and Table 8 lists weight loss as a function of cobalt content for cemented carbides in mineral acids.

Table 6 Corrosion resistance of cemented carbides in various media at room temperature Medium

Corrosion resistance(a)

WC-Co cemented carbides

TiC-Ni cemented carbides

Acid salts in water

E

E

Neutral salts in water

V

E

KOH in water

F-G

F-P

NaOH in water

V

E

NH3 in water

F

E

Weak acids

G

G-E

Distilled water

E

E

Seawater

V

E

Organic solvents, including acetone, alcohols, gasoline, benzene, carbon tetrachloride, and ethylene glycol

E

E

Alkalies

(a) Corrosion resistance: E, excellent; V, very good; F, fair; G, good; P, poor

Table 7 Corrosion resistance of cemented carbides in various media Data for two AISI austenitic stainless steels are included for comparison. Medium

Chemical designation

Concentration, %

Temperature, °C (°F)

pH

Type of cemented carbide/corrosion resistance(a)

AISI stainless steels(b)

WCCo

TiCNiMo

WCNi

WCCoCr

WCTaCCo

Type 302

Type 316

Acetic acid, unaerated

CH3COOH

4

Room

...

C

B

B

B

A

...

...

Acetic acid (glacial), unaerated

CH3COOH

99.8

Room

...

C

C

B

A

A

...

...

Acetone

(CH3)2CO

...

Room

...

A

A

A

A

A

A

A

Alcohols

...

...

Room

...

A

A

A

A

A

...

...

Ammonia, anhydrous

NH3

...

...

...

B

B

B

B

A

...

...

Argon gas

Ar

...

...

...

A

A

A

A

A

...

...

Benzene, liquid

C6H6

...

Room

...

A

A

A

A

A

...

...

Carbon tetrachloride

CCl4

Pure

Room

...

A

A

A

A

A

...

...

Chlorine gas, dry

Cl

...

Room

...

C

C

C

C

B

...

...

Chlorine gas, wet

Cl·H2O

...

Room

...

D

C

C

D

B

...

...

Citric acid

C3H4(OH)(COOH)3

5

Room

1.7

C

A

A

...

...

A

A

Citric acid

C3H4(OH)(COOH)3

5

60 (140)

1.7

D

A

B

...

...

A

A

Copper sulfate solution

CuSO4

0.01

Room

6

C

A

A

...

...

A-C

A-C

Copper sulfate solution

CuSO4

0.01

70 (160)

6

D

A

A

...

...

A-C

A-C

Digester liquor, black

...

...

66 (150)

...

B

B

B

B

A

...

...

Esters

...

...

Room

...

A

A

A

A

A

...

...

Ethanol

C2H5OH

96

Room

...

A

A

A

...

...

A

A

Ethylene glycol

C2H6O2

...

Room

...

A

A

A

A

A

...

...

Ferrous sulfide

FeS

Slurry in water

Room

...

C

C

C

C

A

...

...

Fluorine, liquid

F

...

-188 (-305)

...

50% formaldehyde, 50% alcohol

...

...

Room

...

C

Uncoupled B Coupled C(c)

C

C

A

...

...

Formic acid

HCOOH

5

Room

...

C

A

C

...

...

A

A

Formic acid

HCOOH

5

60 (140)

1.8

D

A

...

...

...

B

A

Freon gas

C2Cl3F3/CH2Cl3

...

Room

...

A

A

A

A

A

...

...

Gasoline

...

...

Room

...

A

A

A

A

A

...

...

Helium, liquid

He

...

-269 (-450)

...

A

A

A

A

A

...

...

Hydrochloric acid

HCl

0.5

Room

1

D

C

C

...

...

C

A

Hydrochloric acid

HCl

0.5

60 (140)

1

D

C

C

...

...

D

A

Hydrochloric acid

HCl

10

Room

...

D

D

D

...

...

D

C

Hydrochloric acid

HCl

37

Room

...

D

D

D

D

A

...

...

B

Hydrochloric acid

HCl

37

100 (212)

...

D

D

D

D

B

...

...

Hydrofluoric acid, anhydrous

HF

...

Room

...

B

B

B

B

A

...

...

Hydrofluoric acid

HF

1-60

Room

...

D

D

D

D

D

...

...

Hydrogen, liquid

H

...

253 (488)

...

A

A

A

A

A

...

...

Kerosene

...

...

Room

...

A

A

A

A

A

...

...

Magnesium bisulfite digester liquor

MgHSO3

...

Room

...

B

B

B

B

A

...

...

Methane, liquid

CH4

...

162 (324)

...

A

A

A

A

A

...

...

Methanol, anhydrous

CH3OH

...

Room

...

A

A

A

A

A

...

...

Methanol, 20% water

CH3OH/H2O

...

Room

...

A

A

A

A

A

...

...

Nitric acid

HNO3

0.5

Room

1.1

D

C

A

...

...

A

A

Nitric acid

HNO3

5

Room

...

D

D

D

D

B

...

...

Nitric acid

HNO3

...

100 (212)

...

D

D

D

D

B

...

...

Nitric acid

HNO3

10

Room

...

D

B

C

...

...

A

A

Nitrogen, liquid

N

...

196 (385)

...

A

A

A

A

A

...

...

Oil, crude (Sand, salt water, high in sulfur)

...

...

Room

...

C

C

C

C

A

...

...

Oxalic acid

(COOH)2·2H2O

5

Room

1

A-B

A

A

...

...

A

A

Oxalic acid

(COOH)2·2H2O

5

60 (140)

1

B-C

A

...

...

...

B

A

Oxygen, liquid

O

...

183 (361)

...

A

A

A

A

A

...

...

Perchloric acid

HClO4

0.5

Room

1.3

C-D

A

C

...

...

D

...

Perchloric acid

HClO4

0.5

60 (140)

1.3

D

A

D

...

D

D

Phosphoric acid

H3PO4

5

Room

1.2

D

B

C

...

...

A

A

Phosphoric acid

H3PO4

85

Room

...

D

C

C

D

A

...

...

Crude phthalic acid and anhydride

C6H4-1,2 (COOH)2/C6H4-1,2 (CO)2O

...

250-280 (480-535)

...

C

C

B

C

A

...

...

Sodium carbonate

Na2CO3

5

Room

12

A

A

A

...

...

A

A

Sodium carbonate

Na2CO3

5

60 (140)

12

A

A

A

...

...

A

A

Sodium chloride

NaCl

3

Room

7

A-B

A

A

...

...

A

A

Sodium chloride

NaCl

3

60 (140)

7

A-B

A

A

...

...

A

A

Sodium cyanide

NaCN

10

Room

...

D

D

D

D

A

...

...

Sodium hydrogen sulfate

NaHSO4

5

Room

1.2

C-D

A

A-B

...

...

D

A

Sodium hydrogen sulfate

NaHSO4

5

60 (140)

1.2

D

C

C-D

...

...

D

A

Sodium hydroxide

NaOH

5

Room

14

A

A

A

...

...

A

A

Sodium hydroxide

NaOH

5

60 (140)

14

B

A

A

...

...

A

A

Sodium hydroxide

NaOH

40

Room

16

A

A

A

...

...

A

A

Sodium hydroxide

NaOH

40

60 (140)

16

A

A

A

...

...

A

A

Steam, superheated

H2O

...

600 (1110)

...

A

A

A

A

A

...

...

Sulfuric acid

H2SO4

0.5

Room

1.2

C-D

A

B-C

...

...

C

A

Sulfuric acid

H2SO4

0.5

60 (140)

1.2

D

D

D

...

...

D

A

Sulfuric acid

H2SO4

5

Room

...

C

B

C

C

A

...

...

Sulfuric acid

H2SO4

5

100 (212)

...

D

C

C

D

A

...

...

Sulfuric acid

H2SO4

10

Room

0

D

D

B

...

...

D

A

Sulfuric acid

H2SO4

10

60 (140)

0

D

D

...

...

...

D

D

Sulfur, liquid

S

100

130 (265)

...

A

A

...

...

...

...

A

Water, boiler feed

H2O

...

66 (150)

...

B

C

A

A

A

...

...

Water, fresh, distilled, purified

H2O

...

Room

...

A

A

A

A

A

...

...

Water, tap

H2O

...

Room

...

B

A

B

B

A

...

...

Water, sea

...

...

Room

...

B

B

B

...

A

...

...

Source: Ref 23, 24 (a) A, highly resistant, negligible attack; B, resistant, light attack; C, poor resistance, medium attack; D, not resistant, not suitable. This table should be used only as a guide. Many factors, such as temperature variations, changes in chemical environment, purity of solutions, and stress or loading conditions, may invalidate these recommendations. Tests under operating conditions should be made.

(b) Results were obtained under laboratory conditions in pure solutions and are classified with reference to corrosion resistance only.

(c) Coupled to brass.

Table 8 Corrosion of WC-Co cemented carbides in mineral acids Corrosion rates for AISI type 304 stainless steel are shown for comparison. Cobalt content, wt %

Weight loss mg/mm2

37% HCl

5% HCl, 10% H2SO4

5% H2SO4

10% HNO3

5% HNO3

Room temperature

100 °C (212 °F)

100 °C (212 °F)

Room temperature

100 °C (212 °F)

Room temperature

100 °C (212 °F)

10 h

100 h

10 h

20 h

200 h

20 h

20 h

20 h

5.5

0.001

0.015

0.05

0.01

0.020

0.10

0.02

0.02

6

0.003

0.02

0.01

0.02

0.030

0.20

0.10

0.15

9

0.005

0.03

0.2

0.08

0.033

0.25

0.20

Destroyed

13

0.01

0.05

0.04

0.12

0.036

0.35

Destroyed

Destroyed

15

0.015

0.13

1.8

0.15

0.040

0.40

Destroyed

Destroyed

Type 304 stainless steel

1.2

Destroyed

Destroyed

1.2

0.18

Destroyed

None

None

Source: Ref 25

In general, it can be stated that the corrosion of cemented WC is fair to good in a limited way in all acids except HNO3. The corrosion resistance of cemented TiC is excellent in phosphoric acid (H3PO4), boric acid, and picric acid and is somewhat better than cemented WC in HCl or sulfuric acid (H2SO4). Cemented TiC is poor in HNO3. As expected, increasing the cobalt content to increase strength significantly decreases the corrosion resistance (Table 8). The same situation exists in virtually all corrosive environments, and because the same effect is seen for abrasion resistance, it is recommended that the minimum cobalt content be used for all wear and corrosion applications. This means that for a given application the hardest grade that will give adequate strength, impact resistance, and resistance to chipping should be chosen. Special Corrosion-Resistant Grades. To obtain corrosion resistance above and beyond that available with the

regular WC-Co and TiC-Ni grades, the special corrosion-resistant grades are used. These always result in a sacrifice in strength, hardness, and/or abrasion resistance, as shown in Table 2. On the other hand, the corrosion-resistant grades do offer significant benefits in corrosion resistance in many media (Table 7). These grades include the WC + Ni binder, the WC + Co-Cr binder, and the so-called binderless WC, which generally contains about 10% TaC and between 1 and 2% Co. In addition, there are other special grades, such as the 0.1 to 1.0% Pt addition patented as an improvement toward ink corrosion resistance in ballpoint pen balls (Ref 26). Sintered cemented carbide compositions based on more than 50% Cr2C3, for corrosion resistance are also mentioned in patents (Ref 27) and the literature (Ref 3, 4, 5). These are generally not commercially viable and are brittle materials; therefore, they cannot compete with the ceramic materials, such as silicon carbide, silicon nitride, aluminum oxide, boron nitride, and the whisker-reinforced ceramics, which have superb corrosion resistance. Where impact and chipping are not

problems, these ceramic materials are a better choice than the cemented carbides. The cemented carbides have the advantage, however, in strength, impact resistance, thermal conductivity, and often greater ease of manufacture. The best recent work showing the performance of the special corrosion-resistant compositions compared to the standard compositions and even some experimental compositions is that done at Metallwerk Plansee (Ref 7, 28). Table 9 lists the properties of these grades; for convenience, the proprietary designations are given, and the grades are also noted by composition, such as WC-10Co-4Cr. Grades are also listed by a grade number that can be used when referring to Fig. 7, 8, 9, 10, 11, 12, 13, and 14. Table 9 Properties of corrosion-resistant cemented carbide grades See Fig. 7, 8, 9, 10, 11, 12, 13, and 14 for the corrosion resistance of 12 of these grades in various media. Proprietary designation

Grade number(a)

Composition symbol(a)

Composition, wt%

Others

WC

Ni

Co

Cr

Hardness

Density, g/cm3

HV (30-gf load)

Converted to HRA

Transverse rupture strength

MPa

ksi

H03T(b)

1

WC-3Co

96.7

...

3

...

0.3TaC

1850

92.9

15.3

1400

203

H10T(b)

2

WC-6Co

94.5

...

5.5

...

(c)

1730

92.4

15.0

1900

276

H30T(b)

3

WC-9Co

90.4

...

9

...

0.2TiC, 0.4TaC

1450

90.7

14.6

2000

290

H40T(b)

...

...

88

...

12

...

(c)

1340

89.7

14.3

2600

377

K701(d)

4

WC-10Co-4Cr

85.8

...

10.1

4.1

...

1645

92.0

14.0

1140

165

WC6Ni(b)

5

WC-6Ni

94

6

...

...

...

1400

90.2

15.0

1500

218

WC9Ni(b)

6

WC-9Ni

91

9

...

...

...

1150

87.6

14.6

1800

261

TCR10(b)

7

WC-6NiCr

94

5.7

...

0.3

...

1520

91.2

14.8

2000

290

TCR30(b)

8

WC-9NiCr

91

8.5

...

0.5

...

1420

90.4

14.4

2500

363

H032(e)

...

WC-10TaC3NiCoCr

87

1.5

1

0.5

10TaC

2000

93.3

15.3

1300

189

H031(e)

13

WC-20TaC3NiCoCr

77

1.5

1

0.5

20TaC

1940

93.1

14.9

1430

207

V492(e)

9

WC-40TaC3NiCoCr

57

1.5

1

0.5

40TaC

2000

93.3

14.9

1400

203

H035(e)

10

WC-40NbC3NiCoCr

57

1.5

1

0.5

40NbC

1870

92.9

11.1

1280

186

V455(e)

11

WC-40TaC9NiCoCr

50

4

4

1

41TaC

1450

90.7

14.2

1850

268

V473(e)

12

WC-40NbC9NiCoCr

50

4

4

1

41NbC

1400

90.2

14.5

1750

254

TWF18(b)

...

...

...

18

...

...

66TiC, 16Mo2C

1470

90.8

6.0

1270

184

Source: Ref 7 and 28 (a) Used to refer to grades in Fig. 7, 8, 9, 10, 11, 12, 13, and 14

(b) Metallwerk Plansee grade designation.

(c) Depending on the reference used, these grades are sometimes shown with small additions of TaC and TiC.

(d) Kennametal designation.

(e) Metallwerk Plansee experimental designation.

Fig. 7 Corrosion resistance of cemented carbides in 22% HCl at room temperature. See Table 9 for properties of these grades, and Fig. 8, 9, 10, 11, 12, 13, and 14 for corrosion resistance in other media. Source: Ref 7 and 28

Fig. 8 Corrosion resistance of cemented carbides in 37.8% HNO3 at room temperature. See Fig. 7 for key to identification and compositions. See also Table 9 and Fig. 9, 10, 11, 12, 13, and 14. Source: Ref 7 and 28

Fig. 9 Corrosion resistance of cemented carbides in 9.8% H2SO4 at room temperature. See also Table 9 and Fig. 7, 8, and 10, 11, 12, 13, and 14. Source: Ref 7 and 28

Fig. 10 Corrosion resistance of cemented carbides in 6% acetic acid at room temperature. See Fig. 9 for key to identification and compositions. See also Table 9, Fig. 7, 11, 12, 13, and 14. Source: Ref 7 and 28

Fig. 11 Corrosion resistance of cemented carbides in 6.5% H3PO4at room temperature. See Fig. 9 for key to identification and compositions. See also Table 9, Fig. 7, 8, 9, 10, and Fig. 12, 13, 14. Source: Ref 7 and 28

Fig. 12 Corrosion resistance of cemented carbides in 4% NaOH at room temperature. See Fig. 9 for key to identification and compositions. See also Table 9, Fig. 7, 8, 9, 10, 11, and Fig. 12 and 13. Source: Ref 7 and 28

Fig. 13 Corrosion resistance of cemented carbides in 2.9% NaCl at room temperature. See Fig. 9 for key to

identification and compositions. See also Table 9, Fig. 7, 8, 9, 10, 11, 12, and Fig. 14. Source: Ref 7 and 28

Fig. 14 Resistance to erosion-corrosion of cemented carbides in a room-temperature slurry of artificial seawater and sand. See Fig. 9 for key to identification and compositions. See also Table 9 and Fig. 7, 8, 9, 10, 11, 12, and 13. Source: Ref 7 and 28

Figure 7 shows the corrosion of the 12 different compositions listed in Table 9 in 22% HCl at room temperature. Grades 1 to 3 (WC-3Co, WC-6Co, and WC-9Co, respectively) illustrate the increase in corrosion rate that results from increasing cobalt binder content. The nickel binders (grades 5 and 6; WC-6Ni and WC-9Ni, respectively) are an improvement, but again, the increase in binder content increases the corrosion rate. Of the more exotic compositions, grades 4 (WC-10Co4Cr) and 7 (WC-6NiCr) are viable choices for limited use in HCl at room temperature. The best of the experimental compositions is grade 9 (WC-40TaC-3NiCoCr); it has greater strength and higher hardness. If additional strength is needed above grade 9, grade 11 (WC-40TaC-9NiCoCr) is a good choice with the increases binder content, but as is generally the case, this results in a loss of corrosion resistance. Figure 8 shows the same type of information for 38% HNO3 at room temperature. In general, corrosion is lower, but again, the higher-cobalt WC-Co compositions (grades 2 and 3) are not suitable, nor is the WC-9Ni composition (grade 6). Grade 5 (WC-6Ni) is marginal in HNO3, but grades 1 and 4 are still better. On the other hand, the commercially available grades 7 and 8 (WC-6NiCr and WC-9NiCr, respectively) show very limited corrosion attack that is virtually equal to that of three of the four experimental grades; the commercial alloys in this case have better strength. The basic cemented carbides are attacked most severely by H2SO4 (Fig. 9). Some of the WC-Ni or WC-NiCr commercial compositions can tolerate limited use. However, the experimental grade 7 (WC-40TaC-3NiCoCr) provides exceptional corrosion resistance. Figure 10 shows that many compositions are available for use in acetic acid with little corrosion. Attack in H3PO4is relatively only on the WC-Co compositions(Fig. 11). Figures 12 and 13 show the suitability of all of the compositions listed in Table 9 in sodium hydroxide (NaOH) and sodium chloride (NaCl). In NaCl, there is significant benefit in choosing a nickel binder cemented carbide (for example, grade 5, WC-6Ni) if the loss in strength can be tolerated. Figure 14 shows the resistance to erosion-corrosion of different cemented carbide compositions in a slurry of artificial seawater and sand. It follows the pattern of benefit for the use of nickel binders in saline applications. The best of the WC-Co compositions is obviously the one with the lowest binder content (WC-3Co; grade 1). It shows a rate, however, more than 10 times greater than the experimental grade 9 (WC-40TaC-3NiCoCr), and both have the same transverse

rupture strength and equivalent hardness. For a commercial composition, the grade 9 (WC-9NiCr) cemented carbide shows excellent performance, with one-half the rate of attack of the low-cobalt composition (grade 1, WC-3Co) and much higher transverse rupture strength. Some of the same data are shown in Fig. 15 and 16 to compare the relative corrosion of the different compositions in various media. These tests were performed at room temperature, and solution concentrations are the same as those in Fig. 7, 8, 9, 10, 11, 12, 13, 14. In Fig. 16, the cemented carbides are also compared to an Fe-20Cr-32Ni alloy; the superiority of the experimental WC-40TaC-3NiCoCr cemented carbide is evident. As with corrosion test data, care must be taken not to extrapolate these to different solution concentrations and temperatures. It would be logical to assume, for example, that the WC-40TaC-3NiCoCr alloy would always outperform WC-3Co in these media at different concentrations and temperatures, but the validity of this assumption must be verified through further testing.

Fig. 15 Corrosion resistance of four commercial cemented carbide compositions in aqueous media at room temperature. Source: Ref 28

Fig. 16 Comparison of the corrosion resistance of a commercial WC-3Co cemented carbide and two experimental compositions in aqueous media. Source: Ref 28

Corrosion in Warm Acids and Bases. The corrosion rate of various cemented carbide compositions in warm (50 °C, or 120 °F) acids is shown in Table 10. The straight WC-Co compositions show rapid attack in dilute H2SO4and HNO3, and little attack in those concentrated acids. Although the corrosion rate is lower in HCl, it is obvious that these compositions are not suitable for use in warm or hot acid solutions. The TiC-6.5Ni-5Mo composition is quite good in H2SO4, moderately good in HCl, and very poor in HNO3. Several of the binderless compositions and the TaC-base cemented carbide show very acceptable corrosion resistance in these warm acids. These results are to be expected, because the cobalt and nickel binders are completely soluble in these acids.

Table 10 Weight losses of cemented carbides immersed in various acids at 50 °C (120 °F) for 72 h Composition

Weight loss, mg/cm2/d

HCl, %

H2SO4, %

HNO3, %

5

10

37

10

50

98

5

10

50

WC-6Co

2.29

2.43

0.79

8.72

2.82

0.72

13.50

1.45

0.16

WC-9Co

2.55

1.96

1.92

12.70

5.05

0.72

25.60

6.48

0.18

WC-8Ni-2Mo-3CR

0.07

0.01

+0.01

5.01

0.76

0.01

5.71

1.23

0.11

WC-5TaC

+0.02

nil

0.05

+1.03

0.23

0.02

+0.02

0.03

0.15

WC-2TaC-3TiC

0.06

nil

+0.02

+1.02

0.33

0.06

0.35

0.08

0.12

WC-47NbC-15TiC-9Ni-4Mo

1.06

0.98

0.14

5.31

0.51

0.52

8.07

1.35

0.24

TaC-4Co-3Ni-1Cr

0.09

0.02

0.22

2.03

0.41

0.51

0.09

0.05

0.01

TaC-23TiC-3Co-2Ni-1Cr

0.26

0.59

1.03

1.98

0.48

0.43

8.04

5.12

6.35

TiC-6.5Ni-5Mo

0.59

1.73

4.91

0.17

0.35

0.39

35.20

19.80

68.2

Source: Ref 29

The corrosion rates of various cemented carbides in basic solutions at 50 °C (120 °F) is quite a different matter, as shown in Table 11. Although corrosion does proceed, it is slow enough to demonstrate the utility of even the WC-Co compositions in such applications as seal rings in these basic solutions.

Table 11 Weight changes of cemented carbides immersed in NaOH, KOH, and NaOCl at 50 °C (120 °F) for 72 h Composition

Weight loss, mg/cm2/d

NaOH, %

KOH, %

NaOCl

5

10

5

10

WC-6Co

+0.75

+0.85

0.39

0.30

1.44

WC-9Co

+0.83

+0.88

0.24

0.28

2.35

WC-8Ni-2Mo-3Cr

+0.09

+0.11

0.08

0.09

1.12

WC-5TaC

+0.89

+0.92

0.18

0.18

+0.15

WC-2TaC-3TiC

+0.87

+0.90

0.20

0.20

+0.13

TaC-4Co-3Ni-1Cr

+0.71

+0.68

0.11

0.14

0.05

Source: Ref 29 Galvanic Corrosion. The resistance to galvanic corrosion of various cemented carbides coupled to AISI type 316 stainless steel has been investigated (Ref 30). Immersion testing of uncoupled specimens was also performed for comparison. Compositions of the materials tested are given in Table 12.

Table 12 Compositions and properties of galvanic corrosion test specimens Specimen

Composition, wt%

Hardness, HRA

Transverse rupture strength

MPa

ksi

WC-6Co

WC-6Co

91.0

2400

348

WC-3TiC-2TaC alloy

WC-3TiC-2TaC

92.9

1200

174

TiC-base cermet

TiC-10TiN-2.5Mo2C-15Ni

91.5

1500

218

Sintered cobalt-base alloy

Co-Cr-W-C

85.5

1400

203

WC-NiCrMo alloy

WC-3TiC-1.5(Cr3C2Mo2C)-15Ni

89.0

2100

305

Source: Ref 30

The apparatus used for the galvanic-corrosion testing is shown in Fig. 17. Figure 18 shows the corrosion rates of the materials in the immersion test. The binderless WC-3TiC-2TaC alloy performed the best, followed by the TiC-base cermet, the WC-Ni-CrMo alloy, the sintered cobalt-base alloy, and the WC-6Co alloy. The logarithm of weight loss plotted against the logarithm of time yielded the linear weight loss curves in this test. Based on this, it was postulated that the movement of electrons between cemented carbide and stainless steel is the rate-determining factor in galvanic corrosion. Table 13 compares the corrosion rates of the materials in the immersion and galvanic tests. For most of the alloy tested, the rate of galvanic corrosion is greater than the corrosion rate in the simple immersion test. It is thought that the larger the potential difference between the cemented carbide and the stainless steel, the greater the difference between the corrosion rates obtained in the immersion test and in the galvanic-corrosion test. Table 13 Corrosion rates of immersion and galvanic corrosion test specimens Specimen

Weight loss, g/m2/d

Immersion test

Galvanic test

WC-6Co alloy

4.2

16.67

WC-3TiC-2TaC alloy

0.6

0.03

TiC-base cermet

0.2

0.58

Sintered cobalt-base alloy

0.3

4.77

WC-NiCrMo alloy

0.2

1.71

Source: Ref 30

Fig. 17 Schematic of experimental apparatus used to study galvanic corrosion of cemented carbides in seawater. Source: Ref 30

Fig. 18 Corrosion weight loss as a function of time for uncoupled test specimens from Ref 30

Figure 19 shows cross sections of specimen rings after the galvanic corrosion test. Corrosion proceeded inward from the surface that contacted the seawater in the WC-6Co alloy (Fig. 19a). The investigators postulated that the electrode potential is large and that electrons would move smoothly between the cemented carbide and the contacting stainless steel; therefore, attack proceeded according to the galvanic-corrosion mechanism. In the case of the binderless WC-3TiC2TaC alloy (Fig. 19b), corrosion is very slight even after 1 year. The TiC-base cermet (Fig. 19c) shows corrosion only on the inner side surface (the side contacting the teflon; see Fig. 17). In the sintered cobalt-base alloy and the WC-Ni-CrMo alloy (Fig. 19c and d), corrosion proceeded from the corner that contacted both the seawater and the stainless steel. It was postulated that the electrode potential and the distance of electron movement were smaller than those for the WC-6Co alloy. Based on the results of these tests, either the binderless alloy or the TiC-base alloy should be acceptable for this type of application.

Fig. 19 Cross sections of galvanic-corrosion test specimens after (left to right) 1 month, 3 months, 6 months, and 12 months. (a) WC-6Co alloy. (b) WC-3TiC-2TaC binderless alloy. (c) TiC-base cermet. (d) Sintered cobaltbase alloy. (e) WC-NiCrMo alloy. Source: Ref 30

Crevice Corrosion. The same investigators also reported on the crevice-corrosion resistance of cemented carbides in

seawater with specimens of type 316 stainless steel, teflon, and silicon carbide adjacent to the cemented carbide specimens (Ref 30). Of the five compositions tested, only the WC-6Co specimen showed any significant attack after 1 year. The attack was moderate and progressed the least against the silicon carbide and the most against the stainless steel (Ref 30).

Oxidation Resistance of Cemented Carbides The ordinary WC-Co cemented carbides are reasonably resistant to oxidation in air up to about 650 to 700 °C (1200 to 1290 °F). The constituent affected the faster is WC, which will oxidize to WO3. In oxygen, the temperature limit is lower, and rapid deterioration will occur at about 500 °C (930 °F). Even in air, however, the practical temperature limit for WCCo compositions for any length of time is 500 to 600 °C (930 to 1110 °F). Nonetheless, these compositions do stand up, for example, in cutting tools in which localized higher temperatures at the cutting tip will be encountered. The addition of both or either TiC or TaC to the WC-Co compositions increases the oxidation resistance somewhat and is undoubtedly also related to the improvement found for these additions for machining steel. In applications in which oxidation resistance combined with wear resistance is required, as in hot glass forming and shearing tools, the addition of TiC and/or TaC to the basic WC-Co is of little benefit. The TiC-Mo2C-Ni compositions have clearly superior oxidation resistance and can be used at temperatures up to 900 °C (1650 °F), at which point they start to oxidize fairly rapidly. At the lower temperature, the TiC-base compositions form a tight adherent oxide film that tends to resist rapid attack. This behavior difference is analogous to the behavior difference between cobalt and nickel alone, but WC is also more readily oxidized that TiC.

Saw Tips and Corrosion Cemented carbides are in widespread use in slitter saws, which are used to saw all types of metals, composites, lumber, and many other materials. Small saw blades are sometimes manufactured from a single piece of carbide; larger blades, which may run up to 2 m (6 ft) in diameter, more commonly use cemented carbide tips brazed onto the steel saw body. The heavy-duty chain saws used in the lumber industry also have carbide teeth. Selection of cemented carbides for these

applications is invariably based on the need for excellent wear resistance and toughness. Basic WC-Co compositions are almost always used. The rapid dulling of saws in such applications, however, is attributable to corrosive as well as abrasive conditions. For example, on investigation studied the corrosion of WC cutting tools used to cut western red cedar (Ref 31). Tests were performed to determine the relative rates of attack of WC and cobalt in substance extracted from western red cedar, which has a higher content of such substances than other commercial lumber species. Because the WC was not attacked, it was concluded that the cobalt binder content should be reduced to minimize attack. Alternatively, the cobalt binder could be replaced with another binder material, such as nickel; however, such a substitution would result in a serious loss of strength. Thus, the solution to this particular problem is not a simple one, and western red cedar is still being sawed primarily with WC-Co cemented carbide compositions. It was also suggested that the carbide be coated with TiC, TiN, or Al2O3 (or a combination of these). To date, these coatings are not used in such applications because of the need for resharpening and because of the difficulties of brazing a coated tip.

Coating of Cemented Carbides This widely used process has been primarily applied to metal cutting tools. Certain special applications can be cited, such as the coating of cemented WC watch cases with TiN to form a hard, corrosion-resistant gold-colored watch case (bezel). Clearly, the potential exists to utilize these thin (2 to 10 m, or 0.08 to 0.4 mil) coatings on wear- and corrosion-resistant parts. The limitation is that the coating must be very thin to avoid spalling or chipping. In addition, because the use of a cemented carbide in a corrosion application will only be in a very high-integrity, relatively costly application, the potential danger of a coating failing or being locally breached rules out consideration in most applications. Despite this, coating of cemented carbides is an important state of the art that must be considered in special applications. Coating is most commonly done by chemical vapor deposition (CVD), and this process gives a wide range of possible coating materials. In addition to the common TiN, TiC, Al2O3, perfectly feasible coating materials include hafnium carbide (HfC), hafnium nitride (HfN), zirconium carbide (ZrC), zirconium nitride (ZrN), TaC, and NbC. The state of the art includes all combinations of TiC, TiN, Al2O3, and titanium carbonitride (TiCN), with limited commercial use of HfN and TaC as coating materials. Chemical vapor deposition is generally performed at 900 to 1100 °C (1650 to 2010 °F). Titanium nitride is coated at lower temperatures, down to perhaps 700 °C (1290 °F), in less used commercial apparatus. Physical vapor deposition (PVD) has the advantage of being done at lower temperatures, down to perhaps 500 °C (930 °F), but it is a line-of-sight process that generally requires rotation of the parts being coated. Deposition rates for PVD are much lower than those of CVD, and PVD equipment is more expensive. Physical vapor deposited coatings have also been limited commercially to TiN, usually at thicknesses of 3 m (0.12 mil) or less. More information on the corrosion and wear resistance of coatings applied by these methods is available in the article "CVD/PVD Coatings" in this Volume. Although there are few applications in which cemented carbides are used solely for corrosion resistance, it is essential to recognize the availability of the coated carbides. Coatings of TiN, TiC, or Al2O3 can impart important corrosion and oxidation resistance to cemented carbides.

Special Surface Treatments Considerable work has been done to enhance the surface properties of cemented carbides (Ref 32, 33, 34), but it generally has been derived from surface modification processes developed for other metals. These surface treatments include boriding, nitriding, and ion implantation. Most of the treatments have been used to enhance resistance to wear, abrasion, or erosion. The benefits, if any, of such treatments in increasing resistance to oxidation and corrosion are not yet well documented. Nevertheless, these processes may have potential in special applications. The article "Surface Modification" in this Volume contains information on the ion implantation and laser surface modification processes and their effects on the surface properties of metals.

References 1.

"Hardmetals--Metallographic Determination of Microstructure," ISO 4499, International Organization for

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

Standardization, 1978 "Standard Method for Determination of Microstructure in Cemented Carbides," B 657, Annual Book of ASTM Standards, American Society for Testing and Materials P. Schwarzkopf and R. Kieffer, Cemented Carbides, Macmillan, 1960 R. Kieffer and F. Benesovsky, Hartmetalle (Hard Metals), Springer Verlag, 1965 C. Goetzel, Treatise on Powder Metallurgy, Vol I to III, Interscience, 1949 H.S. Kalish, Some Plain Talk About Carbides, Mfg. Eng. Mgmt., Vol 71 (No. 1), July 1973 E. Kny, T. Bader, Ch. Hohenrainer, L. Schmid, and R. Glätzle, Korrosions-resistente, hochverschlei ß feste Hartmetalle, Werkst. Korros., May 1986 H. Suzuki et al., Choukougoukin to Syouketu Koushitu Goukin, Tokyo, Maruzen, 1986, p 514 J. Gurland and P. Bardzil, Relation of Strength, Composition, and Grain Size of Sintered WC-Co Alloys, J. Met., Feb 1955, p 311-315 H. Suzuki, Variation in Some Properties of Sintered Tungsten Carbide-Cobalt Alloys With Particle Size and Binder Composition, Trans. Jpn. Inst. Met., Vol 7, 1966, p 112 H.S. Kalish, Carbide Grade Classifications--What They Mean, Mfg. Eng. Mgmt., Vol 76 (No. 1), 1976 R.S. Montgomery, The Mechanism of Percussive Wear of Tungsten Carbide Composites, Wear, 1968, p 309-329 "Application of Carbides for Machining by Chip Removal," ISO R 513, 1st ed., UDC 621.9.027, International Organization for Standardization, Nov 1986 K.J.A. Brookes, "Metalworking Production's Guide to Hardmetals for Machining," Application Code, International Organization for Standardization, 1984 K.J.A. Brookes, World Directory and Handbook of Hardmetals, 3rd ed., 1982 Published and unpublished data, Adamas Carbide Corporation Properties and Proven Uses of Kennametal Hard Carbide Alloys, Kennametal, Inc., 1976 "Evaluating Apparent Grain Size and Distribution of Cemented Tungsten Carbide," B 390, Annual Book of ASTM Standards, American Society for Testing and Materials "Hardmetals--Metallographic Determination of Porosity and Uncombined Carbon," ISO 4505, International Organization for Standardization, 1978 "Standard Test Method for Apparent Porosity in Cemented Carbides," B 276, Annual Book of ASTM Standards, American Society for Testing and Materials R.C. Weast, Ed., Handbook of Chemistry & Physics, 67th ed., CRC Press, 1986 E.K. Storms, Refractory Carbides, Academic Press, 1967 Published data, Sandvik Coramant Published data, Kennametal, Inc. A. Hara and Y. Saito, Corrosion and Oxidation Resistance of "Igetalloy," Sumitomo Elec. Tech. Rev., No. 13, Jan 1970 Corrosion Resistant Binder for Tungsten Carbide Materials and Titanium Carbide Materials, U.S. Patent 3,628,921, 1971 Sintered Alloys of a Chromium Carbide Tungsten Carbide--Nickel System, U.K. Patent 1202844, 1970 E. Kny and L. Schmid, New Hardmetal Alloys With Improved Erosion and Corrosion Resistance K. Takao and K. Terasaki, Chemical Resistance of Various Cemented Carbides, Nipon Tungsten Rev., Vol 10, 1977 Y. Masumoto, K. Takechi, and S. Imasato, Corrosion Resistance of Cemented Carbide, Nippon Tungsten Rev., Vol 19, 1986 E. Kirbach and S. Chow, Chemical Wear of Tungsten Carbide Cutting Tools by Western Redcedar, Forest Prod. J., Vol 26 (No. 3) H. Ito and Y. Mohashi, Corrosion-Resistant Cemented Carbides by Chromium Diffusion Methods, Nippon

Tungsten Rev., Vol 6, Sept 1973 33. H.S. Kalish, Method of Forming a Hard Surface on Cemented Carbides and Resulting Article, U.S. Patent 3,744,979, 1973 34. Materials Development Corporation, Medford, Mass. Selected References • •

• • • •

O.A. Drobysheva and V.N. Latyshev, Interaction of a Hard Alloy and a Cutting Fluid, Fiz. Khim. Mekh. Mater., Vol 8 (No. 3), May-June 1972, p 38-40 A.L. Echtenkamp, "Combating Corrosion/ Wear With the Hard Carbide Alloys," Paper presented at the ASLE/ASME Lubrication Conference, Minneapolis, MN, American Society of Lubrication Engineers, Oct 1978 E. Suganuma, Electrochemical Behaviour of Cemented Carbide, Bull. Yagamata Univ. Eng., Vol 11 (No. 2), March 1971 G. VerWeyst, Corrosion-Resistant Tooling for Metal-Forming Operations, Lubr. Eng., Vol 41 (No. 6), June 1985, p 370-374 V.A. Zhilin and V.M. Druzhinin, Corrosion-Induced Erosion of Hard Alloy Tools During Machining of High-Strength Steels, Fiz. Khim. Mekh. Mater., Vol 6 (No. 4), July-Aug 1970, p 57-62 V.A. Zhilin and V.M. Durzhinin, Electrochemical Corrosion of Hard Alloys, Poroshk. Metall., Aug 1970, p 68-71

Corrosion of Metal Matrix Composites Denise M. Aylor, David Taylor Naval Ship Research and Development Center

Introduction METAL MATRIX COMPOSITE (MMC) materials have developed substantially over the past 20 years. Developmental efforts have involved aluminum-, copper-, magnesium-, titanium-, and lead-base MMCs, but the primary emphasis has been on aluminum-base materials. Interest in the use of these composites is due to the higher attainable strength and stiffness properties as compared to materials prepared by conventional alloying. However, successful application of MMCs in the marine environment requires adequate corrosion resistance. This article will discuss the ambient-temperature corrosion characteristics of MMCs that have potential application in marine environments, with primary emphasis on aluminum-base composites. Structural characteristics, design criteria, and coatings for optimum protection of MMCs will also be discussed. More information on various types of MMCs is available in Volume 1 of the 1st Edition of the Engineered Materials Handbook.

Structural Characteristics Metal matrix composites basically consist of a nonmetallic reinforcement incorporated into a metallic matrix. Reinforcements, characterized as either continuous or discontinuous fibers, typically constitute 20 vol% or more of the composite. Reinforcements in continuous-fiber composites include graphite (Gr), silicon carbide (SiC), boron (B), or aluminum oxide (Al2O3). Fabrication techniques for these composites vary from chemical vapor deposition (CVD) coating of the fibers, liquid-metal infiltration, and diffusion bonding to liquid-metal infiltration and direct casting to nearnet shape. Discontinuous-fiber composites consist almost exclusively of SiC in whisker (w) or particulate (p) form. These MMCs are produced by using modified powder metallurgy techniques (Ref 1). Figure 1 shows cross sections of typical continuous and discontinuous reinforced MMCs.

Fig. 1 Cross sections of typical fiber-reinforced MMCs. (a) Continuous-fiber reinforced graphite/aluminum composite. (b) Discontinuous silicon carbide(p)/aluminum composite. (c) Continuous-fiber silicon carbide/aluminum composite

Corrosion Behavior of Aluminum-Base MMCs Graphite/aluminum composites exhibit accelerated corrosion in marine environments when graphite fibers and aluminum are simultaneously exposed. Assuming that the edges of the graphite/aluminum composite are masked off to prevent exposure of both the graphite and the aluminum, only the aluminum surface foils will initially be exposed to the environment. The aluminum surface foils will pit at an average rate of 0.025 to 0.035 mm/yr (1.0 to 1.4 mils/yr) in seawater and at 0.5 to 0.76 m/yr (0.02 to 0.03 mils/yr) in the marine atmosphere (1100, 6061, and 5000 series aluminum

alloys). Pits may also be present with depths much greater than the average rates reported (Ref 2). Crevice corrosion of the aluminum foils may also occur at the edges because of the crevice formed between the aluminum surface foil and the masking material. The pitting and crevice corrosion processes eventually penetrate the foils and result in exposure of the graphite/aluminum composite matrix below, at which point the corrosion rate becomes extremely accelerated. Corrosion has been shown to proceed preferentially along foil/foil, wire/wire, and wire/foil interfaces in the composite (Ref 3). Severe exfoliation occurs because of wedging of the hydrated alumina (Al2(OH)3) corrosion products within the composite. Figure 2 shows an example of severe graphite/aluminum corrosion (known as catastrophic failure). This catastrophic condition can occur within 30 days in seawater after exposure of the graphite-aluminum matrix. Catastrophic failure in the marine atmosphere and in splash/spray environments is less rapid than in seawater, but can occur within 6 months (Ref 4). This accelerated corrosion is believed to result from the aluminum carbides that are formed at the reinforcement/matrix interface during fabrication, which alter the properties of the aluminum surface film at these locations and render the composite more susceptible to breakdown (Ref 3, 5).

Fig. 2 Catastrophic failure of a graphite/aluminum MMC after 6 months in a marine atmosphere

The aluminum surface foils alone provide reasonably good corrosion protection to the composites. Marine exposure tests of graphite/aluminum MMCs with 6061, 5056, and 1100 aluminum alloy surface foils (graphite/aluminum edges masked) revealed no pitting penetration through the foils to expose the graphite/aluminum composite wires below during a 20month exposure (Ref 4). Pitting of the foils, which occurred on most of the graphite/aluminum panels, ranked as light pitting in the splash/spray zone and marine atmosphere and as localized pitting in filtered seawater. In summary, graphite/aluminum composites undergo extremely severe corrosion in marine environments when the graphite and the aluminum are mutually exposed. Aluminum surface foils have provided 20 months of protection to MMCs, assuming there is no graphite-aluminum exposure. However, the composite will start to fail upon foil penetration by the deepest pit. Service life can be extended by applying corrosion-resistant coatings. Primary emphasis should be placed on preventing exposure of both the graphite and the aluminum, and the graphite/ aluminum composite should be frequently inspected while the component is in service. Silicon Carbide/Aluminum Composites. Marine corrosion of silicon carbide/aluminum composites is much less severe than that observed on graphite/aluminum MMCs. Discontinuous silicon carbide/aluminum MMCs, however, are susceptible to localized corrosion. Mild-to-moderate pitting has been reported on SiC whisker- and particulate-reinforced composites containing 6061 and 5000 series aluminum matrices exposed for a maximum of 42 months in splash/spray and marine atmospheric environments. The degree of corrosion present on the composites is slightly accelerated compared to that on unreinforced aluminum alloys.

Silicon carbide/aluminum composites immersed in natural seawater are susceptible to significantly more severe corrosion than is typical for silicon carbide/aluminum MMCs in the aforementioned environments (splash/spray and atmosphere). Silicon carbide/aluminum panels in seawater undergo pitting, both localized at the edges and distributed uniformly across

the surface. The extent of pitting varies from minimal attack through 33 months of exposure to extensive corrosion that is equivalent to a rate as high as 0.25 mm/yr (9.8 mils/yr). Corrosion rates for silicon carbide/aluminum MMCs in seawater are also generally higher than is typical for unreinforced aluminum alloys. This was documented in Ref 4 for discontinuous SiC in 6061 and 5000 series aluminum matrices and in Ref 6, which reports than silicon carbide/2024 aluminum corroded approximately 40% faster than 2024 aluminum in sodium chloride (NaCl) solution. Contrary to these findings, in another study, little difference in weight loss measurements was noted between silicon carbide/6061 aluminum and 6061 aluminum in NaCl (Ref 7). Discontinuous silicon carbide/aluminum MMCs are believed to corrode at the silicon carbide/aluminum interfaces (Ref 4, 6, 7). Concentration of the corrosion at these interfaces is presumably due to the crevices formed there, which are preferential sites for pitting. Evidence of the pitting concentrated at the silicon carbide/aluminum interfaces in both whisker and particulate composites is shown in Fig. 3.

Fig. 3 Cross sections of discontinuous silicon carbide/aluminum MMC panels. (a) Silicon carbide(p)/6061 aluminum MMC after a 230-day, tidal-immersion exposure. (b) Silicon carbide(w)/6061 aluminum MMC after a 60-day filtered-seawater exposure

Electrochemical studies of discontinuous silicon carbide/aluminum MMCs containing 6061 and 5000 series aluminum alloy matrices demonstrated that the presence of the SiC does not increase the susceptibility of the composite to pit initiation (Ref 5, 8). Research on silicon carbide/2024 aluminum did show a more electropositive pitting potential for the composite relative to the 2024 aluminum (Ref 8); however, this difference in pitting potential may be due to difference in microstructure between the composite matrix and the 2024 aluminum (Ref 1). Continuous-fiber silicon carbide/aluminum composites also undergo localized corrosion (Ref 4). These composites are susceptible to both crevice corrosion and pitting. Seawater entry into the silicon carbide/aluminum composite matrix will result in crevice corrosion at the fiber/matrix interfaces, which accelerates the corrosion rate and eventually results in delamination of the aluminum surface foils. However, the rate of silicon carbide/aluminum corrosion is much less severe

than is typical for graphite/aluminum. Figure 4 contrasts the extent of corrosion evident on the silicon carbide/aluminum panels described above.

Fig. 4 Silicon carbide/aluminum, MMC panels after exposure to filtered seawater. (a) Silicon carbide(w)/6061 aluminum after a 4-month exposure. (b) Silicon carbide(p)/6061 aluminum after a 24-month exposure. (c) Silicon carbide (continuous fiber)/6061 aluminum after a 33-month exposure

In summary, silicon carbide/aluminum MMCs are susceptible to localized corrosion in marine environments. Generally, the susceptibility to pit initiation is similar for composites and unreinforced alloys; however, the rate of pit propagation is higher for composites. Silicon carbide/aluminum corrosion in seawater is increasingly more severe that in the other marine environments. Corrosion-resistant coatings are recommended for these composites to enhance their service lives. Boron/Aluminum Composites. The corrosion properties of boron/aluminum composites are extensively reviewed in

Ref 1. This section will summarize the significant findings. Boron/aluminum MMCs experience severe corrosion in chloride environments and are significantly less corrosion resistant than unreinforced aluminum alloys. The concentration of corrosion in these composites has been found at fiber/matrix interfaces and at the bonds between foils (Ref 9, 10). The accelerated corrosion at these sites has been attributed to imperfect bonding and fissures in the composite and emphasizes the need for eliminating fabrication flaws to reduce corrosion of boron/aluminum MMCs in chloride environments (Ref 9). Corrosion at the fiber/matrix interfaces has also been attributed to the presence of aluminum boride formed during fabrication (Ref 10). Aluminum Oxide/Aluminum Composites. The Corrosion properties of aluminum oxide/aluminum composites are

reviewed in Ref 1. The significant findings are summarized below. To obtain good wettability and bonding in aluminum oxide/aluminum MMCs, the aluminum matrix is alloyed to form a bonding compound between the fiber and the matrix. Corrosion studies of Al2O3/Al-2Li MMCs (containing a Li2O5Al2O3 bond layer) in NaCl solutions indicated no severe attack at the fiber/matrix interfaces. The corrosion rate of the MMCs (based on weight loss measurements) was only slightly higher than for aluminum alloy 6061-T6 (Ref 11). Corrosion evaluations of Al2O3/Al-2Mg MMCs identified pitting at the fiber/matrix interfaces, presumably due to the Mg5Al8 precipitated there during fabrication (Ref 12). Research on aluminum oxide/6061 aluminum MMCs also reported preferential corrosion at the fiber/matrix interface (Ref 1). These findings suggest that the corrosion resistance of aluminum oxide/aluminum composites is highly dependent on the bonding compound formed at the fiber/matrix interface. To date, no severe corrosion problems have been identified with Al2O3/Al-Li composite. The stress-corrosion cracking (SCC) properties of graphite/aluminum MMCs are discussed in Ref 13 and 14.

Based on evaluation of a limited number of specimens, an initial stress-dependent corrosion mechanism was reported for graphite/aluminum, aluminum, which then shifted to a corrosion-dominated failure as the exposure in seawater increased (>100 h) (Ref 13). In another study, a corrosion-dominated mechanism was also noted at longer exposure times, but it was suggested that the failures were creep related as well (Ref 14). Stress-corrosion cracking testing of boron/aluminum MMCs at lower stress intensities (105 K/s) are required to retain the highly metastable glassy state. Although the temperature decrease in quenching from the liquid to the solid is not large, the rate of heat extraction is very high and requires at least one dimension of the resulting alloy to be very thin. Because of this requirement, glassy metals produced by liquid quenching are typically in the form of ribbons, wires, and filaments. Techniques for producing glassy metals can be divided into three main groups (Ref 1). The first, which is termed quenching, involves rapid solidification from the melt under a set of conditions, such as cooling rate and sample dimensions, that precludes the formation of the stable equilibrium structure. The second technique, termed atomic or molecular deposition, involves growth from the vapor phase, such as thermal evaporation and sputtering, or from a liquid phase, such as electroless deposition and electrodeposition, to form the desired alloy. These techniques have higher effective cooling rates than liquid quenching processes; therefore, they allow the formation of glassy alloys that cannot be produced by rapid liquid quenching. The third set of techniques is classified as external action techniques, and they rely on such procedures as solid deformation and irradiation to form the metastable glassy alloy. Ion implantation and ion beam mixing, for example, can produce amorphous surface layers on bulk crystalline substrates. The latter two groups of techniques--molecular deposition and external action techniques--have the advantage of being able to produce considerably thicker alloys, but they typically require considerably more time for completion of the process. Some glassy metals exhibit extremely good corrosion resistance because of several factors. Glassy metals are free from such defects as the grain boundaries and second-phase particles that are present in crystalline metals. Corrosion often occurs preferentially at such sites; therefore, glassy metals might be expected to exhibit better corrosion resistance than crystalline alloys. The galvanic corrosion associated with chemical inhomogeneities, such as secondphase particles, is also impossible in glassy metals. In addition, the passive films responsible for corrosion resistance in crystalline alloys also play a role in glassy metal corrosion. Thus, the effect of the amorphous structure, chemical homogeneity, and unique chemical composition on the formation and stability of the passive film must also be considered. Any corrosion reaction involves two or more partial anodic and cathodic half-cell reactions (Fig. 1). Two sets of curves are shown in Fig. 1. The bottom curves represent metal dissolution and plating, and the top curves are for proton reduction and hydrogen oxidation. The reversible potential, Eo, represents equilibrium between the oxidized and reduced species. For the metal, the reaction is M Mn+ + ne-. The current density M Mn+ + ne- at Eo is termed the exchange current density, io, and is a characteristic of the metal. The cathodic half cell in acid solutions is typically proton reduction, which is represented by the reaction 2H+ + 2eH2. In the case of the cathodic reaction, io is a function of the surface on which the proton reduction reaction occurs. The corrosion rate of a metal in deaerated acid is represented by the intersection of the metal dissolution and hydrogen reduction curves. This intersection establishes the corrosion potential, Ecorr, and the corrosion current, icorr, because the anodic and cathodic partial currents are equal at that point. The net result is that the metal corrodes and hydrogen is evolved simultaneously at the metal surface.

Fig. 1 Schematic Evans diagrams showing the possible influence of alloy structure and composition on the corrosion rate, icorr, and corrosion potential, Ecorr. See text for discussion. Source: Ref 2

There are several possibilities for explaining the difference in corrosion behavior between amorphous and crystalline metals. A metal-metalloid glassy metal typically contains of the order of 20 at.% metalloids. The exchange current density for the hydrogen reduction reaction may be lower on the metalloid surface than on the metal (Fig. 1b). If it is assumed that the anodic kinetics do not change, the Evans diagram predicts a decrease in icorr and a more active Ecorr. A more likely possibility is that the anodic kinetics do in fact change with structure and composition. Because dissolution occurs preferentially at active sites, including kinks, ledges, steps, and grain boundaries, and because there are fewer long-range defects in an amorphous alloy, a lower exchange current density might be expected for metal dissolution in the case of amorphous alloys (Fig. 1c). In this case, the decrease in icorr is accompanied by a shift in Ecorr in the noble direction. If amorphous alloys are compared with crystalline alloys with different chemical compositions ( 20% metalloids for the glassy metal versus none for the conventional crystalline alloy), a shift in the reversible potential for metal oxidation would be expected. If this shift is in the noble direction and if all other factors remain constant, a decrease in icorr would be expected, but a shift in the active direction would tend to increase the corrosion rate. The above analysis is for a metal that is dissolving under activation control and at open-circuit potential. With systems containing film formers, the effect of the passive film and the interaction between metalloids and the film formers must be included, and the above analysis may not be appropriate.

Corrosion Behavior: A Historical Review The first published information on the corrosion behavior of metallic glasses appeared in 1974 (Ref 3), and it concerned the Fe-Cr-P-C alloy system. Figure 2 shows the corrosion rates of Fe70Cr10P13C7 and Fe65Cr10Ni5P13C7 metallic glasses and a typical AISI type 304 stainless steel in hydrochloric acid (HCl) of various concentrations at 30 °C (85 °F). It also includes data from Ref 4 obtained under similar test conditions. The corrosion rates, calculated from gravimetric measurements, were relatively large for the stainless steel because of pitting attack, but the rates for the metallic glasses were so low that they could not be detected even after immersion for 168 h. This early work illuminated the distinct differences in corrosion behavior between crystalline stainless steel and iron-base metal-metalloid glasses.

Fig. 2 Comparison of the corrosion rates of metallic glasses and crystalline stainless steel as a function of HCl concentration at 30 °C (85 °F). No weight changes of the metallic glasses of Fe70Cr10P13C7 were detected by a microbalance after immersion for 200 h. Open/closed circles (Ref 3); open/closed squares (Ref 4)

Other early research includes work that appeared in 1976 (Ref 5, 6). Figure 3 shows the relative corrosion rates in 1 N sodium chloride (NaCl) of crystalline iron-chromium alloys as compared to those of glassy Fe-CrxP13C7 alloys, where x ranges between 2 and 10 at.%. The corrosion rates of the crystalline alloys were about 0.6 mm/yr (24 mils/yr) and were largely unaffected by chromium content, which is the result of pitting corrosion. Conversely, the glassy alloys exhibited a sharp decrease in corrosion rate with increasing chromium content, with an undetectable rate occurring above 8 at.% Cr. Pitting did not occur on the glassy alloys, even those with only a few atomic percent of chromium.

Fig. 3 Comparison of the corrosion rates of glassy Fe-CrxP13C7 alloys and crystalline iron-chromium alloys in 1 N NaCl solution at 30 °C (85 °F). Source: Ref 5

In another study, the anodic polarization behavior of glassy Fe25Ni40Cr15P16B4 and Fe40Ni40P16B4 alloys was compared in sulfuric acid (H2SO4) with and without NaCl additions. The presence of chromium facilitated passivation over a broad potential range. Thermal crystallization of the metallic glasses caused the corrosion rates during anodic polarization to increase sharply, especially in the presence of chloride ion (Cl-). It was concluded that crystallization probably decreased corrosion resistance by introducing chemical and structural heterogeneities into the alloys. Studies such as those summarized above emphasized the excellent resistance to uniform and localized corrosion that could be obtained with certain types of metallic glasses. Results of these studies stimulated additional research into broader compositional ranges. Research during the late 1970s focused primarily on the transition metal-metalloid compositions, although some work was also initiated on metal-metal systems, such as copper-zirconium. Regarding the former compositional class of glassy alloys, research addressed the effects of phosphorus, boron, silicon, and carbon, which are added to stabilize the glassy structure. These additive elements strongly influence the corrosion behavior of glassy alloys, as shown in Fig. 4 for Fe70Cr10B13X7 and Fe70Cr10P13X7 alloys. Specifically, in acidic solutions, the corrosion rates of the alloy system containing phosphorus as the major metalloid are two orders of magnitude lower than those of the alloy system with boron as the major metalloid. In addition, the corrosion rate of the glassy iron-chromium alloy progressively decreased by the addition of silicon, boron, carbon, and phosphorus in 0.1 N H2SO4. The addition of chromium without phosphorus to the glassy alloys is relatively ineffective in reducing corrosion rates, as evident from the Fe70Cr10B13C7 and Fe70Cr10B13Si7 alloys. Thus, phosphorus was identified as the single most effective metalloid element among phosphorus, carbon, silicon, and boron for improving the corrosion resistance of glassy iron-base alloys containing chromium. The combination of metalloids that is most effective in providing corrosion resistance in glassy iron-chromium alloys is phosphorus and carbon.

Fig. 4 Average corrosion rates estimated from the weight loss of amorphous Fe70Cr10B13X7 and Fe70Cr10P13X7 alloys in 0.1 N H2SO4 at 30 °C (85 °F), where (a) X is silicon, boron, carbon, and phosphorus, and (b) X is silicon, boron, and carbon. Source: Ref 7

It was also recognized that the corrosion behavior of glassy alloys is strongly influenced by additions of metallic elements, especially those that form films on the alloy surface, that is, film former additions. Figure 3 shows an early example of the strong beneficial effect of chromium additions to an iron-base glassy alloy. The effect of chromium content on the corrosion rates of glassy Ni-Cr-P15B5 alloys in 10% ferric chloride (FeCl3) is apparent in Fig. 5, which shows that an undetectably small corrosion rate was attained with 7 at.% Cr. A large variety of other metal additions have also been investigated. For example, such elements as titanium, manganese, niobium, vanadium, tungsten, and molybdenum can benefit the corrosion resistance of Fe-Cr3P13C7X alloys in 1 N HCl (Ref 9).

Fig. 5 Effect of the chromium content on the corrosion rates of amorphous Ni-Cr-P15B5 alloys in 10% FeCl3·6H2O at 30 ± 1 °C (85 ± 2 °F). The corrosion rate was estimated from weight loss during immersion for 168 h. Source: Ref 8

Corrosion research involving metal-metalloid system soon led to research with metal-metal glasses. One study characterized the corrosion behavior of copper-zirconium and copper-titanium alloys in H2SO4, HCl, nitric and (HNO3), and sodium hydroxide (NaOH) (Ref 10). In all of the solutions except NaOH, the crystalline and glassy copper-titanium alloys exhibited corrosion rates lower than those of pure copper, and in all cases, the corrosion resistance of the glassy alloy was better than that of the crystalline alloy. The glassy alloys in these compositional systems are not unusually corrosion resistant; in fact, neither the crystalline nor glassy forms of the alloys were more corrosion resistant than pure titanium or pure zirconium. This fact suggests that the corrosion resistance of the glassy alloys is the result of the presence of the passivating element (titanium or zirconium), not the presence of the glassy state. In another investigation, several alloys in the Cu-Zr system were examined in H2SO4 electrolyte (Ref 11). It was shown that the copper-zirconium alloys remained resistant to corrosion regardless of whether they were devitrified to a singlephase or a multiphase equilibrium microstructure; however, the glassy state was about 20% more corrosion resistant than the devitrified state. Since the early work with iron-base metal-metalloid glasses, the field of study has been extended to include many alloy systems. Results with nickel-, titanium-, copper-, and cobalt-base alloy systems, among others, have been reported in the literature. The effect of metalloid additions on corrosion behavior is reasonably well characterized, and theories have been proposed to explain the beneficial effect of phosphorus on corrosion (Ref 7). The influence on corrosion behavior of a wide variety of elemental additions has been evaluated, and many such additions increase corrosion resistance; those with the strongest effect are the classical film formers, such as chromium, titanium, and molybdenum. Research on glassy alloy corrosion in the past few years has expanded to include developing means for using these alloys in practical applications. Therefore, research has accelerated in such areas as laser surface remelting, ion implantation, sputtering, electrodeposition, and chemical vapor deposition. See, for example, the articles "Surface Modification" (ion implantation and laser surface processing are discussed) and "CVD/PVD Coatings" in this Volume.

General Corrosion Behavior Glassy alloys can be grouped into two major categories with intrinsically different corrosion behaviors. The first group includes the transition metal-metal binary alloy systems, such as Cu-Zr, Ni-Ti, W-Si, and Ni-Nb. The second class consists of transition metal-metalloid alloys. These alloys are usually iron-, nickel-, or cobalt-base systems, may contain film formers (such as chromium and titanium), and normally contain approximately 20 at.% P, B, Si, and/or C as the metalloid component. Transition Metal-Metal Binary Alloys. Research in the transition metal-metal systems indicates that corrosion

resistance is primarily determined by the behavior of the more corrosion-resistant component of the alloy. For example, in an investigation of the corrosion behavior of copper-zirconium alloys, the potentiodynamic polarization behavior of the alloys exhibited characteristics of both components, but it was not superior to the more passive material (zirconium) (Ref 2). This work also examined the effect of alloy structure. By choosing the proper alloy composition (Cu60Zr40), the researchers devitrified the glassy alloy to a single-phase Cu10Zr7 and found a slight improvement in corrosion resistance when the material was in the glassy state. They worked with a series of copper-zirconium alloys and found that whether or not the alloy forms a single-phase or multiphase alloy upon devitrification has very little effect on corrosion resistance. This lead to the conclusion that, at least in this alloy system, structure plays a secondary role in establishing corrosion resistance. Other alloys, including copper-titanium, copper-zirconium, and nickel-titanium, have demonstrated that the corrosion behavior of the transition metal-metal class of glassy metals is determined almost completely by the more corrosionresistant component. Thus, the corrosion resistance in the transition metal-metal alloy systems is apparently the result of the presence of a passivating element, not the glassy structure. Transition Metal-Metalloid Alloys. The second class of glassy alloys consists of transition metal-metalloid elements.

This family includes iron-, nickel-, and cobalt-base alloys containing combinations of phosphorus, boron, carbon, and silicon as the metalloid constituents. These alloys can be formed as binary systems, such as Ni-P and Fe-B, or they may be considerably more complex, such as Fe-Ni-P-B quaternary systems. In addition to the base metal, they often contain appreciable concentrations of film formers to promote passivity--for example, the Fe-Ni-Cr-P-B system. They derive their corrosion resistance from the same type of processes as crystalline alloys, namely the development of a passive film. The

significant difference between corrosion-resistant glassy alloys and their crystalline counterparts, such as stainless steels, is that the level of chromium necessary to promote passivity can be considerably less in the glassy alloys. Figure 3 shows a comparison between the corrosion rates of crystalline iron-chromium alloys and amorphous ironchromium-phosphorus-carbon alloys as a function of chromium concentration. At low chromium levels, the amorphous alloy corrodes at a higher rate than the crystalline material. However, at slightly higher chromium levels (4 at.%), there is a significant decrease in the corrosion rate of the glassy alloy, but the crystalline material is essentially unchanged. At an intermediate level of 8 at.% Cr, no corrosion of the glassy alloy was detected by weight loss experiments after immersion for 168 h. It was also found that the concentration of HCl, which has a profound effect on corrosion behavior of crystalline, alloys, had no effect on corrosion of the glassy Fe-Cr-P-C or Fe-Ni-Cr-P-C alloy systems, which exhibited no weight loss after exposure for 168 h (Ref 5). Figure 6 shows the effect of chromium concentration on the corrosion behavior of iron-, nickel-, and cobalt-base alloys. In all cases, the corrosion rate decreases with increasing chromium concentration and becomes vanishingly small at some level of chromium. In addition, the Fe-Cr-P-C alloy system, which exhibits the highest corrosion rate at low chromium contents, exhibits no weight loss in immersion tests with a chromium concentration of as little as 8 at.%. This behavior supports the theory that, when an alloy contains a strong film former, the higher the initial reactivity of the alloy, the more rapidly the film former can be accumulated at the interface and the more rapid the rate of passivation (Ref 9).

Fig. 6 The influence of chromium content on the corrosion rates of iron-, cobalt-, and nickel-base alloys in 1 N HCl. Source: Ref 12

The corrosion behavior of nickel-base (Ref 8) and cobalt-base (Ref 13) glassy alloys is very similar to that of the ironbase systems, and it is also a strong function of chromium concentration. Figure 7 shows the corrosion rate for a nickelbase glassy alloy as a function of chromium concentration. These data, which are very similar to the data for the iron-base system, show that at a concentration of about 7 at.% Cr the alloy is extremely resistant to corrosion in a FeCl3 solution.

Fig. 7 Plot of corrosion rates of metallic glasses of Ni-Cr-P15B5 in 10 wt% FeCl3·6H2O at 30 °C (85 °F) versus chromium content

In one investigation, ion implantation was used to make glassy iron-chromium-phosphorus alloys in which the chromium level varies from 6 to 18 at.% (Ref 14). An interaction between chromium and phosphorus was observed, which suggests a mechanism for the passivation of these amorphous alloys. Specifically, at low chromium levels, phosphorus implantation degrades passivity and induces pitting. At high chromium concentrations, there is a slight improvement in passivation, although the crystalline and amorphous alloys are both spontaneously passive and exhibit current densities of the order of 1 uA/cm2. However, at intermediate chromium levels (8 to 10 at.%), there is a profound benefit from the phosphorus implantation. At these intermediate levels, there is a decrease in current density relative to the unimplanted alloy, namely four orders of magnitude for the Fe-Cr10P alloy. In fact, the Fe-Cr10P alloy exhibits current decay characteristics similar to an Fe-18Cr crystalline alloy. This research indicates that there is a critical chromium concentration required to provide passivity and that below this level the combination of phosphorus and the amorphous structure increases the initial dissolution rate. Also, below their critical chromium level, there is insufficient chromium for passivation; therefore, pitting is observed. Above this critical chromium concentration, phosphorus and the glassy structure increase the initial dissolution rate, cause a rapid accumulation of chromium in the passive film, and result in an increased rate of passivation. One study examined the influence of alloying elements on the corrosion behavior of iron-chromium base alloys (Ref 12). Figure 8 shows current decay transients for glassy Fe70Cr10B13X7 alloys (X is silicon, carbon, or phosphorus) that were potentiostated in the passive region and abraded under potentiostatic control to produce the repassivation transients. The glassy alloy containing phosphorus exhibited the highest initial current, the most rapid repassivation kinetics, and the lowest passive current density.

Fig. 8 Current density transients for glassy Fe70Cr10B13X7 alloys following mechanical abrasion of specimen surfaces during anodic polarization at constant potentials in 0.1 N H2SO4. X denotes minor metalloid content, and potentials (SCE) are indicated in the figure. Source: Ref 12

Thus, the most effective elements in providing corrosion resistance are chromium and phosphorus. Published concepts concerning and phosphorus appear to be consistent with existing data, but the relative effects of structure and composition on corrosion behavior remain to be quantified.

Localized Corrosion Behavior One of the most outstanding characteristics regarding the corrosion behavior of certain metallic glasses is their ability to resist localized corrosion. In this article, the term localized corrosion refers to pitting and crevice attack (stress-assisted forms of corrosion, such as stress-corrosion cracking and hydrogen embrittlement, are discussed in the section "Environmental Cracking Behavior" in this article). In many cases, this resistance to localized attack extends over wide ranges of oxidizing potential and pH and to alloy compositions that would be considered lean in film former elements compared to conventional crystalline stainless steels. Effect of Chromium on Pitting. The work summarized in Fig. 3 and described previously in this article indicates that iron-base glasses with only several atomic percent of chromium very effectively resist pitting in chloride-containing solutions. Polarization curves of glassy alloys obtained in 1 N NaCl do not show a characteristic pitting potential; rather, they exhibit stable passivity until the onset of transpassivity. In addition, results from the study discussed in Ref 6 with Fe25Ni40Cr15P16B4 showed that the passive range in 1 N H2SO4 plus 0.1 N NaCl is not interrupted by pitting but extends to transpassivity.

In another study, increasing the chromium content from 0 to 16 at.% in a series of Fe-Ni-Cr-P-B alloy systems facilitated passivation in acidified 1 N NaCl, but pitting was not observed on any alloy polarized below the transpassive potential region (Ref 15). Polarization at transpassive potentials caused numerous pits to form that penetrated the filament and were noncrystallographic in shape. Chromium was shown to be very effective in conferring pitting resistance, such as for the glassy alloys Fe-CrxB13C7 and Fe-CrxB13Si7 in 3% NaCl (Ref 16). With chromium levels of 2 and 5 at.%, both alloy types pitted at potentials slightly anodic to the free corrosion potential of about -0.6 V (saturated calomel electrode, SCE). The addition of 8 at.% Cr

extended the pitting resistance to about 1 V (SCE), which is an extremely aggressive condition for alloys containing such a low level of chromium. By contrast, type 304 stainless steel contains about 18 wt.% Cr, yet its pitting potential is several hundred millivolts less positive than that of these glassy alloys. About 7 at.% Cr was sufficient to prevent pitting of Ni-Cr-P-B alloy systems in 10% FeCl 3·H2O at 30 °C (85 °F) (Ref 17). Glassy Fe73Cr7P15B5 passivated spontaneously in 1 N HCl. Surface analysis by x-ray photoelectron spectroscopy showed that chromium and phosphorus were enriched and that nickel depleted in the alloy substrate beneath the passive film. Figure 9 compares the effect on corrosion rate of adding chromium and titanium to Ni-X-P20 glasses. Chromium is more effective than titanium in conferring corrosion resistance, and the chromium-containing alloys exhibited a stronger tendency for spontaneous passivation. The corrosion rate decreased approximately logarithmically with increasing chromium or titanium up to about 10 to 7 at.%, respectively.

Fig. 9 Changes in corrosion rates of glassy Ni-Ti-P20 and Ni-Cr-P20 alloys measured in 1 N HCl and 1 N HNO3 at 30 ± 1 °C (85 ± 2 °F) as a function of the titanium or chromium content. Source: Ref 18

Effect of Molybdenum on Pitting. Molybdenum benefits the pitting resistance of glassy alloys and crystalline steels. The addition of molybdenum to glassy Fe-MoxP13C7 alloys suppressed pitting and decreased the critical current density for passivation and the passive current density (Ref 19). As little as 4 at.% Mo prevented pitting in 1 N HCl, and small additions of molybdenum were more effective than chromium in decreasing corrosion rates. Molybdenum has been shown to facilitate the formation of a passive hydrated chromium or iron oxy-hydroxide film through its enrichment in the corrosion product layer during active dissolution (Ref 20). The enrichment assists the accumulation of the passivating species in the film by lowering the dissolution rate of the species; the molybdenum-rich product subsequently dissolves and thus leaves little molybdenum behind in the film. Effect of Other Alloying Elements on Pitting. Titanium, tantalum, molybdenum, and tungsten were incorporated

by high-rate sputter deposition into alloys of the general composition T1-T2 where T1 = titanium, tantalum, molybdenum, or tungsten and T2 = rhenium, iron, cobalt, nickel, or copper (Ref 21). Tungsten-iron and titanium-copper resisted pitting corrosion up to 2.5 V(SCE) in chloride solutions of pH 1 and 7. Addition of tungsten to Fe-WxP13C7 increased the critical pitting potential, Ecrit, to above 2 V(SCE) at x = 6 at.%, but x = 10 at.% caused transpassive dissolution at 1 V(SCE). Addition of tungsten to commercial type 304 stainless steel by sputtering stabilized the glassy structure and increased Ecrit in chloride electrolyte (Ref 22).

One study investigated the effects of the alloying additions titanium, zirconium, vanadium, niobium, chromium, molybdenum, tungsten, manganese, cobalt, nickel, copper, ruthenium, rhodium, palladium, and platinum in the glassy alloy Fe-X-P13C7 (Ref 23). All elements except manganese decreased the corrosion rate in H2SO4, HCl, HNO3, and NaCl solutions. Although the base alloy, Fe-P13C7, did not passivate, additions of any of the preceding elements at levels from 0.5 to several atomic percent enabled passivation to occur during anodic polarization in 0.1 N H2SO4. Chromium was most effective, and molybdenum and titanium also were very beneficial. Pitting was not observed in 3% NaCl for those alloys that passivated. The alloys that did not passivate, such as Fe-Co-P13C7, also did not pit, but they dissolved uniformly. Passivation. It has been proposed that the excellent resistance of certain glassy alloys to uniform and localized

corrosion results from their enhanced chemical reactivity relative to conventional stainless alloys (Ref 24). Transient repassivation experiments with glassy Fe70Cr10P13C7 and crystalline type 304 stainless steel in acidified chloride electrolyte showed a higher initial reactivity on the glassy alloy after abrasion and a more rapid rate of repassivation. These experiments demonstrated that there is a synergistic effect between chromium and phosphorus in transition metalmetalloid glasses such that maximum resistance to localized corrosion exists when these two elements are both present. The excellent resistance to localized corrosion may result from the rapid re-formation of a passive film at regions where it is damaged by mechanical or electrochemical means, combined with enrichment of Cr3+ species in the film. Research cited earlier in this article concerning the use of ion implantation to make iron-chromium-phosphorus alloys (Ref 14) demonstrated that the synergistic effect of chromium and phosphorus is a strong function of the chromium concentration. Potentiostatic polarization experiments in acidic chloride solutions showed behavior that was a strong function of chromium concentration. At low chromium levels, phosphorus implantation induced pitting, but at high chromium levels (18 at.%), a slight improvement in passivation was observed. At intermediate levels (10 to 12 at.%), substantial improvement in the passive film was obtained through phosphorus implantation. It was proposed that when the alloy contains small amounts of chromium, there is not enough chromium to passivate the alloy, but when the phosphorus stimulates the initial dissolution, the alloy becomes susceptible to pitting. As the chromium concentration increases, there is sufficient chromium for passivation, and the phosphorus promotes the accumulation of chromium and a very protective passive film. Crevice Corrosion Resistance. One investigation examined the resistance of glassy alloys to crevice corrosion in

acidic solutions containing Cl- ion. Crevice corrosion was studied as a means of circumventing the need for initiating localized corrosion; that is, crevice corrosion behavior was considered to represent more a measure of the resistance to propagation, rather than initiation, of localized corrosion (Ref 25). Introducing micro-cracks into Fe-Ni-Cr-P-B glassy alloy filaments by cold rolling (Ref 26) was found to cause susceptibility to a transient form of crevice attack; however, the crevices widened into pit-shaped cavities and then passivated spontaneously. Subsequent research with an electrochemical cell consisting of a prepared crevice instrumented with microreference and pH electrodes showed that the glassy alloy possessed a strong tendency to passivate, even under the aggressive conditions of low pH, low dissolved oxygen concentration, and oxidizing potential that prevail within crevices (Ref 27). The conclusion was that the alloy resisted crevice attack because of its strong ability to passivate, which in turn stifled propagation. This resistance to crevice corrosion could be expected to extend to other glassy transition metal-metalloid compositions containing both a film former and phosphorus. Glassy nickel-phosphorus is another alloy system that has been recently investigated and that appears to resist chlorideinduced corrosion (Ref 27). In fact, the potentiodynamic polarization curves are virtually identical in both chloridecontaining and chloride-free electrolytes. A form of chemical passivity has been proposed to explain the corrosion behavior. Passivation in this system is due to the formation of an ionic barrier layer, not to the formation of a classical passive oxide film. This barrier layer consists of hypophosphite ion adsorbed on the nickel-phosphorus surface, which may in turn be hydrogen bonded to an outer layer of water molecules. This barrier layer inhibits the transport of water to the surface and thus prevents hydration of nickel, which is the first step in the nickel dissolution process.

Environmental Cracking Behavior The environmentally induced fracture of glassy alloys, namely hydrogen embrittlement and stress corrosion cracking (SCC), will be discussed in this section. Details on the mechanisms of these phenomena can be found in the article "Environmentally Induced Cracking" in this Volume.

Stress-Corrosion Cracking. One of the first reported experiments on the SCC of glassy alloys concerned

Ni49Fe29P14B6Al2 (Ref 28). Loading to 75% of the fracture stress in air in 3.5 N NaCl solution resulted in a slow, presumably SCC fracture region and a final, fast fracture region. However, another researcher suggested that the fracture of this alloy was actually induced by hydrogen (Ref 29). In another case, the SCC behavior of a glassy Fe32Ni36Cr14P12B6 alloy in boiling magnesium chloride (MgCl2) at 125 °C (255 °F) was studied by means of constant extension rate tensile tests and constant strain tests (Ref 30). Stress-corrosion cracking occurred at the corrosion potential and anodic overpotentials, and slight cathodic polarization prevented SCC. Examination of the fracture surfaces led to the conclusion that localized corrosion enhances hydrogen entry and subsequent embrittlement. The SCC behavior of glassy Fe-Cr-Ni-P-C alloy systems in acidic chloride solutions was investigated with constant extension rate tensile tests (Ref 31). Hydrogen embrittlement occurred at cathodic polarizations up to -300 mV relative to the corrosion potential. In the passive potential region, no cracking occurred in neutral NaCl solutions and in acidic solutions at low Cl- concentrations. Fracture stress decreased only when the specimens were strained in strong acidic solutions containing Cl-, and this phenomenon was also attributed to hydrogen embrittlement. The tendency of glassy Fe40Ni40P14B6 to undergo SCC and hydrogen embrittlement in acidic electrolytes was also studied (Ref 32). Cathodic polarization of elastically stressed specimens in 1 M HCl resulted in failure by hydrogen embrittlement. Specimens immersed in aqueous FeCl3 solution at the free corrosion potential failed by SCC, as did those that were anodically polarized in 1 M HCl. These specimens were covered by an iron oxide film, and selective leaching (dealloying of nickel from pits and cracks occurred. Hydrogen Embrittlement. Although classical SCC (defined as cracking caused directly by anodic dissolution at the

crack tip) of glassy alloys apparently occurs, hydrogen embrittlement is a more common mode of environmentally assisted failure in aqueous electrolytes. Hydrogen embrittlement of glassy alloys has been observed during bending or tensile tests during or after hydrogen charging in the following alloys: • • •

Fe80P13C7 and Fe70Cr10P13C7 (Ref 33) Fe32Ni36Cr14P12C6 (Ref 34) Fe49.5Cr7.5Ni23P13C7 and Fe53Cr7Ni20P14C6 (Ref 35)

As shown in Fig. 10, the stress-strain curve of an Fe49.5Cr7.5Ni23P13C7 glassy alloy exhibits almost completely elastic behavior in air and in various acid chloride solutions. The fracture strain decreased to about 30% of that in air as a result of charging the specimen with hydrogen. It was proposed that local corrosion at the open-circuit and even passive potentials can produce hydrogen and thus create embrittlement; fractographic evidence and the return of ductility by baking after corrosion were cited as evidence for this claim (Ref 29).

Fig. 10 Stress-strain behavior of a glassy Fe49.5Cr7.5Ni23P13C7 alloy at various potentials in air and in solution at a strain rate of 4.2 × 10-6 s-1. Line 1: in air; line 2: in 5 N H2SO4 + 0.1 N NaCl, Ecorr = -20 mV; line 3: 5 N H2SO4 + 0.1 N NaCl, E = +500 mV; line 4: 5 N H2SO4 + 0.1 N NaCl, E = -500 mV. Source: Ref 31

Although it appears that hydrogen embrittlement is more common that SCC as the environmentally assisted failure mode for glassy alloys, the detailed mechanism of the hydrogen embrittlement of transition metal-metalloid alloys is uncertain. At cathodic potentials in deaerated solutions, the principal cathodic reaction produces hydrogen by the following reaction sequence: 2H+ + 2e H2, which can be separated into a proton reduction step and a hydrogen adatom-adatom combination step. Elements such as phosphorus, sulfur, arsenic, and antimony poison the reaction Hads + Hads H2; thus, they increase the concentration of adsorbed (ads) hydrogen on the electrode surface and consequently the flux of atomic hydrogen absorbed through the surface into the bulk alloy. Because phosphorus is commonly found in transition metalmetalloid glassy alloys, it would seem likely that these phosphorus-containing alloys might have a large tendency toward absorbing hydrogen from the electrolyte and consequently a significant tendency toward hydrogen embrittlement. Augmenting this tendency would be the very high strength and limited ductility characteristic of this compositional class of glassy alloys. In this regard, an investigation of hydrogen permeation through glassy phosphorus-containing nickel-base alloys concluded that phosphorus increases the rate of hydrogen absorption relative to that for pure nickel (Ref 2). It was proposed that internal voids in the alloys act as traps for the atomic hydrogen and that this atomic hydrogen may subsequently combine to form molecular hydrogen, which ultimately produces internal pressure that can shatter the specimen. Another researcher characterized the effects of metalloid additions on the susceptibility to hydrogen embrittlement of glassy Fe-Cr5Mo12X and Fe-Cr10Mo12X (X = 18C, 20B, or 13P-7C) in 1 N HCl, 0.5 N NaCl, and 1 N H2SO4 (Ref 29). Although phosphorus is an effective hydrogen recombination poison, alloys containing this element showed a lower susceptibility to hydrogen embrittlement. (The alloys containing carbon were the most susceptible.) This lower susceptibility was ascribed to the higher rate of repassivation of phosphorus-containing alloys; because the corrosion rate was decreased, the amount of hydrogen produced by the open-circuit corrosion reaction and that absorbed into the alloys should also be lowered. The specific effects of each metalloid on hydrogen embrittlement susceptibility are still uncertain. However, the large concentrations of metalloids present in transition metal-metalloid alloys almost certainly influence the high susceptibility of these materials to hydrogen embrittlement.

Applications Although metallic glasses are interesting from a research standpoint, there are several engineering applications in which their unique properties may be important. In considering possible applications of glassy metals, two obvious limitations are apparent. First, all alloys in this class of materials are metastable. If they are subjected to elevated temperatures, devitrification will occur, which normally results in loss of the properties of interest. The glass transition temperature is of course a function of the alloy composition; therefore, some alloys are suitable for use at temperatures substantially above ambient. The second limitation concerns the physical dimensions of the material produced. Because of the high cooling rate required, at least one dimension of the alloy must be very thin. Therefore, the most common forms of glassy metals include filaments, wires, and ribbons. An obvious application of glassy alloys is that of corrosion-resistant coatings or barriers. In certain applications, a thin, highly corrosion-resistant coating may be sufficient, and these coatings permit the use of less-expensive base materials. To make commercial use of glassy alloys, advances are required in several areas. Glassy coatings are very appealing in that they can be applied to engineering parts of complex geometry. However, most of the deposition processes used to make glassy alloys are slow, expensive, or both. Techniques are needed that enable the uniform and rapid coating of large areas. In addition, because most current used deposition techniques tend to be expensive, replacement of corrosionresistant bulk alloys with glassy metal coatings must await the development of cost-competitive procedures. Additional information on the applications of amorphous materials, such as tool applications, aluminum die-casting mold inserts, metal-bonded abrasive wheels, and hardfacing coatings, can be found in Powder Metal Technologies and Applications, Volume 7 of the ASM Handbook.

References 1. 2. 3. 4. 5. 6. 7. 8. 9.

10. 11. 12. 13. 14. 15. 16. 17. 18.

H. Jones, Proceedings of the Second International Conference on Rapidly Quenched Metals, N.J. Grant and B.C. Giessen, Ed., Elsevier, 1976 R.M. Latanision, J.C. Turn, and C.R. Compeau, in Proceedings of the Third International Conference on Mechanical Behavior of Metals, Vol 2, 1979, p 475 M. Naka, K. Hashimoto, and T. Masumoto, J. Jpn. Inst. Met., Vol 38, 1974, p 38 Y. Waseda and K.T. Aust, J. Mater. Sci., Vol 16, 1981, p 2337 M. Naka, K. Hashimoto, and T. Masumoto, Corrosion, Vol 32, 1976, p 146 R.B. Diegle and J.E. Slater, Corrosion, Vol 32, 1976, p 155 M. Naka, K. Hashimoto, and T. Masumoto, J. Non-cryst. Solids, Vol 28, 1978, p 403 K. Hashimoto, M. Kasaya, K. Asami, and T. Masumoto, Corros. Eng., Vol 26, 1977, p 445 K. Hashimoto, M. Naka, J. Noguchi, K. Asami, and T. Masumoto, in Proceedings of the Fourth International Conference on Passivity, R.P. Frankenthal and J. Kruger, Ed., The Electrochemical Society, 1978, p 156-169 M. Naka, K. Hashimoto, and T. Masumoto, J. Non-cryst. Solids, Vol 30, 1978, p 29 J.C. Turn and R.M. Latanision, Corrosion, Vol 39, 1983, p 271 K. Hashimoto, "Supplement to the Scientific Report of the Research Institutes of Tohoku University," A 201-216, Tohoku University, 1980 M. Naka, K. Hashimoto, K. Asami, and T. Masumoto, Proceedings of the Third International Conference on Rapidly Quenched Metals, B. Cantor, Ed., The Metals Society, 1978, p 449 N.R. Sorensen, R.B. Diegle, and S.T. Picraux, J. Mater. Res., Vol 1 (No. 6), 1986, p 752 R.B. Diegle, Corrosion, Vol 35, 1979, p 250 M. Naka, K. Hashimoto, and T. Masumoto, "Scientific Report of the Research Institutes of Tohoku University, A-26, Tohoku University, 1979, p 283 K. Hashimoto, M. Kasaya, K. Asami, and T. Masumoto, Boshoku Gijutsu, Vol 26, 1977, p 445 M. Naka, K. Hashimoto, and T. Masumoto, 1703rd Report, Research Institute for Iron, Steel, and Other Metals, 1980, p 156

19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

M. Naka, K. Hashimoto, and T. Masumoto, J. Non-cryst. Solids, Vol 29, 1978, p 61 K. Asami, M. Naka, K. Hashimoto, and T. Masumoto, J. Electrochem. Soc., Vol 127, 1980, p 2130 R. Wang, Abstract 80-2, in Extended Abstracts, The Electrochemical Society, 1980, p 620 R. Wang Paper presented at the Fall Meeting, Pittsburgh, PA, The Metallurgical Society, Oct 1980 M. Naka, K. Hashimoto, and T. Masumoto, J. Non-cryst. Solids, Vol 31, 1979, p 355 K. Hashimoto, M. Naka and T. Masumoto, "Scientific Report of the Research Institutes of Tohoku University, A-26, Tohoku University, 1976, p 48 R.B. Diegle, Corrosion, Vol 35, 1979, p 250 T.M. Devine, J. Electrochem. Soc., Vol 124, 1977, p 38 R.B. Diegle, Corrosion, Vol 36, 1980, p 326 C.A. Pampillo, J. Mater. Sci., Vol 10, 1975, p 1194 A. Kawashima, K. Hashimoto, and T. Masumoto, Corrosion, Vol 36, 1980, p 577 R.F. Sandenbergh and R.M. Latanision, Corrosion, Vol 41, 1985, p 369 A. Kawashima, K. Hashimoto, and T. Masumoto, Corros. Sci., Vol 16, 1976, p 935 M.D. Archer and R.J. McKim, J. Mater. Sci., Vol 18, 1983, p 1125 M. Nagumo and T. Takahashi, Mater. Sci. Eng., Vol 23, 1976, p 257 R.K. Viswanadham, J.A.S. Green and W.G. Montague, Sci. Metall., Vol 10 1976, p 229 A. Kawashima, K. Hashimoto, and T. Masumoto, Corros. Sci., Vol 16, 1976, p 935

Corrosion Chromium

of

Electroplated

Hard

Allen R. Jones, M&T Chemicals, Inc.

Introduction HARD CHROMIUM plated parts have more than about 1.2 m (0.05 mils) of chromium. Parts that are plated with less than this amount are referred to as decorative applications. Electroplated chromium protects substrates by means of a barrier coating as opposed to a sacrificial coating, such as zinc. Chromium is more electrochemically active than steel; however, it forms a dense self-healing oxide layer on its surface. Chromium can be passivated, or the oxide layer can be formed by exposure to air or by immersion in room-temperature oxidizing acids. Electroplated chromium is chemically resistant to most compounds and offers excellent corrosion protection in various environments. It is especially useful in applications that also require wear resistance.

Electrodeposition Parameters Chromium Plating Baths. Commercially, hard chromium is deposited from four types of baths; conventional,

fluoride, and the two high-efficiency etch-free baths. All of the baths contain chromic acid (CrO3) and sulfate (

).

The ( ) acts as catalyst. Chromium cannot be electrodeposited from an aqueous CrO3 solution unless one or more catalysts are present. Depending on which catalysts are present and the plating parameters, between 10 and 45% of the cathodic current will be used to reduce hexavalent chromium (Cr6+) to chromium metal. The properties of the electrodeposits are influenced by the ratio of CrO3 to the catalysts, plating temperature, and current density. Stress and Microcracks. The tensile stress in most electroplated chromium deposits increases until microcracks are

formed (Ref 1, 2). The microcracks decrease the stress in the deposit as the thickness of the deposit increases. Stress is inversely proportional to the number of microcracks. The number of microcracks is more important in controlling stress than the type of bath chemistry. Crack-free deposits are highly stressed.

Microcracks are present in most electroplated hard chromium deposits. Figure 1 shows a typical microcrack structure. The density of microcracks in chromium deposits varies from 0 to more than 1200 cracks/cm (3000 cracks/in.), depending on bath chemistry, current density, and temperature. The number of microcracks increases with the concentration of catalyst in the plating bath. The depth of a microcrack is less than about 8 m (0.3 mils) on a deposit and that is 130 m (5 mils) thick with crack counts of about 800 cracks/cm (2000 cracks/in.).

Fig. 1 Photomicrographs of chromium deposits (plated in a high-efficiency etch-free bath) after etching. (a) and (b) Deposit plated at 78 A/dm2 (5 A/in.2) and at 55 °C (130 °F). (a) 540×. (b) 2300×. (c) Cross section of a chromium deposit plated at 93 A/dm2 (6 A/in.2) and at 58 °C (135 °F). The specimen was polished before etching. 880×. Both deposits contain 800 microcracks/cm (2000 microcracks/in).

Because chromium protects substrates by forming a barrier, the coatings must be thicker than the microcracks to provide good corrosion resistance. Thin coatings may not form microcracks and can offer as much corrosion resistance as thicker coatings (see the section "Coating Thickness" in this article). Chromium electrodeposits that are about 25 m (1 mil) thick with crack counts of about 400 cracks/cm (1000 cracks/in.) are as resistant to corrosion as deposits with crack counts of about 100 cracks/cm (250 cracks/in.). Deposits with very low crack counts have deeper microcracks than deposits with higher crack counts. Therefore, highly microcracked deposits are as resistant to corrosion as sparsely microcracked deposits. Microcracks are not as detrimental to corrosion resistance as might be expected. There are two reasons for this. First, the microcracks are not voids, but are areas with a structure and composition that are different from those of the bulk. Second, because the microcracks are very narrow (about 0.1 m wide) and because water does not wet chromium, the water does not readily enter the microcracks. Microcrack-free thick chromium deposits can be plated from baths at low current densities and high temperatures. These microcrack-free deposits provide better corrosion protection than microcracked chromium. However, these deposits are highly stressed and are not as hard as microcracked chromium. Crack-free deposits can be used when corrosion protection is the only requirement for the deposit. Postplating grinding or cutting may cause pickout (chromium fracturing from chromium) in highly stressed deposits. The conditions under which some crack-free coatings are deposited will result in a deposition efficiency that is lower than that normally observed for the plating bath. Additional Deposit Properties Influencing Corrosion. Hardness is related to microcracking, which is related to corrosion. Chromium coatings have hardnesses between 850 and 1050 HK (100-gf load). Microcrack-free deposits can have hardnesses as low as 600 or 300 HK (Ref 2). According to one study, as deposit hardness increases or crystal size decreases, the rate of attack by sulfuric acid (H2SO4), hydrochloric acid (HCl), and CrO3 decreases (Ref 3).

Coating Thickness Figure 2 shows that the corrosion resistance of hard chromium plated steel in salt spray undergoes a maximum and a minimum, then increases with the chromium thickness (Ref 4). Figure 2 also shows the average of two panels in a salt spray exposure. Maximum corrosion resistance occurred at a chromium thickness of about 5 m (0.2 mils). As the thickness increased above 5 m (0.2 mils), microcracking occurred and corrosion resistance decreased. When the chromium thickness increased to about 10 m (0.4 mils), the initial cracks were covered by more chromium, there were fewer corrosion paths to the substrate, and the corrosion resistance of the deposit increased. These deposits were plated from a conventional bath containing 250 g/L of CrO3 and 2.5 g/L of H2SO4 at 31 A/dm2 (2 A/in.2); no temperature was specified.

Fig. 2 Chromium corrosion in salt spray versus thickness of deposit. Curve A shows time to general rust; curve B is for time to initial corrosion. Parts were plated in a conventional bath (250 g/cL CrO3 and 2.5 g/L 31 A/dm2, or 2 A/in.2). Source: Ref 4

at

Figure 3 shows additional data on corrosion resistance and chromium thickness. The electrodeposits were prepared from a conventional bath containing 295 g/L of CrO3 and 3 g/L of H2SO4. Data are given for two plating conditions: 30 °C (85 °F) at 20 A/dm2 (1.3 A/in.2) and 60 °C (140 °F) at 43 A/dm 2 (2.8 A/in.2). The first condition produced cold chromium that was crack free and soft. The second condition produced conventional microcracked hard chromium. The cold chromium deposit showed excellent corrosion resistance at thicknesses of 4.8, 9.1, and 12.4 m (0.2, 0.36, and 0.49 mils), but the corrosion resistance was very poor at a thickness of 15.5 m (0.6 mils). The 15.5- m (0.6-mil) coating was not porous, and no reason was given for its poor corrosion resistance. The high stress in the coating and the poor adhesion of cold chromium may have resulted in coating failure. The thinner (1 MeV)

Fracture 99% Intergranular SCC Ductile

Note: NVT = neutrons per unit volume × time.

The second promising technique for mitigating irradiation-assisted SCC is the use of high-purity materials containing controlled amounts of such impurities as sulfur, phosphorus, and silicon (Ref 27). This mitigation technique is based on early laboratory testing in a highly oxidizing medium, such as boiling nitric acid (HNO 3) with Cr6+ ions, and on the highly successful in-reactor performance of high-purity type 348 stainless steel fuel cladding at the La Crosse BWR. In contrast, commercial-purity type 348 stainless steel materials with the nominal commercial levels of sulfur, phosphorus, and silicon installed at the La Crosse BWR suffered irradiation-assisted SCC. Other support for the implementation of highpurity materials is based on in-reactor studies (Ref 29) and on laboratory studies on nonirradiated stainless steel where the presence of these specific impurities increased the intergranular SCC susceptibility (Ref 30). Work is continuing in this area. Conclusions. The results of in-reactor and laboratory investigations suggest the following conclusions concerning

irradiation-assisted SCC in the BWR: •

•

The irradiation-assisted SCC of components in the BWR core was the result of the simultaneous interaction of irradiation-enhanced impurity segregation, tensile stress, and highly oxidizing environment Preliminary studies suggest that hydrogen water chemistry and high-purity materials are promising remedies for mitigating irradiation-assisted SCC

Case History: Corrosion Fatigue in Feedwater Nozzles Fatigue evaluation of nuclear power plant equipment subjected to cyclic loading conditions is of primary concern to equipment designers and stress analysts. Because the number of loadings resulting in significant stresses that occur over the life of light water reactors seldom exceed several thousand, and in numerous cases do not even exceed several hundred, fatigue usage is generally classified as low cycle. However, in this particular case history on feedwater nozzle cracking, both low- and high-cycle fatigue were involved. Figure 23 shows a cross section of a feedwater nozzle that suffered from cracking of the cladding of the feedwater nozzle bed radii, a problem common to a number of BWRs.

Fig. 23 Cross section of feedwater nozzle with cracking location

Field Experience. Metallurgical examination of a boat sample (a wedge-shaped specimen cut out of a component) removed from the cladding of the feedwater nozzle bend radii revealed that the primary cause of the nozzle crack was thermal-induced corrosion fatigue (Fig. 24). It was further determined that there are two corrosion fatigue mechanisms present: a high-cycle corrosion fatigue mechanism, which initiates the cracks, and a low-cycle corrosion fatigue mechanism, which causes the cracks to propagate.

Fig. 24 Example of corrosion fatigue striations in a feedwater nozzle. 1500×

Mechanisms of Corrosion Fatigue in Feedwater Nozzles. The high-cycle mechanism was found to be primarily

caused by leakage flow passing between the thermal sleeve and safe end. This leakage flow, which is at feedwater temperature, mixes in a turbulent manner with hot downcomer flow in the annulus between the nozzle and thermal sleeve. The mixing fluid impinges on the nozzle wall, causing thermal cycling of the metal surface. It has been determined by test and by field measurements at two BWRs that the metal temperature cycling, with leakage present, has a magnitude of up to 50% of the difference in temperature between the feedwater and the downcomer water. The cycling occurs with frequencies between 0.1 and 1 Hz and thus can initiate cracking rapidly with little if any environmental contribution. The exact time to crack initiation depends on several factors, including the duration of operation with low feedwater temperature.

The cracks initiated by the high-frequency cycling described above will arrest at a depth of approximately 6 mm (0.25 in.) from the metal surface. This arrest results from the fact that the high-cycle thermal input induces thermal stresses with steep gradients and shallow depths. Cracks that arrest, as described above, present no problem from an engineering standpoint, because they will not degrade safety or availability and would not have to be repaired under the rules of Section XI of the American Society of Mechanical Engineers Code. There is another mechanism present in the feedwater nozzles that causes the cracks to continue to grow. This mechanism is the combined pressure and thermal cycles imposed by start-up/shutdown and feedwater on-off transients. These transients, although relatively few in number, produce large stress cycles in the nozzle and in time could drive the cracks to significant depths. This low-cycle fatigue crack propagation is environmentally accelerated under BWR conditions. The deepest cracks observed were up to 18% of wall and relatively short. Fortunately, the cracking is readily detectable by dye-penetrant examination and/or ultrasonic examination from the outside of the nozzle. The deep cracks require repair and thus can result in a significant impact on plant availability and operating cost. Leakage between the safe end and thermal sleeve also has an aggravating effect on the crack growth rate because it increases the heat transfer coefficient between the feedwater and the nozzle. The increased heat transfer coefficient increases the stresses in the nozzle during thermal transients, and because crack growth is dependent on stress to the fourth power, a significant effect results. Mitigation of Cracking. Several solutions were derived to mitigate this problem from both a design and a corrosion/metallurgical viewpoint. The solution consisted of three parts: a revised sparger thermal sleeve design, use of unclad feedwater nozzles, and a revised system of configuration/operating procedures that mitigates the conditions tending to produce crack initiation and growth. Together these three elements constitute a solution to the problem with margin for unexpected conditions. The implementation of hydrogen water chemistry would also mitigate this problem because test results have indicated a decrease in corrosion fatigue crack propagation rate in hydrogen water chemistry (Ref 26).

A sparger thermal sleeve design has been developed that meets the following objectives: • • •

It can be installed and removed without cutting feedwater piping It protects the feedwater nozzle against the high-frequency thermal cycles that initiate nozzle cracks through the use of redundant metal O-ring seals It uses materials and processes that make the part immune to intergranular SCC

The resulting design is schematically shown in Fig. 25. The thermal cycling profiles for the previous and the newerdesign feedwater spargers are shown in Fig. 26. The marked reduction in thermal cycling is evident.

Fig. 25 Improved feedwater thermal sleeve design to eliminate corrosion fatigue

Fig. 26 Temperature variations with (a) and without (b) bypass leakage

The main conclusions concerning the corrosion fatigue cracking of BWR feedwater nozzles are:

•

•

The corrosion fatigue of BWR feedwater nozzles was a result of the interaction of high-cycle fatigue crack initiation due to leakage flow passing between the thermal sleeve and safe end and low-cycle corrosion fatigue crack growth due to start-up/shutdown and feedwater on-off transients The cracking problem was mitigated by a redesign of the feedwater thermal sleeve to eliminate leakage and the removal of the nozzle cladding. Hydrogen water chemistry will also provide an additional margin against this phenomenon

Steam Generator Failure or Degradation Stanley J. Green, Electric Power Research Institute

Steam generators in PWR power plants transfer heat from a primary coolant system (pressurized water) to a secondary coolant system. Primary coolant water is heated in the core and passes through the steam generator, where it transfers heat to the secondary coolant water to make steam. The steam then drives a turbine that turns an electric generator. Steam is condensed and returns to the steam generator as feedwater. Two types of PWR steam generators are in use: recirculating steam generators (RSGs) and once-through steam generators (OTSGs). Most of the units are vertical, and this Section will be limited to vertical units. Some of these steam generators have operated with a minimum of problems while other steam generator designs have experienced a variety of corrosioninduced and mechanically induced problems. The discussion will focus more on those designs that have experienced problems and where effort has been expended to correct them. The corrosion problems include denting, wastage, intergranular attack, SCC ad pitting on the outside surfaces of the tubes, and SCC from the inner surfaces of the tube. The mechanical concerns have included water hammer, thermal stratification in feedwater pipes, fretting and water of the tubes caused by excessive tube vibration, and erosion-corrosion. These problems have caused unscheduled outages and expensive repairs. Where most extensively affected, some steam generators have been replaced after 8 to 12 years of operation, which is far short of the expected plant operating period of 40 years.

The scope of this Section will be limited to the corrosion-related issues. A brief summary of other degradation phenomena is given at the end of this Section, and previous summaries of these issues are available in Ref 31, 32, 33, and 34. Also, specific preventive and corrective actions are discussed in Ref 35.

Steam Generator Design A typical vertical recirculating steam generator (feedring-type) is shown in Fig. 27. (Other designs employ preheat features, where the cold feed is introduced into the lower part of the bundle.) Water is fed into the downcomer, where it is mixed with two to five volumes of recirculating water from the moisture separators. The downcomer water flows to the bottom of the steam generator, across the tubesheet, and then upward through the tube bundle where steam is generated.

Fig. 27 Recirculating steam generator

The thermodynamic quality of the water-steam mixture at the top of the bundle is about 17 to 33% when it enters the steam separators, which corresponds to a circulation ratio in the range of 6:1 to 3:1. The pressure on the secondary side is about 4135 to 7240 kPa (600 to 1050 psia). The primary coolant flows through U-tubes at a pressure of about 9655 to 15515 kPa (1400 to 2250 psia). It enters the steam generator at about 310 to 325 °C (590 to 620 °F) and leaves at about

255 to 290 °C (495 to 550 °F). At the primary inlet, the temperature difference across the tube wall is about 35 to 50 °C (65 to 90 °F), corresponding to a heat flux of 315,460 to 441,640 W/m2 (100,000 to 140,000 Btu/h·ft2). At the primary outlet or cold side, the temperature difference between the primary and secondary sides is about 10 to 15 °C (20 to 25 °F), corresponding to a heat flux of about 94,640 W/m2 (30,000 Btu/h·ft2). The Westinghouse, Combustion Engineering, Kraftwerk Union (KWU), Framatome, and Mitsubishi designs have comparable operating parameters, while the Babcock and Wilcox (B&W)/Atomic Energy of Canada Limited (AECL) design operates at lower temperatures and pressures. Knowledge of the materials, water chemistry, and tube support arrangements used in RSGs is required to understand the problems that have occurred. The tubes have been made primarily of Alloy 600, a nickel, chromium, iron alloy. In KWU and later B&W/AECL designs, the tubes have been made of Alloy 800, an iron-base superalloy. The mill-annealing conditions vary among the manufacturers, while B&W/AECL have used stress-relieved (605 °C, or 1125 °F, for 8 h) tubing for their Alloy 600 tubed steam generators. Some of the more recent designs using Alloy 600 tubing have thermally treated the tubing (705 °C, or 1300 °F, for 15 h) to improve resistance to SCC. The tube support structures for most of the early units were made of carbon steel, while later units have switched to type 405, 409, and 410 stainless steels for additional corrosion resistance. Type 347 stainless steel has always been used for KWU steam generator tube support structures. Tube support structures of early units used tube support plates with drilled holes (Fig. 28a and b), plates with broached holes (Fig. 28c), and lattice bars or egg crates (Fig. 28e). When the drilled-hole support plates were found to promote accumulation of corrodents, they were changed to a quatrefoil design hole with lands (Fig. 28d) or to a lattice support structure (Fig. 28e).

Fig. 28 Tube support device designs. (a) Drilled, without flow holes. (b) Drilled, with flow holes. (c) Broachtrefoil. (d) Broach-quatrefoil. (e) Egg crate

In some of the early designs, the tubes were only partly expanded into the lower end tubesheet, leaving a crevice between the outside diameter of the tube and the inside diameter of the hole in the tubesheet, that is, in the upper part of the tubesheet; in other designs, the tubes were expanded into the tubesheet along its full length. The tubes have been expanded into the tubesheet by mechanical, hydraulic, and explosive expansion methods. Generally, the tubes have been expanded for the full length of the tubesheet in the later designs. Most of the early RSGs employed coordinated sodium phosphate water treatment for the secondary side, which is the conventional water treatment method for fossil-fired boilers with similar steam pressures. An OTSG is shown in Fig. 29. Water enters a feed annulus above the ninth tube support plate level. There it is mixed with steam aspirated from the tube bundle area and preheated to saturation. The saturated water flows down the annulus, across the lower tubesheet, and upward into the tube bundle, where it becomes steam. It reaches 100% quality (on the average) in the ninth and tenth support plate region, and achieves about 20 to 35 °C (40 to 60 °F) of superheat at about 6370 kPa (924 psia) at the top of the unit. The superheated steam flows radially outward and then down the annulus to the steam outlet connection. The primary coolant flow is from top to bottom. It enters at about 315 to 325 °C (600 to 620 °F) and leaves at about 290 to 295 °C (555 to 560 °F). The temperature difference between the primary and secondary sides at the bottom of the steam generator is similar to that an the cold leg of a recirculating steam generator.

Fig. 29 Once-through steam generator

Once-through steam generator materials are similar to those used in RSGs. Tubes are Alloy 600 in a mill-annealed plus a stress-relieved condition, which sensitizes the tubing (causes chromium-depleted grain boundaries). Tube support plates are carbon steel. However, the holes are not drilled round holes as in the early RSGs, but trefoil holes with three lands supporting the tubes (Fig. 28c). Tubesheets are low-alloy steel. Once-through steam generators have always used allvolatile water treatment water chemistry.

Prevention of Steam Generator Corrosion Problems As noted above, these steam generators have experienced a variety of corrosion-induced and mechanically induced problems. The types of damage and the number of units affected are presented in Table 3, which indicates that essentially no units have operated trouble-free for more than 5 years. The industry has made a very substantial effort to prevent and minimize these steam generator problems. This includes a major program by the Steam Generator Owners Group, managed by the Electric Power Research Institute (EPRI), and substantial programs carried out by steam generator suppliers and utilities.

Table 3 Units affected by steam generator problems Total units: 121 (65 units having >5 years of operation and 56 units having 5 years of operation)

Units affected (as of 1984) 42 18 31 11 31 33 28 2 30 5 2 10 6 37 1

The approaches used to correct or prevent corrosion for both new and operating units are based on addressing the causes. Thee corrective measures can be divided into the following three categories: • • •

Modifying the environment Modifying the materials Modifying the stresses

The corrosion problems and the types of corrective action developed for each issue are listed in Table 4. Each of the problems and corrective actions is discussed below.

Table 4 Corrective actions for steam generator corrosion problems Problem Tube wastage Denting Inside-diameter SCC Outside-diameter intergranular attack Pitting Corrosion fatigue

(a)

Modify environment Operating New X ... X X ... ... X X X X X X

Modify materials Operating(a) New ... ... ... X X X X X X ... ... ...

Modify stress Operating New ... ... ... ... X X ... X ... ... ... ...

By sleeving

Tube Wastage Causes. Tube wastage, or thinning, was once of the first corrosion problems that occurred in recirculating steam

generators operating with sodium phosphate as a secondary water treatment. Phosphate wastage was first observed in commercial PWR plants when phosphate treatment was changed to a low sodium-to-phosphate molar ratio control, in which the molar ratio of Na/PO4 was maintained at about 2.0. This change was in response to a series of caustic SCC events attributed to operation with uncontrolled high Na/PO4 ratios (above 2.8), from which free caustic could result. The incidence of caustic SCC dropped markedly, but the general corrosion, now known as phosphate wastage, began to be observed within approximately 1 year after the change. Some phosphate wastage observations are discussed in Ref 36; a detailed report of such events at two specific nuclear power installations is provided in Ref 37. The phosphate wastage in PWR steam generators of the Westinghouse design occurred for the most part at the interfaces between hot leg tubes and the tops of sludge piles that accumulated on top of the tubesheet. Where the sludge pile was deep, the zone of wastage extended about 25 mm (1 in.) into the pile; in most cases, it did not penetrate appreciably into the tube/tubesheet crevice. Wastage was first noted in steam generators of the Combustion Engineering design. It was most extensive in the vicinity of antivibration straps, which were relatively wide and oriented in such a way as to define a region of steam blanketing. Corrosion was concentrated at the boundaries of this region, where a liquid/vapor interface presumably fluctuated over a short length of tubing. As recognition grew that this new phenomenon was widespread and of a generic nature, major laboratory investigations were launched by the two United States vendors of RSGs. It was demonstrated in pot boilers that the location of attack was related to the concentration of aggressive species at steam/water interfaces. Emphasis was placed on investigating the chemistry of sodium phosphate solutions in high-temperature water, in addition to corrosion studies and model boiler tests. The laboratory work showed that the corrosivity of concentrated sodium phosphate solutions was related to both concentration and the Na/PO4 molar ratio. The laboratory results led to recommendations for Na/PO4 molar ratio control in the relatively narrow band of 2.3 to 2.6. The lower limit was related to the rapidly increasing rate of wastage at lower ratios, and the higher limit was selected to avoid free caustic and concomitant caustic SCC. A number of plant operators were successful in controlling their steam generator chemistries within this restricted range of compositions, but eddy-current inspections indicated that the rate of attack was only slowed, not stopped. It appeared that attack could not be stopped unless the operators consistently operated within a narrow control band at the high end of the allowed range. Furthermore, a constantly increasing sludge burden, augmented by precipitated phosphate compounds, made such control increasingly difficult. Corrective Actions. The above difficulties led the vendors of the affected plants to recommend that the operators adopt

an all-volatile water treatment based on the use of ammonia and hydrazine. This recommendation was almost universally followed in the United States except for two plants. Steam generators manufactured by KWU continued to use phosphate water treatment. These steam generators, tubed with Alloy 800 material, have experienced a slower rate of wastage corrosion. It is not clear whether the slower rate is a function of the alloy, the low concentration of phosphate used, the relatively low Na/PO4 ratio, or the relatively good impurity control history at several of these units. It is probable that all of these factors are involved. However, KWU has switched to all-volatile water treatment for their new units. Thus, the cause for wastage was determined to be an aggressive environment, and the corrective action involved a change in the environment for operating units from a phosphate chemistry to all-volatile water treatment. However, as discussed below, this change led to other corrosion problems, namely denting.

Denting Causes. Denting was discovered in 1975 when eddy-current probes were prevented from passing through tube/tube support plate intersections by tube diameter restrictions. By 1977, denting had become a widespread problem and had resulted in the formation of the Steam Generator Owners Group as a concerted effort to address the problem (Ref 38).

Denting is a term used to describe the localized tube diameter reduction that occurs when the hole of a carbon steel tube support plate corrodes to the point at which the corrosion products deform the steam generator tubing (Fig. 30). The cause of denting is best explained by Potter-Mann-type linear accelerated corrosion, in which a nonprotective oxide layer is formed as the corrosion progresses. These corrosion products, which have a bulk volume considerably larger than the volume of metal corroded, fill the original tube/tube support plate crevice and deform the tube and the tube support plate simultaneously.

Fig. 30 Denting at tube/tube support plate intersection. (a) Normal tube/tube support plate intersection. (b) Dented intersection. Denting results in a reduction in tube diameter and thinning and cracking of support plate ligaments. 1 in. = 25.4 mm

Sample intersections of tubes and support plates removed from dented steam generators have shown local chloride concentrations of over 4000 ppm in the dented region. The high local chloride concentration is caused by local thermalhydraulic conditions within the crevice between the tube and the tube support plate. The source of chloride is generally condenser leakage, particularly at plants cooled by seawater. Tube deformation has been reported in some steam generators with carbon steel egg crate supports, implying that concentration of chemicals and carbon steel corrosion may also be occurring with this design. Similarly, distorted eddycurrent signals (dings) have been observed at the ninth and tenth support plates of OTSGs that use a trefoil design support plate. This is a most likely region for corrosion because it is the region where dryout and most of the deposition of any chemicals present in the feedwater occur. It is the region corresponding to approximately 95 to 100% quality on the average. However, it is possible that these dings may be caused by tube vibration and impact on the tube support plates.

Denting has also been observed at the top of the tubesheet crevice in several RSGs, but it has progressed at a slower rate. Whether this is due to different geometry the presence of sludge, or different materials (different grades of low-alloy or carbon steel are used in the tubesheet and the support plate) is not known at this time. Denting has not been reported in the KWU steam generators where the support structures are fabricated of type 347 stainless steel or in the B&W/AECL units that have used lattice bars and broached hole plates with both carbon steel and type 410 stainless steel. Several major consequences can result from the uncontrolled progression of denting: tube cracking and leaking at Ubends, tube cracking and leaking of a highly concentrated acidic and oxidizing environment at the constricted crevices, tube support ligament failure, and gross deformation of tube support plates. Laboratory tests proved that denting is caused by the presence of a high concentration of chlorides in the support plate crevice and the presence of an oxidizing environment outside the crevice (Ref 39, 40, 41). This aggressive environment is brought about by the introduction of acid-forming impurities via the feed into the steam generator and subsequent concentration at the tube/tube support crevice. Concentration factors of greater than 20,000 times within the crevice have been observed in laboratory tests. Corrective Actions. Based on this knowledge of the causes of denting, the corrective actions for operating plants have

been to modify the environment and make it less aggressive. This has been accomplished by establishing secondary water chemistry guidelines based on laboratory and field data that recommend reduced levels of impurities in the steam generator (Ref 42) and by recommending methods for achieving these greatly lowered impurity levels. These methods include reducing condenser in-leakage (Ref 43, 44), reducing air in-leakage (Ref 45), producing purer makeup water, and using condensate polishers to purify the water. Another method employed to modify the environment is the addition of boric acid, which inhibits the acid chloride attack (Ref 46, 47). The application of these methods has led to greatly reduced denting rates. The PWR Secondary Water Chemistry Guidelines were prepared by a committee of industry experts under the sponsorship of the Steam Generator Owners Group (Ref 42). The guidelines combine the results of laboratory corrosion studies and evaluation of good PWR operating practice. The PWR Secondary Water Chemistry Guidelines consist of chapters covering: • • • • •

Management Responsibilities Recirculating Steam Generators Once-Through Steam Generators Analytical Methods Data Management and Surveillance

These guidelines have been endorsed by the Steam Generator Owners Group and have been generally adopted by the utilities. They have also been endorsed by the NRC and the Institute for Nuclear Power Operations. For new plants, one vendor minimized the potential of formation of the aggressive environment by switching from drilled to broached support plates having flat lands. With this design, the concentration of chemicals within the tube/tube support crevice is greatly reduced. For new plants, the corrosion was reduced by the use of more corrosion-resistant materials. Types 405 and 409 stainless steels, which are more corrosion resistant, have replaced the use of carbon steel for United States manufacturers of RSGs (Fig. 31). (Type 410 stainless steel replaced carbon steel in the B&W/AECL units.) The differences in corrosion between carbon steel and type 405 stainless steel is less at higher salt concentrations (Ref 48, 49).

Fig. 31 Corrosion of carbon steel and 12% Cr stainless steels in nickel chloride (NiCl2) solutions

In summary, denting, brought about by the corrosion of carbon steel, is caused primarily by concentration in the tube/tube support crevices of impurities in the secondary coolant. The corrective actions include modifying the environment by reducing, inhibiting, or neutralizing the impurities in the steam generator; by using a tube support design that reduces the concentrating mechanism; and by using more corrosion-resistant materials. Inside Surface (Primary-Side) SCC Causes. Primary-side SCC of Alloy 600 steam generator tubing evolved from a laboratory prediction during the 1950s and 1960s to a major degradation mechanism during the 1970s and 1980s in operating steam generators. As early as 1957, cracking of high-nickel alloys in high-purity water at 350 °C (660 °F) was reported (Ref 50). During the following years, numerous laboratory tests were performed in different environments in an attempt to duplicate and explain these observations. This work was summarized in 1975 (Ref 51), in 1981 (Ref 52), and in 1983 (Ref 53). In 1971, the laboratory cracking phenomenon in high-purity water became an in-service degradation mechanism, with the first confirmed primary-side cracking of hot leg roll transition regions at the tubesheet and suspected primary cracking in Ubends (Ref 54). Leakage at U-bends was experienced in Obrigheim steam generators, manufactured by KWU, after only 2 years of operation.

In recent years, cracking of Alloy 600 tubes from the primary side has become a problem of increasing importance. As of early 1985, this type of cracking has been experienced by a number of plants (Table 5) of some steam generator designs.

Table 5 Plants affected by primary-side cracking Type of primary-side attack(a)

U-bend tangent cracks Roll transition cracks Roll expansion cracks Explosive expansion transition cracks Total number of plants affected

(a)

Number of plants affected (as of 1985) 16 16 10 1 29

Excluding those caused by denting

Cracking in the U-bend has occurred mainly on the inner row at the apex and at the tangent points (Fig. 32). Cracks have also occurred in the tubesheet at transition expansion and roll expansion regions (Fig. 33).

Fig. 32 Location of cracking in steam generator inner row U-bends. 1 in. = 25.4 mm

Fig. 33 Details of tubesheet expansion transition region. (a) Partial tubesheet expansion transition region. (b) Full tubesheet explosive, hydraulic, or roller expansion transition region

Primary-side cracking is a form of intergranular SCC. This type of intergranular SCC, in common with other forms of SCC, occurs when certain environmental, tensile stress, and material susceptibility factors are sufficiently severe. Environmental factors influence the initiation of primary-side SCC; of these, temperature must be considered a major factor. The first roll transitions experiencing SCC were located on the hot leg side. Because the hot leg at 320 °C (610 °F) is 30 to 40 °C (55 to 70 °F) hotter than the cold leg at 280 °C (535 °F), it would appear that temperature has a significant influence on SCC initiation, indicating a thermally activated process. The reaction rate can be characterized by an Arrhenius factor, exp (-Q/kT), where Q is an apparent activation energy, k is the Boltzmann constant, and T is absolute temperature. Thus, the influence of temperature can be graphically represented by an Arrhenius plot of reciprocal

temperature versus the logarithm of crack growth rate (Ref 55) or time to failure (Ref 56). At a typical temperature difference of 30 °C (55 °F) between hot and cold legs, this energy of activation could account for a factor of four to five increase in the time to the onset of cracking. Other environmental factors that might influence crack initiation are hydrogen gas and chemical contaminants. Various investigations have shown that dissolved hydrogen gas can make an environment more aggressive (Ref 55, 57, 58). It has also been determined that the addition of lithium hydroxide and boric acid along with hydrogen (simulating PWR primary water) appears to produce results intermediate between pure water (least aggressive) and pure water with hydrogen added (most aggressive) (Ref 55). Primary water is less aggressive than pure water at the same hydrogen concentration. Tensile stresses also have a major impact on the initiation of SCC. The main source of stress is residual stress from tube manufacture and installation. Pressure and thermal stresses also play significant secondary roles. To date, only the most highly strained regions of steam generator tubing (that is, row-one U-bends, roll transition regions, expanded regions, and dented areas) have exhibited SCC. This has also been demonstrated by the results of numerous laboratory tests, which show that plastically deformed samples, that is, split tube U-bends, accelerate crack initiation (Ref 50, 55, 57, 58, 59, 60, 61). Stresses in excess of yield can lead to rapid cracking in susceptible material. A threshold stress, below which cracking will not occur, of roughly 0.8 yield for high-temperature water (Ref 55) and 0.5 yield for concentrated sodium hydroxide (Ref 61) has been determined. Material susceptibility, in combination with the factors mentioned above, can cause SCC. However, it is important to emphasize that there is not a single product called "mill-annealed" Alloy 600 tubing. Each tubing manufacturer employs a different process to produce "mill-annealed" tubing, and the resistance to SCC varies greatly with the process. Some of the mill-annealed tubing has not experienced any SCC over extended periods of operation, which SCC has occurred in other tubing after 1 to 2 years of service.

This microstructural aspect of cracking involves mostly the final mill-annealing temperature and whether precipitation of grain-boundary carbides occurs during the annealing treatment. The most susceptible microstructures are those produced by low mill-annealing temperatures (Ref 57, 62) that develop fine grain size (ASTM 9 to 11) and a copious quantity of intragranular carbide. The grain boundaries of these materials usually contain little, if any, carbide phase (Ref 57, 58, 63). The beneficial influence of grain-boundary chromium carbides on primary-side SCC resistance has received extensive evaluation over the past several years. As far back as 1973, results were published indicating that Alloy 600, given a heat treatment to precipitate grain-boundary chromium carbide, developed improved SCC resistance in high-purity water (Ref 59). Recent work has demonstrated that grain-boundary carbides improve primary-side SCC resistance with or without grain-boundary chromium depletion (Ref 57, 58, 60, 61). Annealing temperatures in the range of 980 to 1010 °C (1800 to 1850 °F) will avoid undue grain growth and provide enough dissolved carbon so that carbide precipitation will occur during cooling and subsequent thermal treatment for 15 h at 705 °C (1300 °F). The 15-h treatment provides enough time for grain-boundary chromium content to recover (Ref 58). Corrective Actions. Preventive measures against SCC are discussed in Ref 64 and are summarized below. Reduction of temperature in the inner row U-bends is one possible corrective action. By orificing the inlet, the flow

to these tubes is reduced, which would cause the primary coolant temperature to reach the cold-side temperature at the apex and thus reduce the time to crack by factors of about two to three. Stress Relief. For operating plants and for plants already built but not operating, reduction of tensile stresses by stress-

relief heat treatment and peening (discussed below) of the inside-diameter surface are the most practical approaches. Stress relief includes the use of resistance heaters, induction heaters, and global heat treatment. Resistance heaters can provide in situ stress relief of U-bends according to laboratory tests; during the stress-

relieving process, the heater is pulled up to and around the bend (Ref 65). It is believed that temperatures in a fairly broad range, even for a short time, would significantly reduce stress and improve resistance to attack. Induction heaters can be used to reduce stresses sufficiently in roll transitions to increase resistance significantly to

primary-side attack (Ref 65, 66). This type of stress relief would have the advantage of decreasing both residual stresses at inner and outer surfaces in the roll transition area without causing significant sensitization. Induction heating equipment has been developed for use in brazing sleeves inside tubes and is thus available for in situ treatment. One question that still needs to be resolved is the extent of high-temperature corrosion that would occur from any residual salts that would

be present in the gap created between the tube outside diameter and the tubesheet in the area that reaches high temperature. According to preliminary evaluation, the gap is acceptable because the crevice depth is relatively shallow (for example, 6 mm, or in.) and comparable to tube support plate crevices already present in the generator. Another question that needs to be resolved is whether unacceptable tube axial stresses could be developed by the stress relief. Global heat treatment as a means of stress relief involves heating the entire tubesheet area to a temperature of about

610 °C (1130 °F) in order to relieve residual stresses and to obtain microstructural improvement (carbide precipitation at the grain boundaries). It would be performed by using heaters placed under the tubesheet and along the outside of the shell up to about one shell diameter above the tubesheet. Separately controlled heaters would be located along the divider plate and around the channel head in order to ensure that large thermal expansion stresses do not develop at divider plate junctions. This procedure would require only a few weeks of critical path time and would minimize the amount of highradiation exposure work required in the channel head. Before application, plant operators should resolve specific questions, such as the effect of residual salts, the factor of improvement expected from the stress relief, and whether the carbide precipitation obtained during the stress relief could cause Alloy 600 to become susceptible to attack by sulfur species such as polythionic acid. This would require increased care in water chemistry control. In the evaluation of various stress-relieving methods, plant operators need to assess the benefits and risk factors connected with the different corrective measures. Shot peening to produce residual surface compressive stresses is a well-known approach for providing resistance to

stress corrosion. Shot peening has been investigated for use inside tubes in the tubesheet area of steam generators (Ref 67) and has been performed in a number of plants. Shot are blow up the tube so that the shot impinge on a conical deflector located in the area to be peened. It is possible that the same approach could be used for U-bends. Before applying this method to the U-bend and/or tubesheet region it is necessary to ensure that a compromise can be reached between the induced inside-diameter compressive stresses and the resultant outside-diameter tensile stresses. Thus, the total tube SCC resistance, inside and outside diameter, would be optimized. Rotopeening has been performed for both radioactive and nonradioactive plants to reduce the stresses in the expansion

transition and expanded areas. It is performed by using beads bonded to fabric flappers that are rotated in a flapper wheel arrangement such that the beads impact the tube inside-diameter surface. The compressive stresses developed on the tube inside diameter by the peening must be balanced by tensile stresses in the remaining wall thickness. Excessive inside-diameter peening could lead to significant tensile outside-diameter surface stresses. These stresses, combined with applied pressure and thermal stresses, might then aggravate attack at the outer surface. This possibility needs to be evaluated in relation to water chemistry and possible sludge pile accumulation. If cracks are already present at the time peening is performed, the effects of the stresses induced by the peening on crack growth need to be assessed. It is suspected that peening will serve to prevent initiation of new cracks, but will not prevent growth of existing cracks through the wall thickness. If this is correct, then peening will be most useful in cases in which most tubes have either no cracks or only very small ones. Sleeving is another repair method for modifying the tubing material in operating plants. Sleeves up 1120 mm (44 in.)

long have been installed in the region between the lower face of the tubesheet and the first support plate as a corrective measure against pitting and intergranular attack at the outer surface. In some cases, even longer sleeves have been installed. The sleeves bridge the damaged area and are attached to sound material beyond either end of the damage. The ends of the sleeves are expanded hydraulically or explosively and are in most cases sealed by rolling, welding, or brazing (examples of sleeves are shown in Fig. 34). A discussion of tube sleeving development for OSTGs is presented in Ref 68.

Fig. 34 Example of sleeve designs for protecting tubing material. (a) Combustion Engineering welded sleeve. (b) B&W regular length sleeve. (c) Westinghouse-type sleeve

For more recent plants, the corrective action has been to use tubing made of thermally treated Alloy 600. Extensive tests have shown this material to be greatly superior to mill-annealed Alloy 600 (Ref 57, 58, 61). The thermal treatment involves a final mill-annealing temperature of 980 to 1010 °C (1800 to 1850 °F), followed by a 705-°C (1300-°F) treatment for 15 h to produce a semicontinuous grain-boundary precipitation (Ref 58). More recent steam generators are being fabricated by using thermally treated Alloy 690. It has been reported that this alloy is superior to thermally treated Alloy 600 from the standpoint of resistance to SCC. Intergranular Attack and SCC at the Outside Surface Causes. Intergranular corrosion in one or another of its various forms has been experienced in many steam generators operating with seawater- or freshwater-cooled locations. The rate of propagation has been shown to vary widely, depending on the form of the attack. Progression rates in some cases have been sufficiently rapid to require mid-cycle inspections and unscheduled outages to plug or repair leaking tubes and to cause significant economic loss to the PWR operator. Fortunately, most of the corrosion has been confined to crevice locations so that leaks have been small and without risk of a large rupture. Intergranular corrosion has been found to take various forms, such as intergranular SCC, intergranular attack, and intergranular penetration. Intergranular SCC in Alloy 600 steam generator tubing is illustrated in Fig. 35. In the case of intergranular SCC, the

corrosion morphology consists of single or multiple major cracks with minor-to-moderate amounts of branching. Cracks propagate intergranularly in essentially all cases in Alloy 600 tubing. Experience suggests that intergranular SCC requires stresses greater than 0.5 yield in order to propagate rapidly. At lower levels, propagation rates may approach 0, or the corrosion may take another intergranular form.

Fig. 35 Intergranular SCC in Alloy 600 C-rings. Source: Ref 69

Intergranular attack, a second form of intergranular corrosion, has been described as general intergranular attack or

volumetric intergranular attack (Fig. 36). Its morphology is characterized by a uniform or relatively uniform attack of all grain boundaries over the surface of the tubing. In the purest case, stress does not contribute to the morphology of intergranular attack, which distinguishes this phenomenon from intergranular SCC. However, the close relationship between intergranular attack and intergranular SCC is apparent in Fig. 36, in which a stress-assisted finger of corrosion penetrated from the layer of intergranular attack into the tube material.

Fig. 36 Integranular attack of Alloy 600 steam generator tubing. Etched sample. Source: Ref 70

Intergranular penetration, the third form of intergranular corrosion, has been variously described as a mixture or a

hybrid of the other two forms. Within the steam generator, intergranular corrosion has been found in a variety of locations. The most important of these locations are the tube/tubesheet crevice (an annular gap remaining after steam generator manufacture), as shown in Fig. 37 , and the tube/tube support plate crevice. Intergranular corrosion at this latter location has been found at several Japanese sites, but has not occurred widely in the United States. Corrosion has also been observed in the sludge region above the tubesheet. All of the steam generators experiencing secondary-side intergranular corrosion are tubed with millannealed Alloy 600. In laboratory tests, however, intergranular corrosion has been produced in mill-annealed Alloys 600, 690, and 800. Based on these findings, there appears to be little difference in performance among these alloys in the millannealed condition (Ref 33, 34).

Fig. 37 Typical steam generator tube showing the tubesheet annular crevice. 1 in. = 25.4 mm

Tests have shown how aggressive chemicals can concentrate in these crevices and in the sludge under heat transfer conditions (Ref 71, 72); alternate wetting and drying is a particularly effective concentration mechanism. Five classes of environmental contaminants have been postulated to explain the occurrence of intergranular corrosion (Ref 33). These include high concentrations of sodium hydroxide (NaOH) and/or potassium hydroxide (KOH), the products from the reaction of sulfate ions with hydrazine or hydrogen (reactive sulfur-bearing species are postulated), the products of thermal decomposition of ion exchange resins (sulfates and organic residuals), highly concentrated salt solutions at neutral or nearly neutral pH (these salt solutions are the natural consequences of condenser leakage concentrated to high levels of salt by the boiling processes in the steam generator), and alkaline carbonates and/or their reaction or hydrolysis products (believed to affect the nature of the passive film on the alloy surface). Corrective Actions. It has been shown with reasonable certainty that increased service stress, dynamic strain, and a

high residual stress level can all be major factors in accelerating initiation or propagation of intergranular SCC. A similar statement for intergranular attack cannot be made, because experimental results have shown that intergranular attack can develop even with compressive stress (Ref 58). Stress-relief treatment of tubing in new steam generators is accomplished by installing thermally treated Alloy 600 tubing that has been aged at 705 °C (1300 °F) for 15 h. Also, one manufacturer uses a full bundle stress relief after steam generator fabrication to provide stress relief. No practical procedure for the application of thermal stress relief to the secondary side of in-service tubing has yet been proposed.

For new plants, modification of the materials by thermal treatment to improve the microstructure can reduce the susceptibility to caustic-induced intergranular SCC. For operating plants, sleeving can be applied to mitigate the effect of intergranular corrosion. The modified material or sleeve should be resistant to faulted secondary environments; for example, sleeves have used a duplex sleeve with a nickel or nickel alloy outside layer (which is considered immune to caustic attack). Modification of the crevice environment appears to be the most straightforward method of preventing or arresting intergranular attack and may apply equally to the intergranular SCC. Modification can include several factors, such as lowering the temperature, adding a pH neutralizer, removing the corrodent by flushing or soaking, and changing the concentration and/or ratio of bulk water contaminants. Laboratory studies have confirmed the benefit of several of these corrective measures; some of the modifications have been applied to operating steam generators. The operating temperature of one unit was lowered for several years, which effectively reduced the rate of progression of the intergranular attack. The hot leg temperature was lowered to approximately the normal cold leg temperature. A necessary side effect of significant temperature reductions is loss of power. Based on the experience of one utility, it is judged that the temperature must be reduced to 300 °C (575 °F) or lower to have a major impact on corrosion rate. Boric acid has been added during off-line crevice-flushing operations to reduce the pH level. The pH neutralizer was used on the basis of limited laboratory data regarding its effectiveness. Tubes subsequently removed from one of the steam generators were found to have boric acid present over the full length of the tube surface within the tubesheet crevice. No estimate can be made as to effectiveness. Equilibrium calculations suggest that high levels of boric acid (equivalent to an Na/B ratio approximately equal to 1) will be required to reduce the pH significantly when the pH is held in the alkaline range by NaOH or KOH (very strong bases). On-line addition of boric acid has been employed at several plants. Flushing or soaking off-line maintenance procedures have been used with varying results at several plants. The

flushing procedures are based on laboratory tests and involve a depressurization that causes boiling within the tube/tubesheet crevice and ejection of concentrated solutions from the crevice. Optimization of crevice-flushing procedures for each unit and periodic repetitive application of the optimized procedures are recommended. It should be recognized that flushing or adding a pH neutralizer may be difficult if denting has occurred at the top of the tubesheet or if the annulus is fouled with corrosion products, thus blocking access to the crevice. The concentration of contaminants accessible to the crevice may be controlled by eliminating or reducing the

entry of contaminants to the steam generator and by controlling the concentrating capability of the sludge pile above the crevice. Reducing the entry of contaminants is best accomplished by preventing condenser leaks, routing drains to the condenser (if condensate is subsequently treated), and properly treating makeup water sources. Use of full-flow condensate polishers for control of ionic species has not been shown to be effective in controlling the species that are

probably responsible for intergranular attack and intergranular SCC, as was shown for control of chloride responsible for denting. Polishers are potential sources of sulfates (SO4) and sodium (Na+) if operated or regenerated improperly, and they do not effectively remove silica or organics. However, some plants have used these condensate polishers very effectively to minimize impurity entry into the steam generators. Control of the sludge pile, which is an effective concentrating mechanism, requires three courses of action:

• • •

Effective, periodic sludge lancing Minimization of particulate transport by preventing air entry and/or providing for feedwater filtration, such as by powdered resin condensate polishers Preventing the entry of chemical species that tend to promote agglomeration

Chemical cleaning has been used to remove the sludge on the tubesheet. One group of utilities has attempted to modify the aggressive environment in the tube/tubesheet crevices by fully expanding the tubes in the tubesheet, thus eliminating the crevice. Pitting Causes. Extensive pitting on the outer surfaces of tubes has been observed in two units. The pitting occurred primarily

on the cold leg between the tubesheet and the first support plate in regions where sludge or tube scale was present. Figure 38 shows a typical steam generator pit. The laminar appearance is caused by the presence of metallic copper layers. In addition, the corrosion deposit is enriched in chromium and depleted in nickel and iron compared to the base metal. Laboratory tests have shown that pits can be formed in the presence of high chloride ion concentration from either seawater or copper chloride (Ref 74). Thus, it is concluded that the pits are caused by chloride, low pH, and an oxidant such as cupric chloride (CuCl2) or oxygen. Temperatures above about 150 °C (300 °F) are required to form pits such as those observed in the operating units. It is further believed that sludge and scale act as a medium in which the bulk impurities are concentrated to higher levels by the boiling action.

Fig. 38 Pitting of a steam generator tube. Source: Ref 73

Corrective Actions. For existing plants, the tubing is repaired and retained in service by the use of sleeves. Sleeves

may be of one alloy or a bimetallic alloy in which the outer alloy is selected for resistance to pitting corrosion and the inner alloy is selected for primary-side corrosion resistance. The principal corrective action is to modify the environment to make it less corrosive. One approach is to reduce the sludge and scale by minimizing the entry of solids (higher pH and reduced air in-leakage), by sludge lancing, and by chemical cleaning. All of these methods are now being used. Another approach is to minimize the entry of soluble contaminants (principally chlorides and oxidants) by following the Secondary Water Chemistry Guidelines (Ref 42), by quickly repairing leaky condensers, and by deaerating auxiliary feedwater used for hot standby. Corrosion Fatigue Causes. Cracking of some tubes from the outer surface has occurred in the upper regions of several OTSGs. It is

believed that these cracks are caused by corrosion fatigue resulting from small-amplitude vibration combined with the transport of impurities into the upper regions of the OTSG units, particularly in the open lane. Laboratory tests have shown a decrease in Alloy 600 fatigue strength in the presence of chemicals that are judged to be present in these upper regions (Ref 75).

In these tests, a substantial decrease in fatigue resistance of Alloy 600 was observed in acid sulfate/silicate solution. The environment was selected to be consistent with that postulated to exist in the region of interest at the top of an OTSG. Figure 39 shows the response of Alloy 600 in an aggressive (pH 2) sodium sulfate/sodium silicate/sulfuric acid environment. The fatigue strength at 107 cycles at 289 °C (552 °F) was found to be approximately 56% of the value measured in air at about the same temperature. Aerated acid sulfate/silicate and alkaline sulfate/silicate environments had a less deleterious effect.

Fig. 39 Corrosion fatigue effects of Alloy 600

Corrective Action. The corrective action is to modify the environment to make it less aggressive. This can be done by

minimizing the entry of these impurities into the steam generator. Also, the use of mechanical flow-diverting lane blockers has been tested and shown to be a useful steam generator modification; they reduce the flow of liquid droplets to the upper regions via the open lane (Ref 76). In new plants, the open lane has been eliminated. Other Phenomena A variety of additional factors have contributed to lower-than-desired steam generator availability or have caused the utility operators to incur repair costs. These phenomena will not be discussed in this section. Some of the mechanical problems and proposed or actual corrective actions are summarized briefly in Table 6. References 77 and 78 provide information that is particularly useful to those utility operators that are assessing whether to repair or replace their steam generators.

Table 6 Corrective actions for steam generator mechanical problems Problem Erosion-corrosion of feedwater nozzles Loose parts fretting and wear Wear at antivibration bar locations Wear in preheat units Water hammer

Corrective action Higher pH coolant; replace with higher chromium tubing Inspection; keep loose parts out Replace with wider bars, improved materials; reduce tube-to-bar gap Improved flow distribution baffle; reduced tube to support plate clearance Split feed flow Modify feed design and feed rings to avoid drainage of feed ring

Corrosion of Zircaloy-Clad LWR Fuel Rods A.J. Machiels, Electric Power Research Institute

Almost all power reactors in the world are cooled by either normal or heavy water and are termed light water reactors (LWRs) or heavy water reactors (HWRs), respectively. The heat-generating fuel elements, or fuel rods, are made of stacks of uranium dioxide (UO2) pellets clad or sheathed in hermetically sealed tubes. Fuel rods are grouped together in assemblies whose designs differ markedly according to the type of reactor (Ref 79). A primary fuel assembly reliability consideration is the ability to maintain optimum power generation over the design lifetime of the assembly without releasing the radioactive by-products of the nuclear fission process into the primary coolant (water). Therefore, the cladding or sheathing tubes separating the coolant water from the nuclear fuel and its byproducts must resist the corrosive attacks of both environments while meeting all mechanical loads present or anticipated during power and fuel-handling operations. Although the first LWRs in the United States used stainless steel as a cladding material, the economics of LWR-generated electricity favor the use of zirconium-base alloy (Zircaloys) wherever possible in the high-flux regions of the reactor core. The reason for this is the low neutron absorption cross section of zirconium. As a result, tubes and other assembly structural components are for the most part made of zirconium alloys. The primary zirconium alloys are known as Zircaloy-2 and Zircaloy-4. They are mostly zirconium ( 98 wt%) with low amounts of alloying elements (tin, iron, chromium, nickel, and oxygen), which confer to the alloys the desirable mechanical and corrosion properties. Zirconium-niobium alloys are also used in Canada and Russia. Zircaloy-2 and Zircaloy-4 were selected during the early days of the U.S. Nuclear Navy program on the basis of relatively few data (Ref 80), and it is now generally accepted that the formation of these alloys is not optimized with regard to their corrosion resistance to water, which is the topic of this Section. The discussion that follows is restricted to the water-side corrosion properties of Zircaloy under normal operating conditions in LWRs. Oxidation during high-temperature transients leading to core damage is beyond the scope of this Section, but was recently reviewed (Ref 81). A brief introduction to corrosion under isothermal conditions in high-temperature water is first presented. The applications to pressurized PWRs and BWRs are then discussed. Also, given the large number of reviews already devoted to the topic, this discussion is limited to the more recent developments that have, or will have, an impact on the corrosion performance of the Zircaloys. For additional information on the corrosion performance of Zircaloy alloys, see the article "Corrosion of Zirconium and Hafnium" in this Volume.

Corrosion in High-Temperature Water Light water reactors normally operate with a coolant temperature between 250 and 350 °C (480 and 660 °F) at a pressure of either approximately 7 MPa (1000 psi) (BWRs) or 15 MPa (2200 psi) (PWRs). Therefore, out-of-reactor corrosion studies are usually conducted in high-pressure autoclaves, in which the water is either in the liquid state at temperatures up to approximately 360 °C (680 °F) or in the gaseous state (steam) at temperatures of approximately 400 °C (750 °F) and higher. Under these conditions, the attack of Zircaloy by water is generally uniform. With steam at high temperature

(typically 500 °C, or 930 °F) and high pressure (typically 10 MPa, or 1500 psi), this uniform attack can be accompanied by a more localized form called nodular corrosion. The reaction between Zircaloy and water can be written as:

Zr + 2H2O 4×H

ZrO2 + 2(1 - x)H2 +

(Eq 1)

where the first and second left-hand terms represent zirconium (of which Zircaloy is mostly made) and water, respectively; the right-hand terms represent zirconium oxide, hydrogen that is released into the corroding water, and hydrogen that is picked up by the Zircaloy; x is the pickup fraction. The pickup of hydrogen by the Zircaloy substrate leads to changes in its mechanical properties and, in particular, to a loss of ductility. The formation of zirconium oxide results in a loss of metallic Zircaloy. Because the Pilling-Bedworth ratio is equal to 1.56 at ambient temperature, a given loss in metal thickness results in a larger gain in oxide thickness (the PillingBedworth ratio is explained in the article "Fundamentals of Corrosion in Gases" in this Volume). In Fig. 40, the positions of the Zircaloy/ZrO2 and ZrO2/H2O interfaces are shown relative to the position of the original Zircaloy/H2O interface.

Fig. 40 Zircaloy/zirconium oxide/water interface

The kinetics of the uniform corrosion reaction can be conveniently divided into two main periods, referred to as pretransition and posttransition. In the first of the periods, a black coherent oxide film is formed, and the corrosion rate diminishes with time according to a rate law given by the simplified form: (weight gain) = constant x (time)n, with the exponent n in the range of 0.25 to 0.5. A transition follows to a period where the rate law is closer to linear (n 1). This transition is eventually accompanied by a conversion of the black oxide to a gray or white oxide. The corrosion kinetics of Zircaloy in high-temperature water actually follow a periodic behavior, and the apparently linear dependence of the posttransition rate is obtained by averaging the time-dependent rates over several periods (Ref 82). A progressive acceleration of the posttransition rate, especially after sufficiently long reaction times, has also been shown (Ref 83). Given that zirconium is a most reactive metal, it should readily corrode according to Eq 1. However, because of the protective nature of at least part of the oxide film, Zircaloy belongs to the group of extremely corrosion-resistant alloys. The film acts as a barrier through which the reactive species (oxygen ions, electrons, etc.) must diffuse to sustain the corrosion reaction (Ref 84). During pretransition, reaction rates are determined by diffusion through a surface film of increasing thickness. At transition, which typically occurs when the oxide layer is 2 to 3 m thick, part of the film undergoes a change in morphology and becomes nonprotective (Ref 85). During posttransition, the thickness of the protective part of the oxide film successively increases and abruptly decreases, yielding cyclically time-dependent kinetics. As discussed in the previous paragraph, linearization over a sufficiently long period of time can be readily performed. This is equivalent to assuming the existence of a protective barrier of constant thickness. Under typical hightemperature water autoclave conditions, this equivalent thickness is about 1 m (Ref 86).

The out-of-reactor corrosion properties of Zircaloys are, in general, strongly dependent on the material microstructure and on temperature. In some cases, these properties also depend on the water chemistry conditions. Zircaloy Material. The corrosion resistance of Zircaloys depends on a number of factors that include the concentration and distribution of alloying elements and impurity elements. These factors in turn depend on the thermomechanical processing history, which is an essential part of the manufacturing sequence. Other manufacturing variables that are of significance in determining corrosion resistance are related to the surface treatment of the final component, such as grinding, grit blasting, or pickling.

The fabrication variables or the manufacturing variables involved in the production of zirconium sponge are not considered in this section. However, corrosion properties are very sensitive to a number of key manufacturing parameters, such as quenching rates and annealing times and temperatures (Ref 87). Moreover, some variability in these parameters always accompanies the large-scale production of Zircaloy components, even under nominally identical conditions. Therefore, the corrosion performance of a given population of Zircaloy components is generally characterized by a range of corrosion behaviors, the extent of which can be very significant. Temperature. Corrosion studies in autoclaves under isothermal conditions have led to a number of mathematical

expressions, such as (Ref 88):

(Eq 2a)

(Eq 2b)

(Eq 2c)

where W is the specimen weight gain (in mg/dm2); Wt is the specimen weight gain at transition (in mg/dm2); t is the total exposure time (in days); tt is the time to transition (in days); R is the gas constant, 1.987 cal/(mol·K); and T is the absolute temperature (in degrees Kelvin). In the temperature range of 280 to 400 °C (535 to 750 °F), the pretransition rates increase by a factor of two for each 45- to 50-°C (80- to 90-°F) temperature increment; more important, the posttransition corrosion rates increase by a factor of two for each 16- to 20-°C (29 to 36-°F) temperature increment. Water Chemistry Conditions. The protective nature of the diffusion barrier can be drastically altered by the presence

of some species in the corroding water. Most autoclave experiments use pure water only. However, because of its relevance of LWR technology, lithium hydroxide (LiOH) is sometimes added to the water. At sufficiently high concentrations, the presence of LiOH can lead to rapid corrosion of the Zircaloys (Ref 89).

Corrosion in LWRs Early experiments on zirconium-alloy corrosion in reactors showed that significant differences in corrosion behavior could occur (Ref 90, 91). Compared to a high-temperature water autoclave, the reactor environment is characterized by the presence of an intense radiation field, the existence of large temperature gradients across the Zircaloy-water contact layers, and the presence of impurities and soluble chemical additives that are usually not present in high-temperature water autoclaves. Subsequent investigations have shown that each of these factors, separately or in combination, can be important. Water Chemistry. Boiling water reactors normally operate with high-purity water. However, radiolytic decomposition

of the water eventually results in the production of stoichiometric amounts of hydrogen and oxygen. Partitioning of those two gases occurs between the steam phase, which is continuously extracted from the system, and the liquid phase, which

is recirculated. Typical steady-state concentrations of hydrogen and oxygen in the recirculation loops are, respectively, 20 and 200 ppb. Pressurized water reactors operate with a hydrogen overpressure sufficient to inhibit the formation of radiolytic oxygen in the coolant, a basic additive (LiOH) to control the release of corrosion products into the coolant and their deposition on fuel surfaces, and a chemical shim (orthoboric acid, H3BO3) to control the nuclear reactivity. Under these conditions, typical hydrogen and oxygen concentrations are maintained at >2000 ppb and less than 5 ppb, respectively. In both systems, the structural components of the reactor system are the source of corrosion products generally characterized by low solubilities. As they are transported throughout the system by the coolant, they tend to form deposits on both in-core and out-of-core surfaces. These deposits, along with the coolant-borne impurities, are referred to as crud. Crud deposits consist mostly of iron oxides with high porosities (65 to 85%). Magnetite (Fe3O4), in which the iron is partially replaced by the other constituents (such as nickel or chromium) of the alloys exposed to the coolant, and hematite (Fe2O3) are the most common crud constituents in PWRs and BWRs, respectively. Formation of crud deposits on fuel element surfaces is largely dictated by solubility and heat transfer considerations. Under conditions that favor negative solubility temperature coefficient (that is, solubility decreases when temperature increases), or the formation of concentration cells created by boiling heat transfer, or both, crud deposition readily occurs. Therefore, the important water chemistry parameters include: • • •

Concentrations of dissolved oxygen and hydrogen Concentrations of chemical additives (PWR technology only) Concentrations and precipitation characteristics of the coolant-borne impurities

The effects of these parameters can be very important in the presence of a reactor radiation field, or a large enough temperature gradient, or both. Effect of Radiation. Radiation can affect the corrosion behavior of Zircaloys by modifying the aggressivity of water

through the formation of highly reactive, oxygenated radicals; by modifying the diffusive properties of the protective oxide layer through the formation of point defects; and by modifying the microstructure of the Zircaloy substrate by changing the concentration and distribution of some elements in the Zircaloy material close to the Zircaloy/zirconium oxide interface. Effects due to the second item appear to be relatively minor. Effects due to the last item are potentially important, but little is known in this area. Therefore, these radiation effects will not be considered any further in this section. As already discussed, the BWR and PWR environments lead to significantly different oxygen concentrations in the coolant. In BWRs, where the oxygen content of the coolant water is classified as high, irradiation clearly enhances the oxidation rate. The formation of a uniform oxide layer proceeds at a rate greater than that measured under similar temperature conditions in an autoclave and is generally accompanied by the appearance of locally thicker patches of zirconium oxide having the form of nodules or pustules (Fig. 41). Initially, the nodules appear as white patches on the black pretransition oxide surface (Fig. 42); with exposure, they grow in diameter as well as in thickness, and they eventually cover the entire exposed surface.

Fig. 41 Photomicrograph showing the uniform and nodular oxides

Fig. 42 White oxide nodules on a black, pretransition oxide

In PWRs, where the oxygen content of the coolant water is very low, only the formation of a uniform film is observed. There remains some disagreement as to whether or not irradiation enhances oxidation under those PWR conditions (Ref 92). Estimates of the magnitude of acceleration generally vary between none and a factor of four. Up to oxide thicknesses of 5 m, no acceleration is detected; however, for oxide thicknesses greater than 15 to 25 m, some acceleration is usually observed. Effect of Heat Flux. The transport of heat from the nuclear fuel pellets to the coolant produces large surface heat fluxes. Both the magnitude of the heat flux and the mode of heat transfer between the fuel element outer surface and the coolant have significant effects on the corrosion performance of the Zircaloy cladding. Temperature gradients in the cladding itself influence the distribution of the hydrogen picked up by the alloy during the corrosion process (Eq 1), and temperature gradients across the cladding/coolant interface increase the temperature of the protective oxide layer.

Acceleration of the corrosion process by the hydrogen produced by the corrosion reaction is a possibility, but the existence and magnitude of this effect under operating conditions applicable to nuclear fuel elements remain to be established. The thermal acceleration obtained by increasing the interface temperature is, in principle, more straightforward to evaluate. With reference to Fig. 43, which schematically represents the temperature profile across the Zircaloy/coolant interface at some axial elevation of a heat-generating fuel element, it can be seen that the average temperature of the protective oxide layer, TI, located at the metal/oxide interface is given by:

TI = TB +

TCO +

TCR +

TOX

(Eq 3)

where TB is the bulk coolant temperature and TCO, TCR, and TOX denote the temperature differences across the crud/coolant interface, the crud layer of thickness , and the oxide layer of thickness , respectively. Because the oxide thickness, , varies between 0 at the beginning of power generation to its maximum value, end of power generation, TOX is a function of time and is given by:

MAX,

at the

(Eq 4) where t denotes time q'' is the surface heat flux, and kOX is the thermal conductivity of the oxide. Assuming steady-state reactor operating conditions resulting in constant TB, TCO, and TCR, that is:

TB +

TCO +

TCR = TO

constant

where TO is the temperature of the outer surface of the oxide layer. Eq 3 can be rewritten as:

(Eq 5)

(Eq 6)

Transforming Eq 2b from weight gain to oxide thickness and differentiating with regard to time leads to:

(Eq 7) where K and Q are two constants that can be readily evaluated. Substituting Eq 6 into Eq 7 yields after some manipulation:

(Eq 8a)

where A(t) is a time-dependent thermal acceleration factor equal to:

A(t) = exp [

O

(Eq 8b)

· q'' · (t)]

with

(Eq 8c)

Therefore, in the presence of a heat flux, oxidation rates are obtained by multiplying the oxidation rate calculated at a temperature TO by a time-dependent thermal acceleration factor that depends exponentially on the oxide layer thickness and on the surface heat flux. TO < TSAT. When TB is significantly lower than the saturation temperature, TSAT, corresponding to the system pressure, as

it is for the major part of a PWR core, TO, is given by:

(Eq 9) where is the convective heat transfer coefficient; is a strong function of the local flow turbulence and assumes typical values of 3 to 6 W/cm2·K (Ref 88). The thermal conductivity of the crud is denoted by kCR and assumes typical values of approximately 1 W/m·K (Ref 93). Equation 9 predicts that large heat fluxes produce large temperature differences; as a result, TO can be significantly higher than TB, especially when substantial amounts of crud deposits are present. Under present recommended operating practices for PWRs, crud accumulation is nominal ( 10 m); therefore, its impact tends to be small. TO

TSAT. When TB is not significantly lower than TSAT, as it is in BWRs and in parts of some PWRs, boiling heat

transfer dominates. Except when dense crud deposits are present, the maximum value for TO (TO,MAX) is given by:

TO,MAX = TSAT +

T1

(Eq 10)

where T1 represents the amount of superheat needed to sustain the boiling process. The evaluation of T1 is complicated by the fact that boiling heat transfer promotes the concentration of coolant impurities and additives, especially when the change of phase occurs at fixed locations along the heat transfer surface. These conditions may exist with noncrudded rods when steam bubbles originate from the same sites (Ref 94), as well as with crudded rods when wick

boiling is present (Ref 93). For example, a simple evaluation of the maximum concentration factor, FMAX, that can be obtain under wick-boiling conditions is given by (Ref 95):

(Eq 11)

where CB is the bulk coolant concentration of a given species, and CMAX is its maximum concentration in a crud of thickness . The diffusivity of the species in water is given by D, p is the crud porosity, LVAP is the latent heat of evaporation, and d is the density of water. Soluble species may lead to an elevation of the boiling point when CMAX is large enough. When this is the case, T1 is given by:

T1 =

TBPE +

T2

(Eq 12)

where TBPE represents the boiling point elevation, and T2 is the superheat calculated from boiling heat transfer considerations. The value T2 is typically small (a few degrees Celsius). The value TBPE is expected to be negligible in BWRs, but not necessarily so in PWRs, because of the additives present in the coolant. Species with a low-to-moderate solubility may precipitate for CMAX large enough. A few impurities, most notably copper, lead to the formation of dense cruds, through which heat can be transferred by conduction only. The value TO,MAX is then given by:

(Eq 13) Large amounts of dense cruds ( 100 m) can result in cladding temperatures that are approximately 50 °C (90 °F) higher than those obtained in the presence of porous cruds for typical heat fluxes in BWRs. Boiling also influences the radiolytic conditions existing in the outer layers of the heat-generating fuel elements. A quantitative description of the effect of boiling on Zircaloy corrosion is proposed in Ref 96. Application to BWRs. In BWRs, the temperature is practically constant along the length of a fuel element, and TO,MAX

is approximately equal to the system saturation temperature, typically about 290 °C (555 °F). Under these conditions, uniform corrosion rates are low, even when an enhancement factor is factored in to take into account the oxygenated conditions of the coolant. The formation of a uniform film is accompanied by the nucleation and growth of oxide nodules. The nodular oxide thickness is mainly a function of burnup with a tendency to saturation (Ref 97). Zircaloy corrosion in BWRs has not been a problem except in reactors characterized by high concentrations of soluble copper in the reactor water. Failures have been observed, in particular with fuel claddings having high nodular corrosion susceptibility (Ref 98, 99). Those claddings develop nodular oxide layers that contain sizable cracks through which heat transfer appears to be occurring by wick boiling. As discussed above, although wick boiling itself leads to negligible superheats, it can produce large concentration effects. When copper concentrations (CMAX, Eq 11) are large enough, copper compounds precipitate and progressively plug the network of cracks necessary to sustain the wick-boiling process. This condition eventually leads to a significant temperature increase of the cladding because TO,MAX is now given by Eq 13 rather than Eq 10. Thermal acceleration of the corrosion reaction can eventually lead to fuel failures. As can be seen from Eq 11, CMAX is linearly dependent on CB, but depends exponentially on the heat flux, q'', and the nodular corrosion properties of the cladding through /p, where and p now represent the thickness and porosity of the cracked nodular oxide. The latter depend on the corrosion properties of the Zircaloy material. The recent implementation of adding hydrogen to the feedwater of several operating BWRs to prevent intergranular SCC in the recirculation system piping has created a water chemistry in the BWR core that is outside the experience base of either BWRs or PWRs. The effects of lowering the concentration of oxygen in the reactor water should be beneficial as

far as oxidation rates are concerned. The effects on hydriding rates are more difficult to assess because the pickup fraction could be directly influenced by the hydrogen dissolved in the water. Overall, the effects of hydrogen addition on oxidation and hydriding are expected to be small (Ref 100). After 1 cycle of hydrogen water chemistry at Dresden-2, the fuel examination results show that cladding oxidation is well below any performance concerns. Also, for the Zircaloy components that were examined, it does not appear that, within the limits of experimental uncertainties, hydriding rates are directly influenced by the presence of hydrogen in the coolant. Additional examinations after 2 and 3 cycles of hydrogen water chemistry are planned (Ref 101). Application to PWRs. Because boiling heat transfer also occurs locally in PWRs--especially in the more recently

designed ones--TO,MAX is given by Eq 10, in which TSAT is typically 340 °C (645 °F). The evaluation of T1 is complicated by the presence of LiOH and boric acid in the reactor coolant. Because boiling promotes the concentration of these additives on the outer surfaces of the fuel elements, some local elevation of the boiling point is likely. Lithium hydroxide has a weak effect on the boiling point elevation of water, but the impact of boric acid may be more significant (Ref 93). Uncertainties associated with the accurate evaluation of the interface temperature have hampered efforts to quantify the effects of other phenomena, such as those resulting from the presence of a radiation field. Zircaloy corrosion in PWRs has become an area of prime concern from a fuel performance point of view. In particular, fuel failures and plant derating due to corrosion have occurred in plants that combine high coolant temperature and high heat flux. Therefore, the full economic benefits associated with high thermal efficiency, high burnup, higher-pH coolant (by adding greater amounts of LiOH), and higher fuel utilization (by fuel rod design changes) may not be obtainable in many existing plants without some improvement in the corrosion properties of the current zirconium alloys. Influence of Corrosion on Radiation Fields Robert A. Shaw, Electric Power Research Institute

Radiation fields exist in nuclear power plants primarily because of the deposition of radioisotopes on the surfaces of pipes and other components. These radiation fields can significantly influence the operation and maintenance of nuclear power plants. Consequently, the control of these radiation fields wields an influence over the cost of the generation of nuclear power. The radiation exposure experienced by the personnel who work in nuclear power plants is a key measure of the effectiveness of radiation field control measures. Figure 44 traces the history of the media of plant exposures for BWRs and PWRs. It indicates a regular, continuing increase until the early 1980s and some reduction after 1983. Significant contributors to the exposure peaks in the early 1980s have been the materials problems experienced in these plants. Pipe cracking and the consequent inspection and replacement of such pipes caused larger-than-normal radiation exposures to be experienced in BWRs. For PWRs, steam generator materials degradation caused significant increase in inspection and, in some cases, replacement of steam generators, causing a similar increase in exposure at roughly the same time.

Fig. 44 Median of U.S. nuclear power plant radiation exposures. Rem, roentgen equivalent man

Concurrent with these problems, which caused increase in exposure, has been the application of a number of techniques for reducing radiation exposure in plants. These have included techniques for reducing the radiation fields such as water chemistry control techniques, materials selection emphasizing replacement of cobalt alloys and decontamination to remove radioisotopes from surfaces. In addition to these techniques for reducing radiation fields, radiation exposures have been reduced through the use of remote equipment, shielding, and extensive planning and training for high exposure tasks. A second measure of expenditure of radiation exposure is presented in Fig. 45 and 46. These diagrams show the man rem/MW·yr, an expression of the radiation exposure that has been required in order to generate electrical energy at nuclear power plants. These two figure compare the experience of the United States with PWRs and BWRs, respectively, to that of other countries. It shows the experience in other countries to be superior to that of the U.S. and stresses the need for effective radiation control programs at U.S. plants. For additional information on techniques to control radiation fields in nuclear reactors, see Ref 102, 103, 104, 105, 106.

Fig. 45 Occupational radiation exposure per unit electricity generated by PWR nuclear power plants

Fig. 46 Occupational radiation exposure per unit electricity generated by BWR nuclear power plants

Radiation Sources There are two primary sources of radioisotopes generated in nuclear power plants. The first is within the fuel itself where the fissioning process creates fission products and their decay products, which are radioactive. In modern nuclear power plants, fission products are effectively constrained to remain within the Zircaloy cladding present on each fuel rod. Occasionally, in some plants, a few rods will experience pin hole or other types of penetrations through the Zircaloy, permitting the release of fission products into the reactor coolant. Two of the major constituents of fission product radioisotopes--radioactive iodine and radioactive cesium--are useful to illustrate the transport processes, which are typical for fission product. Iodine, when released to the coolant, will predominantly shift toward the volatile gaseous species. As such, it will processed and monitored with the gases that are removed from the reactor coolant. Iodine generally makes an insignificant contribution to the total exposure experienced at nuclear power plants, although there have been occasional situations in which iodine, which has dissolved in the coolant at the time of shutdown, has delayed shutdown operations pending its removal from the coolant systems by ion exchange. Cesium, on the other hand, is readily soluble in reactor water, and therefore is removed from this water by the purifications system. These purification systems generally use an ion exchange process, which removes the dissolved cesium from the coolant. Subsequently, the cesium ends up in the radioactive waste, where it is processed and packaged on site and shipped off site for disposal. For each of these particular isotopes, and similarly for most fission product isotopes, due to their chemical behavior and/or their low concentration in the coolant water, they contribute relatively insignificantly to the radiation fields that are present from deposition on the surfaces of the components of the reactor coolant system. The other primary source of radiation fields is from the corrosion products generated at the surfaces of the iron-nickel alloys present as the primary constituent of the pressure boundary. The general corrosion that takes place on these surfaces releases dissolved metallic constituents, such as iron, nickel, manganese, chromium, and cobalt, into the cooling water. Some of these metallic ions are deposited either by particle deposition or by precipitation on hot fuel element surfaces in the core. There they are exposed to neutrons that are generated as a result of the fission process generating the power within the cores. These neutrons will cause the stable radioisotopes of these various metallic constituents to be transmuted into radioactive species. Two of the primary radioactive constituents generated here are 60Co and 58Co. Cobalt-60 is generated as a result of neutron absorption on naturally occurring 59Co. Although cobalt is present generally as an impurity in the iron-nickel alloys used in reactor power plants, its high susceptibility to neutron adsorption and the high energy of the radioactive emissions from 60Co cause it to predominate over other radioisotopes formed from other

elements that are present in higher concentrations. Cobalt-58 is formed as a result of a neutron knocking a proton from the nucleus of 58Ni, a naturally occurring isotope of nickel. Following the generation of radioisotopes from the deposits present on the surface of the fuel, various processes such as adsorption, dissolution, and erosion can cause these isotope to be released from the fuel surface, returning them back to the reactor coolant. Transport with the reactor coolant will allow some of these to be deposited later and then incorporated into the growing corrosion films on the surfaces in the reactor coolant system. To understand the various operational and design parameters that influence the buildup of radiation fields, it is necessary to discuss separately the BWR and PWR. Such influencing features as materials present in the systems, the operational chemistry used in the systems, and the filtration and removal processes used differ sufficiently between these two to warrant separate discussion of these systems.

Boiling Water Reactors In BWRs, the pressure in the reactor coolant system is maintained sufficiently low to allow boiling to take place. Consequently, the steam generated is fed directly to the turbine and is then condensed in the condenser. This condensate, which is collected in the hot well, is then purified through a condensate polisher. Such a condensate polisher can be either a deep bed ion exchange resin or a powdered resin. The former is more effective in removing dissolved material, whereas the latter is more effective in removing particulate material. This polished condensate water is then delivered through the feedwater heater section, where steam from the various turbine sections is used to preheat the feedwater prior to its return to the reactor vessel. Water in the reactor vessel that is not vaporized as steam is recirculated through a recirculation pipe and pump system back to the bottom of the reactor core, where it is fed upward through the core for cooling the fuel elements. In considering the generation of corrosion products within a BWR in Fig. 47, a reasonable starting point is in the hot well. Here in the condensate, the primary species will be Fe2O3, which is produced primarily from steam impingement on the walls of the hot well. In addition, dissolved forms of other metallic species can be present, particularly from the condenser tubes, which may contain zinc, copper, aluminum, and other elements different from those found in the stainless steel used throughout the rest of the system.

Fig. 47 Diagram of BWR circuit showing origin of radioactive contamination. HP, high pressure. LP, low pressure. Numbered areas indicate sequence of events.

The chemistry of the water exerts significant influence over the form and the amount of the corrosion products present. In the BWR, there are generally no additives. The objective is to keep the water as pure and clean as possible. The result is a neutral water chemistry with regard to pH and an intention of keeping the conductivity as low as possible. The latter is a measure of the impurities present in the system. In addition, as the water passes through the core, radiolysis (the dissociation of molecules by radiation) occurs, causing a small portion of the water molecules to be decomposed into its constituent elements--hydrogen and oxygen. The presence of these gases plus small amounts of air in-leakage, which occurs mostly in the region of the condenser and hot well, create an oxidizing condition in the cooling water. The condensate polisher serves to remove impurities that are present in the condensate in the hot well. It generally performs its function quite effectively. The accepted measure of its effectiveness in the BWR is the conductivity of the outlet of the polisher. As the feedwater proceeds up along the feedwater train, some of the impurities will deposit on surfaces and corrosion and other processes will cause materials to be leached from the system surfaces to the water. In addition, in some plants, the condensate in some of the feedwater heater drains is forward pumped into the system. That is, the drains are pumped directly into the feedwater lines at an appropriate point for the purpose of increasing thermal efficiency of the nuclear power plant. In so doing, this forward-pumped water has bypassed the condensate polisher, meaning that its impurities have not been removed by such a system. The result of all these processes is that at the final feedwater location, where the water is pumped into the reactor vessel, there is a concentration of impurities. The primary constituent will still be Fe2O3, but the presence of cobalt, nickel, copper, and zinc can influence the subsequent processes that determine radiation field buildup on pipe surfaces. Hematite is the primary constituent of the deposit on the fuel surfaces. It also acts as the absorption medium for other constituents, including copper, zinc, and cobalt. For the BWR, 60Co is the primary constituents of radiation fields, which generally contributes more than 95% of the radiation fields coming from plant surfaces. In addition to being present as an impurity in stainless steel, which makes up almost all the surface area outside of the Zircaloy core cladding, cobalt is present in small surface area cobalt-base alloy materials used in valves, pumps, and control rods. Table 7 shows the distribution of cobalt sources within a BWR.

Table 7 Estimates of principal sources of 59Co for a BWR System/component/alloy Forward-pumped heater drains Main steam valves + HP Turbine Other valves Nickel in carbon steel Stainless steel Total Feedwater Valves Stainless steel Condensate treatment effluent Total Primary system Control blade pins/rollers Jet pumps Recirculating system valves Stainless steel + nickel-chromium-iron alloy corrosion Control rod drives Total Total Off-line valve maintenance

59

Co input, g/yr

95 22 2 11 130 18 5 6 29 29 23 18 7 8 85 224 30-90

Cobalt-60 has a relatively long half-life of 5.27 years. This means that any 60Co generated in the system that is incorporated in the growing corrosion film on system surfaces will influence radiation fields in the system for quite a few years unless it is removed from such surfaces. The currently accepted model for corrosion product incorporation in a growing corrosion film on stainless steel surfaces is illustrated in Fig. 48. This illustration shows the particles of Fe2O3 and 60Co being deposited on the surface and slowly incorporated in the growing thin film on the stainless steel. Steps 3

and 4 then show the development of the two layers and the corresponding ionic diffusion processes, finally resulting in an outer layer of Fe2O3 with an inner layer of Fe3O4. The Fe3O4 is generated in the oxygen-depleted region near the surface of the stainless steel.

Fig. 48 Crud growth model for BWRs

The 60Co and other impurity ions that are present in very low concentrations could be incorporated into this film in a variety of ways, including adsorption, crystallization, precipitation, and substitution. However, their presence is in such low concentration that chemical analyses have thus far been unable to determine the particular form of the 60Co in this film. It is important to note that the 60Co is not in a volatile form and therefore stays in the water in the region near the reactor core. This means that the turbine, hot well, condenser, and feedwater systems are not significantly contaminated by this radioisotope or any other for that matter. Radiation fields in a BWR are predominantly, and almost exclusively, from the reactor pressure vessel, recirculation pipes and pumps, and the surfaces of the reactor water cleanup system.

Pressurized Water Reactors In the PWR, higher pressure is sustained in the system; this prevents gross boiling so that the system is maintained liquid. The water is cooled by its passage through steam generator tubes. These tubes are made of Alloy 600 in U.S. PWRs.

Secondary-side coolant on the outside of these steam generator tubes boils, creating steam that is used to turn the turbines in the system. In the primary circuit, after the water has passed through the steam generator tubes and through the pump, it is driven back through the core, where it is heated again. The interconnecting piping and the cladding of the pressure vessel are all of stainless steel. Therefore, in the PWR, the major surface areas are the zirconium alloy of the fuel cladding and the Alloy 600 of the steam generator tubes, with a secondary surface area of stainless steel. This system is illustrated in Fig. 49.

Fig. 49 Diagram of a PWR primary circuit showing the origin of plant-surface radioactive contamination. Numbered areas indicate sequence of events.

The water chemistry of the system is distinct from that in the BWR. In the PWR, boric acid is added as a neutronic control, and LiOH is added to create a basic pH condition in the primary loop coolant. In addition, a hydrogen overpressure is maintained on the coolant circuit, which creates a dissolved hydrogen concentration sufficient to recombine with the oxygenating species formed by radiolysis within the core. This creates a reducing condition in the PWR primary coolant. The primary constituents released from the corrosion of the surfaces present in the PWR are iron and nickel. This fact, combined with the reducing condition present in the primary coolant, results in a deposit on the fuel that is nickel ferrite, where nickel replaces one of these iron atoms in Fe3O4 in a fraction of the lattice positions. Nickel will commonly replace one of the iron atoms in Fe3O4 in 30 to 60% of the molecules. Similar to the BWR, cobalt occurs within this deposit on the fuel. As a result of the presence of both cobalt and nickel on the fuel, 58Co and 60Co are both significant sources of radiation in the PWR system. Cobalt-58, however, has a much shorter half-life of 71 days and consequently is not retained from one refueling outage to the next as is 60Co, but will be newly generated from fuel cycle to cycle. The transport processes that conduct the cobalt radioisotopes to the out-of-core surfaces are similar to those for the BWR except that in the PWR, dissolution and precipitation probably play a much more significant role. These radioisotopes, however, will be transported to the out-of-core Alloy 600 and stainless steel surfaces, where they are incorporated in the corrosion product film as it develops on those surfaces. The most significant region in the PWR for radiation control is within the channel heads of the steam generators. This is the region where maintenance and nondestructive examination of steam generator systems takes place. Consequently, this is also the region where most of the radiation exposures associated with primary circuit maintenance occur.

Radiation Control Techniques

There are a number of techniques that have been developed over the last few years that permit the operators of nuclear power plants to effect a reduction in the rate of buildup of radiation fields within their plants. These fall into four categories: water chemistry control, materials selection, surface treatment, and decontamination. Water Chemistry Control. In BWRs, as was previously mentioned, the control of the amount of iron in the coolant is

a key to reducing the deposition on the fuel and subsequent radiation field buildup. Control of the iron in the coolant is determined by the operation of the condensate polisher. The electrical conductivity of the water is used as the measure of determining how effective the condensate polisher is performing and is an indicator of any change in the performance of the polisher. Similarly, the reactor water cleanup system provides a kidney-type purification circuit for the water in reactor vessel. Dissolved oxygen also plays a role in BWRs. It has been determined that extremely low concentrations of oxygen, below of the order of 20 ppb, forms a corrosive that is very easily released form the steel surface. Consequently, it has been found desirable in BWRs to maintain the oxygen in concentration in the feedwater lines of 20 to 200 ppb. Recent information indicates that certain metallic constituents can significantly influence the sites available for cobalt in the growing crystal lattices. In particular, the presence of zinc in the water has been found to interfere with the absorption of cobalt ions into the growing film. Consequently, small concentrations of zinc of the order of 5 to 10 ppb can reduce the deposition of 60Co on pipe surfaces and therefore reduce the radiation field buildup in BWRs. In PWRs where precipitation and dissolution from fuel surfaces have been shown to be a significant influencing factor in radiation field buildup, the pH of the coolant can be controlled in a manner to reduce the amount of precipitate on the fuel surfaces. LiOH that is added to the coolant is controlled so that as the boric acid is reduced through the fuel cycle, the LiOH is similarly reduced to maintain constat pH throughout the cycle. When this pH is high enough, in the region of 7.1 to 7.4 at operating temperatures, the thermodynamics of the solubility of nickel ferrite are such that nickel ferrite will not precipitate on the fuel at this pH. This reduction in the fuel deposition concurrently reduces the generation of cobalt isotopes and the subsequent buildup of radiation fields in the steam generator and on other surfaces. Materials Selection. High-cobalt alloys are present in most nuclear reactor systems, usually in valves, pumps, and

control rods. New materials containing low amounts of cobalt or no cobalt are being developed that are designed to serve the function of the high-cobalt alloys. These materials are currently in testing stages to show that their performance is sufficient for what is needed within the power plants. Some changes have already taken place, such as the replacement of pins and rollers in BWR control blades, which were previously of high-cobalt alloys and are now made of low-cobalt alloys. Such replacements offer significant opportunity for radiation field reduction in the future from 60Co. Surface Treatment. As both BWRs and PWRs are experiencing component replacements (recirculation pipes on the

former and steam generators in the latter), the condition of the surfaces of these new components is also a consideration. Electropolishing has been shown to reduce the effective surface area up to a factor of five on pipe on steam generator surfaces. This is accomplished by trimming off the asperities present on the surfaces, thus creating a more uniform surface finish. Decontamination solvents have been developed that are quite effective in removing the radioisotopes in the oxide

layers that are present in previously operated nuclear power plants. These are used particularly when major maintenance and repair work is conducted on a piece of equipment that has very high radiation fields associated with it. Such decontamination can afford a utility an opportunity to make use of recently developed procedures, such as chemistry control or electropolishing, to reduce the subsequent radiation field buildups on the component. SCC in Steam Turbine Materials Floyd Gelhaus, Electric Power Research Institute

Stress-corrosion cracking is the fracture of a metal that results from the joint action of tensile stresses and a corrosive environment. Crack initiation and propagation are also dependent on metallurgical parameters (Ref 107). Cracking has occurred in both reheat and nonreheat turbines, in quenched-and-tempered 3.5NiCrMoV, and normalized-and-tempered 2.5NiCrMoV steels with a wide range of grain sizes. A typical intergranular crack is shown in Fig. 50. A single

compilation of laboratory and field data, in conjunction with an annotated bibliography of more than 200 references, was published in 1984 (Ref 108).

Fig. 50 Typical keyway intergranular stress-corrosion crack. Sections are orientated normal to the length of the keyway at different distances from the outlet face. (a) 28 mm (1.1 in.). (b) 10 mm (0.4 in.). (c) 75 mm (3 in.). Source: EPRI Report NP-3341

Parameters of Influence Stress. Until the introduction of the 500 + Mg (550 ton) monoblock rotor by Japanese steel manufacturers in the late 1970s, the large low-pressure steam turbine rotors that are utilized in nuclear power plants were fabricated by one of two techniques. Brown Boveri Company in Switzerland pioneered the welded rotor, in which several large disks are joined at the low-stress outer diameter with a sequence of gas tungsten arc and shielded arc welding processes. Other manufacturers, including the General Electric Company (Fig. 51) and the Westinghouse Electric Company (Fig. 52), have used the technique of thermally shrinking disks onto a stepped-diameter shaft, keying each disk/shaft interface to prevent independent rotation (Fig. 53 and 54). In addition to creating a zone of high local stress, the shaft keyways are flow/noflow locations that promote the capture of chemical species. The welded rotor has no keyways. Since the catastrophic rupture of a shrunk-on disk at the Hinkley Point A Nuclear Station in England in September 1969 (Ref 109, 110), considerable attention has been given to understanding disk cracking and this potential failure mode.

Fig. 51 Schematic of General Electric low-pressure turbine rotor. All disks are shrunk on, with one blade row per disk. Source: EPRI Report NP-2429-LD, Vol 6

Fig. 52 Schematic of Westinghouse low-pressure turbine rotor. All disks are shrunk on, and the number of blade rows per disk is as indicated. Source: EPRI Report NP-2429-LD, Vol 6

Fig. 53 Schematic of keyway design reportedly used by General Electric in shrunk-on disks of low-pressure turbine rotors. Source: EPRI Report NP-2429-LD, Vol 6

Fig. 54 Schematic of keyway design reportedly used by Westinghouse in shrunk-on disks of low-pressure turbine rotors. Source: EPRI Report NP-2429-LD, Vol 6

Cracking has occurred not only in keyways but also on the bore surfaces, hub faces, web faces, and in the rim attachment area of shrunk-on disks (Fig. 55). Also, cracks have been found on the web faces and in the rim blade-attachment area of the integral disks of monoblock rotors. Monoblock and welded rotor configurations eliminate the keyway as a crackstarter location, but cracks persist in the rim blade-attachment area (Fig. 56).

Fig. 55 Schematic showing typical locations and orientations of cracks in various U.S. low-pressure rotor disks. This illustration is of a shrunk-on disk with a semicircular keyway and an axial-entry fir-tree rim attachment

configuration. Disks of this type have experienced cracking at all of the locations illustrated except at the central web in the axial-radial orientation and on the web below the rim (location 3). Cracks at the latter two locations have been found in integral disks. Source: EPRI Report NP-2429-LD, Vol 1

Fig. 56 Schematics of rim attachment configurations used in U.S. low-pressure rotors in which cracking has been found. Locations and orientations of rim attachment cracking experienced in various rotors are illustrated. (a) Axial-entry fir-tree. (b) Notch-entry dovetail. Source: EPRI Report NP-2429-LD, Vol 1

Material. Most modern U.S. turbine rotors use 3.5NiCrMoV steel, conforming to the requirements of ASTM A471

(Class 1 through 9). Yield strength variation with tempering temperature is shown in Fig. 57, and the A471 temperature boundary for all classes (>839 K, 565 °C, or 1050 °F) is specifically noted. Chemical compositions for A471 Class 6 and A470 Class 5 are shown in Table 8.

Table 8 Chemical specifications for representative steels used in modern U.S. low-pressure rotors with integral and shrunk-on disks Element, wt%

Carbon (max) Manganese Phosphorus (max) Sulfur (max) Silicon (max)(a) Nickel Chromium Molybdenum Vanadium

ASTM A470 Class 5 steel for integral rotor disks 0.28 0.20-0.60 0.015 0.018 0.10 3.25-4.00 1.25-2.00 0.25-0.60(b) 0.05-0.15

ASTM A471 Class 6 steel for shrunk-on rotor disks 0.28 0.70 max 0.015 0.015 0.10 2.00-4.00 0.75-2.00 0.20-0.70 0.05 min

Source: EPRI Report NP-3634, Vol 2

(a)

Modern U.S. turbine rotor disks are made from vacuum-deoxidized steels for which the silicon limit is 0.10%. For nonvacuum-deoxidized steels. A470, Class 5 allows a silicon range of 0.15-0.30%, and A471, Class 6 allows a silicon range of 0.15-0.35%.

(b)

If desirable because of operating temperatures, a minimum molybdenum content of 0.40% may be specified by the purchaser.

Fig. 57 Yield strength versus tempering temperature for low-alloy steels. Source: EPRI Report NP-4056

Early correlations of crack growth rate data revealed a dependence on yield strength and an inverse dependence on temperature. This equation is generally expressed as:

(Eq 14)

where is the crack growth rate. T is the disk operating temperature, sy is the room-temperature yield strength, and C1, C2, and C3 are fitting constants. A recent analysis identified two new key variables: composition and tempering temperature (Ref 111). The latter variable does not exhibit a strong effect for materials meeting A471 tempering temperature/yield

strength standards, and for these, the correlation with yield strength is sufficient. However, for NiCrMoV and other lowalloy steels, tempering temperature is a key variable, with a stronger but separate effect on crack growth rate as compared to yield strength. Regarding composition, manganese was indicated by this analysis to be a strongly correlated variable, and Eq 14 was modified to include that effect by adding a C4Mn term, using the weight percent of manganese. This manganese dependence should be interpreted as an unresolved manganese/nickel/vanadium effect, because nickel and vanadium were highly correlated with manganese. Sulfur concentration was also identified as a key variable, but no model containing a sulfur term was calibrated to the data. For the two-, three-, and four-variable model, Eq 14 is:

ln = -8.8 - (4040/T) + 0.0231 sy ln = -4.74 - (9270/T) + 0.3337 sy + 4.5Mn

(Eq 15a) (Eq 15b)

ln = -4.74 - (9270/T) + 0.3337 sy + 4.53Mn

(Eq 15c)

ln = -7.04 - (9270/T) + 0.0337 sy + 4.53Mn

(Eq 15d)

where is given in inches per hour, T is in degrees Rankine, sy is in kips per square inch, manganese (Mn) is in weight percent, and Tt, the tempering temperature, is in degrees Kelvin. Three of the Eq 15b constants were forced not to change when including the Tt term; Eq 15c permits a variety of low-alloy steels (single or multi-heat) to be analyzed on a common basis. Laboratory experiments to determine the effect of the segregation of phosphorus to the grain boundaries have shown little effect of this impurity on SCC at high stress levels (Ref 112, 113). All specimens in both experiments failed in less than 500 h. When the test solution is neutral, segregated phosphorus has no effect on the resistance of an alloy to SCC. The response to a highly caustic solution (9 M NaOH) was, however, dependent on the applied potential, changing from a strong increase in the cracking susceptibility at -400 mV (Hg/HgO) to no effect at -800 mV (Hg/HgO). Environment. Potential is a thermodynamic parameter that is a measure of the energy of a chemical reaction. It is a

measure of the oxidizing power of a solution. For a corrosion reaction, potential is a measure of the energy necessary to cause a metal atom to be transferred from the metal lattice into solution as a metal ion. Both temperature and pH can either raise or lower the potential of a corrosion reaction, depending on the species involved and on other factors. With no applied voltage, the rate of metal atom transfer would reach equilibrium, and those chemical processes would establish the free corrosion potential. To investigate the range of possible chemical reactions, a variable voltage is applied, and a polarization diagram is generated. Figure 58(a) shows an anodic polarization diagram for a metal-solution combination that forms a protective corrosion layer (passivation film) over a range of potentials (Epr). Figure 58(b) includes a second area of anodic activity (Ecrit(2)), a phenomenon that can be associated with a redox reaction or with the dissolution of a second element, such as chromium, in low-alloy steels.

Fig. 58 Schematic of anodic polarization curves for (a) a metal that forms a protective layer and (b) an alloy that includes a second passivation peak where Ecorr is the corrosion potential, Ecrit is the critical potential of passivation, Eox is the oxygen evolution potential, and Epr is the passive potential range.

The importance of pH is shown by the curves in Fig. 59(a). The additional protection offered by neutral-to-higher pH in this example generally holds true for all the metals-environment combinations in steam turbines (type 403 stainless steel is used for turbine blades); therefore, the working fluid pH is held neutral for BWRs and between 8.9 and 9.6 for PWRs.

Fig. 59 Anodic polarization curves for type 403 stainless steel showing the effects of solution pH (a), temperature (b), different neutral pH solutions of the same molar concentration (c), sodium chloride (NaCl) solution concentration (d), and sodium sulfate (Na2SO4) solution concentration (e). In each of these figures, the current density is proportional to the corrosion rate.

Figure 59(b) shows the influence of temperature at the neutral pH 7. The corrosion potential becomes more negative with increasing temperature. The lower temperature allows the protective layer to remain intact over a much larger range of potential. It is important to note that, although type 403 stainless steel does suffer pitting (for example, at pH 10 in a solution of NaCl), the alloy tempered at 650 °C (1200 °F) is quite resistant to SCC. The potential is most strongly influenced by the kind and concentration of species in solution, causing marked shifts in the corrosion rate (that is, the current density). Figure 59(c) shows how the polarization curve shifts with species kind, and Fig. 59(d) and 59(e) indicate the influence of concentration for the same species. The different behaviors exhibited by these latter curve sets are evidence of the complexity of the corrosion processes and point out that the behavior of an alloy in service is difficult to state in general terms, because the service environment is often complex and ill-defined. However, the general trends from laboratory tests help define what service conditions are required to get to observed response. For intergranular SCC, polarization curves for bulk alloys are helpful, but the real need is for data showing how grain-boundary compositions behave.

Laboratory Tests Considerable data were obtained by U.K. experimenters following the failure at Hinkley Point, but most of this was generated using 3CrMo steel. Later work in the U.S. focused more on the 3.5NiCrMoV alloy, but much of these data were taking using strong caustic environment (35% NaOH). Table 9 summarizes the results of these earlier experiments. Analyses of these data show that several environments--for example, >1% NaOH, aerated/oxygenated water--produced higher crack growth rates than were estimated from the field data; models that are derived to help a utility predict the lifetime of turbine components (Eq 15a, 15b, 15c, and 15d) have to be based on laboratory data that approximate field conditions.

Table 9 Collation of crack growth rates from precracked wedge-opening load specimens Steel and condition

3CrMo (AOH), embrittled

3CrMo (AOH), deembrittled

Environment

3CrMo (BE) 3.5NiCrMoV

3.5 NiCrMoV

Stress intensity

Crack growth rate

MPa

ksi

mm/yr

in./yr

700 700 700 700 700

102 102 102 102 102

-400(b) -600(b) -800(b) -900(b) -800(b)

18-39 13-18 8-36 18 13-38

16-35 12-16 7-33 16 12-34

50-278 63-83 57-631 71 13-85

1.9-10.9 2.5-3.3 2.2-24.8 2.8 0.5-3.4

700

102

-600(b)

19

17

52

2.05

700

102

-800(b)

7-37

6-34

47-662

1.8-2.6

700

102

-900(b)

20

18

4-39

0.16-1.5

700

102

-800(b)

10-20

9-18

6.3-328

0.25-4.8

710

103

-800(b)

10-20

9-18

6.3-136

0.25-5.4

102 102 102 114 89162 119

-800(b) -800(b) -800(b) -250(b) -250(b)

20 20 20 10-66 30-60

18 18 18 9-60 27-55

35-91 8-14 0.6-2.5 0.4-5 1-1135

1.4-3.6 0.3-0.5 0.02-0.1 0.016-0.2 0.04-44.7

...

16-126

0.3-4.7

0.01-0.18

0.1-0.3

35% NaOH 10% NaOH 4% NaOH 35% NaOH 28% NaOH(c)

115 115 115 115 85115 85115 85115 85115 85115 85115 100 100 100 115 110

240 240 240 240 185240 185240 185240 185240 185240 185240 212 212 212 239 230

25% NaOH

90

195

700 700 700 786 6151115 821

25% NaOH

90

195

653

95

...

32-126

14.5115 29-115

25% NaOH 10% NaOH(d)

90 157

195 315

681 1124

99 163

32-48 22-88

29-44 20-80

1.2-12.6 4-1182

10% NaOH(d)

157

315

876

127

22-88

20-80

3-334

0.01-13

10% NaOH(d)

157

315

820

119

22-88

20-80

4-123

0.16-4.8

10% NaOH(d)

157

315

731

106

22-66

20-60

7-244

0.3-9.6

10% NaOH(d)

157

315

717

104

22-88

20-80

3.6-78

0.14-3.1

10% NaOH(c)

157

315

1124

163

22-88

20-80

18-259

0.7-10.2

10% NaOH(c)

157

315

876

127

22-88

20-80

1-32

0.04-1.3

10% NaOH(c)

157

315

820

119

22-88

20-80

3.6-78

0.14-3.1

10% NaOH(c)

157

315

731

106

22-88

20-80

2-66

0.08-2.6

10% NaOH(c)

157

315

717

104

... -670 to 780(e) -670 to 780(e) -670 to 780(e) -670 to 780(e) -670 to 780(e) 30 to 210(e) 30 to 210(e) 30 to 210(e) 30 to 210(e) 30 to -

0.0040.01 0.05-0.5 0.16-46.5

22-88

20-80

3.6-78

0.14-3.1

35% NaOH 35% NaOH 35% NaOH 35% NaOH 35% NaOH

35% NaOH

3.5NiCrMoV (VCD)

Potential, mV

°F

35% NaOH

2NiCrMoV

Yield strength(a) MPa ksi

°C

35% NaOH

3CrMo (AOH), asreceived 3CrMo (AOH), deembrittled

Temperature

35% NaOH 35% NaOH

(d)

2CrlMo

10% NaOH

157

315

706

102

3Cr0.5Mo

8% NaOH

115

240

73-78

3NiCrMoV

1% NaOH + 0.1% NaCl(d) 1% NaOH + 0.1% NaCl(d) 1% NaOH + 0.1% NaCl(d) 1% NaOH + 0.1% NaCl(d) 1% NaOH + 0.1% NaCl(f) 1% NaOH + 0.1% NaCl(f) 1% NaOH + 0.1% NaCl(f) 1% NaOH + 0.1% NaCl(f) 1% NaOH + 0.1% NaCl(f) Pure water(d)

157

315

505540 1124

163

157

315

876

127

157

315

820

119

157

315

731

106

157

315

1124

157

315

157

3.5NiCrMoV

3.5NiCrMoV

210(e) 30 to 210(e) -354(e)

66

60

7.4

0.3

20

18

0.3-82

0.01-3.2

33-66

30-60

7-78

0.28-3.1

33-66

30-60

2-4.4

0.08-0.17

33-66

30-60

0.9-5.7

0.04-0.22

33-66

30-60

3

0.12

163

-700 to 760(e) -700 to 760(e) -700 to 760(e) -700 to 760(e) 120-170(e)

33-66

30-60

2.2-53

0.09-2.1

876

127

120-170(e)

33-66

30-60

0.5

0.02

315

821

119

120-170(e)

33-66

30-60

...

...

157

315

731

106

120-170(e)

33-66

30-60

0.9-3.1

0.04-0.12

157

315

717

104

120-170(e)

33-66

30-60

0.5-5.8

0.02-0.23

157

315

1124 876

163 127

... ...

22-88 22-66

20-80 20-80

0.7-23 0.4-4.5

821 731

119 106

... ...

22-88 22-66

20-80 20-60

1.3-3.6 0.4-4.2

22-66 66 33-66

20-60 60 30-60

0.5-1.1 4.2 76-197

0.03-0.9 0.0160.18 0.05-0.14 0.0160.165 0.02-0.04 0.165 3-7.8

33-66

30-60

2-16

0.08-0.25

33-66

30-60

3.5-8

0.14-0.3

33-66

30-60

1.6-4.6

0.06-0.18

33-66

30-60

1.6-20

0.06-0.8

25

23

3.2

0.13

33-63

30-57

11

0.4

32-62

29-56

...

...

32-66

29-60

0.9

0.035

32-65

29-59

4.5

0.18

33-66

30-60

13.8-70

0.54-2.8

33-66

30-60

3-7.2

0.12-0.28

33-66 10-110

30-60 9-100

7.2 0.2-0.6 0.63248 0.0619000 0.35-1.4

0.28 0.0080.02 0.02-128

Pure water(d) Pure water(f)

157 157

315 315

717 706 1124

104 102 163

Pure water(f)

157

315

876

127

Pure water(f)

157

315

820

119

Pure water(f)

157

315

731

106

Pure water(f)

157

315

717

104

Pure water(c)

157

315

1124

163

Pure water(c)

157

315

876

127

Pure water(c)

157

315

821

119

Pure water(c)

157

315

731

106

Pure water(c)

157

315

717

104

Pure water

157

315

1124

163

Pure water

157

315

876

127

Pure water Pure water(d)

157 100

315 212

821 760

119 110

... ... 30 to 200(e) 30 to 200(e) 30 to 200(e) 30 to 200(e) 30 to 200(e) 235 to 290(e) 235 to 290(e) 235 to 290(e) 235 to 290(e) 235 to 290(e) 130 to 250(e) 130 to 250(e) ... ...

Pure water(d)

100

212

1220

177

...

10-115

9-105

Pure water

100

212

635

92

...

30-60

27-55

3CrMo (AOH)

Water

90

195

66

60

Water

90

195

105110 132

...

3.5NiCrMoV

722758 910

...

66

60

2CrlMo 3.5NiCrMoV

3.5NiCrMoV

3.5NiCrMoV

2CrlMo 26NiCrMoV127

0.040.05

0.002748 0.0140.06 0.00160.002

3CrMo (AOH)

Steam

90

195

3.5 NiCrMoV

Steam

90

195

3CrMo (AOH)

Steam

90

195

3CrMo (BE)

Steam

90

195

3.5NiCrMoV

Steam

90

195

2NiCrMoV

Steam

90

195

4.5NiCrMoV

Steam

90

195

3CrMoV

Steam

90

195

3.5NiCrMoV 3CrMo 3CrMo

Steam Steam Steam

120 120 120

250 250 250

722758 910

105110 132

722758 745772 910

105110 108112 132

786848 896938 862924 896 772 772

114123 130136 125134 130 112 112

...

11-112

10-102

0-0.9

0-0.035

...

66

60

0.9-1.9

...

66

60

0.9-1.9

...

66

60

1.2-1.7

...

66

60

0.3-1.3

...

66

60

1

0.0350.07 0.0350.07 0.050.067 0.0120.05 0.04

...

66

60

1.9-3.5

0.07-0.14

...

5-110

4.5-100

1.9-3.5

0.07-0.14

... ... ...

66 66 5-110

60 60 4.5-100

3.1 13 1.6-3.2

0.12 0.5 0.06-0.13

AOH, acid open hearth; BE, basic electric, VCD, vacuum carbon dioxide. Source: Ref 114

(a) (b) (c) (d) (e) (f)

Yield strength values were intentionally varied by heat treatment. Potential with respect to Hg/HgO. Oxygenated. Deaerated. Potential with respect to standard hydrogen electrode (SHE). Aerated.

In-Plant Conditions The majority of U.S. PWR power plants control steam generator water pH using phosphates (Na3PO4) or all-volatile treatment, that is, control by ammonia (NH3) or by amines such as morpholine (C4H9NO). The major advantage for phosphate treatment is its capacity to buffer against both acidic and basic upsets, a protection only weakly afforded by the amines. The volatile amines introduce no solids into the system, whereas the salts added in the phosphate method can accumulate and concentrate. No chemical additions are made to control feedwater pH in BWR power plants. Oxygen levels in power plant feedwater are controlled by adding scavenging chemicals (hydrazine, N2H4) or by mechanical means. Only two U.S plants currently use a feedwater deaerator, with most using only air ejectors on the deaerating section of the condensers. In BWRs, the reactor water oxygen level can reach several parts per million during start-up. Chemical upsets can be initiated by a condenser tube failure, and the analyses of chemicals that can be introduced is shown in Table 10. Air in-leaking rates of 2800 L/min can be experienced with severe condenser leaks, while a rate one-tenth that levels is common. With 500 L/min in-leakage, the oxygen concentration in the condensate can increase to 30 ppm. Powder resin filters and deep bed demineralizers are two common full-flow condensate polishing systems used in U.S. plants. However, if not properly operated/regenerated, these systems can become sources of ionic impurities.

Table 10 Typical cooling water analyses Element

Calcium (ppm) Magnesium (ppm) Sodium (ppm) Potassium (ppm) Lead (ppm) Chloride (ppm) Bicarbonate (ppm) Total alkalinity (ppm as calcium carbonate, CaCO3) Fluoride (ppm) Bromide (ppm) Sulfate (ppm) Nitrate (ppm) Phosphate (ppm) Silica (ppm) Carbon dioxide (ppm) Oxygen (ppm) pH

Fresh river water(a) 58 15 13

Fresh lake water(b) 32 11 3.2

(e) (e)

4.8 217 178

0.004 2.1 149 (e)

(e)

0.25

(e)

(e)

45

7 1.6 0.6 5 3.8

(e) (e)

14 (e) (e)

(e)

(e)

(e)

Brackish water(c)

Seawater(d)

44 78 603 20

400 1272 10,561 380 0.21 18,980 142

(e)

1053 68 56 0.08 3.5 220 1.2 (e)

8.6 2.9 6.2 6.2

(e)

3.5 65 2649 10-3 to 7 × 10-1 10-3 to 10-1 0.01 to 7.0 6 5 5

Source: EPRI Report NP 2429, Vol 4

(a) (b) (c) (d) (e)

Mississippi River. Lake Michigan. Estuary on U.S. East Coast. Typical ocean water. Not determined.

Table 11 shows typical power plant data, and such data are used to infer the concentration in the steam reaching the turbine systems. Because the solubility of salts in steam decreases as temperature and pressure are decreased, solubility limits are exceeded at some stage with the expansion of the steam in the turbine. The thermodynamic data required to predict the exact conditions under which various salts will deposit are not available. However, scrapings taken from turbine surfaces indicate that condensation to form concentrated solutions in the crevice areas (keyways, blade attachments) does occur and that SCC is a typical result. Field data on crack growth rate are comparable with lab data using deaerated water, steam, carbonated water, and 1% NaOH + 0.1% NaCl (Ref 53).

Table 11 High-pressure drain (HPD), moisture separator drain (MSD), and condensate (Cond.) chemistry during a power escalation Nuclear PWR with recirculating steam generator Power level, % 20 25 32 35 40 45 58 64 68 80 82 86 90 93 95

Concentration, ppb Na+(a) HPD(b) MSD(c) 10,000 years) (Ref 146, 147). Finally, a 300+ year container could assist any short-term waste retrieval operations, should that option need to be exercised. In the U.S., the Nuclear Regulatory Commission (NRC) requires that the waste be reliably contained for a period of 300 to 1000 years (Ref 148). Other countries have adopted this requirement at least as a goal in their programs. The ultimate application for the container would be for it to function as an absolute barrier for the entire time that significant radiotoxicity is present. Canada and Sweden are investigating this possibility by determining the potential for containing the waste for periods of at least 100,000 years (Ref 147, 149). Because the container must function both as a transportation vessel and as a corrosion barrier, almost all of the materials being considered for its construction are metallic. Given the environments and the candidate metals, the forms of corrosion that are applicable to waste containers include: • • •

General, or uniform, corrosion Localized attack (crevice, pitting, and intergranular attack) Stress-corrosion cracking

•

Hydrogen-assisted cracking

The extremely long lifetime requirement being placed on the container forces the designer to consider only materials that have very predictable and reliable behavior. The rates of the latter three mechanisms are all sufficiently high or unpredictable that any design specifying a material that would fail by these types of nonuniform attack would be impractical. Thus, the challenge is to identify materials that will undergo only general corrosion in a given disposal environment. Materials that are acceptable in one environment may not be usable in another (for example, crystalline rock versus bedded salt). Two basic approaches have evolved for potentially satisfying the lifetime requirement being placed on the waste container: use of a corrosion-resistant material or use of a corrosion-allowance material. For a reasonably designed container to survive for 300 to 1000 years, its corrosion rate must be relatively small. As will be shown, a number of materials have been identified with corrosion rates that are practically nil (micrometers or less per year). Very little corrosion allowance is required in the design of the waste package constructed with these "corrosion-resistant" materials. These metals usually owe their durability to very adherent oxide films, and as such, the major difficulty in their characterization is to demonstrate that the other, more insidious corrosion mechanisms will not be active during the long containment time. Examples of corrosion-resistant materials are titanium alloys, nickel-base alloys, and, in some applications, possibly stainless steels. The alternate approach is to use "corrosion-allowance" materials that, in general, have more predictable corrosion characteristics. Although these materials have much higher general corrosion rates, their susceptibility to nonuniform corrosion is usually very limited due, in part, to a lack of protective oxide containing product layers. The corrosion rate of these active materials is often controlled by the rate of the cathodic process (oxygen reduction and/or hydrogen ion discharge). As will be shown later, the ground-water flow rates (oxidant supply) are very limited, and their pH is usually near neutral. Given these types of conditions, the general corrosion rate may be relatively low and can be measured reliably. Thus, an allowance for the material wastage can be incorporated into the design. Examples of these types of materials include mild steel, cast iron, and copper. Given the above considerations, the materials characterization programs have two primary objectives: • •

To quantify the general corrosion rate To demonstrate that no accelerated attack will occur. This requirement can be satisfied either by showing that the material is not susceptible to these forms of attack or that their time to initiation (the incubation period) is longer than the required lifetime of the container (for example, sensitization of stainless steel)

The impact of these two objectives on the corrosion programs will be evident in the discussion of "Corrosion Results From Site-Specific Studies" provided later in this section. In addition to corrosion characteristics, cost and mechanical properties must also be considered in the selection of container materials. The mechanical properties can be very important when satisfying waste package requirements related to waste-handling and retrieval operations, but this subject is not relevant in this section. Container cost can be a factor. Unit material costs for the corrosion-allowance materials tend to be much lower than those for corrosion-resistant alloys; however, because wall thicknesses must be greater, this cost difference is reduced. Overall fabrication costs are very design dependent, and cannot be generalized. To compare costs properly for different container options, total disposal cost should be considered because the use of a corrosion-resistant material may allow a simple, lightweight design to be used with substantially fewer components, less fabrication, less emplacement-hole mining, and smaller handling equipment. The remainder of the background information needed to evaluate properly the results of ongoing and completed corrosion testing is given in the next two parts of this section. The first, "Considerations for Waste Isolation." describes the unique problems associated with characterizing waste container behavior, and the second, "Environment," summarizes the expected container environments.

Considerations for Waste Isolation The requirement that a waste container survive intact for periods of several hundred years in elevated-temperature irradiated geologic environments has created a difficult problem for materials and design engineers to solve. The unique aspect of this problem is associated with making very reliable predictions about the corrosion behavior or container materials for these extended periods of time. Many of the alloy systems being considered have been in existence for less than 100 years. Thus, a data base concerning long-term behavior is practically nonexistent. Because most countries, including the United States, do not plan to have their first repository operational until the late 1990s, sufficient time (10 to 15 years) exists to allow most of the needed long-term simulations and accelerated testing to be completed. In the U.S., an additional 50 years is available for testing before the repository is actually sealed (the "retrieval period"). Certainly, the important corrosion processes should be identifiable within these time frames. The general problems associated with interpreting results from long-term and accelerated testing have been addressed by many investigators (Ref 150, 151, 152). For example, K. Nuttall discusses many of the considerations that should be given to identifying long-term failure processes and noted that unless the effect of a particular failure mechanism is large, it is difficult to characterize by using short-term (Ref 152). The real concern is that these unknown changes may lead to delayed, catastrophic stress-assisted failures (for example, SCC, hydrogen cracking, hydriding). The materials engineer should remember, however, that a proper waste package design will minimize container stress and thus greatly decrease the potential for these types of stress-related corrosion processes. A detailed methodology for predicting long-term behavior using accelerated testing is proposed in Ref 153. The specific problems associated with predicting material behavior over extended time periods relate to identifying potential changes in both the environment and the metallurgical characteristics of the alloy. To understand these types of problems better, some examples are given below. The important environmental parameters affecting corrosion include container temperature, groundwater chemistry, groundwater flow rate, hydrostatic and lithostatic pressure (influences water phase and container stress), and radiation flux. The predicted history for these parameters must be quantified before accurate and reliable assessments of the performance of the container can be made. The difficulty in easily predicting the effect of these environments on corrosion was demonstrated by some results from initial site studies performed in Canada. The groundwater in the crystalline-rock disposal environment in the Canadian shield was expected to be benign (no halide ion, neutral pH, low ion strength) (Ref 147). During a research-drilling program, however, groundwater containing 5.6 g/L of chloride ion was encountered. The existence of isolated pockets of saline groundwater in the deep rock formations supports the need for careful site characterization studies. Metallurgical changes could be important because of their effect on the structures and properties of passive films, time to inducement of delayed mechanical failure, and time to sensitization of the microstructure. Again, the state of stress in the container could be a major factor that determines whether these metallurgically induced changes will actually lead to a failure. These potential effects are recognized and are being addressed by many investigators (Ref 151, 152, 154, 155). Theoretically, radiation can affect the corrosion behavior of the container by affecting both the container environment and its metallurgical properties. However, the general conclusion reached by most investigators is that the types and dose rates of radiation emitted from decaying wastes are not sufficient to degrade the properties of either the container material or its passivating oxide layer and that the important effect of radiation is the change produced in the external environment due to groundwater radiolysis (Ref 152, 156). The effect of radiation on the environment is covered in the latter part of the next discussion. However, radiation has been shown to produce small changes in the properties of some metals that could lead eventually to increased corrosion rates (Ref 157). Such potential effects of radiation should be identified during longterm testing.

Environment Numerous deep-geologic formations are being considered by the various countries as the media for isolating radioactive waste with the characteristics of many of these formations being quite similar. To aid in these discussions, this similarity allows the formations to be grouped into the following three categories: • •

Crystalline rock (igneous and metamorphic) Salt deposits (bedded and domed)

•

Sedimentary deposits (clay and seabed sediments)

The properties of the geologic formation are very important because of their influence on many of the factors that determine the corrosion behavior of container materials. These factors include the groundwater chemistry, the groundwater flow rate, the hydrostatic pressure, the lithostatic pressure, the phase of the water contacting the container (liquid and/or vapor), the temperature of the container, and the types and concentrations of radiolysis products. The container temperature and radiation output are influenced by the design and loading of the waste package (size, thermal output, radiation output), the rate and density of waste package emplacement, and the thermal properties of the formation. A brief summary of the types of environments being used in the selection of waste container materials follows. Because radiation could have an important and potentially unique impact on the environment of the container through the production of radiolysis products, this discussion of environment is separated into two parts: general characteristics and radiation effects.

General Characteristics Temperature. Because heat is a significant by-product of HLW decay, the temperature of all waste containers will

initially increase and then decrease as the activity of the waste decays. The predicted temperature history for 3-kW waste packages emplaced in a consolidated volcanic ash (tuff) formation in the U.S. is shown in Fig. 84. Typical maximum container temperatures for a number of other repository locations are given in Table 16. Although these maximum temperatures are quite different, the general shape of the temperature curve shown in Fig. 84 should be similar for the other repositories. The variability in maximum temperature is due primarily to design philosophy. The temperature at a given location can be lowered by longer waste aging before emplacement, lower package loading, and lower overall repository loading. The lower temperatures will, in general, enhance the performance of the entire waste package and decrease the impact of emplacing waste on the geologic formation itself. However, a penalty is incurred to reduce temperatures because higher handling and emplacement costs, along with a larger usable area, are also required.

Table 16 Typical peak container temperatures for selected repository locations Host rock

Spent fuel °C °F

Crystalline rock 210 Basalt, U.S. 260 Tuff, U.S. Granite, Sweden 80 Salt 180 United States ... West Germany Sedimentary deposits Subseabed, U.S. . . . Subseabed, U.K. . . .

Reprocessed HLW °C °F

Ref

410 500 175

250 280 80

480 535 175

150 158 149

355 ...

235 200

455 390

159 160

... ...

200 100

390 212

161 162

Fig. 84 The predicted temperature history of waste packages emplaced in a tuff formation at a 50-kW/acre areal loading. Power loading for a spent-fuel package is 3.3 kW and for a reprocessed-waste package is 2.2 kW. Source: Ref 158

Water Chemistry. A summary of the expected composition of the groundwaters associated with several of the candidate formations is given in Table 17. The Eh, or oxidizing potential, of the groundwater in most cases will be determined by the presence of radiolysis products (see the discussion of "Radiation Effects" in this article). However, if radiation shielding is used in the package, conditions can range from slightly oxidizing (unsaturated tuff) to slightly reducing (basalt, granite, sub-seabed). The groundwaters associated with the crystalline-rock formations should all be relatively benign to most materials because of their low ionic strengths, near neutral pH, and low concentrations of halide ions. The corrosivity of these waters could be increased if significant groundwater vaporization occurs during the early times following emplacement when high container temperatures exist. Any brine contacting the container in the salt repositories will be quite corrosive because of the high concentration of halides and the potential for low pH due to magnesium salt hydrolysis (Ref 154). The situation is similar for the chloride-containing seabed environment. The corrosivity of ground-waters in a clay environment should fall between that of seawater and the crystalline-rock waters.

Table 17 Groundwater compositions (in parts per million) for selected repository locations at 25 °C (75 °F) Ion

Basalt (Ref 163)

Tuff (Ref 158)

Granite (Ref 164, 165)

Na+ K+ Mg2+ Ca2+ Sr2+ Fe2+ NH4+ Cl-

250 1.9 0.4 1.3 ... ... ... 148 180

51 5 2 14 0.05 0.04 ... 7.5 22

I-

... 97

Br-

HS FH3 pH

-

Seawater (Ref 154)

Clay (Ref 168)

0-106 ... 0-6 10-40 ... 0.02-5 0.05-0.2 4-36 0.5-24

Salt (highmagnesium brines) (Ref 154, 166, 167) 6,500-42,000 10,500-30,000 35,000-85,000 600-14,700 5 ... ... 19,000-270,000 160-13,000

10,600 380 1,270 400 13 ... ... 19,000 880

63 7.4 3.6 21 ... 189 ... 36 ...

... 120

... 90-275

10 700

0.05 146

... 188

... ...

... ...

... ...

400-2,400 1,200

65 ...

... ...

...

5.6

0.01-0.05

...

...

6

... 37 103

... 2.2 61

0-0.5 0-2 0-19

... ... ...

... ... ...

... 817 8

9-10

7.1

7-9

6.5

8.1

7.4

Water Availability. Because water movement is the primary mechanism for transporting radioactive species away

from the package, an important criterion for selecting geologic formations is a lack of available water. The candidate formations have a low permeability and/or a small quantity of flowing water, resulting in a very low flow rate of groundwater past the waste containers. For crystalline-rock formations, groundwater will flow, because of hydraulic gradients, to the package either through the rock matrix or fractures and fissures. The water flow rate through the granite formation in Sweden and the tuff formation in the U.S. has been estimated to be of the order of 0.1 L/m2/yr (Ref 149, 158). Brine can be transported in the relatively dry (0.05 to 2% H2O) salt formations by thermally induced migration of brine inclusions. This process has been estimated to transport only 8 L of brine to the package in the first 1000 years (Ref 151). Additionally, more significant quantities of brine could contact the package as a result of accidental flooding. Sub-seabed sediments will be saturated with essentially stagnant seawater (water velocity: 250

>10

>250

>10

Vehicle 1, 660 days, 51, 000 km (31 700 miles); vehicle 2, 600 days, 53, 500 km (33, 250 miles). Source: Ref 9

The surface roughness of electrogalvanized steels is advantageous for paint adhesion, however, and electrocoated steels also offer better formability and better weldability than hot dip steels (Ref 1). Electrogalvanized surfaces generally contain fewer defects than hot-dip steels, so it is easier to achieve a Class A paint finish on electrocoated steels. For these reasons, use of electrocoated steels (both pure zinc and zinc alloy coatings) in the automotive industry is on the rise; this trend can be expected to continue if research into more efficient electrodeposition processes proves beneficial and if the cost of these materials is lowered. Zincrometal is also used extensively for outer body panels in automobiles. First introduced in 1972, Zincrometal is a coil-coated product consisting of a mixed-oxide underlayer containing metallic zinc particles and a zinc-rich organic (epoxy) topcoat. It is weldable, formable paintable, and compatible with commonly used adhesives. Zincrometal is used primarily in one-side applications to protect against inside-out corrosion. The corrosion resistance of Zincrometal is not as good as that of hot dip galvanized steels (Ref 4), and its use is expected to decline as more electrogalvanized steels and other types of coatings are employed. Zinc-alloy coated steels have also been developed. Coatings include zinc-iron (15 to 80% Fe) and zinc-nickel (10 to

14% Ni) alloys. These coatings have been developed for the most part by Japanese steelmakers and are applied by electrodeposition. Zinc-iron coatings offer excellent corrosion resistance and weldability. Zinc-nickel coatings are more corrosion resistant than pure zinc coatings, but problems include brittleness from residual stresses and the fact that the coating is not completely sacrificial, as is a pure zinc coating. This can lead to accelerated corrosion of the steel substrate if the coating is damaged (Ref 10). Multilayer coatings that take advantage of the properties of each layer have been developed in Europe. An example of this is Zincrox, a zinc-chromium-chromium oxide coating (Ref 10): The CrOx top layer of this coating acts as a barrier to perforation and provides excellent paint adhesion and weldability (Ref 10). Another relatively new development is zinc alloy coatings is Galfan, a Zn-5Al-mischmetal alloy coating applied by hot dipping. Applications in the United States are limited, but European automakers have used Galfan in such applications as brake servo housings, headlight reflectors and frames, and universal joint shrouds (Ref 11). Galfan is also being considered for oil pans, fuel tanks, and heavily formed body panels. Table 3 compares the undervehicle corrosion resistance of Galfan with that of hot dip galvanized steel. More information on zinc coatings is available in the articles "Corrosion of Zinc," "Electroplated Coatings," and "Hot Dip Coatings" in this Volume.

Table 3 Corrosion of hot-dip galvanized and Galfan coated steel specimens in undervehicle testing Exposed in Buffalo and Detroit for the entire winter of 1981-1982 Material

Buffalo Galfan Galvanized Detroit Galfan Galvanized

Specimen orientation

Coating weight Exposed side Protected side g/m2 oz/ft2 g/m2 oz/ft2

Change in coating weight, %

Pitting

Surface appearance

Horizontal Horizontal

216 174

0.71 0.57

223 272

0.73 0.89

3 36

Slight Severe(a)

Smooth, uniform gray Rough, dirty brown

Vertical Vertical

216 207

0.71 0.68

238 253

0.78 0.83

9 18

Slight Severe(a)

Smooth, uniform gray Rough, dirty brown

Source: Ref 11

(a) Other Coated Steels

Galvanized samples averaged 53 times as many pits per unit area than Galfan samples.

Aluminum-coated (aluminized) steels containing 8 to 12% Si in the coating are used by automakers for applications

involving high-temperature corrosion resistance, such as in exhaust systems, heat shields, and underhood components. The Al-45Zn coated steels are used in similar applications (Ref 1). Long terne coated (lead-tin alloy, usually 3 to 8% Sn) steels offer corrosion protection in gas tanks, fuel lines, and

brake lines and do not contaminate gasoline (Ref 1). Terne is cathodic to the steel substrate and therefore does not offer the sacrificial corrosion protection of galvanized coatings. Terne-coated steels are also sometimes given a thin coating of electrodeposited nickel as an intermediate layer; this material is used in applications similar to those of regular terne (Ref 1). Organic composite coated steels have been developed mainly by Japanese steelmakers in cooperation with

automakers in that country, although development is underway in other countries as well. These coil-coated products generally employ an electroplated zinc alloy base layer and a chemical conversion coating under a thin organic topcoat containing a high percentage of metal powder (Ref 12, 13, 14). The thinnes of the organic topcoat allows for good formability without the risk of damaging the coating. Figure 7 (Ref 12) compares the corrosion resistance of one of these organic composite coated sheet steels to cold-rolled steel and to Zincrometal. Another of these products uses an organic-silicate composite topcoat only about 1 m thick and has corrosion resistance and weld-ability superior to that of Zincrometal (Ref 13). A bake-hardenable version of this material has also been developed (Ref 13). Researchers at a third Japanese steel company have developed a bakehardenable organic composite coated sheet steel with a 0.8- to 1.5- m thick organic topcoat. The material possesses corrosion resistance, formability, and weldability equivalent to that of Zincrometal-KII, which uses a 7- m thick top coat (Ref 14). Production of these composite coated materials in increasing in anticipation of increased demand from Japanese automakers.

Fig. 7 Corrosion of heavily worked samples of a composite coated steel, Zincrometal, and cold-rolled steel in a laboratory cyclic test. Test consisted of 28-min cycles of dipping in 5% saline solution at 40 °C (100 °F), humidifying at 50 °C (120 °F), and drying at 60 °C (140 °F). Source: Ref 12

A similar material has been developed in the United States. This material has an electrodeposited zinc alloy base coat, a mixed intermediate layer of chromium oxide and zinc dust, and an organic topcoat for barrier protection (Ref 15). Figure 8 is a micrograph showing the cross section of the composite coated steel. In salt spray tests comparing this material to electrodeposited zinc-nickel and Zincrometal, zinc-nickel failed after 216 h, Zincrometal at 480 h, and the composite coating at 960 h (Ref 15). This material was developed to have weldability, formability, and adhesive compatibility similar to that of Zincrometal, and developmental work is continuing.

Fig. 8 SEM micrograph of across section through a composite coated sheet steel. Source: Ref 15

Paint Systems The primary function of automotive part systems is to provide a protective barrier against corrosive substances in the outside environment. This is accomplished through the use of a paint system comprising a conversion coating, one or more coats of primer, and a colored topcoat. A typical system might employ a conversion coating, an electrodeposited primer 30 m thick, a base color coat 15 m thick, and a clear topcoat 40 m thick. Because corrosion usually initiates at coating defects, emphasis is placed on obtaining the most defect-free coating system possible. Detailed information on the types of paint formulations available, the application of paints, and corrosion of painted metals is available in the article "Organic Coatings and Linings" in this Volume, and Table 4 outlines the development in the last 25 years of primer and paint systems at one U.S. automaker.

Table 4 Evolution of primer and paint systems at one U.S. automaker Primer or paint

Primers Water-reducible dip primer High-solids spray primer Cathodic electrodeposited primer Color coats Conventional high-efficiency acrylic enamels Basecoat/clear coat acrylic enamels

Year of introduction

Curing temperature °C °F

Minimum thickness mm (mils)

1960 1981 1983

165 165 180

325 325 350

0.015 (0.6) 0.019 (0.75) interior, 0.03 (1.2) exterior 0.013 (0.5) interior, 0.03 (1.2) exterior

1979 1981

120 120

250 250

0.043 (1.7) 0.02 (0.8) basecoat, 0.03 (1.2) clear coat

Surface Preparation Metal Cleaning. Surface preparation for painting begins with cleaning of the metal. Cleaning usually involves the use

of an alkaline cleaner and one or more water rinses to remove dirt and contaminants--for example, oil left from the stamping operation--that can limit adhesion of the subsequent coatings. More information on chemical cleaning of metals is available in the article "Cleaning for Surface Conversion" in this Volume. A phosphate conversion coating is then applied to the clean metal by either spraying or immersion. A typical

process sequence involves application of the phosphate solution to the metal, a cold water rinse, and a chromic acid rinse to seal the conversion coating and to enhance corrosion resistance. More information on the application and corrosion resistance of phosphate conversion coatings is available in the article "Phosphate Conversion Coatings" in this Volume; chromate conversion coatings are discussed in the article "Chromate Conversion Coatings" in this Volume.

The conversion coating enhances the corrosion resistance of the metal surface, but more importantly, the crystalline nature of the coating provides excellent paint adhesion for subsequent application of primer and topcoats. For this reason, a conversion coating with small, dense crystals and minimal porosity is desirable (Ref 4). Primers Primers are applied to body panels to enhance corrosion resistance, to give a better surface appearance to the finished panel, and to provide an adherent surface for subsequent organic coatings. Parts are primed immediately after the conversion coating process. Currently, the most frequently used primer materials are the high-film build epoxy resins, usually with corrosion-inhibiting pigments and a small amount of solvent added to aid in flow during the priming operation (see the article "Organic Coatings and Linings" in this Volume for more information on paint formulations). Most primers are thermosetting compounds that require curing after application. Either single or multiple primer coats can be applied by spraying, dipping, flowing, or the cathodic electrocoat process (Ref 16). Spray application of primer can be accomplished either manually or automatically, depending on the complexity of the parts being processed. Both electrostatic and airless spraying are employed. The spraying process gives uniform coating thicknesses and flexibility (coating thickness and areas to be covered are easily varied) (Ref 16). In dip coating, the parts to be primed are immersed in a large tank containing the primer. Care must be taken when

dipping parts containing inner panels or other features that could result in the formation of air bubbles, which prevent paint coverage in the area of the bubble. Adequate drainage must be provided to prevent collection of excess primer in crevices and shelf areas of the workpiece. Both of these potential problems can be minimized by good part design (for example, adding vent holes to parts with inner and outer panels, such as hoods). In flow coating the primer is dispensed through large nozzles; excess primer is then allowed to drip off of the part and

is collected for reuse. Flow coating usually gives better coverage of multipiece components than spraying, because the primer is allowed to flow around and through the part. Electrocoating. The state-of-the-art in primer application is electrodeposition. Cathodic electrodeposition has become

the dominant method of primer application for automakers around the world. In 1982, 33.4 million of the approximately 37 million cars and trucks built throughout the world used electrodeposited primers; of these, 25.4 million were coated by the cathodic electrodeposition process (Ref 17). In the cathodic electrodeposition process, the workpiece is negatively charged and is immersed in a large tank containing the primer, which is positively charged. The primer is electrically attracted to the metal, resulting in a uniform coating thickness, excellent adhesion, a smooth surface, and excellent corrosion resistance. The electrodeposition process lends itself to automation, saves paint, is free from dripping and sagging, and is extremely reliable. Film thicknesses of 13 to 18 m can easily be deposited, and high-film build primer formulations can allow deposition of thicknesses up to 35 m (Ref 17). Topcoats The colored organic topcoat in automotive applications provides an additional barrier against the outside environment as well as a pleasing appearance. A wide range of topcoat formulations have been used in the automotive industry; nonaqueous acrylic dispersion enamels, high-solids solution enamels, thermoplastic acrylic lacquers, and high-solids basecoat/clear coat enamels are currently used in North America (Ref 1). Topcoats are usually applied by spraying to obtain the best possible finish and high gloss. Electrostatic spraying techniques help to maintain a uniform film thickness and appearance. The topcoats system currently used by one U.S. automaker uses a urethane-acrylic enamel color coat followed by a clear coat. The clear acrylic topcoat is applied to the wet color coat, and it protects the color coat from ultraviolet radiation from the sun, preventing color changes in the color coat pigment and resin. The clear topcoat can be polished to remove dirt and defects; polishing also gives a high gloss to the topcoat system (Ref 4). Trends in topcoat formulation and application are being determined more by stricter environmental regulations and the desire to reduce production costs than by need or improved paint performance (Ref 1). Air quality legislation mandating reduced solvent emissions has resulted in the development of waterborne base coats and high-solids paints; automated

cathodic electrode-position of primers and electrostatic spraying of topcoats has increased production efficiency and reduced costs (Ref 1).

Corrosion in Other Automotive Systems This article thus far has dealt mainly with corrosion and corrosion protection for automotive body materials. This section will discuss other areas of the automobile that are subject to corrosion. Underhood and underbody corrosion affects such vital components as fuel systems, cooling systems, electrical systems, and exhaust systems. Although perhaps not of cosmetic concern, corrosion in these areas can affect the safe operation of the vehicle. Fuel Systems. As mentioned earlier in this article, fuel tanks and lines are generally fabricated in the United States from long terne (lead-tin) coated steels. Terne-coated steel has good overall corrosion resistance in this application, but is subject to pinhole-type corrosion if water is trapped in the fuel tank (Ref 18). Electrogalvanized steel is also used by some overseas manufacturers, and it has good resistance to pinhole-type corrosion. Over time, however, the inner surfaces of an electrogalvanized fuel tank may form white zinc corrosion products (white rust), with subsequent attack of the base metal (Ref 18). For other components of the fuel system, such as fuel lines, hot dip Zn-5Al-mischmetal (Galfan) coated steel is being employed (Ref 11). Cooling Systems. Corrosion of automotive cooling systems is accelerated by dissimilar-metal couples, exhaust gas

leakage, high operating temperatures, aeration, poor quality water, and coolant flow (Ref 19). A wide variety of materials are used in the typical automotive cooling system. Wrought brass or aluminum is used for radiators and heater cores, stamped steel for various small components and housings, and aluminum for parts such as coolant pumps (Ref 19). Radiator tubes can be attacked by general corrosion, and brass tubes are subject to dezincification on both internal and external surfaces. Also, because brass tubes are fabricated by soldering, solder flux may cause stress-corrosion cracking (SCC) (Ref 20). A Cu-35Zn-0.3Al-0.2Sn-0.02P alloy has been developed that gives resistance to dezincification equal to that of arsenical brass in laboratory tests, and it has higher resistance to SCC than other brasses tested (Ref 20). Components such as the coolant pump are subject to cavitation damage from collapsing vapor bubbles in the coolant; excessive coolant flow can also cause impingement damage to radiator tubes. Because of the variety of materials used in the cooling system, galvanic corrosion is of concern when dissimilar-metal parts are placed in electrical contact by the conducting coolant. Both of these problems can be alleviated somewhat by good design practices (Ref 19). Proper maintenance is also important in minimizing cooling system corrosion, but surveys show that proper maintenance practices and manufacturers' recommendations are often neglected in the United States (Ref 21). Use of antifreeze at the proper concentration--U.S. automakers and antifreeze suppliers usually recommend a mixture of 50 to 70% antifreeze with water--is important, and coolant formulation can have an effect on corrosion if proper inhibitors are not used (Ref 21). No single inhibitor can protect all of the metals in an automotive cooling system, so many coolant formulations use a combination of inhibitors (Ref 19). Electrical systems are subject to a wide variety of corrosion problems caused by the severity of the automotive

environment. Corrosion problems in conventional electrical systems can be minimized at the design stage by sealing components whenever possible. A wide variety of organic and inorganic compounds are used for this purpose, including paints and primers, silicone compounds, varnishes, numerous plastics, and oils and greases (Ref 22). The use of electronic components in automobiles has increased dramatically in the past decade. These systems were first used for engine control, but they now perform a wide variety of control and monitoring functions (Ref 23). Electronic components are subject to a variety of corrosion problems, including corrosion-induced leakage and shorts on printed circuit boards, metal migration problems, and corrosion in plastic-packaged devices (see the article "Corrosion in the Electronics Industry" in this Volume). Proper design, manufacturing, and quality assurance procedures can help to minimize these types of problems in automotive electronic systems. Automotive exhaust systems are subject to general and localized external corrosion and to internal corrosion caused

by exhaust gas condensates (Ref 24). Exhaust gas temperatures near the exhaust manifold have been measured at up to 870 °C (1600 °F), with corresponding metal temperatures as high as 595 °C (1100 °F) (Ref 24); therefore, exhaust system components are subject to high-temperature oxidizing conditions.

The materials used to combat corrosion in exhaust systems include aluminized steel and type 409 stainless steel. The use of type 409 stainless steel is increasing because of its relatively low cost and good resistance to corrosion and hightemperature oxidation. The use of catalytic converters, which began in the United States in the 1975 model year, has presented some special problems for automakers. Because of the oxidizing, catalyzed reaction that takes place in the exhaust gas stream in the converter, exhaust gas temperatures and the amount of corrosive substances, such as sulfuric acid, in the exhaust gas stream are increased (Ref 24). The converter itself essentially consists of a noble metal coated ceramic substrate, often housed in a type 409 stainless steel canister. The location of the converter between the exhaust manifold and the exhaust pipe has prompted the use of more corrosion-resistant materials down-stream from the converter. Other Automotive Systems. A variety of coated steels are used for automotive suspension components, including hot-dip galvanized and galvannealed steels, electrogalvanized steel, aluminized steel, and cathodic electrocoated or epoxy powder coated steels. Some suspension components (for example, rear cross members) and structural components (such as bumper reinforcements and body side rails) are being fabricated from high-strength low-alloy steels to reduce weight, but formability problems with these materials have thus far limited their used to structural-type applications (Ref 25). High-strength steels, although thinner in section than carbon steels, are protected in much the same way by a variety of coating materials. More information on the types of microalloyed steels used, the strength ranges employed, and the automotive applications for such materials is available in Ref 26, 27, 28, 29, 30, 31, and 32.

Clad metals are used in the auto industry for their decorative appearance and corrosion resistance, for example, in automotive trim applications (Ref 33). Figure 9 shows a stainless steel clad aluminum alloy used for trim applications. The aluminum inner surface, being more electrochemically active than steel, acts as a sacrificial anode to prevent corrosion of the adjacent steel body panel. The outer stainless steel surface provides corrosion resistance, abrasion and dent resistance, and a decorative appearance. This material is also used for auto and truck bumpers to reduce weight while maintaining corrosion resistance (see the article "Corrosion of Clad Metals" in this Volume).

Fig. 9 Stainless steel clad aluminum alloy used in automotive trim applications. The use of steel clips to attach the trim strip prevents galvanic corrosion of the aluminum in contact with the steel body panel.

Design Considerations As mentioned earlier in this article, design can play an important role in determining the corrosion resistance of an automotive assembly or component. In fact, the configuration of a part or assembly is often the determining factor in the

type and severity of corrosion that occurs in service (Ref 1). Design factors that can influence corrosion resistance are reviewed in Ref 34 and 35, and additional information on this subject is available in the article "Design Details to Minimize Corrosion" in this Volume.

Corrosion Testing The automotive industry employs many common corrosion test methods; for example, laboratory salt spray and electrochemical tests are often used in the development and evaluation of new materials (see the article "Laboratory Testing" in this Volume). Tests that are specific to the industry also are used, such as proving ground vehicle testing, mobile testing using underbody test racks, and field surveys. Recent emphasis has been on the development of laboratory tests that can closely approximate the results of much longer and more costly proving ground or undervehicle tests. This section will focus on both field and laboratory testing methods that are specific to the automotive industry. Mobile testing using test racks mounted under or on the vehicle is commonly used to evaluate as-received precoated steels and primed and painted specimens. Of the commonly used test methods, mobile testing probably comes the closest to simulating actual service conditions (Ref 1), but it requires test periods of up to several years. The measurements taken usually include percent surface area of base metal attacked and depth and density of pitting. The test methods used and the results of two of these tests are documented in Ref 36 and 37, and Ref 38 correlates the results of an undervehicle testing program with those obtained from a laboratory test method. Proving Ground Testing. In this method, prototype or production vehicles are corrosion tested on the proving grounds

of the company. Test cycles that produce accelerated attack are used, with test times varying from 10 weeks to 10 months (Ref 1). Proving ground testing can simulated service conditions for some types of corrosion, but other forms, such as perforation corrosion, are more difficult to accelerate. Field surveys, when properly conducted, offer a direct method of comparing the performance of various types of

coatings and other materials. These surveys often involve the destructive inspection of actual vehicles with well-defined service histories in corrosive environments (Ref 1). The results of one recent survey of 5- and 6-year-old vehicles are documented in Ref 39. Laboratory Tests. Cyclic laboratory tests have been developed in recent years to simulate in a relatively short time the

effects of many years of actual operation. One such test uses three environment chambers to subject the test material to extreme climatic conditions, and it uses other equipment, such as a gravel blower, to simulate actual road use as closely as possible (Ref 40). This test is intended to simulate 6 years of actual service over a 14-week test period. Other such tests have been developed, but correlation to actual service tests has not been good. More development is required to make these tests reliable indicators of the corrosion resistance of materials for automotive use.

References 1. "Cracking Down on Corrosion: Cooperative Efforts Toward Vehicle Durability," American Iron and Steel Institute, 1985 2. J.C. Bittence, Waging War on Rust, Part I: Understanding Rust, Mach. Des., 7 Oct 1976, p 108-113; Part II: Resisting Rust. 11 Nov 1976, p 146-152 3. "US Automotive Market for Zinc Coatings 1984-1986," Zinc Institute Inc. 4. D.J. Bologna, Corrosion Resistant Materials and Body Paint Systems for Automotive Applications, SAE Paper 862015, in Proceedings of the Automotive Corrosion and Prevention Conference, P-188, Society of Automotive Engineers, 1986, p 69-80 5. M. Chilaud, S. Mathieu, P. Pichant, and G. Quinchon, "Hot Dip Galvanized Steel--Product of the Future," SAE Paper 860271, Society of Automotive Engineers, 1986 6. D.F. Baxter, Jr., Developments in Coated Steels, Met. Prog., May 1986, p 31-35 7. Auto Makers Take the Plunge Into Hot Dip Galvanizing, Zinc, No. 1, Zinc Institute Inc., 1985, p 5 8. H. Kunitake, The Challenge of the Automotive Industry, in Proceedings of the 18th Annual Meetings of the IISI, International Iron and Steel Institute, 1984 9. R.J. Neville and K.M. DeSouza. Electrogalvanized or Hot Dip Galvanized--Results of Five Years of Undervehicle Corrosion Testing, SAE Paper 862010, in Proceedings of the Automotive Corrosion and

Prevention Conference, P-188. Society of Automotive Engineers, 1986, p 31-40 10. M. Memmi et al., A Qualitative and Quantitative Evaluation of Zn + Cr-CrOx Multilayer Coating Compared to Other Coated Steel Sheets. SAE Paper 862028, in Proceedings of the Automotive Corrosion and Prevention Conference, P-188, Society of Automotive Engineers, 1986, p 175-185 11. R.F. Lynch and F.E. Goodwin, "Galfan Coated Steel for Automotive Applications," SAE Paper 860658. Society of Automotive Engineers, 1986 12. Y. Shindou et al., Properties of Organic Composite-Coated Steel Sheet for Automobile Body Panels, SAE Paper 862016, in Proceedings of the Automotive Corrosion and Prevention Conference, P-188, Society of Automotive Engineers, 1986, p 81-90 13. M. Yamashita, T. Kubota, and T. Adaniya, Organic-Silicate Composite Coated Steel Sheet for Automobile Body Panel, SAE Paper 862017, in Proceedings of the Automotive Corrosion and Prevention Conference, P-188, Society of Automotive Engineers, 1986, p 91-97 14. T. Mohri et al., Newly Developed Organic Composite-Coated Steel Sheet With Bake Hardenability, SAE Paper 862030, In Proceedings of the Automotive Corrosion and Prevention Conference, P-188, Society of Automotive Engineers, 1986, p 199-208 15. T.E. Dorsett, Development of a Composite Coating for Pre-Coated Automotive Sheet Metal, SAE Paper 862027, in Proceedings of the Automotive Corrosion and Prevention Conference, P-188, Society of Automotive Engineers, 1986, p 163-173 16. W.E. Tudor, A Primer--Automotive Finishing and Corrosion Protection, SAE Paper 780914, in Designing for Automotive Corrosion Prevention, P-178, Society for Automotive Engineers, 1978, p 36-42 17. F.M. Loop, High Film Build Cathodic Electrodeposition Provides Improved Corrosion Protection. SAE Paper 831813, in Proceedings of the 2nd Automotive Corrosion Prevention Conference, P-136, Society of Automotive Engineers, 1983, p 35-44 18. D.J. Bologna and H.T. Page, Corrosion Considerations in Design of Automotive Fuel Systems. SAE Paper 780920, in Designing for Automotive Corrosion Prevention, P-78, Society of Automotive Engineers, 1978, p 65-70 19. E. Beynon, N.R. Copper, and H.J. Hannigan. Cooling System Corrosion in Relation to Design and Materials, SAE Paper 780919, in Designing for Automotive Corrosion Prevention, P-78, Society of Automotive Engineers, 1978, p 56-64 20. J. Miyake, M. Tsuji, and S. Kawauchi, Corrosion Prevention for Automobile Radiator Tubes, SAE Paper 862021, in Proceedings of the Automotive Corrosion and Prevention Conference, P-188, Society of Automotive Engineers, 1986, p 117-121 21. N.R. Cooper, H.J. Hannigan, and J.C. McCourt. A One Thousand Car Assessment of the U.S. Car Population Cooling Systems. SAE Paper 831821, in Proceedings of the 2nd Automotive Corrosion Prevention Conference, P-136, Society of Automotive Engineers, 1983, p 121-130 22. M.M. Jones and E.E. Welker, Electrical Component Corrosion Prevention. SAE Paper 780924, in Designing for Automotive Corrosion Prevention, P-78, Society of Automotive Engineers, 1978, p 107-118 23. J.P. Cook and G.E. Servais, Corrosion Failures in Semiconductor Devices and Electronic Systems, SAE Paper 831830, in Proceedings of the 2nd Automotive Corrosion Prevention Conference, P-136, Society of Automotive Engineers, 1983, p 187-197 24. W.R. Patterson, Materials, Design, and Corrosion Effects on Exhaust System Life, SAE Paper 780921, in Designing for Automotive Corrosion Prevention, P-78, Society of Automotive Engineers, 1978, p 71-106 25. S. Dinda, C. Belleau, and D.K. Kelley, High Strength Low Alloy Steels in Automotive Structures, in HSLA Steels: Technology and Applications, American Society for Metals, 1984, p 475 26. D.A. Wilkinson and D.D. Rogers, A New HSLA Steel for an Automotive Steering Coupling Component, in HSLA Steels: Technology and Applications, American Society for Metals, 1984, p 459 27. R.G. Davies, Forming Problems Encountered in Application of High Strength Steels to Automotive Components, in HSLA Steels: Technology and Applications, American Society for Metals, 1984, p 467 28. H.E. Chandler, High Strength Sheet Forms Like Mild Steel, Met. Prog., Nov 1985, p 63-66 29. M. Takahashi et al., Criteria of High Strength Steels for Applying to Automobile Frame Components, in

HSLA Steels: Technology and Applications, American Society for Metals, 1984, p 493 30. K. Tamura and M. Shiokawa, Application of Higher Strength Steel Sheets and Its Process in Nissan Motor Company, in HSLA Steels: Technology and Applications, American Society for Metals, 1984, p 503 31. G.T. Halmos, Roll Forming HSLA Steels, in HSLA Steels: Technology and Applications, American Society for Metals, 1984, p 515 32. J.C. Kopchick, Automotive Application of Ultra-High Strength Steel Sheet, in HSLA Steels: Technology and Applications, American Society for Metals, 1984, p 523 33. R. Baboian, Causes and Effects of Corrosion Relating to Exterior Trim on Automobiles, SAE Paper 831835, in Proceedings of the 2nd Automotive Corrosion Prevention Conference, P-136, Society of Automotive Engineers, 1983, p 223-227 34. "Prevention of Corrosion of Metals," Handbook Supplement HSJ 447, Society of Automotive Engineers, 1981 35. L.C. Rowe, "The Application of Corrosion Principles to Engineering Design," Paper 770292, presented at the SAE Automotive Engineering Congress, Society of Automotive Engineers, Feb 1977 36. R.D. McDonald and R.R. Ramsingh, Eighteen Months of Underbody Automotive Materials Testing, Mater. Perform., Vol 24 (No. 4), April 1985, p 48-53 37. R.D. McDonald and R.R. Ramsingh, "Corrosion Testing by the Under-Car Method, December 1978 to May 1980," Report MRP/PMRL 81-46(TR), Canadian Centre for Mineral and Energy Technology, Physical Metallurgy Research Laboratories, July 1981 38. R.D. McDonald, "Corrosion of Automotive Steels in Deicing Salt Environments: Comparison of a Laboratory Method With Undercar Testing," Report MRP/PMRL 83-71(J), Canadian Centre for Mineral and Energy Technology, Physical Metallurgy Research Laboratories, March 1983 39. A.W. Bryant and W.C. Oldenburg, 1985 Body Corrosion Survey--5 and 6 Year Old Vehicles, SAE Paper 862025, in Proceedings of the Automotive Corrosion and Prevention Conference, P-188, Society of Automotive Engineers, 1986, p 143-154 40. R. Dietz, A Three-Chamber Corrosion Test Method for Passenger Cars, SAE Paper 831814, in Proceedings of the 2nd Automotive Corrosion Prevention Conference, P-136, Society of Automotive Engineers, 1983, p 47-56

Corrosion in the Aircraft Industry Introduction CORROSION CONTROL is of the utmost concern in the aircraft industry because of its potential impact on human safety and on extremely expensive aircraft. In the military, winning the war on corrosion is essential to military preparedness and national security. Military aircraft are very expensive; some, such as the B-1B strategic bomber, cost over $200 million each. To achieve high levels of payload and performance in military aircraft, materials are chosen more for their mechanical properties than for their inherent corrosion resistance. They depend on coatings and routine maintenance to preserve their integrity. Military aircraft are flown throughout the world and are therefore exposed to the most severe corrosive environments on earth. The contracting agency imposes numerous specifications on suppliers to control on military aircraft. An excellent text on corrosion is MIL-HDBK-729, Corrosion and Corrosion Prevention--Metals. The military services sponsor a Tri-Service Corrosion Conference every 2 or 3 years to provide a forum for corrosion control and experiences with various weapon systems. In addition, the Corrosion Information Analyses Center serves the Department of Defense (DOD) as part of the Metals and Ceramics Information Center. Each of the military branches have corrosion control centers and laboratories serving them. Corrosion control of commercial aircraft is also of paramount importance for similar reasons. Flight safety is essential to the airline industry. Again, commercial airlines are exposed to highly corrosive environments all over the world.

Commercial aircraft represent investments of up to $100 million per unit for some of the widebody aircraft. Fullscale inspections to determine structural integrity often exceed $2.5 million. Even more severe economic consequences could result from lawsuits or the loss of a manufacturer's reputation in the event of a corrosion-related mishap. Consequently, although only a short reference to the need for corrosion control to ensure airworthiness and safety is found in Federal Aviation Administration regulations (Federal Airworthiness Regulation 25.609), commercial aircraft manufacturers take as much care in this area as the military. Commercial airlines, which are responsible for corrosion control once the aircraft is delivered, are provided with detailed documents from the aircraft manufacturers, such as corrosion control handbooks and customer service documents, to ensure proper servicing and inspection. Nevertheless, in spite of extensive corrosion control efforts, problems are experienced. This article will present typical examples of corrosion problems in the aircraft industry to provide additional insight into the causes and corrective actions. Corrosion of Airframes Michael L. Bauccio, The Boeing Company

The potential for corrosion of aircraft structures is a major consideration in the design of the aircraft. Corrosion can be related to various types of material deterioration. This is because the corrosion process can be defined as the degradation of a material or materials by a reaction with the environment. Usually, the reaction is electrochemical, and the material or materials are metallic (Ref 1). Corrosion phenomena often occur on the surfaces of aircraft structures (Ref 2). As shown in Fig. 1, the physical effects of corrosion can be categorized as follows: • • • •

Local pitting or crevice corrosion (one-dimensional local irregularity) General surface roughening due to uniform, filiform, or fretting corrosion (two-dimensional discontinuity) Intergranular and transgranular cracking, which is observed in stress-corrosion cracking (SCC) (twodimensional discontinuity) Degradative transformation of materials that occurs on a larger scale, which is referred to in Fig. 1 as a three-dimensional bulk reduction of material. This is observed in galvanic and exfoliation corrosion

Fig. 1 Schematic showing the physical effects of corrosion on metallic aircraft materials. See text for details.

The detrimental effects of aircraft corrosion are not due to corrosion alone, but to the interaction of corrosion with fatigue, wear (including erosion and fretting), erosion, and stress resulting in premature fracture. The most significant aspects concerning aircraft corrosion that will presented in this section are that: • • • •

•

Airframe corrosion problems vary in severity Cosmetic corrosion, which simply mars the appearance of airframe surfaces, can develop into a widespread form of material deterioration Catastrophic mechanical failures can develop from corrosion that is permitted to spread and cause a significant reduction in structural strength Environmental and mechanical factors combine to produce the types of aircraft corrosion that usually lead to catastrophic in-flight failures. The tropical marine environment is one of the most severe that modern aircraft are exposed to. An example of wing strut deterioration that was produced by this highly corrosive environment is shown in Fig. 2. The corroded wing strut was fabricated from aluminum alloy 2024-T351 (QQ-A-200/3) extruded bar Some of the most critical types of aircraft corrosion than can result in fracture of components or airframe members are SCC, corrosion fatigue, and hydrogen embrittlement (Ref 3). Selected examples of these aircraft fracture failures are illustrated in Fig. 3, 4, 5, 6, 7, 8

Fig. 2 Corroded aluminum alloy 2024-T351 wing strut (a) with attached leading edge. (b) Close-up of corrosion on the wing strut with the leading edge skin removed. (c) Corrosion on the wing strut. 25×

Fig. 3 Cracking of an aircraft pressure cabin. The crack progressed through the underlying frame. Source: Ref 4

Fig. 4 Service failure of an aircraft pivot bracket. Arrow points to fracture origin. Source: Ref 4

Fig. 5 Service failure of a helicopter rotor drive yoke. Source: Ref 4

Fig. 6 Service failure of an aircraft main-plane spar boom. Arrows point to fracture origins. Source: Ref 4

Fig. 7 Service failure of a helicopter rotor blade extrusion. Source: Ref 4

Fig. 8 Fracture of an aluminum alloy 2014-T6 main landing gear actuator from a C-141 military cargo plane. Exfoliation induced by differential aeration was indicated as the cause of this failure, based on the appearance of the fracture near the origin. Source: Ref 5

New Material and Process Solutions From Old Corrosion Problems Of the numerous cases of aircraft corrosion that the author has reviewed, many have appeared in various forms in the past 30 years. However, those corrosion problems that have challenged aircraft materials and process engineers have usually led to positive developments, which have resulted in greater control of aircraft corrosion. Some very recent corrosion cases have also led to significant new corrosion-preventive measures for aircraft applications. For example, in 1985, a C-5B military cargo aircraft was grounded so that about 11,000 aluminum nuts could be replaced by the specified cadmium-plated steel nuts (Ref 6). An additional 40,000 aluminum nuts are to be replaced on four other C-5B aircraft. The aluminum nuts, which were installed despite the engineering drawing requirement for cadmium-plated steel nuts, were discovered during an inspection by the manufacturer. The positive side of this story is that the C-5B manufacturer has developed and patented a magnesium chromate (MgCrO4) sealant that can be used for corrosion prevention in structures that have mating surfaces of aluminum and steel (Ref 6). Aluminum nuts can be used for reduced weight, and steel is used for greater strength on the C-5B, which contains about 4,000,000 structural fasteners. Other corrosion-related incidents that have been investigated by aerospace materials engineers have been reported many years ago and are still being studied today so that improved solutions can be obtained. One such case is concerned with the hydrogen embrittlement and SCC of high-strength steel aircraft components. A monograph on hydrogen stress cracking and hydrogen embrittlement of low-alloy aircraft steel was published in 1956 (Ref 7). The failure of 4140 steel nacelle eyebolts was reported in 1971 (Ref 7). This bolt failure was due to a baking time that was too short to provide the steel bolts with proper hydrogen embrittlement relief, which would have enhanced the mechanical strength of these fasteners. In-service failures due to hydrogen embrittlement and SCC have also occurred in chromium-plated landing gears and in cadmium-plated steel fasteners (Ref 7). Approximately 14 years after the above case histories were published, a report confirmed that baking time is the only significant variable in determining the final hydrogen concentration in cadmium-plated high-strength steels (Ref 8). The importance of baking time for hydrogen embrittlement relief was recognized by several aerospace companies in the early 1960s. The process specifications that had been written by these companies were revised to require 23 h of baking at 190 °C (375 °F), instead of the 3-h baking time that was found in military specifications.

One approach that can be used to prevent hydrogen embrittlement is to avoid the processing of steel components in plating or pickling baths. Another technique is to apply iron vapor deposited (IVD) aluminum coatings. These coatings have recently been reported as an effective method for protection against hydrogen-induced embrittlement, as well as SCC, and can withstand service temperatures up to 495 °C (925 °F) (Ref 9). Ion vapor deposited coatings must meet the requirements of military specification MIL-C-83488 (Ref 10). Aluminum IVD coatings have been applied to the following aircraft structures with favorable results in terms of improved corrosion resistance (Ref 9): • • • •

Steel and titanium fasteners installed in aluminum aircraft structures A fatigue-critical aluminum alloy wing skin A DC-10 aft engine hanger A high-strength steel landing gear

Exfoliation corrosion, which has a long history in connection with airframe deterioration, was observed many years ago where cadmium-plated fasteners were installed in high-strength aluminum alloys. Exfoliation corrosion occurs in metallic materials that have a directionally oriented grain structure. Riveted aircraft structures have a high vulnerability to exfoliation corrosion because the rivet holes provide an unobstructed pathway for corrosive electrolytes to reach metallic airframe materials, especially aluminum, which is the material selected for many aircraft parts (Ref 11). Exfoliation corrosion initiates between bimetallic couples and progresses along grain boundaries as an intergranular crack. This intergranular crack widens into a crack plane and enlarges into multiple crack planes. Corrosive oxides press outward against the adjacent metal, thus producing a pattern of delamination. An illustration of this type of failure, originating at a fastener hole, is shown in Fig. 9. The best alternative available for preventing exfoliation corrosion is to select a corrosion-resistant alloy and heat treatment. Chromate-inhibited elastomeric sealants can also be applied to protect airframe fasteners against corrosion (Ref 12). These fasteners must be wet-installed with the chromated sealant material in order to obtain the highest degree of corrosion protection.

Fig. 9 Schematic of exfoliation in an aluminum aircraft panel. Source: Ref 12

Another example of exfoliation corrosion is illustrated in Fig. 10. The failed airframe structure shown was removed from an aircraft that operated primarily in a marine environment. The structure is a tail plane attachment fitting made of an aluminum alloy (2024-T4) that meets federal specification QQ-A-250/4. The arrow in Fig. 10(a) points to the corrosion. This corrosion problem was primarily caused by inadequate sealing of the bolt hole during installation of the cadmiumplated steel bolt; this allowed seawater to attack the aluminum alloy.

Fig. 10 Example of exfoliation corrosion. (a) Failed aluminum alloy 2024-T4 tail plane fitting. Arrow points to corrosion that was produced by direct contact between a cadmium-plated steel bolt and the aluminum fitting. (b) Exfoliation in the tail plane fitting. 55×

Additional Corrosion Cases There are many more examples of how the scope of aircraft structural corrosion spans long periods of time between the origination of a specific problem and the development of an improved material or process, or both, for mitigating the corrosion problem. Aircraft corrosion cases are often solved within much shorter periods than those problems described above. However, the same immediate, or quick-fix, solutions will be reevaluated and modified over the years by aircraft manufacturers for: • •

Increased reductions in the gross weight and cost of aircraft systems Materials and processes that will meet the increasingly stringent environmental and occupational safety regulations of state, local, and federal government agencies

Corrosion-preventive technology also requires continued study and modification. This is for the purpose of ensuring that corrosion-resistant materials and corrosion control techniques are effectively applied to the diverse variety of airframe components. For example, thick organic coatings usually cannot be used on mating parts that have close tolerances. These parts are often protected by inorganic coatings that are relatively thin (about 1 to 35 m, or 0.04 to 1.4 mils), such as the wear-resistant chromium platings or anodized coatings on aluminum structures, the black oxide or phosphatized coatings on steel parts, and the dichromate and anodic (Dow 17) treatments used on magnesium alloy structures. Some materials problems, such as microbial deterioration of integral wing fuel tanks, have plagued the aircraft industry since the beginning of the jet age (Ref 13). Such ongoing problems illustrate the need for continuing research and development of new materials and protection systems for corrosion control. Aircraft materials that are resistant to corrosive deterioration should be selected during the design phase. This objective can be facilitated by following the general techniques and design rules for corrosion prevention and control presented in Table 1. When selecting aluminum alloys for SCC resistance, aircraft designers should consult Table 2. The process of selecting corrosion-resistant steels for aircraft systems can be facilitated by reference to Table 3, which provides the resistance to general corrosion and SCC of several classes of steel.

Table 1 General techniques and design considerations for minimizing corrosion Item Eliminate areas where trapped moisture is held in contact with metal

Choose nonabsorbent, nonwicking materials Protect all faying surfaces

Use compatible metals

Select proper finishing systems

Suggestions Avoid such features at the design stage by careful attention to design or structure details. Provide properly located drain holes. Minimum hole size should be 3.2 mm ( in.) to prevent plugging. Determine water absorption qualities of materials to be used. Use epoxy and vinyl tapes and coatings, wax, or latex for protective barriers. Avoid, if possible, use of wood, paper, cardboard, open cell foams, and sponge rubbers. Use proper sealing materials (tapes, films, sealing compounds) on all faying surfaces. Use primers. Lengthen continuous liquid path to prevent formation of an electrolytic cell. For magnesium-aluminum couples, 5000- and 6000-series aluminum alloys are the most compatible. For magnesium-steel couples, use tin or cadmium plated steel. For bimetallic couples use metals or alloys in the same group per MIL-STD-889, or as close as possible. Use tapes or primers on faying surfaces to prevent metallic or electrical contact. Choose chemical treatments, paints, plating on basis of service requirements. Service test system before setting up production run. Use past experience in similar applications as guide to choice.

Source: Ref 14

Table 2 Relative resistance of aluminum alloys to SCC Alloy and temper 2014-T6 2024-T3, T4 2024-T6 2024-T8 2124-T851 2219-T3, T37 2219-T6, T8 6061-T6 7049-T73 7149-T73 7049-T76 7x75-T736 7050-T736 7050-T76 7x75-T6 7x75-T73 7x75-T76

Source: Ref 15

Product form Rolled plate Rod and bar Poor Poor Poor Poor ... Good Good Excellent Good ... Poor ... Excellent Excellent Excellent Excellent Excellent ... ... ... ... ... ... ... Good ... Intermediate . . . Poor Poor Excellent Excellent Intermediate . . .

Extruded shapes Poor Poor ... Good ... Poor Excellent Excellent Good Good Intermediate ... Good Intermediate Poor Excellent Intermediate

Forgings Poor ... Poor Intermediate ... ... Excellent Excellent Good Good ... Good Good ... Poor Excellent ...

Table 3 Resistance of some stainless steels to general corrosion and SCC Alloy General corrosion resistance Austenitic grades Excellent Type 301 Excellent Type 316 High Type 347 High A286 High Type 321 Type 304 High (ELC) High Type 302 High Type 304 Excellent Type 310 Martensitic grades Low to moderate 440C Low to moderate; will develop superficial rust film with 420 atmospheric exposure 410 416 Precipitation-hardenable grades Moderate 21-6-9 Moderate PH13-8Mo Moderate PH15-7Mo Moderate PH14-8Mo Moderate 17-4PH Moderate 15-5PH Moderate AM355 Moderate AM350

SCC resistance Excellent Moderate Excellent Excellent Excellent Excellent Excellent Excellent Excellent Susceptibility varies significantly with composition, heat treatment, and product form.

Susceptibility varies significantly with composition, heat treatment, and product form.

Source: Ref 15

Aircraft Corrosion-Related Failures The best way for professional engineers to grasp the extent of aircraft-related corrosion problems and solutions is to become familiar with many of the corrosion-related cases that have been documented. The following discussions in this section provide a comprehensive overview of many aircraft corrosion problems that have been described in the technical literature. For most of these corrosion cases, the successfully applied corrosion-preventive treatments, which were developed to avert future corrosion problems, are described. These airframe corrosion cases were selected because they can be of great assistance to design and corrosion engineers in alleviating possible airframe corrosion problems in the future. Galvanic Corrosion. When dissimilar and unprotected structural materials with different electrochemical potentials are

assembled and exposed to a corrosive environment, galvanic corrosion (dissimilar-metal corrosion) occurs. The more active (less noble) material of the galvanic couple becomes the anode and therefore undergoes dissolution and corrosion (Ref 16). One of the best ways of identifying galvanic corrosion is to examine the severity of corrosion damage at the junction between the dissimilar metals. Galvanic corrosion usually predominates at the point of contact between the two materials. Several examples of galvanic corrosion, which have been described in the literature on aircraft structural corrosion, are provided below (Ref 16). Cadmium-Plated Steel Fasteners. These fasteners were in contact with a 7075-T6 aluminum skin (Fig. 11). Severe corrosion damage occurred on the aluminum skin around the periphery of the fastener heads.

Fig. 11 Galvanic corrosion of aluminum alloy 7075-T6 aircraft skin around cadmium-plated steel fasteners. (a) General view. (b) and (c) Close-ups of individual fasteners. Source: Ref 16

Aluminum Alloy Skin. Galvanic corrosion in this aluminum alloy skin was caused by an electrically conductive

adhesive compound. The corrosion occurred along the edges of an anti-icing boot that was bonded to the wing leading edges. 7075-T6 Kingpin Lug. Galvanic corrosion on this lug was caused by direct contact between this part and steel

bushings. The bushings were used to protect the lug mounting holes from mechanical damage. This failure could have been prevented by the proper application (and continual inspection) of a protective coating on the aluminum alloy, such as a hard anodized coating. Cast Magnesium Flap Control Lever Arm. The lever arm was manufactured from a magnesium casting alloy. A

zinc-chromate primer and an epoxy topcoat had been applied to the lever arm structure for corrosion protection. Galvanic corrosion occurred in the sockets of the levers, where steel balls were located. The movement of these steel balls in the socket caused mechanical deterioration of the protective coating. The highly anodic magnesium, therefore, was in direct contact with the steel balls, causing the galvanic corrosion. Galvanic corrosion also was found around a press-fitted aluminum bushing in the lever arm. Current aircraft design requirements (in most cases) prohibit the use of magnesium castings, especially for components that will be in contact with moving parts. Application of the sealants and organic coating systems according to the guidelines in Ref 17 might have averted this corrosion problem. In summary, galvanic corrosion can be prevented by avoiding dissimilar-metal contact. Whenever joints between dissimilar metals must be made, metals should be selected whose galvanic potentials in seawater (or salt water) are closest in order to minimize corrosion rates. Generally, potential differences of less than 0.25 V will be satisfactory without protection. Where possible, a high ratio of anodic-to-cathodic area is desirable in designing to prevent galvanic corrosion. Therefore, the use of a fastener that is slightly cathodic to the base metal is often desirable. Furthermore, materials that naturally form protective oxides should be selected. These materials, including stainless steel and aluminum, often do not become active in galvanic cells. Other design approaches include eliminating the corrosive environment and providing an organic or inorganic barrier coating between the dissimilar metals. This involves the installation of steel fasteners into aluminum with a wet chromate primer or the use of an anodized aluminum structure in contact with a steel part. More information on galvanic corrosion is available in the section "Galvanic Corrosion" of the article "General Corrosion" in this Volume.

Uniform corrosion, or general corrosion, is a type of material deterioration that develops evenly over large areas of

aircraft structures. Uniform corrosion is produced by many closely spaced anodic and cathodic sites on unprotected or partially protected surfaces (Ref 16). Because large areas are affected in uniform corrosion, this type of damage can be recognized and remedied relatively easily as compared to other types of corrosion. The recognition of uniform corrosion is also enhanced by the unique characteristics of the corrosion products that are produced by various engineering materials (Ref 16). Typical problems and corrosion-preventive treatments pertaining to uniform corrosion that have been described in the technical literature include the following (Ref 16). Adjuster and Eye of a Steel Track Rod. Rust occurred on these components after the cadmium plating on the

threads became damaged during assembly or maintenance. This problem may have been prevented by the application of a protective coating, such as an organic paint primer, over the cadmium plating. Aluminum Alloy Stringer. General corrosion was found on the surface of this component Deterioration occurred in

areas where the chromate pretreatment and primer coating had flaked off of the stringer. Magnesium Alloy Skin of a Helicopter. This case is similar to the previous example. The uniform corrosion on this

component also developed from the failure of the chromate conversion coating that had been applied to the magnesium skin. Aluminum-Honeycomb Structure on a Fighter Aircraft. Water penetrated the inner surface of the aluminum

skin through small holes. A foam rubber filler between the skin and the honeycomb core possibly contributed to the uniform corrosion that was produced in this case. Structural Steel Fasteners. The surfaces of the nuts had the greatest degree of corrosion damage. In some areas, the

corrosion penetrated very close to the bolt surface. In summary, uniform corrosion can be prevented by selecting proper protective coatings, by ensuring proper surface preparation and coating application, and by touching up deteriorated surfaces as soon as possible. Sacrificial coatings, such as alcladding on aluminum, spread corrosion over large areas and thus enable aluminum sheet metal products to retain their mechanical strength. Information on the mechanism of uniform corrosion can be obtained in the article "General Corrosion" in this Volume. Pitting corrosion produces deterioration of airframe structures by forming cavities and oxidation products in small

(localized) areas of the affected components. The severity of pitting corrosion is determined by: • •

The susceptibility of the airframe material to pitting attack. Unprotected, active metals, such as magnesium, are most susceptible. The severity of the environments often cause pitting in airframe structures (Ref 16). This is because the chloride (Cl-) ions in seawater promote the destruction of protective oxide films on the metallic materials that are used in airframes

Damage that as produced by pitting corrosion in airframe components has been documented for many cases that provide good examples of this type of aircraft degradation. Several of these incidents will be discussed for the following airframe structures. Helicopter Structural Cleat (Ref 16). This magnesium alloy casting had been given the corrosion-preventive

treatments of chromating and sealing (Ref 16). Pitting deterioration was enhanced by local damage in the protective coating. The alleged cause of this damage was impact from tools that were used for installing fasteners. Stringer End Cap (Ref 16). This airframe structure was also made of magnesium, and it was chromated and primed.

Moisture exposure contributed to the severe degree of pitting corrosion that occurred in this part. Spring Wire. This hard-drawn high-carbon steel structure was observed to have been damaged by fatigue cracking in an

area that had deteriorated by pitting corrosion (Fig. 12). Cyclic loading that is exerted on aircraft components during

takeoff and landing promotes this type of fatigue fracturing in airframes that deteriorate because of corrosion. This particular case can also be classified as an example of corrosion fatigue (see the discussion "Corrosion Fatigue" in this section).

Fig. 12 Fatigue cracking of a hard-drawn carbon steel spring. Courtesy of Aeronautical Research Laboratories, Australia

Hydraulic Cylinders. An inner landing gear door actuator cylinder that failed on an F-101 aircraft is shown in Fig. 13. This part was fabricated from aluminum alloy 2024-T4. A close-up view of this failed component is shown is Fig. 13(b), which clearly indicates surface pitting corrosion. This pitting contributed to the fracture that is illustrated in Fig. 13(c). From Fig. 13(c), it was determined that exfoliation corrosion was involved in the initiation of the failure process.

Fig. 13 Aluminum alloy 2024-T4 landing gear door actuator (a) that failed because of pitting corrosion. Arrow points to crack. (b) Close-up of crack in landing gear door actuator in (a). (c) SEM of the fracture surface of the door actuator. Source: Ref 5

Cargo Aircraft Brake Lining Carriers. Damage produced by pitting corrosion was observed in the brake lining

carrier structure on the C-5A cargo aircraft (Ref 18). This brake lining carrier serves as a heat sink (for energy dissipation) during aircraft landings. Figure 14(a) shows the entire part, and Fig. 14(b) provides a magnified view of the component. Pitting deterioration was responsible for inducing filiform corrosion in the brake lining carrier, which consisted of the purest available beryllium material--approximately 99% Be. After this case was detected, a chromate conversion coating that conforms to MIL-C-5541C was applied to this part to minimize the probability of similar corrosion problems in the future. However, a 1985, the corrosion-prone beryllium was replaced by a carbon (graphite) composite material in the manufacturing of the later-model C-5B cargo aircraft (Ref 18). The carbon-composite material was selected by the C-5H manufacturer because it provided reduced weight (the new brake lining carriers saved up to 180 kg, or 400 lb, per aircraft in comparison to the beryllium structures), lower cost, and longer life, with better corrosion resistance and durability than the beryllium component.

Fig. 14 Pitted beryllium brake lining carrier (a) from a C-5A transport plane. (b) Close-up showing pitting and filiform corrosion. Courtesy of the National Association of Corrosion Engineers

Tip Tank Latch Knob (Ref 16). Pitting corrosion and fatigue occurred on this AISI 4340 steel structure. Several failures were recorded for this component. Fractures occurred below the head of the latch knob at the curved surfaces (radii). The fracture surfaces are shown in Fig. 15 at magnifications of 2.5 and 5×. Pitting corrosion was determined to be the primary cause of the fatigue failures that occurred with this latch knob. A change of the 4340 steel material to 17-4PH stainless steel was recommended as the long-term solution to this problem. This is because 17-4PH steel has significantly greater corrosion resistance than 4340 steel. However, periodic inspections also were required for increased protection against pitting, crevice, and edge corrosion in a marine environment.

Fig. 15 Pitting corrosion and fatigue failure of a 4340 steel tip tank latch knob. (a) General view of the latch knob. (b) Cross section through the knob showing crack initiation sites. (c) Fracture surface of crack that initiated at site I. (d) Fracture surface of crack that initiated at site II. (e) Fracture surface at site I. (f) Pitting at a fatigue crack initiation site. Source: Ref 16

Wing Flap Hinge Bearings (Ref 16). These structures consisted of chromium-plated type 440C martensitic stainless steel. The inner diameter of the failed bearings had numerous corrosion pits. Fracture occurred intergranularly, either by SCC or by hydrogen embrittlement. An electroless nickel plating was used as a substitute for the chromium plating in order to alleviate this pitting corrosion problem.

In summary, pitting corrosion in aircraft systems is often caused by: • • • • •

Local breakdown in a protective film on an alloy. This is usually accelerated when the material comes into contact with chloride-containing solutions. Alloys susceptible to pitting because of localized impurities in the alloys. These impurities are either anodic or cathodic to the base metal Deposits of heavy metal (from water) on aircraft surfaces Localized damage (holidays) in applied protective coatings The accumulation of deposits, (dirt, dust, grease) on bare aluminum, stainless steel, or steel surfaces

Whenever pitting corrosion occurs, a review of its causes is necessary to determine if a change in material (or materials), design, or protective coatings will most effectively deter future pitting corrosion problems. Because pitting can cause perforation of aircraft parts (especially thin, sheetmetal structures) or fatigue failures, care must be taken to:

• • •

Neutralize pitting corrosion whenever it is detected Make repairs and provide local strengthening by using doublers (small patches of boron fiber reinforced plastic) Provide proper coating protection and regularly scheduled maintenance inspection

More information on pitting corrosion is available in the section "Pitting" of the article "Localized Corrosion" in this Volume. Crevice corrosion occurs on aircraft structures when a corrosive fluid, such as salt spray, enters crevices that are

located in individual parts or in between different components of a structural assembly. An anodic region usually develops at the bottom of the crevice, producing corrosive attack on the structure. Differences in the concentrations of dissolved salts or dissolved oxygen in the corrosive fluid will produce concentration cells and thus promote the development of anodic sites in structural airframe crevices (Ref 16). Typical cases of crevice corrosion have been observed in the following aircraft components (Ref 16). Magnesium Panel. Crevice corrosion perforated this structure, which was in a dismantled joint area from a transport

aircraft structure. Inadequate sealing was determined to be the cause of this corrosion problem. Magnesium Alloy Skin Joint. This example is similar to the previous one. The cause of this problem was attributed to

unsatisfactory sealing practices. Point of Contact Between a Magnesium Alloy Floor Panel and Aluminum Alloy Frames and Angle Plates. Both galvanic corrosion and crevice corrosion occurred in this case. Both of these forms of corrosion could have

been prevented or minimized by a better protective coating, a more effective fluid removal (drainage) system, and better utilization of sealants (Ref 16). Crevice corrosion also occurs in many alloy systems other than magnesium, including stainless steels and aluminum alloys. In summary, two methods are commonly used to prevent crevice corrosion. The first is proper design. This includes provisions for drainage, sealants that are applied to faying surfaces, or beads of sealant that cover crevices formed by mating surfaces. The second method consists of spraying aluminum airframes with water-displacing (penetrating) oils that seep into crevices and prevent water ingestion. The section "Crevice Corrosion" of the article "Localized Corrosion" in this Volume contains more information on this form of attack. Stress-Corrosion Cracking. The majority of the aluminum airframe structures that have been documented as failing

by SCC have been manufactured of alloys that contain aluminum, copper, zinc, and magnesium (Ref 19). These are the 2000- and 7000-series aluminum alloys. Stress-corrosion cracking failures have also been observed in high-strength steels, such as: • • •

A martensitic steel spindle sleeve on the main rotor hub assembly of a military helicopter Large (38-mm, 1.5-in., diam) high-strength steel bolts. When these parts failed, the head of the bolt completely separated from the shaft (Ref 19) A 300M steel landing gear drag strut on a military helicopter (Ref 20)

Stress-corrosion cracking is difficult to recognize in aluminum forgings because there is often no visual indication of surface corrosion products. The cracks can also be very long and deep (Ref 16). Stress-corrosion cracks usually occur in the end grain of the forged component at the parting plane of the forging. The cracking that occurs in airframe stresscorrosion failures is usually intergranular. Many airframe SCC failures have involved structures that were manufactured from aluminum alloys, especially 7079-T6 and 7075-T6. These include the following airframe components that were fabricated from aluminum and that have been observed to fail by SCC (Ref 16) (the specific aluminum alloy is provided in parenthesis): • •

A main landing gear locking cylinder (7079-T6) A main landing gear H-link structure (7079-T6). This damaged component is illustrated in Fig. 16. The

•

• • •

• • •

stress corrosion was induced by the precipitation of magnesium aluminide (Mg2Al3, which caused the grain boundaries in the aluminum forging to deteriorate anodically The front and rear spars of a vertical fin (7079-T6). As shown in Fig. 17, many of the cracks in this failure propagated from fastener holes. These spars has received corrosion-preventive surface treatment. However, some of this protection was inadvertently removed during the installation of bolts. Bare metal, therefore, was exposed to a high-humidity environment and sustained high tensile stresses that were produced by the installation of fasteners into these structure. This problem was remedied by the use of 7075-T73 aluminum forgings for the front and rear spars. The latter material provides greater resistance to SCC than 7079-T6 The bearing housing of a vertical stabilizer beam (7079-T6) A main landing gear bogie, which has the appearance of a beam or strut-type structure (7075-T6) The hydraulic cylinders that serve as actuators for a main landing gear door (7075-T6). Views of the fracture surface, including the appearance of intergranular fracture, are shown in Fig. 18. This problem could have been alleviated by the application of a better corrosion-protective treatment in order to minimize the degree of pitting corrosion that occurred during the storage of these cylinders. The use of the more SCC-resistant aluminum alloy 7075-T73 also would have helped to prevent this failure The fork and strut components of a nose landing gear (7075-T6) A fuselage frame structure, in which SCC occurred in between fastener holes (7075-T6) A nose landing gear strut (7075-T6)

Fig. 16 SCC in an aluminum alloy 7079-T6 main landing gear H-link. (a) Overall view of H-link. (b) Pitting and intergranular corrosion that initiated SCC. Source: Ref 16

Fig. 17 SCC in aluminum alloy 7079-T6 spars of a vertical fin. (a) Cracks in the mating surface of the rear spar. (b) Fracture surface of a statically broken front spar. Source: Ref 16

Fig. 18 SCC of an aluminum alloy 7075-T6 hydraulic cylinder. (a) Section through cylinder showing cracks on the inside surface (arrow). (b) Fracture surface showing three regions of cracking. (c) Appearance of intergranular fracture in region I. (d) Fatigue striations observed in region II. (e) Cross section through the inside surface showing corrosion pit. Source: Ref 16

Stress-corrosion cracking has also been observed in an aircraft aluminum-copper main landing gear forging and in four main landing gear retraction cylinders that were made of 4340 steel (Ref 16). Several views of this main landing gear forging failure are shown in Fig. 19. Deep corrosion pits were observed in the area of stress-corrosion crack initiation (Fig. 19c and d). This main landing gear forging was manufactured from an alloy that contains aluminum, copper, silicon, and manganese. This alloy is similar to aluminum alloy 2017. Aircraft structures that are fabricated from the 2017-T4 and 2017-T451 alloys can fail by SCC when sustained tensile stresses are exerted in the transverse direction relative to the grain structure.

Fig. 19 SCC in a 4340 main landing gear pivot pin. (a) Central portion of the broken pin. (b) One of two fracture surfaces on the piece shown in (a). (c) TEM of the fracture surface. (d) Fracture surface of another specimen that failed during a sustained load test

A representative example of SCC in an aircraft tubing structure is shown in Fig. 20. This part was fabricated from AM350 precipitation-hardened stainless steel. An intergranular type of cracking can be seen in the photomicrograph in Fig. 20(b). Fracture was intergranular (Fig. 20c).

Fig. 20 Intergranular SCC in an aircraft pipe structure fabricated from AM-350 precipitation-hardened stainless steel. (a) Overall view of pipe. (b) Micrograph showing intergranular nature of cracking. (c) SEM of crack shown in (b). 400×

Good examples of SCC in airframe fasteners have been presented in the literature. Most of these fractured parts were H11 steel bolts manufactured according to AMS 6487. The usual corrosion-protective treatment for these bolts was either: • •

A coating of fluoborate cadmium that complies with NAS 672 (for fasteners that have minimum tensile strengths of 1517 MPa, or 220 ksi) A vacuum-deposited cadmium finish, as specified by MIL-C-8837 (for fasteners that have minimum tensile strengths of 1793 MPa, or 260 ksi)

As shown in Fig. 21, the initiation of SCC in a failed aircraft alloy steel bolt occurs at a surface corrosion pit. Propagation of this corrosion-induced crack gradually continues until the bolt fractures, because the stress-corrosion crack significantly reduces the amount of mechanical stress the bolt can sustain.

Fig. 21 SCC of an alloy steel aircraft bolt. (a) Fracture surface showing origin at a corrosion pit on the bolt surface. (b) SEM showing brittle, intergranular SCC near the origin. (c) SEM of the region of fast fracture showing ductile nature of fracture in this area. Source: Ref 21

Two SCC case histories involving H-11 alloy steel fasteners have resulted in beneficial airframe design improvements. These incidents, and the design solutions, are described below. In the first case history, cadmium-plated H-11 fasteners became pitted and eventually failed by SCC on a widebody transport (cargo) aircraft. These fasteners were installed in the engine-to-pylon attachment structures. The cadmiumplated steel fasteners had been installed into a titanium airframe component. This coupling of dissimilar airframe materials produced galvanic corrosion, severe pitting, and the eventual failure of these fasteners by SCC (Ref 21). As a result of this incident, all of the fasteners that were considered to be critical to the engine-to-pylon attachments were replaced by cold-worked and aged A286 stainless steel and cold-worked and aged Inconel alloy 718. Generally, neither cadmium- nor silver-plated parts should be used in contact with titanium under compressive loading. This condition causes the cadmium or silver to migrate into the intergranular structure of the titanium and induce cracking. This migration of cadmium or silver into the intergranular structure of titanium is commonly referred to as poisoning. Although suitable for installation into steel structures, cadmium- and silver-plated parts are to be avoided in titanium assemblies that will be subjected to high compressive loads, heat, or both. In the second case history, SCC failures of 1517 MPa (220 ksi) H-11 fasteners occurred in the wing attachment area on several general-purpose aircraft. Alloy steel fasteners installed in this wing attachment structure also fractured. These alloy steel fasteners had maximum tensile strengths of 1103 and 1240 MPa (160 and 180 ksi). The design solution to this problem was to have all of the noncorrosion-resistant steel fasteners, including the alloy steel ones, replaced by 1240 MPa (180 ksi) and 1517 MPa (220 ksi) Inconel 718 fasteners (Ref 21). There are a few important guidelines that can be followed in order to minimize the probability of fastener failure due to SCC or other types of corrosive deterioration (Ref 22). First, fasteners should not produce adverse effects on the structures that they are joining. These fasteners also should not be installed into certain materials that would make these fasteners susceptible to corrosion. Second, fasteners should be selected that are slightly cathodic to the parts they will join. This is especially advisable in aggressive environmental conditions, such as in tropical-marine areas. Materials that are susceptible to hydrogen embrittlement, such as high-strength steels, generally should not be used for fastening applications in aircraft structures. Finally, critical joints should not be connected with fastening materials that are susceptible to corrosive deterioration in the environment in which the actual aircraft system will be operating. Aircraft service safety and airworthiness can be significantly improved by selecting airframe materials that are resistant to SCC and brittle fracture (Ref 23). Aircraft design engineers can minimize airframe failures due to SCC by ensuring that: •

Materials that are selected for aircraft applications have low rates of SCC (Ref 23). The T6-tempered 2000-series aluminum alloys (except for 2024 aluminum) should be avoided. The 7000-series aluminum

• •

• • •

alloys should be in the T-73 or T-76 tempers. Similarly, stress-corrosion-resistant tempers of steel alloys should be used Cracks in selected airframe materials are detected prior to a significant reduction in the static strength of the airframe part (Ref 23) Tensile and residual stresses, which may become significant during fabrication and assembly, are minimized. This can be done by selecting stress-relieved tempers for specific alloys (Ref 24) and by minimizing assembly stresses by shimming, avoiding interference fits, and shot peening Compressive stresses are properly applied on machined surfaces by shot peening Organic coatings are used, whenever possible (Ref 24) Inorganic corrosion-protective methods are used, such as cladding on aluminum or electro-plating on steel, whenever possible (Ref 24)

More information on the mechanism of SCC is available in the section "Stress-Corrosion Cracking" in the article "Environmentally Induced Cracking" in this Volume. Intergranular Corrosion. A severe case of intergranular corrosion was described in the beginning of this section (see

Fig. 9 and 10). This discussion will provide some additional corrosion problems and solutions involving intergranular corrosion and the more damaging form of intergranular corrosion--exfoliation. The purpose of this discussion is to provide a more detailed explanation of how these forms of corrosion are interrelated. In intergranular corrosion and in exfoliation corrosion, the grain boundaries of the corroded metal become anodic. The bulk material in between the grain boundaries is not affected and therefore is cathodic. Corrosion products and, occasionally, cracking are produced on the surface of materials that corrode intergranularly. This form of corrosion sometimes penetrates underneath the metal surface, making it difficult to detect the damage without the aid of a microscope (Ref 16). Intergranular corrosion can occur either alone, in conjunction with pitting corrosion, or with exfoliation corrosion. When intergranular corrosion occurs along grain boundaries that are parallel to the plane of the material, such as in a flat plate, the corrosion proceeds in the short-transverse direction along these elongated grain boundaries. The combination of the corrosion and the entrapped corrosion products causes the plate to delaminate. When intergranular corrosion occurs, a network of fine surface cracks can develop. This fine cracking pattern has been observed in an aircraft hydraulic valve made of a forged aluminum alloy that deteriorated by intergranular corrosion (Ref 16). An example of exfoliation corrosion of an aluminum alloy stabilizer bracket from a light aircraft is shown in Fig. 22. This deterioration started as intergranular corrosion but gradually became more severe and propagated as exfoliation corrosion. The horizontal surface of the stabilizer bracket had been exposed to atmospheric moisture and contaminants, which collected at the interface between the bracket and a nylon bushing. No corrosion was found on bracket surfaces that were protected by a chemical conversion coating. This problem could have been prevented by effective sealing of the bracketto-bushing interface, along with regular inspections (Ref 16). Some typical, documented examples of intergranular corrosion and exfoliation corrosion have also occurred on the following structures.

Fig. 22 Exfoliation corrosion of an aluminum alloy stabilizer bracket. (a) Heavy surface corrosion on the stabilizer bracket. (b) Cross section through the bracket showing corroded surface grains and corrosion at grain boundaries of elongated grains. Source: Ref 16

Wing Box Lower Panel of a Fighter Aircraft (Ref 16). This panel was made of aluminum alloy 7075-T6. Most

of the corrosion occurred around fastener holes. Extensive intergranular cracking was observed. The report on this case also indicated that pitting occurred in the bores and countersinks of the fastener holes. Filiform corrosion was also detected in the fastener hole areas. This problem was solved by applying a conversion coating to the fastener hole bores and countersinks. Next, the fasteners were wet assembled using a strontium chromate primer and an acrylic topcoat. Another part of the solution in this case was the development of a new aluminum alloy, 7475-T761 (Ref 16). This material has a high level of resistance to exfoliation corrosion and therefore has been considered by aircraft designers as a favorable replacement for the 7075-T6 alloy. Main Rotor Blade of a Helicopter (Fig. 23). Intergranular and exfoliation corrosion predominated in the area

between the leading edge spar and the surface skin of the blade. Extensive corrosion accumulated at the leading edge, causing the skin of the blade to lift off of the spar. The leading edge spar was manufactured from aluminum alloy 2024. During a metallurgical examination, copper aluminide was found in the grain boundaries of this material. It was therefore determined that the 2024 aluminum structure was improperly heat treated. More information on this form of attack is available in the section "Intergranular Corrosion" of the article "Metallurgically Influenced Corrosion" in this Volume.

Fig. 23 Corrosion of an aluminum alloy 2024 helicopter rotor blade. (a) Leading edge at the blade tip showing three areas of severe corrosion. (b) Corrosion in the aluminum alloy skin at area. (c) Rupture of the surface skin at area 3 due to buildup of corrosion products in the underlying spar. (d) and (e) Intergranular corrosion in the spar. (f) Exfoliation in the surface skin. Source: Ref 16

Filiform Corrosion. One of the more insidious types of electrochemical degradation in aircraft structures is filiform

corrosion. This form of corrosion usually occurs beneath the cladding of clad aluminum alloys, under organic coatings applied to airframe surfaces, and at fastener holes (Ref 1, 16). Filiform corrosion has been described as a type of anodic undermining, with reference to its occurrence on metallic structures that are protected by organic primers and topcoats (Ref 1). Filiform corrosion usually begins as a shallow corrosion pit, which continues its attack on the base metal by spreading laterally along the surface of the structure. This appears as an irregular pattern of thin filaments of corrosion products on the affected metallic material. A schematic illustration of filiform corrosion is presented in Fig. 24.

Fig. 24 Schematic of the development of filiform corrosion on an aluminum alloy. Source: Ref 25

One theory that is applicable to aircraft systems proposes the following series of events that promote filiform corrosion on coated metallic airframe parts (Ref 25): •

•

• • • • •

In-flight, minor relative motion (vibration) produces very thin discontinuities, which also have been called hair cracks, in the airframe paint coating. Salts that primarily consist of sodium and calcium are present in the areas where these small cracks originate These salts (inside the hair cracks) absorb moisture from the atmosphere by osmosis. High humidity accelerates the propagation of filiform corrosion (Ref 16). Chlorine also can activate this type of corrosion (Ref 25) Moisture absorption at the airframe structure produces an aqueous electrolyte, which causes anodic and cathodic reactions to take place The anodic reaction produces pitting that does not penetrate very deeply into the aluminum airframe material Aluminum hydroxide (Al(OH)3) is precipitated because of the formation of hydroxyl (OH-) groups The head and body of the corrosive filament develop and separate from the areas where the original cracks were formed The corrosive filamentary head continues to deposit corrosion products because water is replenished in the head by osmosis

Because filiform corrosion propagates on structural areas that are either clad or coated with organic paint, this type of material deterioration can spread extensively before it is detected by aircraft maintenance personnel. In typical cases, filiform corrosion develops around fastener holes on airframe sheet structures. Paint blistering around the rivet holes is a characteristic feature of this type of corrosion. Some of the materials that are known to have been affected by filiform corrosion are magnesium, aluminum, steel, and chromium-plated nickel. Documented case histories of filiform corrosion of various components include the following. Fuselage Skins. On a Boeing 707 aircraft operated by a major commercial airline, filiform corrosion occurred on

fuselage skins along rows of fasteners (Ref 26). Paint blistering is produced on airframe aluminum sheetmetal structures when this form of attack occurs (Ref 26). Areas Around Steel Fasteners. Filiform corrosion was observed in the areas around steel fasteners, which originally

were affected by intergranular corrosion. The corroding fasteners were installed in the lower wing skins of the Boeing 707. Figure 25(a) illustrates this deterioration prior to paint removal, while Fig. 25(b) shows the filiform corrosion damage after the paint coating was stripped from the airframe surface.

Fig. 25 Filiform corrosion of an aluminum aircraft skin around steel fasteners. (a) Before paint removal showing paint cracking and blistering. (b) After paint removal. Courtesy of Qantas Airways and the Society for the Advancement of Material and Process Engineering

Lower Wing Skin. When a Boeing 747 aircraft was first placed into service, filiform corrosion was detected on the lower wing skins of one of these aircraft (Ref 26). This corrosion developed from intergranular corrosion around titanium fasteners that were inserted into the airframe structure. Pylon Tank. Filiform corrosion caused the perforation of one area of an aluminum alloy 6061-T6 pylon tank (Fig. 26).

Pitting and intergranular corrosion were also detected on the pylon tank during the investigation of this problem (Fig. 26c). The aircraft that operated with this tank had been flying in the hot and humid Mediterranean environment. The proper application and maintenance of epoxy or polyurethane paint finishes would have minimized the amount of deterioration on this structure.

Fig. 26 Filiform corrosion of a fighter aircraft pylon tank. (a) Overall view of the tank showing uniform corrosion (open arrows) and penetration (solid arrows). (b) Indications of filiform corrosion. (c) Pitting and intergranular corrosion. Source: Ref 16

Other Components. A structural engineer for a commercial airline reported that filiform corrosion was detected on

horizontal and vertical stabilizers, trailing edge flaps, and an aluminum metal sprayed surface on fiberglass wing-body fairing panels (Ref 26). For the areas on the Boeing 707 that were affected by filiform corrosion, a corrosion-inhibiting compound called LPS-3 was found to be a successful treatment for the prevention and control of corrosive (filiform) growth. This compound is a greasy hydrocarbon-base preservative that minimizes oxidation by displacing water away from the head of the corrosive filament (Ref 26). The following sequential treatment has been determined to be effective in alleviating filiform corrosion of airframes: • • • •

Glass bead blasting to remove observable corrosion products Chemical conversion coatings Boeing Material Specification (BMS) 10-79 strontium chromate epoxy primer BMS 10-60II flexible polyurethane enamel

The commercial airline engineer that reported the above airframe treatment for filiform corrosion stated that this treatment provides 5 years of service life with only minor maintenance (Ref 26). A polysulfide-calcium-strontium rubber primer covered with polyurethane paint, which has been used for corrosion protection on military aircraft, has also been found to be effective in deterring filiform corrosion (Ref 26). In one case, for example, filiform corrosion was observed on some aluminum alloy 2024 aircraft ailerons and flaps (Ref 16). This largely superficial corrosion initiated at the surfaces around fastener holes. The recommended solution was to replace the nitroalkyd resin paint coating with a finishing system that consisted of an epoxy primer and a polyurethane topcoating. The section "Filiform Corrosion" of the article "Localized Corrosion" in this Volume provides more information on filiform corrosion. Fretting corrosion occurs on airframe structural surfaces that move against each other in an environment that is

conducive to corrosion. Often, the relative movement is barely discernible, but results in significant deterioration. This type of corrosion is very significant in operational aircraft systems because airframes are subject to high vibration levels and because airframe components must also endure various types of mechanical stress other than vibration, resulting in small relative movement between parts (Ref 27). Some of the airframe parts that have been documented as being susceptible to fretting corrosion are (Ref 27): • • • • • • • • •

Hinge point bearings on ailerons, elevators, rudders, and flap structures Control pulley bearings Universal joint bearings Propeller and propeller control bearings, housings, and shafts Instrument bearings Spline connections Pin joints Clamps Riveted lap joints

Fretting actually appears as a combination of wear and corrosive deterioration. Fretting corrosion occurs in the presence of oxygen, which oxidizes the small wear particles that form in structural contact areas. These small particles develop because of the oscillatory motion that exists in contacting metallic materials. The hard oxides then become trapped in the point of contact and further promote intense wear on the two rubbing surfaces (Ref 28). Concerning the development of aluminum helicopter rotor blades, it has been noted that fretting produces a significant reduction in the fatigue strength of the root retention area of the blade (Ref 28). With little or no fretting, the maximum allowable alternating stress, at 107 cycles, is about 21 MPa (3000 psi). Fretting causes the maximum allowable stress to

drop to between 8 and 10 MPa (1200 and 1500 psi) in joints where there will be high stress concentrations (Ref 28). Representative descriptions of documented cases of fretting corrosion in aircraft structures include the following. Flying Control Hinge Pins. Fretting deterioration on these cylindrical structures appeared as broad bands of accumulated corrosion products. Severe (deep) pitting corrosion was also observed in this case (Ref 16). This damage possibly occured as a result of both wear (due to large and small amplitude displacements) and fretting corrosion (resulting from oxidation). Propeller Shaft Ball Bearings. Fretting corrosion was observed in the inner and outer races of these propeller aircraft

bearings. The bearings were manufactured from 51100 bearing steel, and they were heat treated to produced a fine tempered martensite (hardness 63 HRC). The locations on the outer bearing race at which fretting corrosion occurred appeared to be slightly brighter than the unaffected regions of the structure. The fretted areas were also damage by pitting corrosion (Ref 16). Lower Boom of the Wing Forward Spar. This structure was fabricated from an aluminum alloy 2014 forging (Ref

23). A fatigue crack started at the edge of a bolt hole in this bomber aircraft component. This crack extended to only 0.5% of the total cross-sectional area of this part. The initiation of this fatigue crack, which eventually caused the in-flight failure of the aircraft, was attributed to fretting (Ref 23). Engine Fuel Tanks. Failures in these components occurred because of fretting that caused poorly soldered lap joints to

separate (Ref 29). Mechanically locked joints were the recommended solution for alleviation of this problem. The redesign of these fuel tanks also included use of a thicker terne (lead-tin) plate on tank inner surfaces (Ref 29). One of the most widely used techniques for averting fretting corrosion is to apply a continuous film of lubrication between the contacting surfaces (Ref 28). A second method consists of increasing the load (pressure) at the point of surface contact. This will reduce the amount of slip (relative, oscillatory motion) at the contact point. Yet another method involves selecting materials with a high resistance to fretting corrosion. Metallic combinations that have high, medium, and low resistances to fretting corrosion are presented in Table 4. There results are from a study on fretting wear and corrosion that was conducted by the Massachusetts Institute of Technology. Fretting tests were performed under dry conditions both in air and in a vacuum. In this investigation, low fretting corrosion resistance was attributed to the high hardness of metal oxides that were formed when certain pairs of metals were moved against each other. For example, the low resistance to fretting corrosion of the aluminum-to-aluminum combination was probably a result of the high hardness of aluminum oxide, which is formed during fretting and has a Moh's hardness of 9 (the maximum Moh's hardness value is 10) (Ref 28). Preloading assembled joints and the use of sealants can reduce this problem in aluminum-to-aluminum joints. Anodizing of mating parts is also effective because a continuous, hard aluminum oxide is produced on the mating surfaces.

Table 4 Resistance to fretting corrosion of various material couples under dry conditions Couple Steel on steel Nickel on steel Aluminum on steel Antimony plate on steel Tin on steel Aluminum on aluminum Zinc-plated steel on aluminum Iron-plated steel on aluminum Cadmium on steel Zinc on steel Copper alloys on steel Zinc on aluminum Copper plate on aluminum Nickel plate on aluminum Iron plate on aluminum Silver plate on aluminum Lead on steel

Resistance to fretting Low Low Low Low Low Low Low Low Medium Medium Medium Medium Medium Medium Medium Medium High

Silver plate on steel Silver plate on aluminum plate Steel with conversion coating on steel

High High High

Source: Ref 28 Corrosion Fatigue. Aircraft structures are subjected to in-flight mechanical stresses as well as to corrosive environments. The mechanical stresses are cyclic in nature; that is, periods of high stress, or loading, are followed by relatively lower stresses, or unloading (Ref 1). This cyclic stress results in fatigue of airframe structures.

When airframe structures are adversely affected by the combined effects of corrosion and fatigue, th accelerated deterioration that occurs is referred to as corrosion fatigue. Airframe structures fracture in a shorter time because of fatigue that occurs in combination with corrosion (Ref 16). Further, corrosion reduces the fatigue (endurance) limit of airframe components. Corrosion fatigue is most severe in those airframe parts that are most vulnerable to both fatigue and corrosion, such as fastener holes (Ref 30). A typical example of the corrosion fatigue failure of a military helicopter rotor assembly is discussed in Ref 16, and more information on this failure is also provided in Example 16 in the section "Appendix: Case Histories and Failures" in this article. This rotor assembly consisted of several components--a horizontal hinge pin, a nut, and a locking washer--that were manufactured from 4340 and 4130 steels (Fig. 27a to c). Fracture was observed on the horizontal hinge pi. A flat beach mark on the hinge pin indicated the location of crack initiation in the structure. Figures 27(d) and 27(e) show the respective fracture surface and beach mark. One side of the threaded area of the hinge pin had a dent (Fig. 27f). Corrosion pits were discovered close to the fracture initiation site. Cracks that ran parallel to the primary fracture surface appeared to be emanating from small pits in the hinge pin (Fig. 27g). An embedded thread in the hinge pin was formed by contact of the 4130 low-alloy steel nut with the 4340 pin (Fig. 27h).

Fig. 27 Corrosion fatigue of a 4340 steel helicopter rotor assembly. (a) Horizontal hinge pin. (b) Nut. (c) Locking washers. (d) and (e) Views of the hinge pin fracture surface. A beach mark is visible in (e). (f) Dent on one of the threaded surfaces of the horizontal hinge pin. (g) Corrosion pits and cracks that emanated from pits in the hinge pin. (h) Embedded thread from the nut in the hinge pin threads. Source: Ref 16

The failed hinge pin had been cadmium plated. Pitting corrosion in the hinge pin promoted the deterioration of the cadmium plating and the exposure of the 4340 steel base material. This corrosion fatigue failure led to a complete inspection and overhaul of all components in this fleet of helicopters. The inspected (undamaged) hinge pins were stripped of the cadmium plating, shot peened, and replated to a minimum thickness of 12.5 m (0.5 mil) (Ref 16). Other typical corrosion fatigue failures have been documented, including the following(Ref 16): • •

A main landing gear wheel, made of magnesium alloy AZ91C A main landing gear wheel manufactured from magnesium alloy QE22A. Figure 28 illustrates the deterioration that occured in this case. The fracture of this part started on the outer surface and was probably caused by a defective corrosion-protective coating system. Regular inspection and repair of the

•

paint system that was applied to this structure could have prevented this damage Rivets fabricated from aluminum alloy 5052. These rivets were used for the installation of a helicopter filter housing. The damaged rivets are shown in Fig. 29. Figure 29 also presents views of the fracture surface and the grain structure of the rivet material

Fig. 28 Corrosion fatigue failure of a magnesium alloy QE-22A main landing gear wheel. (a) General view of wheel. Crack is shown by the dashed line. (b) Fracture surface. Origin is indicated by the arrow. Source: Ref 16

Fig. 29 Corrosion fatigue of aluminum alloy 5052 rivets. (a) Rivets removed from the structures for examination. (b) Fracture surface of a rivet showing three zones of fracture 20×. (c) Intergranular fracture observed in Zone A. 1000×. (d) Mud crack pattern in zone A. 500×. (e) Fracture surface in zone B. 2000×. (f) Grain structure of the rivets. The deformation is caused by tearing in an overload region. 500×. Source: Ref 16

In summary, special corrosion prevention and control treatments must be provided to aircraft that operate in environments conducive to corrosion fatigue. Fatigue resistance can be enhanced in critical areas by designing parts with generous corner radii and by shot peening. Adequate corrosion protection must be used and then followed up by continuous inservice inspections. Pitting must be removed from critical parts because corrosion fatigue often nucleates in corrosion pits. More information on corrosion fatigue is available in the section "Corrosion Fatigue" of the article "Mechanically Assisted Degradation" in this Volume. Microbiological Corrosion. This type of airframe corrosion is also called biological corrosion (Ref 31).

Microbiological corrosion was discovered in the mid-1950s in aircraft integral fuel (wing) tanks (Ref 13). The microbes used kerosene fuels and salt water as growth media (Ref 16, 31). Certain types of synthetic rubber have also been found to promote microbial growth (Ref 14, 32). This support of microbial growth by synthetic rubber was discovered in cases of microbiological corrosion on the military C-130 transport aircraft (Ref 32). Most of the observed airframe corrosion has been produced by a group of microbes called fungi. The greatest number of cases of aircraft integral fuel tank corrosion have been attributed to the fungus Cladosporium Resinae (Ref 16). This microorganism was recently identified by some Argentinian investigators in an integral fuel tank corrosion incident (Ref 33). The fuel tank material was aluminum alloy 2024. The fungus was most active (and, therefore, most corrosive) at the fuel/water interface inside of the 2024 aluminum fuel tank.

Microbiological corrosion can also be caused by bacteria and yeasts (Ref 16). The bacterium Pseudomonas aeruginosa has been observed in aircraft fuel tank corrosion cases, but much less frequently than Cladosporium resinae. Aircraft that operate in tropical environments appear to be most vulnerable to microbiological corrosion (Ref 16). Pitting corrosion is often present in microbiological corrosion cases, with pit depths ranging from 1.5 to 3.2 mm (0.06 to 0.125 in.) (Ref 13). One report on microbial corrosion, involving commercial transportation aircraft, indicated that the structural fastener areas had the highest level of deterioration, based on the severity and frequency of corrosive attack (Ref 13). Microbiological corrosion problems have been observed both in military and commercial aircraft (Ref 34). One case that involved a DC-9 wing fuel tank led to the conclusion that microbiological corrosion can develop on airframes protected by the polyurethane coating that conforms to military specification MIL-C-27725 (Ref 34). One researcher confirmed that microorganisms can penetrate polyurethane coatings in a report on C-130 integral wing tank corrosion problems (Ref 32). The remedy that was selected for prevention and control of C-130 microbiological corrosion was to remove the synthetic rubber lining and to substitute it with a polyurethane coating that contains a biocidal green dye (Ref 32). A commercial airline system that started a fuel quality improvement program in the early 1970s has reported that microbiological corrosion problems in its fleet of aircraft have been almost completely eliminated. Therefore, it can be concluded that two actions are necessary for the prevention and control of airframe microbiological corrosion problems. These are the selection of the proper structural materials, including biocidal protective coatings, and the implementation of stringent controls on the quality of fuel that is used in aircraft systems. More information on attack by microorganisms is available in the section "General Biological Corrosion" of the article "General Corrosion" and in the section "Biological Corrosion" of the article "Localized Corrosion" in this Volume. Corrosion in Aircraft Bilges. Corrosion often occurs in aircraft bilge areas, which are sinks (or sumps) into which

waste fluids and solids collect. These materials--especially oil, water, and dissimilar metallic drill chips--can set up electrochemical (galvanic) cells in the bilge (Ref 35). Because these corrosion reactions occur in areas that normally are not accessible to aircraft maintenance personnel, the potential for extensive airframe deterioration due to corrosion is significant (Ref 36). One of the best ways to control bilge area corrosion in aircraft systems is to provide access through stairways or rampways so that mechanics can conveniently inspect and maintain these areas (Ref 36). Additional measures that can be taken, especially during the aircraft design phase, are to (Ref 36): • •

•

Consider using aluminum alloys that have greater resistance to general corrosion and SCC, such as 7075-T73 and 7475-T761 Use an external organic primer and topcoat finishing system. This system can, for example, consist of the MIL-P-23377 epoxy-polyamide primer (one coat) and the MIL-C-83286 polyurethane topcoat (two coats) Seal all faying surfaces of the airframe structure. A sealant meeting MIL-S-81733, which contains corrosion-inhibitive chromates, is often used

Hydrogen-induced failures of aircraft equipment result from a chemical and mechanical interaction with hydrogen.

The hydrogen that enters metallic airframe structures causes mechanically induced deterioration by applying very high pressures within small cavities in the affected airframe structure. Another term that is used for hydrogen-induced fracture is hydrogen embrittlement. Hydrogen embrittlement can cause dangerous and sometimes catastrophic failures in some steel parts that are used in aircraft. Two possible processing sources of hydrogen embrittlement are pickling solutions, which are used for scale and rust removal, and electroplating solutions (Ref 16). Hydrogen embrittlement may also result in areas where steel acts as the cathode during galvanic (dissimilar metal) corrosion. Hydrogen-induced cracking is one of the more severe forms of deterioration because it can result in delayed fracture under sustained stress. Hydrogen embrittlement has been observed (and has caused the failure of) some of the highstrength steel alloys used in aircraft landing gears (Fig. 30). Representative cases involving this type of corrosion were discussed in the introduction to this section. Additional examples of hydrogen-induced airframe corrosion are discussed below for various aircraft structures.

Fig. 30 Hydrogen embrittlement failure of a high-strength steel aircraft landing gear component

Landing Gear. Several C-141 landing gear cylinders failed because of hydrogen embrittlement (Ref 37). Hydrogen entered these 4340 steel cylinders during electroplating operations. Subsequently, 300M steel cylinders, coated on the interior surfaces with a hard chromium plating, have been used as a result of a redesign and retrofitting of the C-141 landing gear cylinders (Ref 37). Propeller Retaining Bolt (Fig. 31). The composition of this steel part--which failed by hydrogen embrittlement--

was Fe-0.5C-2.5Ni-0.75Cr. This steel bolt had been heat treated to a strength of 1379 MPa (200 ksi), and a bright electroplated cadmium coating had been applied.

Fig. 31 Hydrogen embrittlement of a high-strength steel propeller retaining bolt. (a) Overall view of bolt. (b) Cross section on the bolt. 215×. Courtesy of Aeronautical Research Laboratories, Australia

Nose Gear Strut. Hydrogen embrittlement was the primary cause of the failure of this steel aircraft component (Ref 38). The strut separated during the taxiing of a small bomber. The severely corrosive Southeast Asian environment probably abetted the ultimate failure of this structure. Flap Control Return Spring (Ref 16). The internal crack in this carbon steel part indicated that the failure resulted

from hydrogen embrittlement. This flap control return spring was cadmium plated, but there was no evidence that the part had been baked to remove hydrogen. Precautions were subsequently taken to ensure that proper baking is performed. Main Landing Gear Pivot Pins (Ref 16). The intergranular fracture of these 4340 steel components was caused by

hydrogen penetration into the metallic structures. Sustained loads were a major contributing factor in this failure. The pivot pins were chromium plated, and it was determined that the hydrogen came from the plating process. Therefore, the baking procedure, which is performed after electroplating, was modified to enhance the degree of hydrogen removal. Main Landing Gear Drag Link Bolt (Ref 16). This 4340 steel part had been cadmium plated. Fracture occurred in

the threaded region of the bolt. Hydrogen tolerance for steels is inversely proportional to their strength levels. Hydrogen embrittlement can be avoided by using lower-strength steels, by limiting immersion times in acid or plating baths, and by in-process (intermediate) baking. Plated parts are usually baked at 190 °C (375 °F) to provide hydrogen embrittlement relief. A safer baking duration for high-strength steels (>1100 MPa, or 160 ksi) is 23 h. Longer baking times may be required for parts that have large cross sections, bright cadmium treatments, or very high strength. More information on hydrogen-induced failures is available in the section "Hydrogen Damage" of the article "Environmentally Induced Cracking" in this Volume. Corrosion of Bonded Airframe Structures. Adhesive-bonded airframe structures can deteriorate by delamination

and electrochemical corrosion of one or both of the metallic structural face sheets that eventually separate from the bondline. The long-term environment susceptibility of bonded structures to this type of corrosion is known to be primarily caused by (Ref 39):

•

•

Water or moisture intrusion into the adhesive bondline. Perforated honeycomb core sections of these bonded parts promote the collection of moisture (Ref 40). This leads to deterioration of the honeycomb core due to freezing-thawing cycles, corrosion, or both (Ref 41) Deficient surface preparation of face sheet and core materials that are used in the complete adhesivebonded part

Improper surface preparation was the documented cause of the failure of a C-141A wing trailing edge upper surface panel. This was caused by a separation of the outer face sheet because of poor adhesion to a corner section of the wing upper surface. Corrosion of the aluminum face sheet started after the face sheet separated from the bonded structure. The disbonded aluminum face sheet had been prepared by using the sodium dichromate-sulfuric acid etch process, which is also called the Forest Products Laboratory (FPL) etch. The failure of the C-141A outer wing panel was attributed to inadequate process control while the FPL etching process was being accomplished (Ref 39). Another reported adhesive bonding failure occurred on the C-141A petal door inner skin assembly. This delamination and corrosion also occurred because of poor surface preparation and moisture penetration between the bonded structural panels. Extensive corrosion was noted near the hinge fitting of this skin assembly (Ref 39). Delamination and corrosion of adhesive-bonded airframe components have also been reported on the C-141 main landing gear doors (Ref 41) and on the more highly stressed aircraft control surfaces, such as spoilers, on one fleet of commercial airliners (Ref 42). These latter parts consisted of alclad aluminum alloy 7075 bonded to aluminum honeycomb with an adhesive that cured at 120 °C (250 °F). A significant contribution to the engineering effort to mitigate these delamination and corrosion problems has been the development of the phosphoric acid anodizing process (Ref 39, 40, 41, 42). This solution to adhesive structural disbonding resulted from 10 years of research at the Boeing and Douglas aircraft companies. Phosphoric acid (H3PO4) anodizing enhances the durability of adhesive-bonded structures by improving the resistance of the bondline too hydration and by reducing the extent of corrosion of exposed alclad aluminum surfaces (Ref 42). In addition to H3PO4 anodizing, the following techniques have helped to reduce delamination and corrosion of bonded airframe structures: • • •

Curing the film adhesive at 175 °C (350 °F) instead of at 120 °C (250 °F) (Ref 42) Coating the surfaces with a corrosion-inhibiting primer that cures at 120 °C (250 °F) (Ref 42) Better-quality adhesive bonding through the use of adhesives that have greater long-term resistance to plastic deformation. This increased toughness is obtained by the crosslinking of these two-part 100% solids paste adhesives during room-temperature curing (Ref 40)

Because corrosion of alclad aluminum spreads in the plane of the sheet, corrosion of alclad airframe parts results in the delamination of bonded honeycomb panels. Therefore, alclad sheet is often avoided for honeycomb facesheets. Another effective corrosion control approach is to apply only nonperforated honeycomb core. This prevents the moisture penetration that can occur in cellular honeycomb parts. Sealing is performed on the edges of honeycomb panels to protect the core from moisture ingestion. Erosion-corrosion is corrosion that is accelerated by the impingement of moving fluids, which may contain solid

particles, onto the surfaces of materials. Erosion-corrosion in aircraft is most severe when the fluid impinges upon the aircraft surface at a high velocity and when the moving fluid contains abrasive solid matter, especially sand. High temperatures also accelerate the deterioration process when erosion-corrosion occurs. One reported case in which erosion-corrosion was recorded in aircraft structures involved damage to the undercarriage parts of a transport aircraft. This large air vehicle had been operating on a desert landing strip. The erosion-corrosion damage on this structure, therefore, was accelerated by extremely high temperatures and impact by sand (Ref 16).

Prevention and Control of Airframe Corrosion The cost escalation attributed to corrosion problems in the United States has been significant--from $5.5 billion in 1947 to $167 billion in 1985 (Table 5). Therefore, it is important that the manufacturers and users of aircraft systems implement an effective corrosion prevention and control program.

Table 5 Cost escalation for corrosion in the U.S. from 1947 to 1985 Year 1947 1965 1967 1975 1982 1985

Cost, billions of dollars 5.5 >6 >10 70 126 167

Source: Ref 43

There are many techniques that are available for preventing and controlling airframe corrosion. Some of these corrosion control methods have been used for many years, such as the application of greases to bearings in aircraft control mechanisms, wheels, rudder posts, and other components (Ref 44). Other relatively new approaches have been utilized, including the application of advanced composite materials to secondary airframe structures (Ref 45). In cases in which carbon composite materials have been applied, the proper corrosion prevention and control methods had to be developed and practiced for preventing galvanic corrosion between the carbon composites and other less noble materials, especially aluminum. In addition to noncritical (secondary) structures, airframe manufacturers have also applied reinforced plastic and carbon composites to primary structures, such as the fuselage, wings, and empennage, which have a greater degree of safety-offlight importance than secondary structures. On the C-130, for example, the wing contains about 80% advanced composite materials (Ref 45). Advanced composites are also used in 71% of the structure of one business aircraft (Ref 45). Advanced composite materials are being more widely used in aircraft systems, primarily because of their strength at increasing aircraft velocities (Ref 46). Corrosion-preventive technology must be properly and regularly applied to airframe components to ensure that the airworthiness and flight safety of operational aircraft are not jeopardized. The corrosion cases that have been presented in this section demonstrate that the objectives of airframe corrosion prevention and control can be attained by implementing new and improved engineering design solutions. The essential elements of an airframe corrosion prevention and control program are proper material selection, an adequate finish specification, a thorough plan for effective maintenance, inspection, and repair (Ref 47). During the material selection phase, considerable trouble can be avoided if aluminum alloys and tempers are selected based on SCC resistance, exfoliation corrosion resistance, fracture toughness, and mechanical strength properties. Tempered steel having a yield strength below 1379 MPa (200 ksi) should be selected whenever possible in order to avoid hydrogen embrittlement and SCC problems. The corrosion prevention and control specification must ensure that these materials receive satisfactory hydrogen embrittlement relief during processing. Precipitation-hardening steels that have been heat treated for maximum resistance to SCC should be used. The task of preparing and adequate corrosion prevention and control plan is facilitated by feed-back from aircraft operators and maintenance technicians. The engineers and other specialists who provide feedback to corrosion engineers and designers should be thoroughly familiar with the operational profile of the aircraft system for which the specific corrosion prevention and control plan is being written (Ref 48). Feedback will not only assist airframe designers in writing a plan and specifications for corrosion prevention and control but will also enable these professionals to understand how and where corrosion problems tend to occur in particular airframe structures and to determine the most useful (and most cost-effective) applications for airframe corrosion-preventive technology. Determining the most useful applications for corrosion-preventive technology will help to minimize repeated occurrences of airframe corrosion problems. An example of this is a significant metallurgical contribution to the C-130 airframe

corrosion prevention and control program--the use of aluminum alloy 7075-T73 (instead of the less corrosion resistant 7075-T6 aluminum alloy) in the upper-center wing and other components of the C-130 (Ref 32). The main structural (aluminum) parts of the C-130 are provided with additional corrosion protection by anodizing and then coating them with a polysulfide primer (Ref 32). An adequate finish specification must be prepared and enforced in the early stages of aircraft design. During the design period, qualified corrosion engineers should review the initial drawings; this would prevent potential corrosion problems from occurring. Several documents that provide comprehensive requirements for corrosion prevention and control have been published. These standards were written for military airframe systems, but they can also be applied to corrosion protection on commercial airframes with little or no modification. The following United States DOD specifications can be used as a guide in selecting or preparing an adequate finish specification: • • • • •

Military Specification MIL-F-7179: "Finishes and Coatings: General Specification for Protection of Aerospace Weapons Systems, Structures, and Parts" Military Standard MIL-STD-1568: "Materials and Processes for Corrosion Control" Standardization Document SD-24: "General Specification for Design and Construction of Weapon Systems" Military Specification MIL-S-5002: "Surface Treatments and Inorganic Coatings for Metal Surfaces" Aeronautical Design Standard-13: "Air Vehicle Materials and Processes"

The corrosion prevention and control guidelines include in the above specifications have been widely used by government and industry. One of the most significant of these guidelines is the use of compatible materials in dissimilar metal couples, as summarized in Table 6.

Table 6 Metals and alloys compatible in dissimilar-metal couples Source: Ref 17

In addition to having a written plan, it is important to have designated corrosion and design engineers that are appointed to corrosion prevention teams or advisory boards. These groups have the responsibility of establishing materials and process requirements and design improvements for ensuring effective corrosion protection on specific aircraft systems. These requirements are generated from documented cases in which corrosion has been observed on airframe components, especially those parts that have deteriorated on earlier models of the same or similar aircraft systems. Airframe maintenance and repair should be performed to alleviate airframe corrosion. Examples of these actions include: • • •

The application of corrosion-preventive (water-displacing) compounds to airframe surfaces (Ref 49) Use of corrosion-inhibitive elastomeric sealants (Ref 50) Repair of stress-corrosion cracks on internal aluminum wing structures by applying bonded doublers consisting of boron fiber reinforced plastic (Ref 51)

Another important step in the corrosion control effort is the drafting of a preventive maintenance and inspection plan (Ref 47). This plan provides an outline of corrosion maintenance, inspection, and repair requirements that include the locations and frequencies of inspections and the corrosion-contributing factors that should be considered by professionals who

inspect aircraft structures for deterioration. Some of the factors that can induce the corrosion of airframes include the environmental conditions to which the airframe is (or will be) exposed and corrosive spillage (usually alkaline or acidic chemicals) from galleys and toilets. The inspection and maintenance plan must be defined and administered as early as possible. This is very significant because early corrosion control actions are crucial for the avoidance of costly aircraft damage or failures.

Corrosion in the Aircraft Industry

Corrosion of Powerplants Michael L. Bauccio, The Boeing Company

Aircraft powerplants have the important function of providing thrust, or propulsion, to enable aircraft to take off and remain in flight. Powerplant systems also provide reverse thrust to enable aircraft to land safely. The development and proper utilization of corrosion-preventive technology has enabled aircraft powerplants to fulfill these significant functions. In aircraft powerplant systems, including engine and hydraulic power units, several types of corrosion have been observed. Among these types of powerplant corrosion (but not the only kinds) are: •

•

• •

High-temperature corrosion. This has also been referred to as sulfidation, or hot corrosion. Hightemperature corrosion of powerplant components is also manifested as hot-salt-induced SCC and as very rapid oxidation (fire). Both of the latter forms of high-temperature corrosion have been observed primarily in titanium structures Cold corrosion. This type of powerplant corrosion originally referred to all types of corrosion other than hot corrosion, which primarily includes pitting and fretting corrosion of aircraft powerplant components. The cold form of pitting and fretting corrosion develops at considerably lower temperatures than those attained under hot corrosion conditions Chemical corrosion Erosion corrosion

The objective of this section is to focus on the above generic of aircraft powerplant corrosion. The problems and corrosion-preventive solutions that have proved to be significant developments for minimizing future occurrences of these corrosion problems will also be discussed. The thesis of this section is that powerplant corrosion problems can be minimized by regularly scheduled maintenance inspections and by effective applications of corrosion-preventive treatments to powerplant components. Most of the literature on the corrosion of aircraft powerplants is concerned with high-temperature corrosion, which occurs above approximately 315 °C (600 °F). This is because aircraft powerplants operate at extremely high temperatures. Some aircraft powerplant sections--such as the turbine rotor inlet--reach temperatures of about 1240 °C (2260 °F) (Ref 52). These high-temperature operating cycles have produced hot corrosion in powerplant components that did not have adequate corrosion protection. Some reciprocating parts, such as exhaust valves, are protected against high-temperature corrosion by coatings of high-temperature alloys (Ref 2). These metallic coatings are welded onto the valve faces and then machined to the required angle of either 30 or 45° (Ref 53). Corrosion of powerplants has also been found in many cases at much lower temperatures than those attained in hot corrosion. One of the most cold corrosion prone areas of the aircraft--the engine air inlet, or frontal, area and the cooling air vents--has experienced corrosive deterioration at relatively low temperatures. The frontal areas are vulnerable to erosion that is caused by airborne particulate matter, such as dirt, sand, and gravel. This abrasive action removes protective treatments and coatings from the frontal engine structures. This is eventually followed by corrosive deterioration in and around the areas that had been mechanically damaged by airborne solids (Ref 54). Certain engine parts, including reciprocating engine cylinder fins and accessory mounting bases, have small, unpainted areas that are susceptible to corrosion by airborne salts and moisture (Ref 54). The most effective technique for preventing and controlling corrosion in engine intake areas is to clean, examine, and refurbish these areas regularly, especially when the aircraft system must operate under aggressive conditions, such as in a tropical marine environment (Ref 55). The materials that are usually affected by the highly corrosive powerplant environment are metal alloys that are designed for long-term service at high temperatures. These alloys consist primarily of nickel, cobalt, and titanium. Some of the older nickel-base superalloy materials, such as Udimet alloy 700, have a low resistance to fatigue, which compromises the high-temperature static strength of these alloys (Ref 56). Therefore, powerplant corrosion can be minimized by the

increased use of materials that have the best combination of fatigue strength and corrosion resistance at elevated temperatures.

Case Histories Many typical case histories, or lessons learned, have been described in the literature of powerplant corrosion. Discussions of the actual forms of deterioration, as well as the corrosion-protective measures that were implemented to control or avert future problems, are presented below. These corrosion problems have led to significant improvements in the materials and processes used for the protection of aircraft powerplant components and for the assurance of aircraft flight safety High-Temperature Corrosion There are three predominant types of high-temperature corrosion that have caused mechanical failures of aircraft powerplant components. These are sulfidation, hot-salt SCC, and fires. Sulfidation has occurred in gas turbine parts that were made of nickel and cobalt alloys. Hot corrosion was observed in the 1960s on gas turbine blades that were manufactured from Nimonic 105 (Ni-20Co-15Cr-5Mo-4.5Al-1.4Ti) and GMR235 (Ni-15.5Cr-10Fe-5.2Mo-3Al-2Ti). This corrosive deterioration was called black plague (Ref 57). Some of the cobaltbase superalloys, which are also used in gas turbine powerplants, have also required advanced protective coatings for the mitigation of hot corrosion (Ref 58). The nature of these coatings will be discussed later in this section. Hot-salt SCC has been found primarily in titanium-alloy powerplant parts, such as the compressor blades of military supersonic trainer engines (Ref 59). A case history concerning hot-salt SCC in a titanium powerplant component is given in the discussion "Hot-Salt SCC" in this section. Fires have occured in powerplant components that are fabricated from titanium and other materials. This phenomenon is an extremely rapid form of oxidation that can cause extensive damage within a brief period in aircraft powerplant. In the case of titanium fires, this is a very critical problem because titanium oxidation occurs so quickly that is produces combustion. The deterioration that is caused by combustion spreads throughout the powerplant because titanium is widely used in aircraft powerplants because of its high strength-to-weight ratio, ability to maintain adequate strength a moderately high temperatures, and good corrosion resistance at ambient temperatures. Hot Corrosion (Sulfidation). An example of hot corrosion in the rotating blades of an aircraft powerplant is shown in Fig. 32. This part was manufactured from the nickel-base superalloy 1N-738. The creep failure of these blades--which also are called buckets--was initiated sulfidation. Magnified cross sections of these blades at 75 and 130×, respectively, are shown in Fig. 32(b) and 32(c).

Fig. 32 Aircraft powerplant turbine blades (a) that were damaged by hot corrosion. (b) Higher-magnification view of corrosion. 35×. (c) Corroded turbine blade. 55×

Hot corrosion problems have also been found in gas turbine stator (stationary) vanes (Ref 59). In gas turbine blades and vanes that were made of the IN-724 nickel-base alloy, hot corrosion has produced deterioration that appeared as cracking, swelling, or spalling. Oxides have been observed below the surface of components that are subjected to high-temperature corrosive environments. The subsurface oxides consist of nickel and large quantities of chromium, aluminum, titanium, and niobium. Nickel-rich metallic material has been found to lie adjacent to these oxides (Ref 59). The dynamic nature of the environment in aircraft powerplants promotes this type of high-temperature oxidation (sulfidation) of the main burner components and of downstream parts, such as turbine blades, vanes, and seals. Very high thermal and mechanical strains are produced on these parts because of powerplant operation at elevated temperatures (Ref 60). The corrosive ash that is usually present in this dynamic-oxidation environment is sodium sulfate (Na2SO4), which reacts with powerplant materials and causes a depletion of the alloying elements required to form a protective oxide film on the surfaces of these parts. The formation of a protective film can also be inhibited by consumption of the alloying elements by means of either acidic or basic oxide fluxing reactions (Ref 52, 60). Hot corrosion protection has been obtained primarily by the use of advanced-technology coatings and materials that are specifically designed for long-term endurance in dynamic corrosion environments. One such coating is a proprietary mixture of cobalt, chromium, aluminum, and yttrium (CoCrAlY). Aluminum is the most significant element in this coating because it reacts with oxygen to form a thin layer of corrosion-resistant aluminum oxide on the surface of the powerplant parts (Ref 61). Coatings of this type have been applied by electron beam plasma vapor deposition techniques.

Hot corrosion problems in aircraft powerplants have also been alleviated by the application of pack cementation type coatings (Ref 58). These coatings consist of either of the following elements: • •

Aluminum only, or a combination of aluminum and other elements, such as chromium and silicon Chromium alone, or chromium added to other elements, such as tantalum and aluminum

Another type of advanced coating that has been investigated for protection of powerplant components against hot corrosion is the thermal-barrier coating, which consists of various combinations of ceramic materials. These coatings are plasma sprayed, and they are composed of a specific ceramic material that applied over a layer of an oxidation-resistant metal. Accelerated hot corrosion tests on ceramic thermal-barrier coatings have determined that the failure mechanism of these coatings involves delamination, which occurs prior to surface cracking or spalling (Ref 62). Despite the failure of some thermal-barrier coatings, the aluminum-ceramic coatings have been successfully applied in the last 20 years to various gas turbine components, including blades and vanes for corrosion protection to temperatures of about 650 °C (1200 °F) (Ref 63). Other materials have been developed and used for improving the resistance of powerplant components to hot corrosion, such as oxide dispersion-strengthened alloys (Ref 64), which include the following materials: • • •

Inconel alloy MA-754, which was developed for use in gas turbine vanes Incoloy alloy MA-956, which was developed for powerplant combustion chamber burner cans Alloy MA-6000E, which was designed and produced for application in gas turbine blades

In addition to the above oxide dispersion-strengthened materials, the following materials have been designed for mitigating high-temperature corrosion in powerplant systems: • •

Nickel-, cobalt-, and iron-base superalloys that contain platinum and platinum-group metals. Alloying with platinum group metals provides a great improvement in overall hot corrosion resistance (Ref 65) Directional solidified alloys. These materials include columnar-grained and single crystal forms (Ref 56). In addition to having a significantly higher resistance to hot corrosion, these materials (such as the directional solidified MAR-M200 nickel-base alloy) have up to 100 times the high-temperature fatigue life of the polycrystalline forms of the same alloy (Ref 56)

Hot-salt SCC has been produced in the laboratory in titanium alloys that are susceptible to embrittlement and fracture.

The conditions required to induce this type of failure are a combination of the following (Ref 66, 67, 68): • • •

High Stress High temperature (approximately 345 °C, or 650 °F) Exposure to a high concentration of chloride salts, which is anticipated in a marine environment

Hot-salt SCC usually produces intergranular-type corrosion in titanium alloys. In cases in which hot-salt SCC occurs because of the exposure of titanium powerplant parts to solid chloride salts, the minimum temperature at which hot-salt SCC develops is approximately 290 °C (550 °F) (Ref 68). Hot-salt SCC of titanium powerplant components has been reported as the cause of powerplant deterioration in one incident that occured during ground testing. In this case, a compressor rotor assembly in a turbine engine failed during an overspeed acceleration test. Two compressor disks, manufactured from titanium alloy Ti-7Al-4Mo, fractured in several locations. Cracking occured at the bolt holes, where dissimilar-metal (A-286) bolts were in contact with the titanium disk. This failure is illustrated in Fig. 33. The two failed disks had large silver deposits in the tie bolt holes, which was indicated by surface attack at the crack initiation sites. Silver chloride was formed during the test cycle by a chemical reaction between silver and chlorine, which were present in the high-temperature environment. Silver had been used as an

antifretting coating on the tie bolts. As a consequence of this failure, the silver coating was replaced by a molybdenum disulfide dry-film lubricant (Ref 67, 69).

Fig. 33 Cracking of a Ti-7Al-4Mo aircraft powerplant compressor disk

Rapid Oxidation. Corrosion problems produced by rapid oxidation (combustion) in aircraft powerplants have

predominantly involved titanium. In some isolated cases of rapid oxidation, the damaged parts were fabricated from steel. The total number of known occurrences of titanium combustion in powerplants has been reported to be 144. In 85 of these incidents, the combustion did not penetrate the case, or skin, of the powerplant. Details of the deterioration that occurred in the 144 documented cases of titanium combustion are available for only two of these incidents (Ref 70). The ignition temperature of titanium in air is approximately 1625 °C (2960 °F). This is significantly higher than the maximum service temperature for titanium. Therefore, either a mechanical or aerodynamic upset of the operation of the powerplant must occur in order to raise the temperature of titanium components to the ignition point. Several events can cause mechanical rubbing or jamming in powerplant structures, including (Ref 70): • • • •

Interference by loose solid particles with rotating and static structures. These particles can bend blades, causing the blades and vanes to rub against each other continuously. Rotor imbalance, which can cause severe rubbing. This develops from the failure of rotating disks, spools, or blades Radial or axial displacement of the rotor, which can be caused by a failed bearing Displacement of the rotating powerplant blades into stationary parts, which produces compressor stall

The most significant of these causes of powerplant titanium combustion are trapped blades and radial displacement of the rotor (Ref 71). Only a small number of titanium combustion incidents have occurred because of aerodynamic heating, stall, or surge (Ref 71). Powerplant case penetration has been observed in 57% of the titanium combustion incidents involving high-bypass ratio turbofan powerplants. Case penetration is illustrated in Fig. 34. In most of the case penetration incidents, small holes were

observed in various locations on the case. Some incidents of case penetration have been characterized by complete deterioration of the case that covers the compressor section of the powerplant (Ref 71).

Fig. 34 Case penetration produced by titanium combustion in a high-bypass turbofan aircraft powerplant. Courtesy of C.W. Elrod, Wright-Patterson Air Force Base

Titanium fires that are caused by blade tip rubbing can be prevented by (Ref 71): • •

•

Increasing the clearances (tolerances) between the powerplant blade tips and case interior surfaces, although this is in opposition to the current trend toward tighter tolerances for improved performance Applying blade-tip materials that are resistant to wear. Abrasion-resistant materials have been investigated for this purpose. A plasma-sprayed Co-30WC tip treatment compound provided the best results in one study (Ref 72) Protecting the entrance edges of bleed air manifolds with steel grommets (Ref 70)

As previously mentioned, steel components have also failed by rapid oxidation in aircraft powerplants. These powerplant corrosion problems are described as breech chamber failures. Breech chambers are used in various types of aircraft engines. These chambers are part of the cartridge, or pneumatic, starters (Ref 5). Combustion products from the breech chambers are exhausted toward the turbine blades to start the engine rapidly. Typical breech chamber failures are shown in Fig. 35. These components were fabricated from 4340 steel heat treated to between 40 and 45 HRC. As indicated in Fig. 35, the breech chamber dome bursts at a point that is opposite to the hot gas discharge nozzle. These failures occur where stress concentrations are the highest, which is where the dome contacts the breech chamber body (Ref 5).

Fig. 35 Typical failures of 4340 steel aircraft powerplant breech chambers. Source: Ref 5

A fracture at the hot gas nozzle of the breech chamber is shown in Fig. 36. At the base of the nozzle (beyond the location of the fracture), a pattern of trenches was observed on the external chamber surface (Fig. 36b). The most probable explanation for these breech chamber failures is a defective coating system. There was no finish on the internal surfaces of these failed chambers. The domes were coated with electroless nickel, which caused galvanic corrosion on the more active (anodic) steel chambers (Ref 5). Future breech chamber failures can be avoided by coating the entire chamber with one of the high-temperature resistant coatings mentioned in the discussion "Hot Corrosion (Sulfidation)" in this section. The electroless nickel finish, applied directly to the steel chamber dome, should not be used, because of the possibility of galvanic corrosion at the nickel and steel contacting surfaces.

Fig. 36 Fracture in the hot gas nozzle (a) of a powerplant breech chamber. (b) Trench pattern at the base of the breech chamber hot gas nozzle shown in (a). Source: Ref 5

Cold Corrosion Fretting corrosion is a very important design consideration for titanium parts in aircraft powerplants. This is because fretting corrosion can significantly reduce the fatigue strength of titanium alloy Ti-6Al-4V (Ref 73). Fatigue strength reductions of greater than 50% have been documented (Ref 67).

Fretting-induced deterioration has occurred in aircraft powerplant control bearings (Ref 67). The components that are most susceptible to significant fretting deterioration are the fan roller bearings on aircraft turbofan engines (Ref 74). Fretting is also a significant factor in the design of powerplant compressor blades fabricated from titanium. Fretting corrosion can occur on the root and midspan shroud of these blades unless adequate design precautions are taken to minimize the potential for fretting to occur (Ref 67).

In addition to powerplant bearings and compressor blades, the following components of aircraft engines are known to be susceptible to fretting corrosion (Ref 74): • • • • • • • • • • • • •

Airfoil roots in rotors and stators Stator vanes and shroud interfaces Splines and root assembly stackup interfaces Piston ring secondary seals Torque pins for seal mountings Fasteners, including threads, the surfaces of bolted joints, and disconnect rings Metal static seals Engine mounts and thrust links Slip joints, which are used in burners and tubing assemblies Disk spacers Planetary gear shafts (Fig. 37). These parts have incurred fretting damage primarily in the gear bore. A smaller amount of fretting deterioration has also been observed in the bearing bore Clutches Joints and interfaces between titanium, aluminum, and austenitic stainless steel parts

The most significant fretting problems in aircraft powerplants have occurred in seals, splines, blade mountings, and blade dampers (Ref 74).

Fig. 37 Fretting corrosion of a planetary gear shaft from an aircraft powerplant. Courtesy of the Advisory Group for Aerospace Research and Development

The damage produced by fretting corrosion usually appears as pitting, polishing, or both. Galling is produced in powerplant structures when fretting continues over relatively long periods of time (Ref 74). Both galling and fretting have been observed on oil-film dampers that are coupled with thrust washers. Additional powerplant components that have been subject to fretting are discussed below. Over-Running Clutches. These components are used in helicopter drive trains (Ref 74). Figure 38 shows the fretting

damage that has been observed in the powerplant components. This problem is caused by vibration (and rubbing) of the balls in the inner and outer races of bearings that support the rotor shaft. These inner and outer bearing races rotate at the same speed when the over-running clutches are engaged (Ref 74).

Fig. 38 Fretting corrosion of the bearing race of an aircraft powerplant over-running clutch. Courtesy of the Advisory Group for Aerospace Research and Development

Splines. In one case involving fretting on a compressor-turbine shaft spline, the spline failed and resulted in an engine

fire, which caused the aircraft to crash. A spline structure surface that suffered fretting corrosion damage is shown in Fig. 39.

Fig. 39 Fretting corrosion on an aircraft powerplant spline structure. Courtesy of the Advisory Group for Aerospace Research and Development

Fretting of splines is inhibited by maintaining a lubricating oil film on the surfaces of these parts. Solid film lubrication has proved to be an effective antifretting treatment for splines. The use of beryllium-copper material instead of SAE 4140 steel has also enhanced the fretting resistance of powerplant spline structures (Ref 74). Applications of vacuum coatings, such as by sputtering or ion plating, have also proved to be effective in alleviating spline fretting corrosion problems (Ref 74). Components that have been subject to fretting corrosion are discussed below. Powerplant Rotor Blade Root Mounting Surfaces. An illustration of this fretting corrosion problem is provided

in Fig. 40. Fatigue cracks can propagate as a result of this fretting damage, leading to eventual failure of these blades. Plasma-sprayed coatings, which have good adherence and friction-reducing properties, can be applied to blade surfaces in order to minimize the deterioration and subsequent mechanical failures caused by fretting.

Fig. 40 Fretting corrosion on the root surface of an aircraft powerplant compressor blade. Courtesy of the Advisory Group for Aerospace Research and Development

Helicopter Reciprocating Engine Connecting Rod. This component failed in flight by fatigue that was caused by

fretting corrosion on a small area on the bore of the connecting rod (Ref 55). Fretting corrosion developed in this connecting rod because the shell bearing, inside of the large-end bore of the rod, moved in a rapid rotational-oscillatory manner. The design solution to this problem was to increase the circumference of the shell bearings and to apply greater torquing force onto the bolts. This sufficiently reduced the rotational-oscillatory movement that had caused this problem, and further fretting corrosion was prevented (Ref 55). Low-Pressure Compressor Casing. This part was made from magnesium alloy EZ33. Fretting corrosion occurred

on this component because of mechanical rubbing action by a row of steel stator vanes on an annular ring. Because this fretting damage resulted from contact between dissimilar metals (steel against magnesium), this incident is an example of fretting corrosion that occurred in a dissimilar-metal couple (Ref 55). The softer metal in the couple (EZ33 magnesium) was damaged more severely than the harder metal (steel). This problem can be alleviated by changing the EZ33 magnesium casing material to a harder, more wear-resistant alloy or by applying a wear-resistant (hard-facing) coating to the areas of the EZ33 magnesium casing that will come into contact with the steel stator vanes. Stress-corrosion cracking has been reported as the dominant type of deterioration in the following failures of aircraft

powerplant components. Engine Exhaust Tailcone Assembly (Ref 5). A large section (about one-third) of this component fell away from

an A-7D aircraft during flight. Circumferential resistance-welded stiffeners (Fig. 41) were used for stabilizing the tailcone structure on the outer surface of the assembly. The material used in the manufacture of this powerplant assembly was corrosion-resistant steel. In addition to SCC, intergranular degradation was observed where the tailcone was welded to the circumferential stiffener. During the analysis of this failure, the A-286 alloy was found to be prone to intergranular corrosion due to the formation of precipitates in the grain boundaries of the A-286 material.

Fig. 41 Circumferential stiffener (a) from a tailcone assembly that fell off an aircraft during flight. (b) Resistance weld on the stiffener in (a). Source: Ref 5

High-Pressure Powerplant Compressor Vane (Ref 39). This failure occurred on a military fighter aircraft. The

SCC originated at a point on the vane where pitting corrosion had developed. Several actions have been taken to prevent future powerplant vane failures, including changing the vane material, coating the vanes (nickel plating was one of the evaluated coatings), and periodically washing the compressor with water. Pitting Corrosion. Cases involving pitting corrosion in powerplant components have developed over a relatively long

period of time as compared to other types of ambient-temperature corrosion. The pitting corrosion of rotor and stator blades made of low-alloy steels and martensitic stainless steels is a significant powerplant deterioration problem that has produced fatigue failures and compressor stalls (Ref 75). The severity of these pitting problems was reduced by using better coatings and by applying improved corrosion control procedures, such as washing of the unprotected parts with water and inhibitors (Ref 76). Several types of coatings have been applied for the alleviation of pitting corrosion in these steel parts, including (Ref 77, 78): • • • •

Paint Overlay coatings Metal-ceramic barrier coatings Nickel-cadmium diffusion coatings

A proposed solution to this problem was to change the blade material from steel to a nickel-base superalloy, IN-718 (Ref 77). The cost of this material change, however, would be significantly greater than the alternative of applying a specific coating to these parts. Powerplant blades and vanes can be coated with aluminum for pitting corrosion protection. Aluminum coatings can be applied by using the following techniques: • • • •

The slurry method (Ref 79) Pack diffusion (Ref 79) Hot dipping and diffusion treatments (Ref 79) Ion vapor deposition (Ref 80)

Pitting corrosion has occurred in powerplant compressor stator vanes despite the use of a ceramic barrier coating material (Ref 81). Severe pitting was noted in this case. Some of the pits developed at critical locations, where the strength of the stator vane could be adversely affected. Inadequate process controls during the application of the ceramic slurry (bisque) to the stator vanes contributed to the rapid acceleration of the pitting corrosion that occurred in this case (Ref 81). Additional incidents involving pitting corrosion in powerplant components are discussed below. AM-355 Stainless Steel Compressor Blades (Ref 81). Severe pitting corrosion was observed in the dovetail

section, which is where the blade meets the compressor wheel. This problem was solved by replacing the pitted blades

and by applying corrosion inhibitors to the newly installed parts. Pitting corrosion was prevented by the application of two different corrosion-preventive compounds (Ref 81). Thrust Reverser Doors (Ref 55). Pitting corrosion caused the in-flight detachment of the thrust reverser doors of a commercial aircraft. One of the thrust reverser door driver links fractured because of fatigue. The fatigue crack started where pitting corrosion developed on the surface of the driver link, which was fabricated from type 422 stainless steel. The driver link had been coated with a dry-film lubricant consisting of molybdenum disulfide. Pitting damage in the stainless steel part was caused by the accumulation of fuel combustion by-products and moisture in localized areas where the coating was defective. Preventive measures that can be taken to eliminate this problem and similar problems consist of: (1) reducing the high stress level in the thrust reverser door drive link (this is a longer-term solution because it would require a redesign of the drive link); (2) recoating the links with a more protective material, which will prevent the future development of pitting corrosion; and (3) establishing a continuous inspection program to ensure that corrosion is identified early, if it occurs. Engine Exhaust Valve (Ref 55). This fatigue failure occurred during the flight of a single-engine aircraft. The

fatigue-induced fracture started at a local area on the valve stem that was severely damage by pitting corrosion. Pitting reduced the section diameter of the exhaust valve (near the location of the fracture) by about 0.25 mm (0.01 in.). This pitting deterioration was produced by acidic compounds in old engine oil that had not been drained and replaced with new oil. Timely draining and replacement of engine oil is required to maintain the required concentration of corrosion inhibitors in the oil for preservation of the engine components. Erosion-Corrosion and Cavitation. Erosion-corrosion occurs when solid particles collide at high velocities with

powerplant components. This action removes the protective coatings and oxide films, resulting in corrosion of the unprotected metallic parts. Cavitation produces deterioration of power-plant components by the high-velocity impingement of water droplets on these parts. Titanium is one of the best materials to use in turbine blades for resistance to cavitation (Ref 67). Severe erosion problems have developed in magnesium splitter vanes (Ref 75). The Naval Air Development Center has studied the use of explosively welded aluminum claddings as a possible solution to this problem. Aluminum alloy substitutions have also been considered as a solution to erosion-corrosion problems in magnesium air inlet housing (Ref 76). Erosion attack has been especially noted in the air stream where protective coatings had worn away (Ref 76). Hot desert climates can cause very severe erosion-corrosion problems in aircraft that operate in this type of environment (Ref 55). Dust and sand, with high salt concentrations, can erode surface finishes on powerplant components. Corrosion is accelerated by the extremely high temperatures that are reached in desert areas. Erosion-corrosion has a high probability of occurring on various powerplant parts, including the following: • • • • •

Impellers Compressor blades Turbine blades Turbine nozzles Turbine guide vanes

These powerplant components are vulnerable to erosion-corrosion that is caused by solid particles that impact these parts in the air intake area or in the hot gas stream. Powerplant components can be protected from erosion-corrosion by coatings that consist of hard, erosion-resistant materials. Many new erosion-resistant coatings are currently being evaluated (Ref 55). Miscellaneous Types of Powerplant Corrosion This discussion will cover various case histories regarding powerplant corrosion. The types of deterioration that caused these corrosion incidents are general corrosion; chemical corrosion, which is produced by chemical spills; intergranular corrosion; and corrosion fatigue.

General corrosion was found on aluminum-brazed titanium tailpipe extensions that were attached to the engine of a

commercial aircraft (Ref 82). This deterioration was observed at the conclusion of a 3-year in-service evaluation of these tailpipe extensions. Among the factors that contributed to this corrosion problem were operation of the aircraft in rainy and humid conditions and a hydrophilic and sooty coating on the inner surfaces of the tailpipes. The skins, or outer surfaces, of the tailpipe extension surrounded a honeycomb core. The skins were perforated for sound attenuation. The perforation resulted in atmospheric exposure and subsequent corrosion of the aluminum-brazed titanium component. The hydrophilic (sooty) coating and the general corrosion are illustrated in Fig. 42. As a result of this case, it was concluded that aluminum-brazed titanium is not a suitable material for jet engine tailpipes that have perforated skins and honeycomb cores.

Fig. 42 Schematic showing an individual cell of a honeycomb sandwich engine tailpipe that failed because of corrosion

Chemical corrosion has been documented for components of the F-14 aircraft radar antenna hydraulic drive system (Ref 83). Corrosion on the internal surfaces of these parts was produced by contamination of the hydraulic fluid by water and halogenated solvents, primarily trichlorotrifluoroethane. The concentrations of these liquid contaminants exceeded the limits of the hydraulic oil specification (MIL-H-5606). These limits were 25 ppm for halogenated solvents and 150 ppm for water (Ref 83). The parts that were damaged consisted of 4130, nitralloy, and tool steels.

This corrosion did not lead to a failure of the hydraulic drive system. However, the long-term effect would have been a reduction in the reliability of the system. This problem was eliminated by ensuring that periodic draining and refilling with fresh hydraulic fluid was performed. This use of vacuum degassing procedures and desiccant filters also helped to alleviate this deterioration problem (Ref 83). The emergency power unit of the F-16 military aircraft contains mixed hydrazine fuel consisting of 70% hydrazine and 30% water (Ref 84). Although this fuel can cause serious corrosion problems, no such cases have been published. Hydrazine-induced deterioration has been prevented on the F-16 emergency power unit by the development and application of an epoxy-type sealant that is resistant to hydrazine (Ref 84). Intergranular corrosion was found to be the cause of a failure of a light aircraft engine exhaust pipe (Ref 55). The exhaust pipe separated from the aircraft engine during flight. The fracturing of the austenitic stainless steel pipe had occurred near its flange support. Intergranular corrosion cracks produced stress concentration sites. These stress raisers initiated fatigue due to vibration of the exhaust pipe and a muffler that was attached to the pipe. The muffler was cantilevered and unsupported--a design that produced high cyclic stresses, which was a major cause of the exhaust pipe failure.

The severely corrosive environment produced by exposure of the pipe to combustion by-products and condensates at high and low temperatures also played a major role in initiating the fracture of the exhaust pipe. It is possible that the exposure of the austenitic stainless steel to hot exhaust gases caused the sensitization of the steel. When sensitization occurs, the

grain boundaries become susceptible to intergranular corrosion. A greater degree of microstructural stability can be achieved in austenitic steels by reducing the carbon content of these materials (see the article "Corrosion of Stainless Steels" in this Volume). This intergranular corrosion problem was solved by adding a structural support for the muffler in order to minimize the level of cyclic stress that this component would have to endure. Corrosion fatigue has been reported as the cause of failure of an aircraft powerplant rocker-arm journal bearing. The bearing surface was examined, and deep pits and score marks were found (Ref 55). Fatigue cracks were observed to propagate radially from cavities along the surface of this part. This journal bearing was determined to be cast aluminum alloy 242. It was determined that there was gross porosity in the bearing casting. The material was therefore unsatisfactory for fatigue-critical applications. Proper quality control during the metallurgical processing of the casting to prevent porosity would have helped to produce a stronger, more fatigue-resistant part. General corrosion abetted the failure process.

Corrosion fatigue has also been reported to be the cause of steel engine shaft failures. These failures can originate at bolt holes or at the roots of gear teeth. The original corrosion damage then propagates by fatigue (Ref 85). The Air Force Wright Aeronautical Laboratories' Materials Laboratory Structural Failure Analysis Group is responsible for performing metallurgical failure analyses on a variety of aircraft structural components. This includes airframes, propulsion and missile systems, and ground support equipment. Some of these analyses concern failures that occurred by corrosion mechanisms. This number has averaged approximately 15% of the total analyzed failures for the last several years. Most of these incidences are either manufacturing/quality control related or involve materials and/or heat treatments used in older systems. The following discussions represent a number of case histories of failures associated with corrosion and the recommended corrective actions to preclude repetitive occurrences.

Examples of Aircraft Structural Corrosion Failures Example 1: Aircraft Attachment Bolt Failure. During a routine inspection on an aircraft assembly line, an airframe attachment bolt was found to be broken. The bolt was one of 12 that attach the lower outboard longeron to the wing carry through structure. Failure occurred on the righthand forward bolt in this longeron splice attachment. The bolt was fabricated from PH13-8Mo stainless steel heat treated to have an ultimate tensile strength of 1517 to 1655 MPa (220 to 240 ksi). A water-soluble coolant was used in drilling the bolt hole where this fastener was inserted. Investigation. Figure 43(a) shows a bolt shank with corrosion evident on surfaces in contact with the splice joint, while

Fig. 43(b) shows the fracture and initiation site of the bolt failure. Surface pitting on the bolt shank and subsequent corrosion cracking are shown. Scanning electron microscopy (SEM) examination of the fracture revealed the topography to be intergranular at the initiation site and to a depth of 8.4 mm (0.33 in.) (Fig. 43c). The remainder of the fracture showed a mixed topography of cleavage and ductile dimples (Fig. 43d).

Fig. 43 Parallel lines of corrosion (a) on the shank of a PH13-8Mo stainless steel aircraft attachment bolt. (b) Close-up of fracture surface of bolt showing corroded area. Arrows point to one possible crack arrest line. (c) SEM fractograph of area in B in Fig. 2. Note corrosion product (left) and ductile dimples in the center. 265×. (d) SEM fractograph of area C in (b). Area of fast fracture shows cleavage and dimples. 265×

Examination of corrosion products on the fracture by Auger emission spectroscopy and secondary imaging spectroscopy showed the presence of elements typically found in tap water. Chemical analysis of the bolt material showed the composition to be within specification limits for PH13-8Mo stainless steel. Rockwell C hardness measurements taken on the bolt produced values ranging from 47 to 48 HRC, which would correspond to the specified ultimate tensile strength of 1517 to 1655 MPa (220 to 240 ksi). Conclusion. It was concluded from the study that failure of the attachment bolt was caused by stress corrosion. The source of the corrosive media was the water-soluble coolant used in boring the bolt holes. Recommendations consisted of inspecting for corrosion all the bolts that were installed using the water-soluble

coolant at the spliced joint areas, rinsing all machined bolt holes with a noncorrosive agent, and installing new PH13-8Mo stainless steel bolts with a polysulfide wet sealant.

Example 2: Failure by SCC of an Ejection Seat Swivel. A routine examination on a seat ejection system found that the catapult attachment swivel contained cracks on opposite sides of the part. This swivel, or bath tub, does not experience any extreme loads prior to activation of the catapult system. Some loads could be absorbed, however, when the aircraft is subjected to G loads. The bath tub is fabricated from aluminum alloy 7075-T651 plate. Investigation. Visual examination of the part revealed that cracks were positioned near the base of the bath tub

configuration and extended through the wall thickness. One of the cracks was opened (Fig. 44a); this indicated that the fractures initiated on the inner walls of the fixture.

Fig. 44 Opened crack (a) in aluminum alloy 7075-T651 ejection seat swivel fixture that failed by SCC. Note crack propagation markings that suggest the crack initiated on the inside wall of the fixture and woody appearance of the fracture. (b) Higher-magnification view of fracture surface from (a). Note woody appearance, which indicates a precrack mechanism. The inner wall of the fixture is at the top. 5×. (c) SEM fractograph of area A from (b). Note slight intergranular appearance of structure. 425×. (d) SEM fractograph of area B from (c) showing intergranular facets on the fracture surface. 1060×

Electron optical examination of the fracture at low magnifications revealed a woody appearing topography (Fig. 44b). Further electron optical examination of the fracture at 800 and 2000× (Fig. 44c and d) showed that the cracking pattern initiated and progressed by an intergranular failure mechanism. This fracture topography indicated that cracking was due to stress corrosion. Examination of the microstructure near the fracture revealed that the crack was progressing parallel to the transverse grain flow direction and further suggested SCC. Chemical analysis and hardness tests conducted on the submitted material showed it to be within specification requirements for 7075-T651 aluminum base material. Conclusion and Recommendation. It was concluded that failure of the catapult attachment swivel fixture occurred

by SCC, and it was recommended that the 7075 aluminum ejection seat fixture be supplied in the T-73 temper to minimize susceptibility to SCC.

Example 3: Cracking of an Aircraft Wing Bracket. During an inspection cycle, cracking was detected in a wing fillet flap bracket. The cracking was located on the end of the bracket, as shown in Fig. 45(a). The configuration of the end of the bracket suggested that a bushing and rod were integral working components to the bracket hardware.

Fig. 45 End of aluminum alloy 7178-T6 aircraft wing bracket (a) showing cracking. (b) View of bracket showing symmetrical indentations on the top surface. Arrow shows a pit on the inner wall. 1×. (c) Close-up of indentations showing deformed surface (arrow) indicating directional movement of the bushing. 4×. (d) Failed fracture surfaces of bracket showing the woody fracture appearance characteristic of exfoliation. (e) Cross section of bracket showing delamination caused by exfoliation. 105×

Investigation. Visual examination of the bushing seat are showed the presence of surface corrosion pits (Fig. 45b).

Also shown are six symmetrical indentations that were produced during staking to prevent shifting of the bushing. Further visual examination revealed deformation adjacent to the indentations (Fig. 45c), indicating that the bushing had deformed the material. Optical examination of the opened fracture showed a woody, delaminated, fibrous-textured fractured surface (Fig. 45d). These fracture characteristics indicated that failure progressed by exfoliation corrosion. A cross section taken through the fracture surface revealed the presence of delamination due to exfoliation, as shown in Fig. 45(e). A chemical analysis of the material revealed it to be within specification for the required aluminum alloy 7178 except for a slightly lower than required zinc percentage. A hardness survey taken on the hardware found the values to range from 85 to 87 HRB and suggested that the material was in the T-6 condition. Conclusions. From this analysis, it was concluded that failure of the wing fillet flap bracket was due to surface

corrosion pits on the extrusion bracket hole wall surface. Crack progression occurred by exfoliation corrosion and was aided by a contributing stress introduced by movement of the bushing. Recommendations. It was recommended that a material substitution be made of the 7178-T6 material because of its

susceptibility to exfoliation corrosion. Candidate replacements included aluminum alloys 7175, 7050, or 7049.

Example 4: Failure of an Aircraft Controller Diaphragm. The diaphragm from a side controller was found during a preflight inspection to be broken. The controller diaphragm was fabricated from 17-7PH stainless steel in the RH 950 heat treatment condition. Failure occurred by cracking of the base of the flangelike diaphragm. The crack traveled 360° around the diaphragm. Investigation. Examination of the submitted piece by SEM revealed that the failure occurred by a brittle intergranular

mechanism and indicated a failure mode of selective grain-boundary separation (Fig. 46a and b). The diaphragms are heat treated in batches of 25. An improper heat treatment could have resulted in the formation of grain-boundary precipitates.

These grain-boundary precipitates would include chromium carbides. A loss of chromium from the matrix adjacent to the grain boundaries increases the susceptibility of this material to grain-boundary attack by oxidizing media. This action could result in intergranular corrosion and/or SCC if the part was placed under load in an oxidizing medium.

Fig. 46 SEM of fracture surface (a) from a failed 17-7PH stainless steel aircraft controller diaphragm showing intergranular fracture indicative of SCC. 170×. (b) SEM fractograph of area adjacent to that shown in (a) showing intergranular fracture, secondary cracking, and little or no evidence of ductility; this suggested a brittle fracture mechanism. 170×. (c) Microstructure of diaphragm taken perpendicular to the fracture face. The carbide network at the grain boundaries is evident. 210×

Metallographic examination of a section taken perpendicular to the fracture revealed a martensitic structure with grainboundary attack on the outer surfaces. The grain-boundary attack most probably occurred during heat treatment. Figure 46(c) shows a microstructure that was etched to reveal the presence of carbide precipitates. The grain-boundary network of carbides strongly suggested that the material was exposed to an adverse heat treatment with chromium depletion of the matrix adjacent to the grain boundaries. Conclusions. It was concluded that failure of the diaphragm was due to a combination of sensitization caused by

improper heat treatment and subsequent SCC. Recommendations. It was recommended that the remaining 24 sensor diaphragms from the affected batch be removed

from service. In addition, it was recommended that a sample from each heat treat batch be submitted to the Strauss test (ASTM A 262, practice E) to determine susceptibility to intergranular corrosion. It was also recommended that a stress analysis be performed on the system to determine whether or not a different heat treatment (which would offer lower strength but higher toughness) could be used for this part.

Example 5: Fuel Line Corrosion. Inspections revealed fuel line corrosion beneath ferrules (Fig. 47). The cause of the corrosion was traced to the fuel line marking process, which involved electrolytic labeling of ferruled aluminum alloy 6061-T6 tubes. Although subsequent rinsing of the fuel lines washed off most of the electrolyte, some was trapped between the 6061-T6 tubing and the ferrule. This condition made corrosion of the fuel lines inevitable.

Fig. 47 Corrosion (a) of aluminum alloy 6061-T6 aircraft fuel line (arrow). (b) Close-up of corrosion on fuel line. Note pitting and corrosion products. (c) Intergranular corrosion of the fuel line at area A from (a)

Investigation. Microstructural analysis revealed extensive intergranular corrosion of the 6061-T6 tubing beneath the

ferrule (Fig. 47c). This attack caused grains to become dislodged, giving the appearance of pitting. Corrosion penetrated approximately 0.13 mm (0.005 in.) into the tubing. In an attempt to determine if the corrosion products were active, two specimens from the corroded fuel lines with corrosion products were mounted and soaked in distilled water at room temperature for 2 and 4 days. The 2-day exposure resulted in a localized intergranular corrosion on the inside diameter of the tubing, while the 4-day exposure resulted in extensive intergranular corrosion of the tube cross section from the inside diameter to the outside diameter. Corrosion products from beneath the ferrules were placed on a piece of uncorroded 6061-T6 tubing in an attempt to substantiate further whether or not the corrosion products were active. Electrical tape loosely applied to the specimen held the products in place while the test specimens were submersed in distilled water for 5 days. Subsequent inspection of the specimen revealed that corrosion did not occur during the 5 days. Emission spectroscopy of the corrosion products showed that small amounts of aluminum (4%), sodium (3%), cobalt (2%), chromium (0.35%), boron (0.25%), and iron (0.05%) were present. The remaining 90% of the material analyzed was nonmetallic. Conclusions. It was concluded that the marking electrolyte used for labeling was trapped between the 6061-T6 tubing

and the ferrule. This fostered intergranular corrosion. Experiments indicated that the corrosion products were inactive. Recommendations. It was recommended that another marking process be used that does not involve corrosive

materials. The prevention of electrolyte from being trapped between the tubing and ferrules by using a MIL-S-8802 sealant was recommended.

Example 6: Corrosion of Aluminum Alloy 7075-T6 Wing Panel. New aircraft wing panels extruded from 7075-T6 aluminum were reported to be discolored, exhibiting an unusual pattern of circular black interrupted lines (Fig. 48a). The black marks were coherent with the metal and could not be removed by scouring or light sanding. The panels, subsequent to profiling and machining, were required to be penetrated inspected, shotpeened. H2SO4 anodized, and coated with MIL-C-27725 integral fuel tank coating on the rib side.

Fig. 48 Aluminum alloy 7075-T6 aircraft wing panel (a) showing unusual surface appearance. (b) SEM of the panel surface showing cracked anodized coating. 160×. (c) SEM showing the anodized coating flaking away and corrosion deposit under the coating. 85×. (d) Cross section of corrosion site on panel showing depth of intergranular attack. 265×

During processing, the extrusions are machined on the flat side, oiled, deburred, hot formed, cleaned, penetrant inspected, covered with oil, and then shotpeened. They are then recoated with oil, shipped to a second vendor, handwiped with a solvent, alkaline cleaned, acid desmutted, sulfuric acid anodized, and hot water sealed. Investigation. The panels were studied using the scanning electron microscope and microprobe analysis. Both

conventional energy-dispersive and Auger analyzers were employed. Figures 48(b) and 48(c) illustrate the contention that the anodic coating was applied over an improperly cleaned and contaminated surface. It was evident that the expanding corrosion product had cracked and in some places had flaked away the anodized coating. The corrodent had penetrated the base aluminum in the form of subsurface intergranular attack (Fig. 48d). The depth of attack was measured to be 0.035 mm (0.0014 in.). Microprobe analysis of the corrosion product did not reveal any clues concerning the reason or origin of the corrodent. A high sulfur concentration was found associated with the corrosion product and on surface areas away from the products. It was suspected that the origin of the sulfur was the hydrocarbon oil. When the anodized layer was stripped from the panels using a phosphoric-chromic acid solution, the evidence of sulfur disappeared. This same stripping procedure did not remove the black corrosion product. Energy-dispersive analysis of the corrosion product revealed the presence of iron, calcium, phosphorus, and chromium in excess. No chlorides were detected. Auger spectroscopy revealed the presence of large amounts of carbon and nitrogen. The MIL-C-27725 coating was removed from a portion of the rib side by using a paint stripper. No corrosion or discoloration of the aluminum was observed. Conclusions. It was concluded that the corrosion of the anodized panel most probably resulted from improper and

insufficient cleaning prior to anodizing. The preservation oils used during the various steps of manufacture and their incomplete removal prior to anodizing were highly suspect. Recommendations. It was recommended that a vapor degreaser be used during cleaning prior to anodizing. A hot

inhibited alkaline cleaner was also recommended during cleaning prior to anodizing. The panels should be dichromate sealed after anodizing. The use of deionized water was also recommended during the dichromate sealing operation. In addition, the use of an epoxy primer prior to shipment of the panels was endorsed. Most importantly, surveillance of the

anodizing process itself was emphasized, including continual monitoring of bath acid concentration, solution cleanliness, temperature control, and voltage/amperage control.

Example 7: Cracked Aircraft Wing Spar. A crack (Fig. 49) was found in an aircraft main wing spar flange fabricated from aluminum alloy 7079-T6 during a routine nondestructive x-ray inspection after the craft had logged 300 h.

Fig. 49 Aluminum alloy 7079-T6 aircraft wing spar (a) showing crack (arrow). (b) Fracture surfaces of opened spar crack. Note clamshell marks at termination of the crack (left). Suspected multiple initiation sites are located between arrows. 1.5×. (c) Section of flange with surface at right. Grain flow in this area was at an angle to applied stress, which resulted in end grain exposure. 105×. (d) SEM fractograph taken between the arrows in (b). Note intergranular fracture pattern indicative of SCC. 95×. (e) SEM taken near the termination of the fracture showing the crack still progressing by SCC. 190×. (f) SEM showing fatigue striations near the crack origin. 235×. (g) Fatigue striations near the termination of cracking. 30×

Investigation. Visual examination of the crack edge shown in Fig. 49(a) revealed that the installation of the fasteners

produced a fit up stress, as indicated by the approximate 0.75-mm (0.03-in.) springback of the flange after the crack propagated through the hardware. Further inspection of the opened fracture (Fig. 49b) showed that the crack had been present for some time because a heavy buildup of corrosion products was seen on the fractured surface. The fracture initiated at multiple origins between the arrows shown in Fig. 49(b). Metallographic examination of the flange in the area of fracture initiation showed the presence of end grain exposure (Fig. 49c), which would promote SCC. Electron optical examination of the fracture shown in Fig. 49(b) produced the scanning electron fractographs shown in Fig. 49(d) to (g). Figures 49(d) and 49(e) show an intergranular topography, while the fractographs in Fig. 49(f) and 49(g) reveal fatigue striations. This clearly shows the flange was cracking by a mixed mode of stress corrosion and fatigue. Chemical analysis of the flange showed that the material met compositional requirements for 7079 aluminum base material. Hardness measurement of 85 HRB showed the material was in the T-6 heat treat condition. Conclusions. It was concluded that the cracking of the flange occurred by a combination of stress corrosion and fatigue.

The cracking was accelerated because of an inadvertent fit up stress during installation. The age of the crack could not be established. However, a reevaluation of prior x-ray inspections in this area would result in some close estimate of the age of the crack. End grain exposure further promoted SCC.

Example 8: SCC of Pitostatic System Connectors.

Pitostatic system connectors were being found cracked on several aircraft. The cracks were not restricted to any particular group of aircraft. Two of the cracked connectors were submitted for failure analysis. Both were reportedly made of 2024T351 aluminum. The connectors had cut pipelike threads that are sealed with teflon-type tape when installed. Investigation. Longitudinal cracks were located near the opening of the female ends of each connector (Fig. 50a). Both

connectors had the same size female end but different size male ends. The connector with the large diameter and longer male end had two cracks, while the connector with the small diameter and shorter male end had only one crack. This size difference was believed to have had no bearing on the cracking.

Fig. 50 Top view (a) of cracked aluminum alloy 2024-T351 pitostatic connectors. Arrows indicate cracks. (b) Cross section of one connector showing elongated grains that were cut to form connector threads. 25× (c) Cross section showing intergranular cracking with multiple branching in one connector. 105×. (d) SEM fractograph showing intergranular cracking and separation of elongated grains. 130×. (e) Cross section of connector threads showing incomplete thread form resulting from improper tapping

The connector with the large male end was sectioned, and part of the fracture was metallographically examined. The connector exhibited an elongated recrystallized grain structure with cut threads (Fig. 50b). A cross section through the fracture showed intergranular cracking and branching of the crack (Fig. 50c), characteristic of SCC. Corrosion deposits were chemically removed from one section of the fracture surface, and the surface was examined in the scanning electron microscope. The fracture surface exhibited intergranular cracking of elongated grains (Fig. 50d). A section of the connector with the large male end and some thin transparent film found on the threads of the connector were chemically analyzed. The connector was determined to be either 2014 or 2017 aluminum alloy, and the film was determined to be fluorinated hydrocarbon teflon-type tape. Hardness checks on both connectors showed the large male end connector to be 75 HRB and the small male end connector to be 77 HRB. Electrical conductivity checks on both connectors showed the large male end connector to have a conductivity of 31% IACS (International Annealed Copper Standard) and the small male end connector to have a conductivity of 27.5% IACS. The threads of all connector components were incompletely formed with a bottom tap and therefore produced a tapered or pipe-type thread. The large male end connector had only one to two threads cut full depth (Fig. 50e). Conclusions. It was concluded that the pitostatic system connectors failed by SCC. The corrodent involved could not be conclusively determined. The stress was caused by forcing the improperly threaded female nut over its fully threaded

male counterpart to effect a seal. The pipelike, incomplete threads produced high hoop stresses when torqued down over a fully formed thread. The one connector tested for chemical composition was not made of 2024 aluminum alloy as reported but of 2017 aluminum. Hardness and conductivity data on both connectors were compatible with a T351 condition for a 2024 alloy. Recommendations. It was recommended that the pitostatic system connector manufacturing process be revised to

produce full-depth threads rather than pseudo pipe threads. It was also recommended that the wall thickness be increased to increase the hoop stress bearing area if pipe threads were to be used. A determination of proper torque values for tightening the connectors was also suggested.

Example 9: Cracking of Aircraft Pylon Strut. A pylon strut was submitted for failure analysis. Cracks were found in two locations on the ears of the strut (Fig. 51a). Because the part was still intact, the cracks had to be forced open so that the fractures could be examined.

Fig. 51 Cracked aluminum alloy 7075-T6 aircraft pylon strut (a) with arrows indicating cracks. (b) SEM of crack C from (a) showing the mud crack pattern indicative of a corrosion mechanism. 820×

Investigation. Chemical analysis and hardness measurements indicated that the strut was 7075-T6 aluminum.

Scanning electron microscopy of the opened cracks showed that the crack surfaces were covered with a mud crack pattern suggestive of SCC (Fig. 51b). The T6 temper is susceptible to SCC. Conclusions. It was concluded that cracking of the strut could have been aggravated by the hard landing experienced by

the aircraft. The strut, however, contained stress-corrosion cracks which were present before the landing. Recommendation. It was recommended that an inspection for SCC be made of all pylon struts with a similar service

life.

Example 10: Corrosion of a Ballast Gas Elbow Assembly. A cadmium-plated 4340 steel ballast elbow assembly (Fig. 52) was submitted to the laboratory for failure analysis. It was requested that a determination be made regarding the element or radical present in an oxidation product found inside the elbow assembly.

Fig. 52 Cadmium-plated 4340 steel ballast gas elbow assembly (a) with arrow showing hole where corrosion was found. (b) SEM of corrosion products on inside hole surface. 430×

Investigation. It was determined by energy-dispersive x-ray analysis in the SEM that iron was the predominant

species, presumably in an oxide form. No cadmium was present inside the hole where oxidation occurred. However, there was cadmium on the outer surface. The inside surface had the appearance of typical corrosion products (Fig. 52b). Hardness measurements indicated that the 4340 steel was heat treated to a strength of approximately 862 MPa (125 ksi). Conclusions. From these analyses, it was concluded that the oxide detected on the ballast elbow was iron oxide. The

possibility that the corrosion products would eventually create a blockage of the affected hole was great considering the small hole diameter (4.2 mm, or in.). It was recommended that a quick fix to stop the corrosion would be to apply a corrosion inhibitor inside the hole. This, however, would cause the possibility of inhibitor buildup and the eventual clogging of the hole. Recommendations. A change in the manufacturing process to include a cadmium plating on the hole inside surface

was recommended. This was to be accomplished in accordance with MIL specification QQ-P-416, Type II, Class 1. A material change to 300-series stainless steel was also recommended. Its strength is slightly lower than that of the 4340 material, but its corrosion resistance is superior, and no surface treatment would be required.

Example 11: Failure of Aircraft Wing Leading Edge Panel. Cracks were found on the wing leading edge of a test aircraft. The cracks were located on the inboard side of the No. 2 and No. 3 engines. Crack lengths were approximately 230 mm (9 in.) long on the left side and approximately 130 mm (5 in.) long on the right side. The cracks ran parallel to the leading edge. The 230-mm (9-in.) crack was received for examination. Investigation. Visual examination of the submitted panel revealed two cracks. One crack ran through six adjacent

fastener holes (Fig. 53a and b). A close-up of the fastener holes shown in Fig. 53(c) and (d) revealed that sections of the beveled edges of the holes were missing; corrosion was evident. Further visual examination of the fastener holes after separation of the crack showed that the fracture faces were corroded.

Fig. 53 Overall view (a) of cracked magnesium alloy AZ31B aircraft wing leading edge panel. Arrows show the length of the crack. (b) Other side of panel shown in (a). A denotes the primary crack; B shows a second, smaller crack. (c) Close-up of fastener holes through which the crack progressed. Note that the two bottom holes are not beveled. (d) Close-up of fastener hole showing cracking at three sites (arrows)

Optical examination of either side of the middle group of fastener holes showed that the area of suspected crack initiation had suffered excessive corrosion. Although a slight beach mark appearance was seen adjacent to the holes, a definite cause for the cracking could not be established. Examination of the holes on the end of the crack showed fracture characteristics typical of fatigue and/or corrosion fatigue. A chemical analysis of the plate material showed the composition to be an AZ31B magnesium alloy. Conclusion and Recommendation. It was concluded that crack propagation of the fracture in the wing panel occurred by a combination of corrosion and high-cycle fatigue in the end fastener holes. Conclusive proof of the cause of failure initiation in the middle fastener holes was masked by excessive corrosion. It was recommended that future panels be manufactured of 2024 aluminum.

Example 12: Failure of an External Tank Pressure/Vent Valve. The pressure/vent valve shown in Fig. 54(a) was submitted for laboratory analysis. The external tank pressure/vent valve regulates the external tank fuel feed system, which transfers fuel under pressure to the internal tanks of the aircraft. It is stated in the technical order that when the solenoid of the pressure/vent valve (part A, Fig. 54b) is energized, service air at 448 kPa (65 psig) shifts the dual-position valve (part B, Fig. 54b).

Fig. 54 Overall view (a) of external tank pressure/vent valve. (b) Partially disassembled valve. A, solenoid switch; B, dual-position valve; C, valve housing. (c) Segment of air line showing residue on check valve poppet (arrow). (d) Close-up view of check valve poppet

It was reported that the dual-position valve was found to be sticking at the intermediate positions. In addition, it was found that service air check valves located on the incoming lines contained poppets that were being stuck in a closed or partially closed position because of suspected corrosion product (Fig. 54c and d). These two factors prevented full pressurization of the external tank and caused a subsequent degradation in fuel flow to the internal tanks. Investigation. Residue taken from the check valve poppet and from the dual-position valve was chemically analyzed.

Only small samples (103 mg) of residue could be obtained from the parts for analysis. For this reason, each of the samples could be analyzed only for a single element. Chloride was selected for the analysis and was present in both samples. The residue found inside the check valve had a reddish color, indicating that the chloride-containing compound reacted with the anodized, dichromate sealed check valve housing. It was noted that the dry-film lubricant applied to the dual-position valve was of different consistency and was heavier than that found on an identical valve assembly also submitted to the laboratory. In addition, the coating of the failed valve was flaking off. Although it did not cause the initial malfunction of the pressure/vent valve assembly, subsequent application of the lubricant without the complete removal of previous coats was a potential for longer-term binding problems. Conclusions. It was suspected that moisture entering the service air lines left a chloride-containing compound upon

evaporation within the air check valves and pressure/vent assembly. This compound subsequently reacted with the check valve housing to lock the check valve poppets in a closed or partially closed position, decreasing the actual pressure being supplied to the pressure/vent valve. Recommendations. It was recommended that an inspection be conducted to ensure that the service air check valves

are operating properly prior to removal and servicing of the pressure/vent valve assembly. It was also recommended that the dry-film lubricant be checked to ensure that it meets specifications for the pressure/vent valve assembly.

Example 13: Corrosion Fatigue of Aircraft Nose Wheels. Four nose wheel failures were received by the laboratory for determination of the cause of failure. The wheels were fabricated from 2014-T6 aluminum material and were cold worked at the flange. Investigation. Visual examination showed that the failure started in the tube well area on the wheel with serial number

31. The failure initiated in the flange fillet on wheels with serial numbers 67, 217, and 250. Figure 55(a) shows a typical example of these failures. Further visual examination of the wheel fractures indicated that failure progressed because of

fatigue (Fig. 55b and c). There was a superficial indentation adjacent to the origin on wheel 31 (Fig. 55d), and there were superficial periodic blemishes on the fillet of nose wheels 67, 217, and 250 (Fig. 55e). The indentation on wheel 31 could have contributed to the cracking found in the tube well; however, the blemishes at the fillet of wheels 67, 217, and 250 were merely superficial and were not thought to be deleterious.

Fig. 55 Aluminum alloy 2014-T6 aircraft nose wheel (a) that failed at the flange. (b) Close-up of tube well on wheel 31. (c) Appearance of flange failure on wheel 67. The topography is typical of other flange failures. (d) Close-up of wheel 31; note indentation (arrow). (e) Close-up of wheel shown in (a); note surface blemishes (arrow). (f) SEM of typical initiation site showing an angular, blocky structure indicative of a corrosion-related failure mechanism. 780×

Scanning electron microscopy examination of the fractures showed that failure initiated by SCC or a corrosion pit on all failures examined. Figure 55(f) shows a typical example. The failures then progressed by fatigue. A chemical analysis conducted on the submitted wheels showed that the wheels met the composition requirements for 2014 aluminum base material. A hardness survey indicated the wheels were in the T-6 tempered condition. The wheels were examined by dye penetrant to determine if the remaining sections contained additional flaws. No additional flaws were seen on the wheels that had failed in the flange area. There was, however, one flaw area in the flange of the wheel that failed in the tube well. This flaw resembled a corrosion pit. Conclusions. It was concluded that failure of nose wheels 67, 217, and 250 was caused by cracking due to SCC or pitting. The failures progressed by fatigue. Because failure occurred in the same general area on all three wheels, these locations are suspect as being underdesigned. Recommendations. It was recommended that consideration be given to redesign of the nose wheel and that additional

service data be accumulated in order to understand the contributing factors that result in failure of the wheel.

Example 14: Failure of Nose Gear Door Bolts. Nose gear door securing bolts were reported to be failing; three separate incidents were cited. One of those bolts, measuring 25 × 32 mm (1 × 1 in.), was submitted for analysis. The bolt was a cadmium-plated, countersunk head type with a common screwdriver slot. Investigation. Figure 56(a) shows the fracture face of the submitted bolt with the nut still attached. The fracture

originated at a thread root and propagated across the cross section. An arrest pattern characteristic of fatigue that terminates at the smooth, featureless final overload area can be seen. The topography of the fracture was excessively rough and more granular than would be expected from pure mechanical fatigue; this indicated an allied corrosion mechanism.

Fig. 56 Fracture surface (a) of failed cadmium-plated 1040 steel nose gear door bolt. The crack propagation pattern (arrow indicates the origin) and topography suggest both fatigue and corrosion. (b) Head of bolt showing cracking (arrow) that would lead to separation in a short time. 3×. (c) SEM of bolt surface showing extensive intergranular SCC. 650×. (d) SEM fractograph showing fatigue striations (arrow) interspersed with secondary cracking and evidence of SCC. 650×

Cracks other than the one leading of failure were observed in the bolt. The large crack at the bolt head shown in Fig. 56(b) was near separation. Metallographic examination of the bolt cross section showed many cracks typical of stress-corrosion damage. The scanning electron micrograph (Fig. 56c) of the bolt surface clearly showed extensive intergranular SCC. Conclusion. The bolt failed by a combination of SCC and fatigue. The dual interaction of stress corrosion and mechanical fatigue is exemplified by the scanning electron photograph of the fracture surface (Fig. 56d), which shows characteristics of both cyclic fatigue and stress corrosion. Recommendations. It was recommended that aerospace-quality fasteners meeting NAS 7104, NAS 7204, or NAS

7504 be used to replace the currently used fasteners.

Example 15: Failure of Tool Steel Pylon Attachment Stud. The failed pylon attachment stud illustrated in Fig. 57(a) was reportedly found during a routine walk-around inspection. Half of the stud was found lying on the apron under the aircraft.

Fig. 57 H-11 tool steel pylon attachment stud (a) that failed by corrosion. (b) Gross pitting corrosion on the stud surface near the fracture site. Note the irregular, mottled appearance of the coating. (c) Fracture surface of the stud showing extent of corrosion within the stud (dark area). The arrow points to one major crack caused by stress corrosion. (d) SEM showing stress-corrosion damage on the fracture face. 315×

Investigation. The stud exhibited gross localized corrosion pitting at several different areas on its surface. Light

general rust was also evident. Severe pitting near the fracture location is illustrated in Fig. 57(b). The extent of the corrosion damage into the stud is shown on the fracture surface (Fig. 57c). The relatively clean and recent area of the fracture face is not entirely overload failure mechanism, but contains evidence of intergranular SCC as well as ductile dimples. Figure 57(d) is a scanning electron micrograph showing intergranular SCC on the fracture face. The bolt is specified to be H-11 tool steel heat treated to 46 to 49 HRC (ultimate tensile strength: 1517 to 1655 MPa, or 220 to 240 ksi). Hardness measurements of 47 HRC confirmed the heat treatment to be correct. The protective coating was found to be an inorganic water-base aluminide coating having a coating thickness of 7.5 to 13 m (0.3 to 0.5 mil). It was noted that the coating was of a nonuniform mottled nature. Conclusions. It was concluded that the failure of the pylon attachment stud was caused by general corrosion followed

by SCC. The stud was not adequately protected against corrosion by the coating. Recommendations. It was recommended that the coating be applied to a thickness of 38 to 75 m (1.5 to 3 mils) to provide long-time corrosion resistance. The coating must be either burnished or cured at 540 °C (1000 °F) to provide cathodic protection to the steel. Other coatings, such as cadmium or aluminum, were also recommended if a thinner coating is needed.

Example 16: Corrosion of a Laser Mirror. A failed laser mirror and another complete mirror of the same construction were submitted to the laboratory for analysis. The laser mirror consisted of three layers of material brazed together to form channels through which the cooling water flows. Portions of the top and middle layers from an area of the mirror that was visually determined to be the most damaged were included in the parts forwarded for analysis (Fig. 58a).

Fig. 58 Section of the most damaged area (a) of a failed laser mirror. (b) Section of area shown in (a) showing corrosion product and deterioration of the base metal. 5×. (c) Corrosion product and nodules in the cooling water channels. 30×. (d) SEM showing corrosion product and intergranular attack on the base metal. 40×. (e) Nodule on upper surface of middle layer of laser mirror showing failure in a ductile/tensile mode. (f) Nodule from lower surface of middle layer of the mirror showing failure due to compressive/shearing stresses

Investigation. Samples were analyzed with light optical and scanning electron microscopy. Portions of the mirror were

sent out for analysis of the base material and visually identified corrosion products. A portion of the middle layer from area A was noted to exhibit the highest corrosion attack, with many visible areas where the corrosion had completely perforated the base material (Fig. 58b). The corrosion product, which appeared under the SEM as distinct granular particles (Fig. 58c), was shown by chemical analysis to contain molybdenum and copper with a trace of gold. The base material was analyzed as molybdenum with negligible alloying additions. The primary mode of corrosion attack on the base material appeared to be intergranular (Fig. 58d), although uniform corrosion was also evident as a general thinning of the material cross section. The corrosion product was shown to adhere loosely to the base material, and it was found that a gentle wash with acetone could dislodge the particles. This finding gave rise to another possible source for material degradation--that of erosion-corrosion, which may in part attribute to the worn, rounded appearance of the modules under the SEM (Fig. 58c). Scanning electron microscopy examination of the portion of the sample with nodules on both sides (middle layer) showed that the nodules on one side failed in a ductile/tensile mode (Fig. 58e), while those on the opposite side appeared to have failed in a compressive/shearing mode (Fig. 58f). These findings are consistent with the concept of micro-bowing

expected to be present in this type of failure, that is, the water pressure forcing apart the two surfaces in the upper layer (tensile) while imparting compressive stresses to the layers below. Portions of the mirror analyzed were seen to have areas in which the brazing alloy was corroded but the base material was relatively unaffected. This finding suggested the possibility of a galvanic couple between the brazing alloy and the base metal. The sample mirror was radiographed and ultrasonically inspected. The radiograph did not show any areas of corrosion, with the ultrasonic c-scan able to image the rectangular pattern of pin-type spacers. Conclusions. It was concluded that the corrosion attack noted sufficiently weakened the base material and the brazed joints, allowing catastrophic failure of the mirror due to the pressure of the cooling water. Recommendations. It was recommended that the mirrors be cleaned of all corrosion products present as a result of

past service conditions. The mirrors should be proof tested to determine if the residual structural integrity is sufficient for future operational requirements. It was recommended that the water system consisting of deionized water and formaldehyde be replaced with water having a low oxygen content and a cathodic inhibitor (oxygen scavenger).

Example 17: Failure of a Helicopter Rotor. Several rotor blade components were received for laboratory analysis. These included the horizontal hinge pin and the associated nut the locking washer (Fig. 27a, b, and c in the section "Corrosion of Airframes" in this article). Investigation. Visual examination of the submitted parts revealed that the hinge pin, fabricated from 4340 steel, was

broken and that the fracture face showed a flat beach mark pattern indicative of a preexisting crack (Fig. 27d and e). Also noteworthy on the hinge pin was a ding or dent in the threaded area that occurred on only one side (Fig. 27f). Optical examination of the hinge pin showed that the beach mark or preexisting crack had progressed to approximately one-fourth the total cylinder cross-sectional area. Further examination of the area at the fracture initiation site revealed corrosion pits and secondary cracking on the outer circumference of the pin adjacent to the primary fracture face. The cracks ran parallel to the primary fracture and propagated from small pits (Fig. 27g). Close scrutiny of the threaded area of the pin revealed an embedded thread that did not appear to come from the pin (Fig. 27h). A chemical analysis was conducted on the embedded thread and on an associated attachment to determine the origin of the thread. This analysis showed that the thread and nut were 4140 steel. Scanning electron fractographic examination of the fracture initiation site strongly suggested that the fracture progressed by fatigue. Hardness measurements taken on the pin produced a value of 41 HRC, which was within the 39 to 43 HRC drawing requirements. Conclusions. It was concluded that the failure of the horizontal hinge pin initiated at areas of localized corrosion pits.

The pits in turn initiated fatigue cracks, resulting in a failure mode of corrosion fatigue. The embedded thread was the same material as the associated nut and probably came from the attachment nut. Recommendations. It was recommended that all of the horizontal hinge pins be inspected to determine the current

existence of cracks, corrosion pits, and general corrosion. Those pins that are determined to be satisfactory for further use should be stripped of cadmium, shot peened, and coated with cadmium to a minimum thickness of 0.0127 mm (0.0005 in.).

Example 18: Corrosion Failure of Wing Flap Hinge Bearings. Three wing flap hinge bearings were received by the laboratory for analysis (Fig. 59a). The bearings were fabricated from chromium-plated type 440C martensitic stainless steel.

Fig. 59 Cracked type 440C stainless steel (a) aircraft wing flap hinge bearings. (b) Crack configuration of bearing 1 from (a). (c) Crack configuration of bearing 2 from (a). (d) Fracture surface of second crack in bearing 1. Arrow shows the probable fracture origin. 2.5×. (e) Fracture surface of bearing 2. The crack front was abraded, so the area of origin would not be clearly defined. 2.5×. (f) Corrosion pits on the inside surface of one of the bearings. All three bearings submitted had this type of damage. 10×. (g) Cross section of corrosion pit on one of the bearings; note the intergranular attack. 265×. (h) SEM fractograph of bearing 1 showing intergranular mechanism of fracture. 2690×. (i) SEM fractograph of bearing 2 showing intergranular fracture mechanism. Corrosion products are evident on the intergranular facets. 2690×

Investigation. Visual analysis of the hardware showed that two of the three bearings were cracked. One of these two

bearings, designated bearing 1 in Fig. 59(a), contained two separate cracks. One of these cracks took the path shown in Fig. 59(b), while the second crack progressed straight across the bearing cross section. Bearing 2 in Fig. 59(a) exhibited the crack path seen in Fig. 59(c). Further visual examination of the bearing crack fractures showed that the cracks were flat and slightly fibrous (Fig. 59d and e). The fracture face of bearing 2 differed slightly from the two fractures seen on bearing 1 in that there was a chip missing on the outer edge of the bearing. Optical examination of the bearings revealed numerous corrosion pits on the inner diameter of the failed bearings and on bearing 3, which had not failed (Fig. 59f). Metallographic examination showed that the pitting progressed in an intergranular corrosive pattern (Fig. 59g). The microstructure was acceptable for type 440C martensitic stainless steel. Electron optical examination of the fractures by transmission electron microscopy produced fractographs that depicted an intergranular failure mechanism (Fig. 59h and i). Chemical analyses and hardness measurements of bearings 1 and 2 showed the material to be within compositional requirements for type 440C stainless steel and heat treated to 54 to 55 HRC. Conclusion. The intergranular fracture pattern seen in the electron fractographs, coupled with the corrosion pits

observed on the inner diameter of the bearings, strongly suggested that the failure initiated by pitting and progressed by SCC or hydrogen embrittlement from the plating operation.

Recommendations. It was recommended that the extent of the flap hinge bearing cracking problem be determined by

using nondestructive inspection because it is possible to crack hardened type 440C during the chromium plating process. An inspection for pitting on the bearing inner diameter was also recommended. It was suggested that electroless nickel be used as a coating for the entire bearing. Electroless nickel, when heated to 345 °C (650 °F) for 1 h after plating, has wear resistance equivalent to that of chromium. A review of the chromium plating and baking sequence was also recommended to ensure that a source of hydrogen is not introduced during the plating operation.

References 1. M.L. Bauccio, Case Studies of Material Deterioration and Failure, in Materials Selection for Military Equipment (A Handbook of Critical Properties and Applications), National Association of Corrosion Engineers, 1987, p 1 2. L.M. Bland, Aircraft Structural Life Monitoring and the Problem of Corrosion, in Aircraft Structural Fatigue, Structures Report 363, Materials Report AR-000-724, Proceedings of a symposium held in Melbourne, Oct 1976, Aeronautical Research Laboratories, 1977, p 227-253 3. V.S. Agarwala, D.A. Berman, and G. Kohlhaas, Causes and Prevention of Structural Materials Failures in Naval Environments, Mater. Perform., Vol 24 (No. 6), 1985, p 9-16 4. W.T. Kirkby, Examples of Aircraft Failure, in Fracture Mechanics of Aircraft Structures, AGARDograph 176, H. Liebowitz, Ed., Advisory Group for Aerospace Research and Development, 1974, p 8-13 5. W.R. Coleman, R.J. Block, and R.D. Daniels, Corrosion Problems in Aircraft Components--Case Studies of Failures, in Proceedings of the 1980 Tri-Service Corrosion Conference, AFWAL-TR-4019, Vol, II, Air Force Wright Aeronautical Laboratories, 1981, p 241-270 6. Lockheed Grounds C-5B to Replace Fasteners, Aviation Week Space Technol., Vol 123 NO. 13), 1985, p 26 7. W. Beck, E.J. Jankowsky, and P. Fischer, "Hydrogen Stress Cracking of High Strength Steels," Report NADC-MA-7140, Naval Air Development Center, 1971 8. D.A. Berman, The Effect of Baking and Stress on the Hydrogen Content of Cadmium Plated High Strength Steels, Mater. Perform., Vol 24 (No. 11), 1985, p 36-41 9. D.E. Muehlberger and J.J. Reilly, "Improved Equipment Productivity Increases Applications for Ion Vapor Deposition of Aluminum," Report 830691, Society of Automotive Engineers, 1983 10. "Coating, Aluminum, Ion Vapor Deposited," MIL-C-83488B, U.S. Air Force Materials Laboratory, Dec 1978 11. D.P. Lahiri and A.V. Reddy, Exfoliation Corrosion in Aircraft Structural Member Made of Al-Cu-Mg Alloy, Trans. Indian Inst. Met., Vol 35 (No. 5), 1982, p 456-460 12. R.N. Miller, Inhibitive Sealing Compounds and Coating Systems Solve Aircraft Corrosion Problems, SAMPE J., April/May 1970, p 54-58 13. M.H. Trimble, The Need for Improved Materials in Integral Aircraft Fuel Tanks, in Materials and Processes in Service Performance, Society for the Advancement of Material and Process Engineering, 1977, p 3-8 14. N. Kackley and M. Levy, "Combatting Corrosion in Army Aircraft," Paper presented at Corrosion/87, San Francisco, CA, National Association of Corrosion Engineers, March 1987 15. "Materials and Processes for Corrosion Prevention and Control in Aerospace Weapons Systems," MILSTD-1568, Revision A, Air Force Wright Aeronautical Laboratories, Oct 1979 16. Aircraft Corrosion: Causes and Case Histories, Vol 1, AGARD Corrosion Handbook, AGARD-AG-278, Advisory Group for Aerospace Research and Development, 1985 17. "Finishes for Ground Electronic Equipment," MIL-F-14072, U.S. Army Electronics Research and Development Command, June 1986 18. M.L. Bauccio, "Properties and Applications of Beryllium and Beryllium-Copper Aerospace Materials," Paper 109, presented at Corrosion/84, New Orleans, LA, National Association of Corrosion Engineers, 1984 19. J. Fielding, Stress Corrosion Cracking and the Aircraft Industry, J. Inst. Met., Vol 101 (No. 9), 1973, p 238240

20. M.L. Bauccio, Corrosion Problems in the U.S. Army BLACK HAWK Helicopter, in Proceedings of the 1980 Tri-Service Corrosion Conference, AFWAL-TR-81-4019, Vol I, Air Force Wright Aeronautical Laboratories, 1981 21. T.A. Roach, Aerospace High Performance Fasteners Resist Stress Corrosion Cracking, Mater. Perform., Vol 23 (No. 9), 1984, p 42-45 22. V.R. Pludek, Design and Corrosion Control, Macmillan, 1977, p 68 23. W. Barrois, Service Failures and Laboratory Tests, in Fracture Mechanics of Aircraft Structures, AGARDograph 176, H. Liebowitz, Ed., Advisory Group for Aerospace Research and Development, 1974, p 325-345 24. P.M. Toor, Inhibition of Stress Corrosion Cracking in the Design of Aircraft Structures, in Materials and Processes--Continuing Innovations, 28th National SAMPE Symposium and Exhibition, Society for the Advancement of Material and Process Engineering, 1983, p 973-981 25. H. Lajain, Corrosion Protection Schemes for Aircraft Structures: Some Examples for the Corrosion Behavior of Al-Alloys, in Aircraft Corrosion, AGARD-CP-315, Proceedings of the 52nd Meeting of the AGARD Structures and Materials Panel, Advisory Group for Aerospace Research and Development, 1981, p 14-1 to 14-16 26. W.M. Ryan, Filiform Corrosion on Painted Aluminum Alloy Surfaces, in The Enigma of the Eighties: Environment, Economics, Energy, Vol 24, Book 1 of 2, Proceedings of the 24th National SAMPE Symposium and Exhibition, Society for the Advancement of Material and Process Engineering, 1979, p 638-648 27. R.B. Waterhouse, Fretting Corrosion, Pergamon Press, 1972, p 1-5 and 36-43 28. Aeronautical Information Report 47, Society of Automotive Engineers, 1956 29. M. Levy and P.A.M. Farrell, The CPC Mission, U.S. Army Corrosion/Deterioration Problems, U.S. Army Man Tech J., Vol 10 (No. 4), 1985, p 5-14 30. R.J.H. Wanhill, J.J. DeLuccia, and L.B. Vogelesang, Environmental Fatigue of Aluminum Alloy Structural Joints, in Seventh International Light Metals Conference (Leoben, Vienna), June 1981, p 92-93 31. Military Standardization Handbook, Corrosion and Corrosion Prevention, Metals, MIL-HDBK-729, Department of Defense, U.S. Army Materials Technology Laboratory, 1983, p 67 32. R.N. Miller, The Evolution of the Corrosion Free Airplane, Mater. Perform., Vol 25 (No. 3), 1986, p 57-59 33. B. Rosales and M. Del Carmen, "The Predominance of Microbial Growth Versus Metallurgical Characteristics of the Corrosion of 2024 Al Alloy Through Electrochemical Data," Paper 124, presented at Corrosion/86, National Association of Corrosion Engineers, 1986 34. Wing Tank Microbial Growth and Corrosion, Proceedings of the Boeing/Airline Regional Conference, Boeing Commercial Airplane Company, 1980 35. "Corrosion Control for Aircraft," Advisory Circular 43-4, U.S. Department of Transportation, Federal Aviation Administration, 1973, p 10 36. "Lessons Learned Bulletin: Corrosion," Directorate of Systems Support, Air Force Acquisition Logistics Center, 1985, p 83 37. L.D. Griffin and D. Latterman, C-141A Service Experience--Materials and Processes, SAMPE J., Vol 14 (No. 2) 1978, p 9-19 38. J.S. Leak, Corrosion--A Study of Recent Air Force Experience, in Proceedings of the NACE 26th Conference, National Association of Corrosion Engineers, 1970, p 497-503 39. L.D. Griffin and D. Latterman, C-141A Service Experience--Materials and Processes, SAMPE J., Vol 14 (No. 2), 1978, p 9-16 40. V. Wigotsky, Metal Bonding Slowly Winning Over Reluctant Designers, Aerospace Am., Vol 22 (No. 9), 1984, p 82-86 41. .R.E. Horton, Demonstration of an Improved Method for Repair of Bonded Aircraft Structure, in The Enigma of the Eighties: Environment, Economics, Energy, Vol 24, Book 1 of 2, Proceedings of the 24th National SAMPE Symposium and Exhibition, Society for the Advancement of Material and Process Engineering, 1979, p 659-668

42. M.H. Kuperman, Structural Adhesive Bond Repair of Aircraft Flight Control Surfaces, in The Enigma of the Eighties: Environment, Economics, Energy, Vol 24, Book 2 of 2, Proceedings of the 24th National SAMPE Symposium and Exhibition, Society for the Advancement of Material and Process Engineering, 1979, p 1126-1139 43. H.J. Singletary, "Aircraft Service Life Extension Through Corrosion Control," Paper presented at Corrosion/87, San Francisco, CA, National Association of Corrosion Engineers, March 1987 44. "Corrosion Preventive Characteristics of Aircraft Greases," Aeronautical Information Report 40, Society of Automotive Engineers, 1955 45. J.M. Margolis, Advanced Composites for Primary Exterior Aircraft Structures, Plast. Des. Forum, Vol 10 (No. 5), 1985, p 78-84 46. M.A. Steinberg, Materials for Aerospace, Sci. Am., Vol 255 (No. 4), 1986, p 67-72 47. Corrosion Conference, Boeing/Airplane Regional Conferences, Boeing Commercial Airplane Company, 1982, p IV.9-IV.10 48. C.R. Pye, Recorder's Report--Session II, in Aircraft Corrosion, Conference Proceedings 315, Advisory Group for Aerospace Research and Development, 1981, p R2-1-R2-2 49. A.E. Hohman, Design and Manufacturing Practices to Minimize Corrosion in Aircraft, R-714, Advisory Group for Aerospace Research and Development, March 1984, p 7.1-7.5 50. J.J. DeLuccia and S.F. Saletros, Naval Air Systems Command Corrosion Control Program, in Proceedings of the 1980 Tri-Service Conference on Corrosion, Vol II, F.H. Meyer, Ed., AFML-TR-81-4019, Air Force Materials Laboratory, Nov 1980 51. A.A. Baker and M.M. Hutchison, Fibre Composite Reinforcement of Cracked Aircraft Structures, in Aircraft Structural Fatigue, Structures Report 363, Materials Report 104, Proceedings of a symposium held in Melbourne, Oct 1976, Aeronautical Research Laboratories, 1977, p 419-457 52. M.L. Bauccio, "Hot Corrosion: A Review of Chemical Mechanisms and Protective Coatings," Paper 109, presented at Corrosion/82, National Association of Corrosion Engineers, 1982 53. Airframe and Powerplant Mechanics Powerplant Handbook, U.S. Department of Transportation, Federal Aviation Administration, 1976, p 20 54. "Corrosion Control for Aircraft," Advisory Circular 43-4, U.S. Department of Transportation, Federal Aviation Administration, 1973, p 14 55. Aircraft Corrosion: Causes and Case Histories, Vol 1, AGARD Corrosion Handbook, AGARD-AG-278, Advisory Group of Aerospace Research and Development, 1985 56. M. Gell, G.R. Leverant, and C.H. Wells, The Fatigue Strength of Nickel-Base Superalloys, in Achievement of High Fatigue Resistance in Metals and Alloys, STP 467, American Society for Testing and Materials, 1970, p 113-153 57. P.R. Belcher, R.J. Bird, and R.W. Wilson, "Black Plague" Corrosion of Aircraft Turbine Blades, in Hot Corrosion Problems Associated With Gas Turbines, STP 421, American Society for Testing and Materials, 1967, p 123-145 58. H. Grimm and M.P. Malik, Corrosion of Aircraft Structures: Significant Problems, Control Methods, and Possible Improvements, in Proceedings of the Sixth European Congress on Metallic Corrosion (Eurocorr '77), Society of the Chemical Industry, 1977, P 595-606 59. P.A. Bergman, Corrosion Problems in Aircraft Gas Turbines, in Proceedings of the Air Force Materials Laboratory Fiftieth Anniversary Technical Conference on Corrosion of Military and Aerospace Equipment, AFML-TR-67-329, Air Force Materials Laboratory, 1967, p 1413-1431 60. R.R. Dils and P.S. Follansbee, Dynamic Oxidation and Corrosion in Power Generating Units, Corrosion, Vol 33 (No. 11), 1977, p 385-402 61. New Coating Combats "Hot Corrosion" in Gas Turbine Buckets, Anti-Corros. Methods Mater., Vol 27 (No. 12), 1980, p 7 62. R.A. Miller and C.E. Lowell, "Failure Mechanisms of Thermal Barrier Coatings Exposed to Elevated Temperatures," Technical Memorandum 82905, National Aeronautics and Space Administration, 1982 63. D.F. Baxter, Jr., Automotive Technology: Developments in Coated Steels, Met. Prog., Vol 129 (No. 6),

1986, p 31-35 64. J.H. Weber, High Temperature Oxide Dispersion Strengthened Alloys, in The 1980's--Payoff Decade for Advanced Materials, Vol 25, Proceedings of the National SAMPE Symposium and Exhibition, Society for the Advancement of Material and Process Engineering, 1980, p 752-764 65. D.R. Coupland, C.W. Corti, and G.L. Selman, The PGM Concept: Enhanced Corrosion Resistant Superalloys for Industrial and Aerospace Applications, in Proceedings of the Conference on Behavior of High Temperature Alloys in Aggressive Environments, Petten, Oct 1979, p 525-536 66. H.R. Gray, Hot-Salt Stress-Corrosion of Titanium Alloys as Related to Turbine Engine Operation, in Proceedings of the Second International Conference on Titanium Science and Technology, Vol 4, 1973, p 2627-2638 67. J.R. Myers, H.B. Bomberger, and F.H. Froes, Corrosion Behavior and Use of Titanium and Its Alloys, J. Met., Oct 1984 68. S.P. Rideout, M.R. Louthan, Jr., and C.L. Selby, Basic Mechanisms of Stress Corrosion Cracking of Titanium, in Stress-Corrosion Cracking of Titanium, STP 397, American Society for Testing and Materials, 1966, p 137-151 69. R.E. Duttweiler, R.R. Wagner, and K.C. Antony, An Investigation of Stress-Corrosion Failures in Titanium Compressor Components, in Stress-Corrosion Cracking of Titanium, STP 397, American Society for Testing and Materials, 1966, p 152-178 70. T.R. Strobridge, J.C. Moulder, and A.F. Clark, "Titanium Combustion in Turbine Engines," Reports FAARD-79-51 and NBSIR 79-1616, U.S. Department of Transportation, Federal Aviation Administration, July 1979 71. C.W. Elrod, The Combustion of Titanium in Gas Turbine Engines, in Proceedings of the 1980 Tri-Service Conference on Corrosion, AFWAL-TR-81-4019, Vol II, Air Force Wright Aeronautical Laboratories, 1981, p 55-92 72. B.A. Manty, V.G. Anderson and H.M. Hodgens, Blade Tip Treatment--Titanium Alloy Compressor Blades, AFWAL-TR-80-4149, Air Force Wright Aeronautical Laboratories, 1980, p 119 73. R.B. Waterhouse, Fretting Corrosion, Pergamon Press, 1972, p 36 74. R.L. Johnson and R.C. Bill, Fretting in Aircraft Turbine Engines, in Fretting in Aircraft Systems, AGARDCP-161, Advisory Group for Aerospace Research and Development, 1975 75. R.G. Mahorter, Cold Corrosion in Aircraft Engines, in Proceedings of the 1976 Tri-Service Conference on Corrosion, S.J. Ketcham, Ed., Naval Air Development Center, 1976, p 257-274 76. R.G. Mahorter, Current and Anticipated Materials Problems of Aircraft Engines in Marine Atmospheres, in Proceedings of the 1974 Conference on Gas Turbine Materials in the Marine Environment, MCIC-75-27, Metals and Ceramics Information Center, 1975, p 1-10 77. E.J. Hammersley, Corrosion in Airframes, Power Plants and Associated Aircraft Equipment, in The Theory, Significance, and Prevention of Corrosion in Aircraft, AGARD-LS-84, Advisory Group for Aerospace Research and Development, 1976, p 4-11 78. M.P. Malik, Corrosion of Aircraft Structures--Significant Problems, Control Methods and Possible Improvements, in Proceedings of the 6th European Congress on Metallic Corrosion (Eurocorr '77), Society of the Chemical Industry, 1977, p 602-603 79. R. Baboian et al., Aluminum Coating of Steel, in Surface Cleaning, Finishing and Coating, Vol 5, 9th ed., Metals Handbook, American Society for Metals, 1982, p 339 80. D.E. Muehlberger, Presentation (seminar) given on the technology of ion-vapor-deposition of aluminum. Bell Helicopter TEXTRON, Inc., May 1986 81. J.E. Newhart, "Cold" Corrosion Evaluations Resulting from Full-Scale, Controlled Environment, Turbine Engine Testing, in Proceedings of the 1972 Tri-Service Conference on Corrosion, MCIC-73-19, M.M. Jacobson and A. Gallaccio, Ed., Metals and Ceramics Information Center, 1973, p 237-238 82. S.D Elrod. "Service Evaluation of Aluminum-Brazed Titanium (ABTi) Jet Engine Tailpipe Extensions," NASA Contractor Report 3617, National Aeronautics and Space Administration, Scientific and Technical Information Branch, 1982

83. L.C. Lipp, "Halogenated Solvent-Induced Corrosion in Hydraulic Systems," Preprint 78-AM-4A-2, American Society of Lubrication Engineers, 1978 84. H. Weltman, Corrosion Characteristics and Control of Mixed Hydrazine Fuel, in Proceedings of the 1980 Tri-Service Conference on Corrosion, AFWAL-TR-4019, Air Force Wright Aeronautical Laboratories, 1981, p 343-358 85. B. Cohen, Corrosion Fatigue in the Aerospace Industry, in Corrosion Fatigue: Chemistry, Mechanisms and Microstructure, National Association of Corrosion Engineers, 1972, p 65-83

Corrosion in the Aerospace Industry Introduction CORROSION CONTROL is an essential part of the design of aerospace hardware, because a corrosion failure could result in the loss of a satellite, a booster, or the space shuttle orbiter. Of the three, satellites experience the most benign environments; they are usually handled in clean rooms with controlled humidity from the time they are built until they are launched. In low earth orbit, the corrosion effects of oxygen and moisture no longer exist, although some surface erosion from atomic oxygen can destroy optical (thermal control) surfaces. At the other extreme, a booster, such as the solid rocket motor case of the space shuttle system, becomes immersed in seawater after launch and must be pulled from the ocean and thoroughly cleaned to avoid salt-water corrosion. A solid rocket case is expected to be reused up to 19 additional times. The space shuttle orbiter was originally designed to be maintenance free for 10 years. From a corrosion standpoint, this is a rather severe requirement for the highly corrosive Cape Canaveral environment. Because of unexpected delays in launching schedules, there is a high probability that shuttle orbiters will see service in excess of 20 years. Corrosion and structural inspection plans have recently been initiated to ensure the safety of the airframe as well as the systems, which are subjected to highly corrosive fluids. In these systems, corrosion failure from perforation could result in explosions or fire, or corrosion deposits in valves could cause malfunctions, possibly resulting in the loss of the spacecraft. Therefore, control of corrosion is considered to be vital for the safety and success of aerospace hardware. Applicable specifications imposed by the National Aeronautics and Space Administration (NASA) on subcontractors to control or restrict material selection and to control corrosion are: • • • • •

SE-R-0006: NASA-JSC Requirements for Materials and Processes MSFC-SPEC-250(1): Protective Finishes for Space Vehicle Structures and Associated Flight Equipment, General Specification for MSFC-SPEC-522A: Design Criteria for Controlling Stress Corrosion Cracking NHB-8060.1: Flammability, Odor, and Offgassing Requirements and Test Procedures for Materials in Environments That Support Combustion JSC-30233: General Specification--Space Station Requirements for Materials and Processes

The following sections will detail the steps taken by engineering, design, and material and process groups to avoid spacecraft corrosion. Case histories of failures and their solutions will also be reviewed. Corrosion of Manned Spacecraft L.J. Korb, Rockwell International, Space Transportation & Systems Division

The prevention or control of the corrosion of manned spacecraft presents a wide variety of challenges to the corrosion engineer for at least four reasons. First, a manned spacecraft, such as the space shuttle orbiter, has structures and systems

operating at temperatures ranging from -253 °C (-423 °F) for liquid hydrogen tanks for approximately 1455 °C (2650 °F) for critical metallic pressure ports of the nose cap. Several hundred different alloys are required, including the light metals, such as beryllium, magnesium, aluminum, and titanium; steels, such as low-alloy, tool, corrosion resistant, precipitation hardenable, and maraging; nickel and nickel-base alloys, including pure nickel, Monel alloys, Inconel alloys, and other nickel- and cobalt-base superalloys; the refractory metals, principally niobium and molybdenum; the copperbase alloys, including pure coppers, beryllium coppers, bronzes, and brasses; the precious metals; and a variety of plating alloys. Quite often, intimate knowledge of the corrosion behavior of these alloys in the intended environment is simply lacking. Second, because weight is so critical for manned spacecraft, the alloys were often chosen for their inherent high strengthto-density ratios rather than for their corrosion resistance. Early studies indicated that the savings per pound of weight in the orbiter vehicle structure of systems amounted to well over $30,000 during the mission life. Therefore, emphasis was directed toward corrosion control coatings for airframe structures rather than the selection of corrosion-resistant alloys. In fluid systems, however, material compatibility played an important role. Third, the various parts of a spacecraft are subject to a wide variety of environments in addition to the temperature regimes discussed above. The seacoast exposure at Cape Kennedy, for example, is very severe because of the heat, high humidity, salt air, and the daily condensation of dew onto the structures. Fluid systems, on the other hand, are exposed to very aggressive chemicals: hot gaseous oxygen, nitrogen tetroxide, high-pressure hydrogen, hydrazine, and ammonia. Finally, corrosion protection often must be subordinated to other critical functions, such as the spacecraft electrical grounding requirements. Often, where a corrosion engineer may wish to isolate a galvanic couple electrically by using an insulator, the requirements for grounding to ensure lightning protection and control of electromagnetic interference must overrule. Corrosion-protective measures must then be taken for the area of the grounded connection. Early manned spacecraft, such as the mercury, Gemini, Apollo, and Sky Lab programs, were each designed for a single launch. The Mercury, Gemini, and Apollo capsules parachuted into the ocean and could not be refurbished economically. The Skylab, having no reentry capability, broke up and burned at the end of its useful life. The major emphasis in corrosion protection for these vehicles was to ensure freedom from general corrosion and stress corrosion for the 3 to 5 years preceding launch. In the case of the Apollo, and in-depth age life and corrosion reevaluation was made to ensure a 10-year operational life from its manufacture to its use on Skylab missions. The space shuttle orbiter was originally designed to avoid corrosion for a 10-year life, and it was anticipated that it would fulfill its 100-mission launch capability during this period. Because of unforeseen launch problems, some orbiters have experienced less than 10 missions while approaching their 10-year design life. Reevaluations will be made to extend this life to 20 years and to maintain the corrosion protection systems as necessary. Because of its multiple-launch capability and its long design life, the space shuttle orbiter is unquestionably the most challenging of the manned spacecraft from a corrosion control standpoint. Its corrosion protection and the problems experienced will be discussed in detail. However, because other manned space vehicles have experienced some unusual corrosion problems, many of which involved unknown incompatibilities at that time, this article will also cover representative case histories of corrosion of several manned spacecraft.

Operational Mission of the Space Shuttle Orbiter The space shuttle system is designed to provide an economical system for transporting personnel and supplies into low earth orbit. The system (Fig. 1) consists of four major elements: the solid rocket boosters, the space shuttle main engines, the external tank, and the space shuttle orbiter. Elements other than the orbiter itself are discussed in the section "Corrosion of Space Boosters and Space Satellites" in this article. The system is capable of launching up to 29,500 kg (65,000 lb) of payload into orbit and returning up to 14,500 kg (32,000 lb) of payload from orbit (Ref 1).

Fig. 1 Space shuttle system. The orbiter weight includes the 165-klb orbiter (including engines) plus 37-klb crew, consumables, and propellants. The external tank diameter includes external insulation. klb = 1000 lb

Nearly 89% of the thrust for launch is provided by the solid rocket boosters. The remaining 11% is delivered by the main engines, which burn hydrogen with oxygen supplied by the external tank. The orbiter, launched vertically in a piggyback position on the external tank, houses the astronauts, the main engines, and the payloads. The orbiter must function as both a spacecraft and an aircraft. During entry from orbit, it must be protected from temperatures exceeding 1260 °C (2300 °F) on its lower fuselage and in excess of 1455 °C (2650 °F) along the leading edges and nose cap. At an altitude of approximately 46 km (150,000 ft), the orbiter will slow to about eight times the speed of sound and will pass through its maximum heating. At 15 km (50,000 ft), the orbiter will be maneuvered aerodynamically to land as a glider. A typical mission profile is shown in Fig. 2.

Fig. 2 Profile of shuttle mission. Each shuttle orbiter can fly a minimum of 100 missions and carry as much as 29,500 kg (65,000 lb) of cargo and up to seven crew members into orbit. It can return 14,500 kg (32,000 lb) of cargo to earth. klb = 1000 lb; klbf = 1000 lbf

The requirement to achieve a minimum-weight orbiter (68,000 kg, or 150,000 lb, dry weight) has necessitated use of the most efficient structural materials and processes. The requirement for 100-mission reuse has extended advancements in thermal protection materials well beyond the state-of-the-art existing at the inception of the design.

Corrosion Control Program: Space Shuttle Orbiter The key to a successful corrosion control program for the space shuttle orbiter was to develop sound technical and management programs. Although the major structural parts of the orbiter, such as the wings, tail, fuselage, and cabin, were manufactured by only a few companies, it was estimated that more than 20,000 suppliers were responsible for providing systems and parts for the vehicle. It was necessary to review and control all orbiter parts to provide the high levels of reliability required. The material and process management program consisted of the following key elements:

• • • • • •

A material and process group in engineering A drawing review system requiring sign-off by a materials engineer A tracking system for all materials An orbiter materials and process control specification A corrosion-control and finishes specification A stress-corrosion control plan

Each material application was reviewed by a qualified material and process engineer who had sign-off authority on drawing, engineering orders, and material rework dispositions. A material tracking system was set up at the inception of the program to prevent 12 material-related hazards from occurring on the orbiter. These include controls for atmospheric corrosion and stress corrosion, fluid and propellent incompatibilities, age life, flammability, toxicity, offgassing, and condensation of volatile condensible matters. Materials and finishes were identified, evaluated, and, when accepted, entered into a computer. Material identification even included those materials used in minute quantities, such as the ink used to stamp part numbers or the nearly invisible cetyl alcohol lubricants on fasteners. A master directory (index) of the behavior of each material in the 12 hazardous categories was maintained and used as a reference. In each hazardous category, a series of encoded, acceptable engineering approaches for each "buyoff" was listed to assist the engineer. For example, a part may be made from a material having a stress-corrosion threshold of 50% of its tensile yield strength, yet be acceptable because: • • •

It is adequately coated It experiences no significant tensile stresses in the critical stress-corrosion direction (including residual and installation stresses) It is in a benign environment, such as the cabin

In a few cases, the complexity of the part, such as a motor, precluded a separate evaluation of each material, and the entire configuration was qualification tested in its intended-use environment to avoid these hazards. A Material and Process Control Specification (MC999-0096) was placed on all major subcontractors. A similar specification controlled parts that were designed and manufactured in-house. These specifications included controls for fluid systems compatibility, stress corrosion, atmospheric corrosion, and galvanic corrosion. The controls imposed are summarized below. Control of Fluid Systems Compatibility. A fluid systems compatibility analysis is required that covers all fluids

and materials used in the system, such as testing, processing, inspection, and operation, along with known or expected trace contaminants. Fluid system compatibility refers to interaction problems involved with materials and the liquid or gaseous subsystems. The problems experienced generally fall into the following categories: • • • •

• •

Autoignition: Spontaneous ignition of the material or the fluid Impact ignition: Ignition brought about by shock or impact within the fluid Catalytic reaction: Reactions such as the catalytic decomposition of the fluid Material degradation: This includes such phenomena as chemical attack, corrosion, galvanic corrosion, stress corrosion, hydrogen embrittlement, and crack growth acceleration with metals and includes embrittlement, abnormal swelling, leaching of plasticizers, ultraviolet degradation, and so on, with nonmetallic materials Fluid degradation: Reactions in which the physical or chemical characteristics of the fluid are altered Potential ignition: Ignition due to proximity to electrical ignition sources

Materials selection was required to minimize the compatibility problems with the fluid systems. Material-fluid combinations that result in autoignition, impact ignition, or another catastrophic mode of failure were not permitted.

The use of electrical and electronic components exposed to nondielectric fluid systems was avoided. Buyer approval was required prior to using electronic components in hazardous fluid systems. Metallic materials listed in Appendix I of MC999-0096 are rated for compatibility with gaseous oxygen (GOX), liquid oxygen (LOX), nitrogen tetroxide (N2O4), hydrazine (N2H4), monomethyl hydrazine (MMH), and low-pressure ( 3.1 MPa, or 450 psi) and high-pressure (>3.1 MPa, or 450 psi) hydrogen. Nonmetallic materials listed in Appendix II of MC999-0096 are rated for compatibility with low- and high-pressure GOX, LOX, N2O4, N2H4, MMH, liquid hydrogen, and hydraulic fluid. Materials that are compatible and noncompatible with titanium are listed in MF0004-018. Lubricants for static service with special fluids (application to elastomers, metals, and threads) are:

Fluid Ammonia Deionized water Freon 21 FC 40

Lubricant Krytox 240 AC; Braycoat 3L-38RP; Braycoat 815Z oil Krytox 240 AC; Braycoat 3L-38RP; Braycoat 815Z oil DC F-6-1101 DC F-6-1101

Lubricants for dynamic service with special fluids must be resolved on an individual basis. Use of the above lists did not absolve the seller of full responsibility for verifying compatibility under the particular design conditions used by his fluid system. Control of Stress Corrosion. The subcontractor was required to prepare a stress-corrosion plan utilizing MSFC Specification 522A as a guideline for controlling stress corrosion and to take the actions necessary to prevent such failures. Wherever possible, the supplier was required to select materials that are either not susceptible to stress corrosion or have a high resistance to stress corrosion in the anticipated life cycle environment. Where susceptible materials were used, the supplier was required, at the minimum, to take the following actions to reduce stress-corrosion problems to the extent feasible:

• • • • • • • •

Select less susceptible alloys, tempers, or clad products Reduce sustained stress levels on the part below stress-corrosion threshold levels, especially in the more susceptible short-transverse grain direction Protect the part from the detrimental environment by hermetically sealing or coating the part or by inhibiting the environment (closed system) Avoid or reduce residual stresses in parts or assemblies by stress relieving, by avoiding interference fits, or by shimming assemblies Avoid galvanic couples, which may tend to accelerate stress corrosion Provide for regular inspection of parts to determine surface flaws and cracking during the life cycle of the part Improve the surface quality by reducing surface roughness and/or increasing surface compressive stresses Avoid the use of titanium in contact with silver, silver-plated material, or silver-plated fasteners, such as silver-plated A-286 nuts

Control of Galvanic Corrosion. Dissimilar metals were not to be used in intimate contact unless they were suitably

protected against galvanic corrosion. Because of the seriousness of galvanic corrosion, every effort was made to avoid the use of dissimilar metals, to exclude moisture or other electrolytes from the system, and to protect metal surfaces in the contact area. Metals were considered compatible if they were in the same grouping as specified in MSFC-SPEC-250, Class II, or if the difference in solution potential was 0.25 V or less.

Control of Atmospheric Corrosion. All parts, assemblies, and equipment, including spares, were finished to provide

protection from corrosion in accordance with the requirements of MSFC-SPEC-250, Class II, as a minimum. All organic finishes and anodized aluminum that contact titanium were limited to surfaces not normally exposed to propellants. A finish specification delineating the finishes used on each specific material in any particular application and corrosion control procedure was prepared by the orbiter contractor. The finish specification and related procedures from the subcontractor were required to provide for in-process corrosion control. Specific requirements were also given for: • • • • •

Surface preparation for adhesive bonding Finish systems for interior and exterior surfaces (including those surfaces to which the thermal protection system was to be bonded) Fastener installations Joints and faying surface sealing Protection for parts to be shipped for vehicle final assembly

Designing to Control Corrosion of the Space Shuttle Orbiter From a corrosion control standpoint, it is convenient to separate the space shuttle orbiter into four categories; primary structure, fluid systems, mechanical systems, and avionics systems. Each of these areas has its own unique problems. Primary Structure Weight and cost both dictated that the primary structure of the orbiter be made from aluminum. The majority of this structure was made from heat-treated alloys of the 2000 and 7000 series (Fig. 3). However, some 5000- and 6000-series alloys were also used.

Fig. 3 Typical materials of construction for the space shuttle orbiter

Aluminum Airframe. Prior to alloy selection, two surveys were conducted. The first was to identify steps to be taken to avoid stress-corrosion problems with aluminum alloys, and the second to identify a corrosion protection system that could survive the unique spacecraft environment.

The stress-corrosion survey, conducted in the early 1970s, indicated that virtually all stress-corrosion failures in service occurred in the 2000-series alloys in the T3, T4, and T6XX tempers and in the 7000-series alloys in the T6XX tempers and perpendicular to the short-transverse direction. Only alloys 2024, 2124, and 2219 have high SCC resistance in the T6XX

tempers. In forgings, stress corrosion occurred in end grain runout along the forging parting lines. In extrusions and plate, failures occurred where parts were severely formed, where interference fits had occurred, or where assembly stresses were high. Theoretically, stress corrosion will not occur until all three of the essential elements are present: a susceptible microstructure, a corrosive environment, and surface tensile stresses. If any one of these was eliminated, stress corrosion should not occur. The approach used on the orbiter, however, was to eliminate or minimize all three of these conditions to the maximum practical extent. First, only aluminum alloys were permitted with a minimum 170-MPa (25-ksi) stress-corrosion threshold (according to the supplier's standard test methods: 30 days in salt spray) in all directions. This eliminated the T3, T4, and T6XX tempers of nearly all the 2000- and 7000-series alloys, whose stress-corrosion thresholds could be as low as 48 MPa (7 ksi) or less in the short-transverse direction. The preferred microstructure of the -T73 and -T76 tempers was chosen for 7000-series alloys, and although some weight penalty was incurred, program reliability was well served. Because most of the aluminum structure is designed for compressive loading (buckling, crippling) or shear, a very small weight penalty actually resulted. For the 2000-series alloys, the T8XX tempers were used predominantly; however, in a few cases, the T6 or T62 tempers were used for alloys 2124 and 2219. Second, every effort was made to reduce residual stress levels. Mill products were ordered in stress-relieved tempers (for example, T651 or T851) wherever possible to reduce machining distortion and susceptibility to stress corrosion. Interference fits were limited to stress levels below 67% of the stress-corrosion thresholds. Tables were prepared to allow the materials and process engineer to ascertain stress levels resulting from interference fit pins and bushings into various size lugs. Residual stresses in assembly were minimized by shimming. Forming by bending, which put the shorttransverse direction into tension, was not permitted. Finally, the corrosion-protective paint system (described below) was applied to all aluminum parts. As of 1986, no incidents of aluminum stress-corrosion failures have been reported on the orbiter since its inception in 1972. The second survey conducted in the early 1970s involved the selection of a paint system that would meet the unique requirements of the shuttle. The system had to provide protection to aluminum from corrosion for a minimum of 10 years of seacoast exposure, without touchup, because it must also serve as the base to which the thermal protection system (TPS) tiles of the shuttle are bonded. Unlike commercial aircraft, the external surfaces could not be washed, repainted, or protected from water intrusion and crevice corrosion by using water-displacing chemicals. The paint system had to endure temperatures of 175 °C (350 °F) during entry and landing because heat from reentry soaked back into the structure. It had to be capable of surviving space vacuum and low temperatures (-155 °C, or -250 °F) without degradation. Minimum offgassing was desirable to avoid giving off toxic fumes inside the cabin (crew hazard) or the condensation of volatile material on windows or optical (thermal) control surfaces. The need for exceptional corrosion resistance was further mandated by the floating bilge; that is, the orbiter is stacked and launched in a vertical attitude, operates in zero gravity, and reenters and lands horizontally. It was not possible to ensure that all water drains out of the structure in all attitudes. The paint system chosen was a chromate-inhibited epoxy polyamine primer. This system was tough, abrasion resistant, and durable. Surfaces to be painted were either anodized according to MIL-A-8625 type II, Class 1, or chemically filmed according to MIL-C-5541, class 1A. Each coat of paint was 0.015 to 0.023 mm (0.6 to 0.9 mil) thick. A single coat of the chromated epoxy polyamine primer demonstrated 1500 h of salt spray protection without corrosion, even in areas scratched through to bare aluminum. The surface to which the external TPS was bonded had a single coat of the chromated epoxy polyamine paint. It achieved additional corrosion protection from a room-temperature vulcanized (RTV) adhesive layer 0.13 to 0.23 mm (5 to 9 mils) thick used to bond the tiles. In the cargo bay area, the single coat of epoxy polyamine primer was overcoated with one coat of polyurethane (MIL-C-83286 or MIL-C-81773) to achieve the required optical properties--absorptivity ( ), emissivity ( ), and the proper / ratio for heat control. The interior of the cabin required the use of nonglare coatings and selected colors. Again, a single undercoat of the chromated epoxy polyamine primer was coated with polyurethane. In this case, the polyurethane not only provided a durable color but also acted as a barrier to unacceptable offgassing products of the primer. Parts were painted as details, drilled and assembled, and then repainted upon assembly to coat the fasteners. Although it was desirable from a corrosion standpoint to install all rivets wet, practical manufacturing considerations did not permit it.

Automatic riveting machines, which were used to install nearly 90,000 rivets in the wing, could not use wet rivet installations. The weight reduction demands of the program resulted in the elimination of the use of two coats of paint on interior surfaces. More than 500 kg (1100 lb) of paint were used to cover 8175 m2 (88,000 ft2) of surface. By substituting an anodize coating for one coat of primer, significant weight savings were realized. Consequently, the finish system for Discovery and Atlantis followed the general scheme:

Area Exterior TPS surface Exterior non-TPS surfaces Interior surface Crew compartment

Coating Anodize + 1 coat chromated epoxy polyamine primer + 1 coat RTV adhesive Anodize + 1 coat chromated epoxy polyamine primer Anodize + either 1 coat chromated epoxy polyamine primer or 1 coat polyurethane Chemical film or anodize + 1 coat chromated epoxy polyamine primer + 1 coat polyurethane or anodize only

The forward fuselage was fabricated as a sheet metal skin-stringer design in aluminum alloy 2024-T6. Suspended inside the forward fuselage was an all-welded aluminum pressurized cabin made from aluminum alloy 2219 using the T6 and T8 tempers. It was approximately conical in shape, about 5 m (17 ft) long and tapering from 5 to 2.4 m (17 to 8 ft) in diameter at its forward end. The mild and aft fuselage structures were machined from aluminum alloy 2124-T851 plate. Major frames were of aluminum alloy 7075 in the T76 or T73 tempers. Aluminum honeycomb sandwich was extensively used in the wings and body flap area. Corrosion protection systems must prevent corrosion of the thin (0.025 to 0.075 mm, or 1 to 3 mil) honeycomb core (usually aluminum alloy 5056H39) and delamination of the skins. Face sheet skins could not be alclad, because corrosion would proceed in the plane of the sheet, resulting in delamination of the bond line. To prevent corrosion, all aluminum cores were protected with conversion coatings and were nonperforated; the face sheets used corrosion-resistant adhesive primers, and the sandwich assemblies were sealed at the edges to prevent water entry. Structural Joints and Fasteners. The corrosion engineers favored the assembly of structural joints with RTV faying surface sealants; however, the electrical bonding requirements or grounding of each panel eliminated this approach. Electrical bonding requires a maximum dc resistance of 2.5 m across joints requiring lightning protection or radio frequency grounding for electrical or electronic equipment. For a typical faying surface joints, local removal of the paint or anodize on the detail (down to bare metal) is required in the area of the fastener. Bare aluminum surfaces are subsequently coated with a chemical film (MIL-C-5541, class 1A). The joint is then bolted with stainless steel fasteners using stainless steel washers under the bolthead and nut to protect aluminum surfaces during application of torque. Fasteners are installed with the shank portion wetted with chromated epoxy polyamine primer. Joints are subsequently touched up with a chemical film and a coat of chromated epoxy polyamine primer around the washers. Faying surfaces are sealed with a continuous fillet of RTV 577, which is a white, thixotropic silicone rubber material.

There are approximately 30 different kinds of electrical grounding joints available to suite various designs on the orbiter; the grounding techniques used include jumpers, spot welds, staples, metallized tape, and the method described above. Even an adhesively bonded edge member or a T-section bonded to a honeycomb face sheet must be provided with a ground to the face sheet itself. Dissimilar-metal joints are permitted on the orbiter without additional galvanic protection if they fall within the range shown in Table 1. Table 1 should be used only a guideline, and such factors as cathode-to-anode area ratios, corrosive environments, and other detrimental factors must be evaluated.

Table 1 Metals and alloys compatible in dissimilar-metal couples

The fasteners chosen for the spacecraft are all basically corrosion resistant, but care must be taken with dissimilar-metal combinations. Bolts are typically made of alloys A-286 (965 to 1380 MPa, or 140 to 200 ksi), Inconel alloy 718 (1240 MPa, or 180 ksi), and MP35N (1655 MPa, or 240 ksi). For applications up to 870 °C (1600 °F), Udimet 500 is used (1035 MPa, or 150 ksi). Of these materials, only A-286 has shown any corrosion in service. Thecorrosion is only superficial and of on real concern. It is removed only for cosmetic reasons. None of these alloys is susceptible to hydrogen embrittlement in orbiter vehicle service. Nuts are made from A-286 and Inconel alloy 718 and are lubricated with a thin (0.005 to 0.01 mm, or 0.2 to 0.4 mil) silver plate. Bolts and nuts are always installed with washers. The aluminum, therefore, never contacts the silver plate. Because the hole, as previously mentioned, is coated with wet chromated epoxy polyamine primer, no moisture can penetrate between the stainless or nickel shank and the aluminum hole. Stainless steel washers are separated from the aluminum surface with a dry coat of primer (where electrical grounding is not required) or a chemical film coating plus a

touchup of the primmer around the washer (where electrical grounding is required). No problems with galvanic corrosion are experienced in these installations. Inserts, made from A-286, are silver plated and must also be installed into aluminum wetted on their exterior with the chromated epoxy polyamine primer. Where nuts are used in contact with titanium, only molybdenum disulfide-type dry film lubricants are used. Experience has shown that silver in contact with titanium at approximately 265 °C (500 °F) or above can bring about stress-corrosion cracking (SCC) of titanium (Ref 2, 3, 4). Titanium contact with silver is also prohibited by MIL-S-5002. Titanium pin and collar fasteners are used for shear applications. To save weight, the orbiter uses aluminum alloy 2024 collars rather than A-286. Again, the holes are wet coated with the chromated epoxy polyamine primer by applying primer to the fastener shank away from the threads. Both ends of the fastener are subsequently touched up with the primer. Where such fasteners are used on the graphite cargo bay doors, it is necessary to use RTV rubber as a corrosion barrier, thus completely encapsulating the aluminum collar to prevent corrosion. Rivets used on the spacecraft are made from aluminum alloy 2219-T62. These rivets provide good shear strength while avoiding the need to "ice box" rivets after solution treating, as is required with aluminum alloy 2024 rivets. Further, aluminum alloy 2024 rivets would undergo aging at entry temperatures of the orbiter (175 °C, or 350 °F), resulting in an increased susceptibility to corrosion as grain-boundary precipitation initiated. Aluminum alloy 2219-T81 rivets, although also commercially available, lack sufficient ductility to prevent cracking of driven heads. (The widely used aerospace aluminum alloy 7050-T73 rivets were unavailable when the shuttle orbiter was being built.) Other Structural Alloys. No corrosion protection (except for passivation treatments after fabrication) is considered

necessary for stainless steels. Stainless steels, particularly the precipitation-hardenable grades, will often display light surface corrosion products after extensive exposure. A light abrasion will remove the corrosion. No effort is made to passivate stainless parts as installed, because chemical spillage is considered more detrimental to the structure than any enhanced corrosion protection gained from passivation. Nickel alloys such as Inconel alloy 718 and Inconel alloy 625 are used for elevated-temperature service with no corrosion protection. Inconel alloy 718 brazed honeycomb panels are used for the conical seals on the vertical stabilizer and for outboard elevon rub panels and flipper door panels. Inconel alloy 625, made as a resistance-welded sandwich, is used to temperatures of 870 °C (1600 °F). The surface of the sandwich is coated with a wear-resistant high-emittance chromium oxide coating. Titanium also requires no further corrosion protection. Titanium, principally as Ti-6Al-4V, is widely used as forging, bar, and plate products throughout the spacecraft. Many other high-strength titanium alloys are also used. The major structural members of the aft thrust structure are made of Ti-6Al-4V. These transmit the thrust of the liquid rocket engines to the orbiter structure. Titanium honeycomb sandwiches, made by the liquid interface diffusion (LID) bonding process, are used as inboard elevon and flipper door panels. Although the honeycomb has a perforated core, no corrosion is experienced. Because of prior experience in which processing and testing solutions had resulted in SCC of titanium alloys, a control specification, MF0004-018, has been imposed. This specification defines a list of fluids that are suitable for titanium and the specific conditions under which their contact is appropriate. Beryllium alloy S-65 (99% Be min) is used structurally for the external tank door and for windshield retainers. Beams providing structural support in the windshield area use either S-65 or CIP HIP-1 (Ref 5), which is also nonstructurally for the navigational base and the star tracker boom, as well as for heat sinks. Beryllium must be protected in service. The beryllium is anodized according to a Rockwell internal specification and is painted with one coat of chromated epoxy polyamine paint or chemically filmed in a manner similar to that used for aluminum (MIL-C-5541). Two coats of the chromated epoxy polyamine paint are then applied. The anodized coating (0.05 mm, or 0.2 mil, minimum) reveals no corrosion when tested in 168 h of salt spray according to ASTM B 117. Steels must be protected in service from the seacoast environment. Steels are often painted with chromated epoxy polyamine paint or plated with nickel or chromium, depending on the service. Cadmium plating is not used except under rare circumstances, because it can easily sublime in space and redeposit on cooler adjacent surfaces. To avoid problems with SCC and hydrogen embrittlement, steel alloys are restricted to 1380 MPa (200 ksi) or less in tensile applications, and precipitation-hardenable steels are restricted to the H1000 or higher-temperature tempers. Steels with tensile strengths as

high as 2070 MPa (300 ksi) can be used for applications that involve bearing, compressive, or shear loads; such applications include ball or roller bearings,valve seats, and springs. A more complete description of the corrosion protection of steel alloy parts that are in moving contact with each other can be found in the discussion "Mechanical Systems" in this article. Niobium is used for low-stress applications in the orbiter airframe structure. Tubes and nozzle parts fabricated from niobium alloy C-103 (Nb-10Hf-1Ti-0.5Zr) are used in the reinforced carbon-carbon nose cap for the shuttle entry air data system (SEADS) program involving measurements of aerodynamics pressures. These parts have a VH109 silicide coating to prevent high-temperature oxidation. The coating was chosen because of its performance at the design temperatures (1455 °C, or 2650 °F) (see the discussion "Case Histories" in this section). Niobium alloy C-103 parts are used as closeout members in elevon seals to shield the hot plasma from the interior structure and mechanisms. These parts are coated with an R512E silicide coating. They were designed for service at maximum temperatures of 1370 °C (2500 °F). Silicide coating systems are ceramic in nature and may be chipped by impact on edges or surfaces. Tests were conducted to verify that parts with coating damage down to bare metal could still function for a limited number of flights (see the discussion "Case Histories" in this section). Composite materials are widely used on the orbiter and present no corrosion problems except for graphite epoxy

structures. Although graphite is compatible with titanium, corrosion-resistant steels (A-286 and 300-series stainless steels), nickel, and cobalt-base alloys, the galvanic potential between graphite and aluminum or graphite and steel requires special design considerations. Suitable galvanic isolation is accomplished by using a layer of titanium foil, Tedlar, Kapton, or type 120 glass fabric with suitable resin plus two coats of chromate epoxy polyamine primer. All edges of the joints between the graphite and the aluminum or steel are sealed with RTV silicone to preclude moisture intrusion. More than 300 boron/aluminum composite tubes with diffusion-bonded Ti-6Al-4V clevises are used on the orbiter, principally in the mid fuselage to stabilize frames or as pressure vessel supports. The aluminum portion is painted with chromated epoxy polyamine primer. These present no special corrosion design problems. Also, there are no corrosion problems with boron epoxy-bonded reinforcements on titanium thrust structure tubes in the aft thrust structure. Fluid Systems The space shuttle orbiter fluid systems must provide for the storage, transfer, and regulation of 17 different fluids, as shown in Table 2. The fluid systems can be grouped into major functional areas: • • • • • • •

Environmental control and life support system (ECLSS) Electrical power system (EPS) Reaction control system (RCS) Orbital maneuvering system (OMS) Main propulsion system (MPS) Auxiliary power units (APU) Hydraulic system (HYD)

Table 2 Fluids used on the space shuttle orbiter See also Fig. 4. Index number

Fluid/gas

Location

System

Quantity

Explosive limit

1

Ammonia (NH3)

Aft fuselage

ECLSS

16%

2

Mid fuselage

ECLSS

3

Breathing oxygen (GOX) Freon-21

Mid and aft fuselage

ECLSS

4

Freon-1301

Crew compartment

ECLSS (fire extinguisher)

44.27 kg (97.6 lb) 32.21 kg (71 lb) 272.16 kg (600 lb) 5.17 kg (11.4 lb)

Threshold limit value, ppm 25

Remarks

Two tanks

NA(a)

(b)

One tank

NA

1000

System

NA

1000

Three tanks

5

Fluorinert, FC-40

Mid fuselage

EPS

6

Helium

Forward RCS module OMS/RCS modules

Forward RCS OMS

Aft fuselage

MPS

Mid fuselage

MPS

Aft fuselage

Aft RCS

7

Hydrazine (N2H4)

Aft fuselage

APU

8

Hydraulic fluid

Forward, mid, and aft fuselage

HYD

Landing gear struts

LDG

Aft fuselage

MPS

Mid fuselage

EPS

Aft fuselage

MPS

Mid fuselage

EPS, LSS

9

10

Liquid hydrogen (LH2)

Liquid oxygen (LOX)

11

Lubricating oil

Aft fuselage

APU

12

Monomethyl hydrazine (MMH)

Forward RCS module Aft RCS modules

Forward RCS Aft RCS

OMS modules

OMS

13

Nitrogen

Mid fuselage

ECLSS

14

Nitrogen tetroxide (N2O4)

Forward RCS module Aft RCS module

Forward RCS Aft RCS

OMS

OMS

Crew module

ECLSS

Aft fuselage

ECLSS

15

Water (deionized)

16

Water (potable)

Lower crew module

LSS

17

Water (waste)

Lower crew module

LSS

(a) (b) (c) (d)

17.5 kg (39 lb) 3.81 kg (8.4 lb) 44 kg (97 lb) 529.52 L (18.6 ft3) 1843.4 L (65.1 ft3) 7.62 kg (16.8 lb) 131.99 kg (291 lb) 344 L (90.9 gal) 13.6 kg (30 lb) 142.43 kg (314 lb) 83.47 kg (186 lb) 2219 kg (4892 lb) 708.5 kg (2162 lb) 8.16 kg (18 lb) 480.75 L (127 gal) 961.5 L (254 gal) 4845.31 L (1280 gal) 103.42 kg (228 lb) 465.6 L (123 gal) 935 L (247 gal) 4845.31 L (1280 gal) 29.5 kg (65 lb) 194.14 kg (428 lb) 289.4 kg (638 lb) 72.7 kg (159.5 lb)

NA

...

NA

(c)

NA

(c)

Two tanks

NA

(c)

Four tanks

NA

(c)

Six tanks

NA

(c)

Four tanks

4.7%

1

Three tanks

204 °C (400 °F)

(d)

Three systems

110 °C (230 °F) 4%

(d)

Nose and main gear Feedlines and main engine

4%

(c)

Two tanks

NA

(b)

NA

(b)

Feedlines and main engine Two tanks

245 °C (475 °F) 3.0%

(d)

Three systems

0.2

One tank

3.0%

0.2

Two tanks

3.0%

0.2

Two tanks

NA

(c)

Four tanks

NA

5

One tank

NA

5

Two tanks

NA

5

Two tanks

NA

None

NA

None

Two cooling loops Three tanks

NA

None

Four tanks

NA

None

One tank

(c)

NA, not applicable. No threshold limit value; upper limit is 6 h at 1 atm of pressure, lower limit is 19%. Simple asphyxiant, no threshold limit value. No threshold limit value; inhalation of vapors not encountered in normal use.

Fuel cell coolant loops Two tanks

The locations of the pressure vessels used to contain these fluids are shown in Fig. 4. Design information covering these pressure vessels is given in Table 3.

Table 3 Major shuttle orbiter pressure vessels Fluid/gas

System

Alloy

Number per vehicle

Ammonia (NH3) Breathing oxygen Freon-21

ECLSS

Ti-6Al-4V

ECLSS

Helium

Forward RCS Aft RCS OMS MPS MPS APU

Kevlar/Inconel alloy 718 Aluminum 6061T6, AM350 bellows contain the fluid Kevlar/Ti-6Al-4V(a)

Hydrazine (neat) Hydraulic fluid

ECLSS

HYD

2

Size, diameter or length × diameter mm in. 437 17.2

1

660

Operating pressure

Proof pressure

Burst pressure

4.0

MPa 2.2

psi 320

MPa 4.3

psi 625

MPa 8.9

psi 1290

1.5

22.8

3300

29.2

4240

34.1

4950

2

2

686 × 27 × 330 13 (accumulator) 475 18.7

2.0

1.6

230

2.4

345

3.2

460

1.5

27.6

4000

36.6

5310

41.4

6000

Kevlar/Ti-6Al-4V(a) Kevlar/Ti-6Al-4V(a) Kevlar/Ti-6Al-4V(a) Ti-6Al-4V(a) Ti-6Al-4V(a)

4 2 7 2 3

475 1024 663 249 711

18.7 40.3 26.1 9.8 28

1.5 1.5 1.5 4.0 3.35

27.6 33.6 31.0 5.9 2.4

4000 4875 4500 850 355

36.6 44.9 42.7 11.7 6.7

5310 6510 6190 1700 970

41.4 50.5 46.5 23.4 7.4

6000 7325 6750 3400 1070

Chromium-plated 4130 steel cylinder, 2024-T851 aluminum piston

3

12.0

1.7

250

20.7

3000

41.4

6000

4.0

20.7

3000

41.4

6000

82.7

12,000

1.5

2.2

320

2.4

350

3.3

480

26

Factor of safety

Liquid hydrogen (LH2) Liquid oxygen (LOX) Monomethyl hydrazine (MMH)

EPS

Aluminum 2219-T6

2-4

635 × 25 × 86 3.4 (accumulator) 635 × 25 × 86 3.4 (accumulator) 1214 47.8

EPS

Inconel 718(b)

2-4

973

38.3

1.5

7.1

1035

8.6

1240

109

1575

Forward RCS Aft RCS OMS

Ti-6Al-4V(b)

1

991

39

1.5

2.4

350

3.2

465

3.6

525

Ti-6Al-4V(b) Ti-6Al-4V(b)

2 2

39 96 × 49

1.5 1.5

2.4 2.2

350 315

3.2 2.4

465 345

3.6 3.2

525 470

Nitrogen

ECLSS OMS OMS Forward RCS Aft RCS OMS

Kevlar/Ti-6Al-4V(a) Ti-6Al-4V(a) Ti-6Al-4V(a) Ti-6Al-4V(b)

4 2 2 1

991 2438 × 1245 660 84 132 991

26 3.3 5.2 39

1.5 2.5 40 1.5

22.8 3.1 20.7 2.4

3300 450 3000 350

28.8 6.2 41.4 3.2

4175 900 6000 465

34.1 7.4 82.7 3.6

4950 1080 12,000 525

Ti-6Al-4V(b) Ti-6Al-4V(b)

2 2

1.5 1.5

2.4 2.2

350 315

3.2 2.4

465 345

3.6 3.2

525 470

ECLSS

6061-T6 aluminum shell Inconel alloy 718 bellows, and Inconel alloy 625 end fittings contact water Same as above

3

991 39 2438 96 × × 49 1245 737 × 29 × 396 15.6 208 × 8.2 × 152 6 (accumulator)

2.0

0.3

37

0.4

58

0.5

74

2.0

0.6

90

0.9

135

1.2

180

4

902 × 394

2.0

0.1

20

...

...

0.3

40

Crew module

Same as above

1

902 × 394

2.0

0.1

20

...

...

0.3

40

APU

Ti-6Al-4V(a)

1

432

4.0

3.0

435

4.5

655

12.1

1755

Nitrogen tetroxide (N2O4)

Water, deionized

ECLSS

Water, potable and waste

Water,

Crew module

2

35.5 × 15.5 35.5 × 15.5 17

cooling

APU

Ti-6Al-4V(a)

1

244

9.6

8.0

0.7

100

4.1

600

5.5

800

Note: The pressure vessels listed in Table 3 may differ slightly from Table 2, because Table 3 includes major accumulators, pressure vessels integral to a particular hardware system, and extra pressure vessels needed for some missions.

(a) (b)

Annealed. Solution treated and aged.

Fig. 4 Locations of liquids and gases in the space shuttle orbiter. Numbers correspond to the fluid index numbers used in Table 2.

Plumbing Lines. The philosophy followed in corrosion control for fluid systems and pressure vessels was to select materials that were compatible with the fluids without protective coatings and to control the fluid chemistry, not only as purchased but also during vehicle loading and operations. Stainless steel lines are used for all systems except hydraulic fluids and hot gaseous oxygen. Hydraulic fluids are contained in Ti-3Al-2.5V lines. Inconel alloy 718 is used for hot gaseous oxygen lines.

Because many of the fluids are hazardous (toxic, explosive), metallurgical joints were used in all permanent connections of stainless and Inconel alloy lines. Type 304L stainless steel was used for service with helium, hydrazine, liquid hydrogen, liquid oxygen, monomethyl hydrazine, nitrogen, and nitrogen tetroxide. Type 304L was selected to avoid potential sensitization. Most of the permanent joints were automatically welded by orbiting arc equipment, using an external sleeve of the same alloy to supply reinforcement to the weld bead, as shown in Fig. 5(a). The bead geometry on the inside of the tube is smooth and free from crevices that may initiate chemical attack. The stainless steel tubing, in many cases, is in the one-eight hard condition, and the sleeve also provides added strength to compensate for the localized annealing at the welds and heat-affected zone (HAZ); this results in the full efficiency of the one-eighth hard material.

Fig. 5 Stainless steel and Inconel alloy fluid system permanent joints. (a) Automatic weld, inert gas, tungsten arc. (b) Braze joint

Alloy 21-6-9 stainless steel was used in lines carrying ammonia, breathing oxygen, freons, oxygen, hot gaseous hydrogen (280 °C, or 540 °F), nitrogen, and waters. The stainless steel was usually brazed with gold alloy Nicoro 80 (81.5 Au16.5Cu-2Ni) in a configuration shown in Fig. 5(b). This configuration also eliminates an internal crevice in the lines, because the brazing alloy flows along the capillary between the tube and sleeve, seals it, and forms a fillet along the periphery of the sleeve. There were two concerns with this braze combination. First, would the braze alloy create a galvanic-corrosion problem? Testing of the electrode potentials and exposures with actual fluids showed no galvanic effect that could be detected. Second, would the copper in the braze alloy cause liquid-metal embrittlement? The literature indicated the 21-6-9 alloy to be particularly susceptible to copper embrittlement. Joints were made and sectioned to reveal the microstructure of the brazed joint. Intrusion of the brazing alloy into grain boundaries did occur, but never exceeded 0.05 mm (2 mils) even after four induction brazing cycles. The intrusion did not affect the static or fatigue strength of the joint; in fact, it appeared to give a superior attachment to the substrate. In many cases, stainless steel lines must connect pressure vessels made from titanium alloys such as Ti-6Al-4V. Typically, a bimetallic joint is used in which a steel tube is joined to a titanium tube or fitting, which is then welded to the pressure vessel. The steel-to-titanium tube joints are made by coextrusion or, in the case of the APU water tank, by swaging. Aluminum-stainless steel joints are required for attachment of plumbing to the aluminum alloy 2219-T6 tank. These joints are made friction welding. No galvanic-corrosion problems have been experienced with the fluids involved. For hot gaseous oxygen lines operating to 31 MPa (4500 psi) and 300 °C (570 °F), it was more efficient to use Inconel alloy 718 to obtain the required strength at temperature. Inconel joints were welded in the same manner as the type 304L stainless steel joints described above. The Inconel alloy joints also had a type 304L stainless steel sleeve, which provided sufficient reinforcement to ensure that nearly the full heat-treat properties of the Inconel alloy 718 tubing were realized (ultimate tensile strength: 1240 MPa, or 180 ksi; tensile yield strength: 1035 MPa, or 150 ksi). Initial attempts were made to braze this alloy, but because of its tenacious high-temperature oxide film, repeat braze-debraze cycles (even over nickel-plated ends) could not be achieved. The environmental control and life support system was designed to provide:

• • • •

Atmospheric control to the pressurized crew cabin (oxygen, nitrogen, carbon dioxide, water vapor, odor) Pressure control to the crew cabin Thermal control to the crew cabin and avionics boxes Potable water and waste management control

The temperature of the crew cabin is maintained between 16 and 32 °C (61 and 90 °F). Oxygen partial pressure is maintained at 22,000 N/m2 (3.2 psi), and nitrogen is added to achieve pressures of 70,500 N/m2 (10.2 psi) for space operations and 101,000 N/m2 (14.7 psi) for launching conditions. Relative humidity is controlled to prevent condensation of moisture. Therefore, the cabin atmosphere is a benign environment from a corrosion standpoint.

Aluminum cold plates are used to remove heat from the electronic boxes in the mid and aft fuselages. The cold plates are fluxless brazed with aluminum alloy 6951 face sheets and aluminum alloy 6061 cores. Freon 21 (dichloromonofluoromethane) is the coolant. Stainless steel cold plates are used in the crew cabin to carry heat from avionics boxes within the cabin. These coldplates are brazed from AISI type 304L sheet using Amdry 930 (Ni-22.5Mn5Cu-7Si). Deionized water is used in the stainless steel cold plates. A water loop transfers the excess heat from the cabin and cabin avionics equipment to the freon cooling loop by way of the cabin heat exchanger. The freon cooling loop delivers this heat, together with the heat from the fuel cells, payloads, and mid and aft avionics equipment, to a large aluminum radiator (111 m2, or 1195 ft2) where the heat is radiated into space. When the cargo bay doors are closed during ascent of immediately before reentry, an active thermal rejection system, the flash evaporator, is employed. Heat is rejected by the boiling of water in this system. Stainless steel was chosen for water lines because of the prior difficulties encountered on the Apollo program with aluminum lines and cold plates. Despite the use of inhibitors such as triethanolaminephosphate (TEAP) and sodium mercaptobenzothiozole (NABT) on Apollo aluminum systems, serious corrosion problems were encountered. Solutions had to be continuously circulated to avoid solution concentration gradients leading to localized corrosion (pitting) of the aluminum lines. Pits as deep as 0.7 mm (28 mils) were found in aluminum lines after limited service. To preclude problems with corrosion in water systems on the space shuttle orbiter, it was decided that stainless steel would be used for all water (cooling and potable) lines. Each material in the water path was identified, evaluated and accepted or changed, and tested in a control loop at NASA (Houston) for up to 2 years of service. Orbiter water chemistry was controlled to limit oxygen content to 0.5 ppm (to avoid localized oxygen concentration cells), conductivity was limited to 3.3 × 10-6 -1 · cm-1, and the pH was controlled within the 6.0 to 8.0 range. After more than 5 years of service, no corrosion problems have been experienced with stainless steel. Waste water is handled by 21-6-9 stainless steel lines and is stored in aluminum alloy 6061-T6 tanks. Waste water tends to be extremely corrosive to aluminum. However, coatings such as Tufram, a tetrafluoroethylene-impregnated anodize, have been used effectively to protect aluminum tanks from urine. The ammonia (NH3 boiler provides for heat rejection at altitudes below approximately 30.5 km (100,000 ft); at these altitudes, the cargo bay doors are closed, and the boiling of the flash evaporator can no longer provide sufficient cooling to the freon. The ammonia boiler is a shell and tube heat exchanger with a single pass on the ammonia side and two passes for each Freon-21 coolant loop. The ammonia flows through the bank of 77 small-diameter stainless steel tubes, and the Freon-21 flows over the exterior of the tubes. Because these tubes have such thin walls (0.2 mm, or 8 mils), perforation by corrosion is a major concern. Corrosion has been experienced as isolated areas of intergranular attack due to misprocessing of type 304L stainless steel tubing. Very small amounts of carbon residue left on a thin-wall tube such as this will result in sensitization during its brazing cycle. A change has recently been made to a stabilized grade (type 347 stainless steel) to avoid these problems. Ammonia is stored on a titanium Ti-6A1-4V pressure vessel. No corrosion problems have been experienced in this application. Nitrogen is stored at 22.8 MPa (3300 psi) in Ti-6A1-4V pressure vessels that are filament overwrapped with Kevlar 49 aramid. It presents no corrosion problems. Gaseous breathing oxygen is stored in a pressure vessel made from Inconel alloy 718 that is also overwrapped with Kevlar 49 aramid filament. It also operates at 22.8 MPa (3300 psi). Extreme care must be taken when designing either liquid or gaseous oxygen systems because of the dangers of ignition with metals and organic materials (Ref 6). Titanium and magnesium are not used in oxygen systems for this reason and are considered to be highly reactive under impact. Materials that are considered satisfactory for use must pass an impact test consisting of 20 consecutive impacts at an energy level of 98 J (72 ft·lb) when tested in the Army Ballistic Missile Agency (ABMA) Impact Tester according to NHB 8060.1, Test 13. Further, when gaseous oxygen pressures exceed 6.9 MPa (1000 psi), materials must also pass dynamic qualification (pneumatic impact) according to NHB 8060.1, Test 14. Although extreme care is taken in material selection for oxygen systems, this is not enough to ensure an ignition-free system (see the discussion "Case Histories" in this section). Because an oxygen ignition is a catastrophic event that results in an explosion and significant molten metal, it is not always possible to reconstruct the precise cause of a failure. Metallic materials that pass the above tests can still ignite if design or manufacturing operations result in: • •

Energetic particle impact, caused by particles accelerated to sonic velocities Contamination

• • • •

Pneumatic shock from rapid valve opening across large pressure differences Fretting or galling, generating particles and localized high temperatures Frictional heating, such as by flowing past a feather edge Gaseous heating, such as by adiabatic compression and Helmholz resonance (resonance in blind columns) (Ref 7, 8)

Of the various engineering metals tested, Inconel alloy 718 and Monel alloy 400 offer superior resistance to ignition problems and are preferred in valves in which dynamic problems can occur. When properly designed and manufactured, stainless steel has been successfully used in oxygen and LOX valves. Aluminum, stainless steel, and Inconel have been employed without incident in numerous oxygen pressure vessels. The electrical power system uses fuel cells to generate electricity by combining gaseous hydrogen and oxygen. The

electrolyte in the fuel cell--potassium hydroxide--is contained between gold-plated magnesium electrodes. For corrosion protection, the magnesium substrate is plated with zinc, copper, and nickel beneath the gold. The hydrogen and oxygen are stored as supercritical gases in double-wall pressure vessels. The hydrogen is contained in an aluminum alloy 2219-T6 inner shell that has been electron beam welded; the oxygen is stored in an inner pressure vessel of electron beam welded Inconel alloy 718. Both tanks are suspended within pressure-tight external shells made of aluminum alloy 2219-T6. A series of 12 S-glass/epoxy straps suspend the inner tank from the forged girth ring of the outer shell. The annulus between the vessels contains multilayered reflective insulation. A pressure level 1 × 10-5 torr is maintained by a vacuum ion pump. Aluminum alloy 2219-T6 is very susceptible to intergranular pitting corrosion, and care must be taken during fabrication to prevent dirt particles and/or moisture from coming in contact with it (see the discussion "Case Histories" in this section). The reaction control system uses rocket engines burning monomethyl hydrazine (MMH) with N2O4 to achieve the

desired attitude control of the orbiter while in orbit. The 38 reaction control engines are each capable of providing 3870 N (870 lbf) of thrust. Six vernier RCS engines allow fine tuning of the orbiter attitude. They develop 111 N (25 lbf) of thrust each. Both N2O4 and MMH are stored in pressure vessels made of Ti-6A-4V. Several precautions must be taken in the case of N2O4. First, ingestion of moisture will result in the formation of nitric acid, which can be corrosive to some elements of the system. Second, N2O4 spills can be dangerous because of toxicity and can be destructive to spacecraft hardware. Nitrogen tetroxide will readily corrode nickel, strip off protective paints, and dissolve nylon. In addition, because it is an aggressive oxidizer, N2O4 can react violently and ignite organic materials. Third, N2O4 will cause SCC of titanium in the absence of a trace of nitric oxide (NO) (Ref 9, 10, 11, 12) (see the discussion "Case Histories" in this section). Current specifications call for 1.5 to 3% NO; but repeated loading causes volatile loss of NO, and storage may cause stratification. Although only a trace of NO (perhaps as little as 0.2%) will prevent SCC, care must be taken to ensure at least 0.6% NO to be safe. Fourth, impact ignition of titanium can occur in N2O4. Tests have shown that threaded fasteners can suffer localized melting and that even an impact of inert material, such as glass or sand, on the titanium surface can result in localized melting in N2O4. Unlike the oxygen reaction, the reaction is quickly quenched and occurs as low as the 54- to 68-J (40 to 50-ft-lb) level on the ABMA Impact Tester. Therefore, only aluminum fasteners threaded into titanium N2O4 tanks are permitted on the space shuttle orbiter, this ensures that designs are free of contamination and potential impacts. Finally, one of the major problems with N2O4 is not the problem of spacecraft corrosion but the deposition of corrosion products (picked up during storage) into valves, preventing valve closure and restricting flow. Nitrogen tetroxide dissolves small amounts of iron (a few parts per million) from storage tanks. The solubility of the iron is a function of temperature, water content, and NO content. Proper conditioning of N2O4 prior to loading will precipitate out complex iron nitrate compounds; this will prevent problems caused by corrosion product deposition (see the discussion "Case Histories" in this section).

Monomethyl hydrazine itself can cause problems. Although it is not as unstable or reactive as neat (pure) hydrazine, which will be covered in detail in the discussion "Auxiliary Power Unit" in this section, it must be handled in essentially the same way as pure hydrazine. Rocket chambers used in the reaction control system are made from niobium alloy C-103. These are coated with an R512A silicide coating that prevents oxidation in high-temperature service (up to 1315 °C, or 2400 °F). No oxidation problems with reaction control engine chambers have been experienced in service. The primary chambers are film cooled with hydrazine. Burnthrough has occurred in laboratory testing under off-limit conditions (engine instability); however, current design modifications prevent further structural damage to the spacecraft by using automatic sensing devices that shut off fuel and oxidizer valves if penetration of chamber walls occurs. The vernier engines have also shown localized burnthrough during extensive laboratory testing when hot oxidizer impinges on the silicide coating. This condition is aggravated by the thousands of thermal cycles (thermal fatigue) required by this engine, the inability of the coating to accommodate minor machining offsets or discrepancies without fracture, the lack of a fuel-cooling film along the inside of this chamber design, and a doublet-type injector that limits the mixing of the fuel and oxidizer (see the discussion "Case Histories" in this section). Helium pressure vessels (28 MPa, or 4000 psi) made of annealed Ti-6Al-4V liners overwrapped with Kevlar 49 aramid filament provide the pressure necessary to feed the propellants. Some corrosion problems have been encountered in helium systems. The orbiter has experienced problems with hydrazine vapors migrating into the fine orifices of helium valves. Reactions with surface contaminants have resulted in plugging of orifices due to complex hydrazine deposits (see the discussion "Case Histories" in this section). Helium systems must always be designed to be compatible with the fuels and oxidizers with which they are used, because back migration (diffusion) of these substances will occur. The orbital maneuvering system provides the propulsion to insert the shuttle orbiter into earth orbit, to change orbit, to rendezvous, and to deorbit. As with the reaction control system, it also uses the storable propellants N2O4 and MMH, as well as a helium pressurant system. The two OMS engines are capable of providing 267,000 N (6000 lbf) of thrust. The propellant tanks have the capacity to provide a change in velocity of 300 m/s (1000 ft/s) when carrying a full payload of 29,500 kg (65,000 lb).

The nozzles are made of niobium alloy FS-85 coated with an R512E silicide coating that is used for the RCS rocker chambers. These nozzles are used for service to 1350 °C (2480 °F) (Ref 13). Injectors are made from diffusion-bonded 300-series stainless steel platelets. Platelets have injector hole patterns etched in them by a photographic etching process. The pressure vessels for the fuel and oxidizers, operating at 2.2 MPa (315 psi), are made from annealed Ti-6Al-4V. The helium pressure vessels (33 MPa, or 4800 psi) are made from annealed Ti-6Al-4V overwrapped with Kevlar 49 aramid filament. The fuel and oxidizer present the same types of problems as those experienced in the RCS, except for the nozzle extensions. High-temperature oxidation of the nozzle extensions has occurred in areas where mechanical deformation (buckling) caused cracking and spalling of the silicide coating and where thermocouple attach brackets were broken off in service (see the discussion "Case Histories" in this section). The main propulsion system provides the vacuum-jacketed lines for carrying liquid hydrogen and liquid oxygen

from the external tank to the shuttle main engines. These lines are also used to carry high-pressure high-temperature gaseous oxygen and hydrogen back to the external tank as a pressurant to expel the liquid hydrogen and liquid oxygen. The compatibility problems with gaseous oxygen have been covered in the discussion "Environmental Control and Life Support System" in this section. With gaseous hydrogen, the compatibility issue concerns the embrittlement of metals by hydrogen under certain specific conditions. The extent of embrittlement is a function of hydrogen pressure, strain level, and, probably, time of exposure. the embrittlement is thought to occur from rupture of the protective oxide film of the metal, followed by some mechanism by which atomic (nascent) hydrogen enters the metal. In mild cases, this embrittlement takes the form of a reduction in notched strength in hydrogen compared with specimens exposed to helium or air. In more severe cases, the ductility or tensile strength of the alloy changes significantly. In the case of titanium, embrittling hydrides are formed; within a few minutes, these hydrides can result in the destruction of the entire cross section of a part (see the discussion "Case Histories" in this section). For this reason, exposure of titanium to hydrogen is totally avoided.

With other metals, it is important to verify compatibility at the maximum service pressures. Most metals do not show significant property changes below 3.45 MPa (500 psi). At hydrogen pressures of 13.8 MPa (2000 psi), materials such as Inconel alloy 718 show measurable loss of ductility, and at 69 MPa (10,000 psi), even austenitic stainless steel is affected. No accepted criterion exists for the use of metals in gaseous hydrogen. At the Rockwell Space Transportation System Division, for example, an alloy is used in gaseous hydrogen if its sharp notched (Kt 17) to unnotched strength ratio at maximum design pressure does not fall below 1.0 and if a factor of safety of four is maintained in the system. At the Rockwell Rocketdyne division, materials are strain limited, and materials susceptible to high-pressure gaseous hydrogen are copper plated (0.1 to 0.25 mm, or 4 to 10 mils) to ensure that no adverse reactions occur. Welds in nickel alloys are overlayed on the root side by two layers of Inconel alloy 903. Leakage of hydrogen is a major concern in design because hydrogen can form an explosive mixture with air at concentrations between 4 and 96% hydrogen. The aft areas of the orbiter are extensively purged with helium before launch to avoid this problem if a leak occurs. The vacuum-jacketed lines carrying LOX and LH2 are of two size--305 mm (12 in.) and 430 mm (17 in.) in diameter--and are made of welded Inconel alloy 718. Inconel alloy 718 is ideal for this service because it is compatible with both liquid and gaseous hydrogen (at these pressures) and maintains high ductility below -253 °C (-423 °F). The vacuum-jacketed lines must accommodate expansion and contraction and are designed to do so by articulation and angulation at joints. Bellows of Inconel alloy 718 accommodate angulation while containing the cryogens. The articulation is provided by gimbal rings and internally supported ball-strut tie-rod assemblies. The hardened balls, ranging in size from 32 to 57 mm (1

to 2

in.) in diameter, are made of a tungsten carbide (Stoody 2) alloy.

Inconel alloy 718 (see the discussion "Plumbing Lines" in this section) is used for the high-pressure high-temperature oxygen lines (31 MPa, or 4500 psi, at 280 °C, or 540 °F) that deliver pressurant to the external tank. Helium pressure vessels are made from annealed Ti-6Al-4V and annealed Ti-6Al-4V overwrapped with Kevlar 49 filament. The auxiliary power unit provides hydraulic pressure for the actuation of a number of flight control systems,

including the rudder, speed brake, body flap, elevons, landing gear, and brakes. Power is obtained by vaporizing and decomposing neat (pure) hydrazine in a catalyst bed and passing it through a two-stage turbine, which in turn drives a hydraulic pump. Each of the three APUs develops 100 kW (135 hp), and the pump delivers 18 MPa (3000 psi) to the hydraulic system. The compatibility of hydrazine with materials in a system must consider several points. First, hydrazine can be unstable and will decompose into N2, H2, and NH3, causing a rapid pressure rise and, if not properly vented, an explosion. Neat (pure) hydrazine tends to be more reactive (unstable) than monomethyl hydrazine. Decomposition is catalyzed by metal surfaces and/or contaminants left on surfaces, even after cleaning. The data base in aerospace is large and highly inaccurate regarding metal-hydrazine compatibility for several reasons: • • •

Many of the early investigators rated compatibility as satisfactory based on appearance of the metal specimen exposed, not the fluid reaction No uniform cleaning method or decomposition criteria existed when data were generated No uniform controls were used in testing

The current approach at the Space Transportation and System Division is to use materials that have a history of satisfactory hydrazine service at the temperatures expected. Where a new material is used, pressures rise tests of the new material versus time are made, along with controls of compatible materials cleaned in the same manner. Literature data, because of their unreliability, are not used except to indicate which materials should be tested. Second, a metal should also be checked with hydrazine to determine its autoignition temperature (AIT). This is done to ensure the AIT is safely below operating temperature. No standards are available for autoignition testing. Frequently, the fluid is allowed to drip onto a heated plate of the test metal, and the temperature of the test metal is slowly raised until ignition occurs.

Third, hydrazine, with even short exposure to air (a few seconds), will react with CO2 to form carbazic acid--a very viscous, sticky compound which can clog lines, orifices, or valves. It can also aggressively attack certain metals, such as cobalt and nickel. Hydrazine systems must always be kept under an inert gaseous blanket whenever opened. Fourth, hydrazine (and carbazic acid) will dissolve or selectively leach certain metals. This may later result in malfunction of a valve by the flow-decay phenomenon, in which hydrazine salts of these metals end up precipitating at valve seat areas (see the discussion "Case Histories" in this section). Fifth, every effort must be taken to avoid hydrazine spills, because hydrazine is very toxic (Table 2) and highly flammable in air if spread out over a large area (large surface-to-volume ratio). Ignition has occurred when hydrazine was spilled or orbiter thermal protection system tiles made from sintered pure SiO2 filaments. The metals used in the APU in contact with hydrazine include 300-series stainless steels, precipitation-hardening steels, tungsten carbide (for valve seats), and Hastelloy alloy B. After the hydrazine has been decomposed by the catalyst bed, its major compatibility problem is with the formation of nitrides by the hot ammonia gas formed. The catalyst bed exceeds 925 °C (1700 °F), and the turbine operates at 595 °C (1100 °F). Alloys are chosen both for high-temperature properties and resistance to nitriding. The turbine wheel and blades are made from René 41. The wheel has a circumferential Inconel alloy 625 shroud welded with Hastelloy alloy W wire to the blade tips. The injector section and catalyst bed housing are from Hastelloy alloy B. The hydraulic system uses predominately type 300 series stainless steel valves and components attached to Ti-3Al-

2.5V lines. The titanium lines were chosen because they saved approximately 270 kg (600 lb) over stainless steel Permanent joints are externally "swaged" with Permaswage fittings that also incorporate RTV 630 rubber seal rings as a backup to the metal-to-metal seal. In the presence of hydraulic oil, the rubber expands, ensuring a tight joint capable of sealing hydraulic oil. Hydraulic oil per MIL-H-83282A is used throughout the spacecraft except in the landing gear struts, where oil per MIL-H-5606C is used. Control of the chlorine content of hydraulic oil is the most important approach to ensuring corrosion-free systems. (Chlorine concentrations above 100 ppm can cause corrosion problems.) No major corrosion problems have been experienced in the shuttle orbiter hydraulic systems. Mechanical Systems Mechanical systems include primarily mechanical devices, pyrotechnical devices, and landing gear systems. These systems mainly employ high-strength steels. Inconel alloys, and some aluminum alloys. Mechanical Devices. Steels, because of their high strengths and hardnesses, are often preferred for mechanical

devices. Higher loading can be directed into a smaller space volume than with aluminum or titanium, which are less dense. Typical applications involve 4130 and 4340 low-alloy steels for mechanical devices and 4340 or alloy steel carburized grades for gears. Low-alloy steel parts must be protected from corrosion. Depending on the design and service, the following types of systems are used: • • • • • • • • •

Paints, such as chromated epoxy polyamines Phosphate coating (DOD-P-16232) with oils Cadmium-titanium or cadmium plating plus an overcoating Electroless nickel Chromium , electroplated Black oxide Nickel/tin, electroplated Electroless nickel plus electroplated chromium Vapor-deposited aluminum

In addition, lubrication coatings such as Braycoat grease (space compatible), molybdenum disulfide in an organic matrix, or Vitrolube NPI-1220 (a high-temperature vitreous-base lubricant) are used. Where sufficient corrosion resistance cannot be achieved with steels, corrosion-resistant alloys are used, such as Inconel alloy 718, titanium Ti-6Al-4V or Ti-6Al-2Sn4Zn-6Mo, and precipitation-hardenable steels (such as 17-4PH, 15-5PH, AM350, AM355, and Custom 455).

For bearings, type 440C stainless steel and 52100 bearing steel area used, protected with Braycoat grease. High-strength springs are made of type 301 stainless steel in the condition B spring temper, Elgiloy, and 17-7PH in the CH900 condition. High-strength Belleville washers, using 6150 low-alloy steel, exhibit the best resistance to hydrogen embrittlement if they are coated with vapor-deposited aluminum (see the discussion "Case Histories" in this section). Some use is also made of maraging steels and 9Ni-4Co-0.3C steel for such applications as hydraulic cylinders. Pyrotechnic devices are used to separate the external tank from the orbiter, to release umbilicals, and to open

emergency escape hatches or panels. Pyrotechnic devices operate in a number of different ways. In some designs, the explosive device is within the bolt or nut, forcing its separation. Other designs may cut a panel with a linear-shaped charge or blow a panel joint apart far enough to break bolts at a prenotched section. Explosively actuated guillotines have been used on the Apollo to cut cables or shrouds and have been used in some shuttle applications. Pyrotechnic devices have been used to deploy parachutes or to release landing gear. On the shuttle orbiter, most pyrotechnic devices are made of Inconel alloy 718. This includes the frangible nuts attaching the external tank to the orbiter in the aft section, the explosive bolts that mate the external tank with the orbiter in the forward section, and the crew escape and emergency egress hatches. Guillotine blades have been made of Inconel alloy 718, A-286, and corrosion-protected tool steels. No corrosion problems have occurred in these areas. The orbiter landing gear system is a conventional aircraft tricycle configuration with steerable nose gear and the

main left and right landing gear. The major parts of the system include the shock strut assembly, wheels and tires, axles, brakes and antiskid controls, and nose wheel steering and damping controls. Initiation of the system hydraulically releases uplock hooks that permit the landing gear to free fall into the extended position. Springs and hydraulic actuators assist in the free fall. Pyrotechnic actuators may also unlock the uplock hooks if the hydraulic system malfunctions. When fully extended, the gear is locked by spring-loaded bungees. Minimum weight was a prime consideration in the design of the landing gear. This resulted in the need to use highstrength materials, highly efficient braking materials, and minimum thicknesses for tires and wheels. Tires were designed for 2 landings, brakes for 5 landings, and wheels for 100 missions, although increased loading has reduced wheel life somewhat. The steel used for parts carrying high loads was primarily 300M at a 1895-MPa (275-ksi) strength level. The use of high-strength steels in this manner represented a deviation from limitations placed by the Material and Process Group on steel strength levels throughout the spacecraft (see the discussion "Primary Structure" in this section). At this high strength level, steel is notch and impact sensitive, has limited toughness, is prone to stress-corrosion problems, and is highly susceptible to hydrogen embrittlement. Its selection, however, was based on a history of satisfactory use in landing gear applications for both civilian and military aircraft over a 20-year period. Nevertheless, serious embrittlement problems were encountered during development and in early orbiter service. The 300M steel was corrosion protected on functional surfaces by either cadmium alone or by chromium plating. Nonfunctional surfaces were coated with a cadmium-titanium plating plus a chromated epoxy polyamine paint primer and a polyurethane topcoat. Normally, cadmium plating would be avoided because of its tendency for sublimation and redeposition in the space environment, but the environment of the landing gear in the orbiter wheel wells precluded space exposure and sublimation problems. No corrosion problems have been identified with the 300M finish system in service; however, several misprocessing problems with chromium plating and plating of the cadmium-titanium finish have resulted in hydrogen embrittlement failures (see the discussion "Case Histories" in this section). The cadmium-titanium plating is considered to be a low-embrittlement plating process and is currently covered by MIL-STD-1500. The current brake designs use thermal grade beryllium heat sink disks (both rotors and stators) with reinforced carboncarbon linings. They are chemically filmed in a manner similar to that described for aluminum in MIL-C-5541 and are unprotected on their mating (working) surfaces. Temperatures during braking can be very high locally because of uneven pressure distributions and have occasionally resulted in beryllium carbide formation (>1095 °C, or 2000 °F) during highenergy stops in early brake designs. Carbide formation results in embrittlement, cracking, and potential failure of the braking system. Unfailed parts showing this condition are scrapped. Galvanic corrosion between the carbon and beryllium would appear to be a concern, but such a problem has never occurred. Newer brake designs will use structural carboncarbon to replace the beryllium; this will result in longer life (lower lifetime costs) and higher energy absorption capability, but higher axle and wheel temperatures are expected. Axles were made of 300M steel. The nose landing gear axle was later changed to Inconel alloy 718 for improved fatigue life. The cadmium-titanium coating on the 300M main landing gear axles will not exceed 230 °C (450 °F) during normal operation and is safe from liquid cadmium embrittlement. In a rare, high energy level emergency stop, this temperature

could be exceeded, resulting in a potential for cadmium embrittlement. This would occur as heat from the brakes soaks back into the axle well after the vehicle has stopped. Under these conditions, the axle would be replaced. Wheels are currently made of die-forged halves of aluminum alloy 7049-T73 that are bolted together with high-strength MP35N fasteners. The main wheels are chemically filmed according to MIL-C-5541, followed by a chromated epoxy polyamine primer (MIL-P-23377) and a topcoat of polyurethane according to MIL-C-83286 for corrosion protection. The nose wheel substitutes anodizing for the chemical film treatment; otherwise, it is identical. No corrosion problems have been experienced. Aluminum alloy 7075-T73 is used for hydraulic actuators. Exterior surfaces are protected by chromate conversion coating plus chromated epoxy polyamine primer and polyurethane topcoat. Interior surfaces are immersed in hydraulic fluid and are not coated. No corrosion problems have occurred with the aluminum acuators. Avionics All orbiter avionics systems are located within the spacecraft structure or pressurized cabin; therefore, from a corrosion standpoint, they experience a controlled and relatively benign atmosphere that is free of rain and salt spray. Electronic equipment located within the orbiter cabin or cargo bay areas, because of the controlled humidity, cannot experience condensation of moisture, but equipment within the aft equipment bays can experience potential condensation. Based on the types of equipment, corrosion protection can be further examined in three areas: • • •

Black boxes Electronic circuits Electrical connectors

Black boxes are typically made of the relatively corrosion-resistant 6000-series aluminum alloys, either welded or dip

brazed and resin sealed. The boxes are painted or anodized according to MIL-A-8625 on the exterior. They are chemically filmed according to MIL-C-5541 on the interior and on areas that contact coldplates to ensure maximum heat transfer and electrical grounding. The paint system is not fixed; it may be the chromated epoxy polyamine primer used for the spacecraft or a system chosen by the manufacturer; however, it must meet the requirements for exposure to humidity and salt spray of Federal Test Method Standard 141, Methods 6201 and 6061, respectively. Some boxes are sealed with a gasket; others are environmentally sealed and filled with an inert gas. Air-cooled black boxes located in the crew cabin are unsealed. All electrically active surfaces of electrical and electronic circuits, including solder joints, printed circuits, and wire terminations, are conformally coated. The coatings used include Dow Corning 3140 and 3145, Columbia Technology Hysol and Furane Plastics polyurethanes, and General Electric RTV 560 and RTV 566. Printed circuit conformal coating thicknesses range from 50 to 250 m (2 to 10 mils). Coatings on other surfaces range from 100 to 375 m (4 to 15 mils). Electrical connector metallic parts are made of various aluminum alloys of 300-series stainless steels. All aluminum parts are nickel plated. Plating thickness is not always specified, but conectors must pass salt spray corrosion tests per MIL-STD-202, method 101, test condition B.

Connector pin and socket contacts are electrolytically gold plated over a copper and nickel strike. A gold plate thickness of 1.25 m (50 in.) is considered to be borderline between porous and nonporous plating. All orbiter connector contacts include a minimum gold plating thickness of 2.5 m (100 in.). When mated, the connectors are environmentally sealed by peripheral gaskets. No major problems have occurred with corrosion of avionics devices. Nitrogen tetroxide spills in the OMS pod areas and forward RCS areas have required in-depth evaluation of the suitability of connectors and pin contacts; however, no abnormal functions were noted even though connector bodies turned green.

Case Histories Despite extensive efforts to anticipate and avoid corrosion problems with manned spacecraft, such problems will inevitably occur. Corrosion problems, as a group, have clearly become the most critical and costly problems with metals used on space vehicles. Because many corrosion problems can result in unexpected or catastrophic failures, especially those of SCC, hydrogen embrittlement, and metal ignition (in oxidizers), their impact is usually severe. Extensive efforts must be made to identify the precise cause of failrue and to provide suitable inspections and rework for existing hardware; the lives of the astronauts cannot be jeopardized nor can the risk be taken of severe damage to a $2 billion orbiter. As a consequence, some corrosion problems have resulted in rather expensive launch delays, and the inevitable questions are asked: "Why did this occur? Why was this not foreseen?" There are no simple answers. Field experience is unparalled in revealing differences in behavior from laboratory test results or the deficiencies in the hardware designs. The reasons for corrosion problems discussed in this section can be broadly categorized as follows: • • • • • • •

Lack of adequate protection of parts during the manufacturing cycle Failure to remove processing chemicals or fluids completely Failure to provide an adequate hydrogen embrittlement relief Use of improper or contaminated fluids during manufacturing or testing Failure to control fluid chemistry within a spacecraft fluid system Inadequate corrosion protection Unknown reactions or unforeseen problems

In this section, a variety of different corrosion problems that have occurred will be presented. These embrace many metal systems, such as aluminum, stainless steel, low-alloy and precipitation-hardenable steels, nickel, titanium, and niobium, as well as several different types of corrosion attack. The descriptions that follow will identify the specific causes of the corrosion, its characteristics, and how it was corrected. Pitting Attack Pitting attack of metals is caused by local breakdown in a protective oxide film due to local surface inclusions or defects, local changes in microstructure (such as by precipitates that strengthen the metal), or local corrosion cells brought about by deposition of more cathodic materials on a surface. When a pit occurs, the bottom of the pit is anodic and often continues to grow until perforation occurs. Pitting attack of the 2000- and 7000-series aluminum alloys often occurs during the manufacturing cycle if parts are not properly protected with oil or chemical film (MIL-C-5541). Many parts are transported to machining sources, chemical milling plants, and heat-treating companies for processing, often in open-top trucks in California. Occasionally, parts are not unloaded over the weekend and are left exposed to the elements. In other cases, parts may be stored outside, perhaps protected by plastic or a tarp, or even stored inside next to doors; these methods of storage alloy condensation to form on the parts. No outdoor storage is satisfactory. Dust that collects on the parts absorbs moisture from the air, resulting in significant pitting--in some cases within 24 h. Moisture condensation under plastic in contact with the part results in severe attack. Aluminum Spacecraft Structural Parts. Typical pitting attack (Fig. 6) occurred on an aluminum alloy 2014-T6 tests part during the Apollo program. The maximum depth of attack was 0.1 mm (4 mils), about the same as the diameter. The pitting is intergranular. If the quench rated is sufficiently rapid during the heat treating of aluminum, no intergranular corrosion occurs within a pit (see the article "Heat Treating of Aluminum Alloys" in Heat Treating, Volume 4 of ASM Handbook). Because pits such as that shown in Fig. 6 are only slightly larger in diameter than a human hair, they may be missed unless close surface examination of aluminum is made at a magnification of 5 to 10×. This has made it difficult to convince in-house manufacturing personnel or subcontractors of the seriousness of protecting surfaces, and it is only when a gross area of pitting becomes obvious that arguments cease. Another Rockwell facility in Tulsa insisted that the pitting problem was characteristics of the California environment and even ran 6-month exposure tests of unprotected aluminum to prove it. Results are shown in Fig. 7. The coupon shown in Fig. 7(a) was heat-treated aluminum alloy 2024 exposed in the metal-processing area, and the one shown in Fig. 7(b) was heat-treated aluminum alloy 2014 exposed in the manufacturing area.

Fig. 6 Pitting corrosion of an aluminum alloy 2014-T6 sheet. Pitting occured during the manufacturing cycle. Note the intergranular nature of the pit. 150×.

Fig. 7 Heat-treated aluminum pitted from 6 months of in-plant exposure. (a) Aluminum alloy 2024 exposed in the metal-processing area. (b) Aluminum alloy 2014 exposed in the manufacturing area. Note the intergranular nature of the pits. Both 285×

Typical pitting on improperly protected spacecraft parts is shown in Fig. 8. The pits shown are typical of those found in a 7075-T6 radial shear beam of the Apollo Service Module after a fatigue test failure. Although the fatigue test represented an overtest of the structure, pitting and extensive end-grain attack from the chemical milling process contributed to the failure. Figure 9 shows typical pitting corrosion of an aluminum alloy 2024-T62 fitting resulting from inadequate protection during manufacturing.

Fig. 8 Pitting corrosion of an aluminum alloy 7075-T6 aluminum radial shear beam from the Apollo program. The beam is machined and chemically milled from a 64-mm (2.5-in.) think plate to a final nominal thickness of 0.45 mm (0.018 in.). Pitting occurred from improper protection either during manufacturing or in service. 30×

Fig. 9 Pitting corrosion of an aluminum alloy 2024-T62 structural fitting used on the space shuttle orbiter. (a) Fitting that was pitted from lack of interim protection after machining. 0.25×. (b) Enlargement of pitted surface. 4×

Pitting of structures may lead to premature fatigue failure in critically loaded or cycled parts; however, in many cases, it is not critical. Pitting of honeycomb sandwich face sheets may undermine the integrity of the sandwich by moisture ingestion. Pitting of gear teeth or springs often leads to failure in mechanical systems. Pitting in fluid-containing hardware (pressure vessels, tubes) may lead to perforation and leakage. By the time pitting is discovered on structural parts, the part may have already cost $25,000 because of extensive machining, or it may be part of a welded assembly worth over $1 million dollars; therefore, every effort is made to save the part. Pitting corrosion is aluminum must be deactivated, or pits will continue to grow as the aluminum oxide that is formed continues to absorb moisture from the air. Pits can be deactivated in place by using a proprietary deoxidizer, followed by a deionized water rinse, wipe, and pH test of the pits. Local masking is used. Pits are then brush chemically filmed according to MIL-C-5541 and painted. Where stress analysis permits, pits can be sanded flush or removed with a dental drill or hole drill. In some cases, bonded doublers are added to restore strength. Filiform Corrosion Filiform corrosion is a special form of corrosion that occurs underneath a protective film. It is a moving oxygen concentration cell. Corrosion takes the form of threadlike or filamentary trails and proceeds along the metal surface rather than penetrating through the thickness. Significant filiform corrosion can occur in a matter of hours or days. It develops in the presence of relative humidities as low as 60%. A key condition for the development of filiform corrosion is that the film is semipermeable, permitting oxygen as well as humidity to pass through it. Filiform corrosion, therefore, is essentially a form of crevice corrosion in which one member forming the crevice (the protective film) is semipermeable. In filiform corrosion, an anodic head, typically 0.08 to 0.13 mm (3 to 5 mils) wide, advances and dissolves the metal in its path. The pH of the head is highly acidic (often as low as 1), and the tail or trail increases in pH, often exceeding a pH of

8. As the anodic head advances, the cathodic region behind it fills with corrosion products. Filiform corrosion can be recognized by the following characteristics: • • •

Relatively shallow corrosion of the metal surface Meandering, filamentary pathways Occurrence below a paint or other protective film

To prevent filiform corrosion, coatings with lower permeability to water and oxygen are required. The following is an example of failure of a pressure vessel by filiform corrosion. Liquid Hydrogen Pressure Vessel. During modification of the Columbia vehicle, an aluminum alloy 2219-T6 liquid hydrogen tank used for supplying hydrogen to the fuel cells was removed from the vehicle and stored inside the plant for a 3-month period; the tank was inadvertently exposed to the atmosphere. When this was discovered a moisture test was conducted of the air inside the vessel. It indicated higher than acceptable levels. The exterior of the pressure vessel was covered by a vacuum jacket (see the discussion "Electrical Power System" in this section) and protected from corrosion. Boroscopic examination of the interior of the tank revealed a corroded zone completely encircling the vessel on both sides

of the electron beam girth weld. The zone was approximately 4.7 mm ( the edge of the girth weld (Fig. 10).

in.) wide and centered 9.5 mm (

in.) from

Fig. 10 Filiform corrosion in an aluminum alloy 2219-T6 hydrogen tank used on the space shuttle orbiter. Attack was due to atmospheric humidity. Small spherical beads in (a) are splatter from electron beam welding. (a) Root side of weld showing filiform corrosion beyond the HAZ. (b) to (d) Enlargements of the corrosion attack at 60×, 235×, and 1180×, respectively

It was not possible to verify the depth of the attack or to deactivate the corrosion in place. Therefore, it was necessary to remove the tank from service and examine it metallographically. Although the attack turned out to be superficial, the fatigue life of the vessel could not have been evaluated nondestructively. At the time of examination, it was believed that the corrosion zones on both sides of the welds were brought about by a preferential anodic phase in the weld HAZs or a preferential anodic zone caused by differences in permeability and thickness of the protective oxide coating. If moisture had condensed in the tank, corrosion would have taken place at the

tank bottom. Because the corrosion path encircled the weld, it was suspected that corrosion occurred in the water vapor phase. Surprisingly, metallographic examination revealed that the corroded areas were entirely free of corrosion products (Fig. 10). This indicated that the corrosion had occurred at an earlier stage of manufacture and was passivated and removed by chemical cleaning. Although searches of manufacturing records could not verify it, the most probable cause of corrosion was believed to be filiform corrosion occurring under a tape used to attach a protective paper or film over the surfaces prepared for welding. No other explanation adequately accounted for the lack of corrosion products, the shallowness of attack (0.1 mm, or 0.4 mil), and the filamentary network observed. Galvanic Corrosion Galvanic corrosion results from or is accelerated by dissimilar-metal contact. The more electronegative metal becomes the anode and is corroded. The more electropositive metal becomes the cathode and is not attacked. The severity of galvanic corrosion depends on the flow of current; therefore, corrosion rates tend to be accelerated when larger differences in potential exist between the two metals. Corrosion damage to the anode becomes more severe as the cathode-to-anode ratio increases. Corrosion rates also increase as solution conductivities increase. To prevent galvanic corrosion, parts can be isolated from each other electrically, can be coated, or the anode can be cathodically protected by impressed current or other sacrificial, more anodic materials or coatings. If paint is used, the most effective member to coat is the cathode because the reduction in the cathode-to-anode ratio is extremely effective. If the anode alone were coated, corrosion would rapidly occur at coating defects accelerated by the very large cathode-to-anode area ratios. When galvanic couples occur in spacecraft structures that exceed the requirements given in Table 1, the accepted design practice is to use two coats of paint as a moisture barrier and for electrical isolation. Designs should prevent entrapment of water and should provide for sealing of crevices. In fluid systems, detrimental galvanic couples are avoided. In the first 6 years of service of the space shuttle orbiter, four major galvanic problems arose: three with structural components and one in a fluid system. Forward Wing Spar. Attached to the forward wing spar on the space shuttle orbiter are A-286 stainless steel fittings that support the reinforced carbon-carbon leading edges of the wing. Leading edge temperatures exceed 1260 °C (2300 °F), and the wing spar made from an adhesive-bonded aluminum sandwich structure must be insulated to keep its temperature below 175 °C (350 °F). The spar face sheets are made of aluminum alloy 2024-T81 and have been chemically milled down to 0.35 mm (0.014 in.) in some areas. The insulation blanket material is a Dynaflex (alumina silica chromia fibers) felt with a density of 130 to 390 kg/m3 (8 to 24 lb/ft3). It was packed in an embossed, resistancewelded Inconel alloy 601 foil that is 0.1 mm (0.004 in.) thick.

During inspection of the wing of the Columbia vehicle after one of its early flights, blisters were seen in the chromated epoxy polyamine paint used to protect the spar (Fig. 11). The blisters occurred mostly along the edge pattern of the Inconel blankets. Removal of the paint revealed highly localized pitting with depths ranging from 12 to 350 m (0.5 to 14 mils), perforating the face sheets in the thinnest areas.

Fig. 11 Galvanic corrosion of aluminum alloy 2024-T8l sheet of orbiter front wing spar from the Columbia orbiter. (a) Corrosion appeared as aluminum oxide deposits (arrow) along the edge of an Inconel alloy 601 foil covered insulation blanket. (b) Open pits (arrow) in an area that was in contact with the foil insulation blanket. (c) Oxide buildup under paint in a similar contact area causes formation of "bubbles" (arrow)

The interface between the Inconel foil package and the wing spar permitted capillary moisture entrapment, especially while the orbiter was vertical on the launch pad. Although it was protected with two coats of the chromated epoxy polyamine paint, the galvanic current was able to perforate the paint. Because the foil blankets were unpainted, the cathode-to-anode ratio at paint flaws was extremely large. Accelerated salt spray testing showed that the corrosion could appear within 300 h with two coats of paint, but aluminum was fully protected from the galvanic corrosion with a 75- m (3-mil) coating of RTV 560 or three coats of paint. The pits on the orbiter were deactivated. Pit depths were measured. More than 700 local areas exceeded acceptable pit depth criteria and required the bonding of approximately 200 doublers to prevent fatigue or perforation into the honeycomb core. The spar was refinished with three coats of paint; the edges of the foil blankets were coated with 75 m (3 mils) of RTV 560. Elevons. Similar galvanic corrosion of aluminum alloy 2024-T81 occurred in the elevons under flipper doors and rub

panels where Inconel alloy 601 foil insulation blankets rested against painted aluminum honeycomb face sheet. Corrosion was deactivated; however, in this case, a nylon-type surface insulation was used to replace the foil blankets, eliminating the galvanic couple. Orbital Maneuvering System Pod Structure. Galvanic attack similar to that discussed above occurred in the OMS

pod of the Columbia, where goldized Kapton multilayer reflective insulation was in contact with painted aluminum. In this case, a faulty environmental seal permitted water entry, but no provisions were available to drain the water. The pits

were sanded, ground out, or deactivated. The area was chemically filmed and painted with two coats of chromated epoxy polyamine paint and then recoated with 75 to 125 m (3 to 5 mils) of RTV 560. Organic-coated aluminized Kapton insulation was also substituted for the goldized Kapton insulation. Environmental Control and Life Support System Outlet Water Valve. After the sixth flight of the orbiter, a

potable water valve was discovered to be leaking. The leak emanated form a zone of corroded material that traversed the length of the valve forging. Although the valve body was manufactured from a type 304L stainless steel forging, the material within the corroded zone was found to be low-carbon steel. It was learned that the carbon steel had been accidentally introduced in the ingot in the final stages of solidification. Subsequently, the ingot was rolled into bar and then forged. Because any of more than 50 other valves from that heat could have the same defect, a complete review of all parts was made. The defect could be detected easily by the copper sulfate test according to Method 102 of MIL-STD753A. It was recommended that this test be performed followed by repassivation. Figure 12(a) shows the tube section with a longitudinal streak through it (appears similar to a crayon mark). Figure 12(b) shows the corroded area that laked. Attack is seen completely through the tube wall in Fig. 12(c). The microstructure was then etched with 3% nital to bring out the carbon steel (Fig. 12d). Figures 12(e) and 12(f) show x-ray dot maps of nickel and chromium concentrations, respectively, through the inclusion shown in Fig. 12(d). Figures 12(b) to (d) show that the carbon steel is anodic to the stainless steel in water and that corrosion was accelerated by this galvanic couple. Although this problem could not have been anticipated, the supplier would have recognized the problem if he had passivated and inspected the stainless steel properly.

Fig. 12 Galvanic corrosion of forged type 304L stainless steel ECLSS outlet water valve used on the shuttle orbiter. (a) Tubular section of water valve with embedded longitudinal inclusion (between arrows). 2.5×. (b) Enlargement of inclusion revealing perforation. 125×. (c) Section through perforated area. 45×. (d) Section through inclusion showing beginning of galvanic attack. Nital etch (3%) used to bring out structure of carbon steel. 85×. (e) Electron dot map of nickel content through the inclusion. 55×. (f) Electron dot map of chromium content through the inclusion. 55×

Intergranular Corrosion Intergranular corrosion results where the microstructure of an alloy provides a preferred corrosion path along grain boundaries. Often, the material adjacent to the grain boundary is depleted of a particular element that precipitates out as an intermetallic compound at the grain boundary. In the case of aluminum alloys of the 2000 series, the precipitation of

CuAl2 at grain boundaries occurs. The CuAl2 is more noble than the adjacent aluminum, and intergranular corrosion of the adjacent aluminum takes place. Artificial aging of aluminum causes precipitates to occur throughout the grains as well as in the grain boundaries, thus minimizing localized corrosion. Stainless steels, when heated into or cooled through the 425- to 870-°C (800- to 1600-°F) range, will precipitate chromium carbides preferentially at grain boundaries, resulting in anodic paths adjacent to these grain boundaries. Lowering carbon levels to below 0.03% (to avoid a continuous carbide grain-boundary network) or adding niobium or titanium (to precipitate carbides) will permit the chromium in the grain boundaries to keep the steel from becoming sensitized. Intergranular corrosion is insidious in that little actual corrosion is required before the structural integrity of a part or the perforation of a fluid container occurs. A knowledge of metallurgy is important in preventing intergranular corrosion attack because microstructure is sensitive to composition, temperatures, cooling rates, and surface contamination. The ammonia boiler, as previously described, is a heat exchanger using Freon-21 and ammonia. The Freon-21 picks up heat from the orbiter electronic boxes, the crew cabin, and the payloads, and the heat is dissipated by boiling ammonia when the orbiter radiators cannot be deployed, such as during the final stages of reentry. Accidental contamination of these tube surfaces prior to brazing of the heat exchanger resulted in the sensitization of some of the thin-wall (0.2 mm, or 8 mil) type 304L stainless steel tubes. The sensitized tubes were attacked and perforated by the fluid (Fig. 13).

Fig. 13 Intergranular corrosion of a type 304L stainless steel tube in a shuttle orbiter ammonia boiler. (a) Test performed to show tube ductility. 1×. (b) Cross section through the thin-wall (0.2 mm, or 8 mils) tube revealing sensitization on outside diameter due to carbonaceous deposit formed during brazing. 75×. (c) Surface SEM showing grain-boundary carbides are being removed from outside diameter during corrosion. 980×

The corrective action was to change the tubing to type 347 stainless steel to prevent chromium carbide precipitation at grain boundaries if parts were accidentally contaminated prior to brazing. In addition, procedures to prevent contamination in the future were upgraded. Particular emphasis was placed on removal of carbonaceous, drawing compounds from tube surfaces before brazing. Selective Leaching Selective leaching (also called parting, dealloying, or demetallification) occurs when the corrosion process removes one or more elements from the alloy matrix. Specific categories of selective leaching often carry the name of the dissolved element in their title, such as dezincification, dealuminification, denickelification, or decobaltification. In the case of gray cast iron, selective leaching is called graphitic corrosion. The process can occur in single-phase alloys, for example, with brasses having high zinc content such as a 70Cu-30Zn alloy. It can also occur in multiphase alloys. In the selective leaching process, typically one of two mechanisms occurs; alloy dissolution and replating of the cathodic element or selective dissolution of an anodic alloy constituent. In any case, the matrix that is left is spongy and porous and has very little strength or integrity. In some cases, a large, localized area of metal (a plug) is attacked. Selective leaching can be prevented by proper alloy selection, that is, matching the alloy system to a particular environment. Changing ratios of alloying elements and adding inhibition elements are often satisfactory approaches. In

aerospace, selective leaching is a rare problem. Where it occurs, a proper alloy selection or changing the processing medium is the key to prevention. Gas Generator Valve Module (GGVM) Valve Seats. The APU of the shuttle orbiter contains a GGVM that

utilizes four tungsten carbide valve seats to regulate hydrazine flow to the catalyst bed. The 5-mm (0.2-in.) diam valve seats have sealing lands that are only 115 to 150 m (0.0045 to 0.0060 in.) wide. The seats are manufactured from sintered KZ-96 tungsten carbide containing 5% Co as binder. In mid-1985, valve leakage problems during acceptance testing were traced to a breakdown of the valve seat sealing lands. The problem became widespread, with a high rejection rate during acceptance testing procedures (ATP). It was determined that revised cleaning procedures that had been implemented recently to correct a GGVM contamination problem early in the program were subjecting the seats to long periods of exposure to deionized water and ultrasonic cleaning. Tests proved that such exposure produced leaching of the cobalt binder from the sintered tungsten carbide alloy. Subsequent impacting of the valve seats by the poppets (during ATP) resulted in breakdown of the seats and the noted leakage failures (Fig. 14).

Fig. 14 Selective leaching of a tungsten carbide valve seat in a shuttle orbiter APU gas generator valve module. Leaching of the cobalt binder was caused by excessive exposure to water during ultrasonic cleaning and hot water rinsing. (a) Valve seat showing narrow sealing surface 0.1 to 0.15 mm (4.5 to 6 mils) wide. 8×. (b) Loss of sealing surface due to selective leaching and poppet impact. 45×. (c) SEM cross section showing depth of leaching. 1600×

As a by-product of the testing, it was found that a small amount of leaching of the cobalt also occurs from exposure to decomposition by-products of hydrazine, the system fluid. This leaching action is several orders of magnitude less than that from deionized water. Tests were underway in the time period from 1985 to 1986 to identify alternative tungsten carbide alloys that might be acceptable substitutes for KZ-96. Alloys under investigation included those with different binders and those with finer grain sizes. The corrective action taken during the manufacturing cycle was to substitute isopropyl alcohol for deionized water in all operations. Crevice Corrosion Crevice corrosion results from a concentration cell formed between the electrolyte within the crevice, which is oxygen starved, and the electrolyte outside the crevice, where oxygen is more plentiful. The material within the crevice acts as the anode, and the exterior material becomes the cathode. This is similar to pitting, in which the base of the pit becomes the anode. The author finds it convenient to view a crevice as a plane of corrosion, that is, a two-dimensional pit. The resistance of materials to crevice corrosion varies widely. Those metals whose protective oxides films result from oxygen adsorbed to the surfaces and are in dynamic equilibrium with the outside environment, such as stainless steels, appear to suffer most when shut off from oxygen by crevices. In good corrosion design practice, crevices are avoided whenever possible. Crevices not only trap water or the chemical being processed but also become sumps for the other contaminants in the system, often resulting in major corrosion problems. Spacecraft structural faying surfaces, for example, either include a faying surface sealant or the edges are fillet

sealed (see the discussion "Structural Joints and Fasteners" in this section). Sometimes, however, a crevice cannot be avoided, as shown in the cases below. Therefore, additional care is required to avoid detrimental corrosion. Anodized Aluminum Window Frames. The exterior window frames on the space shuttle orbiter are made of aluminum alloy 2124-T851 that has been anodized according to a Rockwell specification by using a sulfuric acid anodize, a black dye, and a sodium dichromate seal. The window frames are protected from heat during spacecraft entry by pure silica tiles. Two grooves within each window frame contain the fibrafax thermal seal (to prevent hot gas plasma flow) and the Viton pressure (environmental) seal. The side windows, when the orbiter is stacked for launch on the pad, provide for possible water entrapment in a small portion of the periphery of the windows. Rain can wash down window surfaces, picking up salt deposits from seacoast exposure. The aluminum window frame peak temperature is approximately 55 °C (130 °F).

After the eighth flight of the space shuttle orbiter, two side windows were removed from the Challenger vehicle for examination. The window surfaces had appeared to have a hazy opacity even after polishing. The opacity was due to microscopic erosion of unknown origin. Examination of the window frame showed several localized areas of corrosion through the anodized coating in and adjacent to the seal grooves. The areas away from the crevice were uncorroded (Fig. 15). The recommended corrective action for new window frames was to add a coat of chromated epoxy polyamine primer, followed by a coat of polyurethane over the black anodize.

Fig. 15 Crevice corrosion of an anodized aluminum alloy 2024-T851 window frame from the space shuttle Challenger. Corrosion occurred along both thermal and environmental sealing grooves. (a) Window frame showing locations of corrosion (arrows). (b) Enlargement of (a) showing corrosion in Viton seal area (arrows). Rain water carrying dissolved salt deposits from the window was the corrosive medium.

Aluminum Brazed Joints. The nature of the brazing process is to provide a gap between faying surfaces that will act as a capillary for braze alloy flow. Selection of the braze processes used for the plumbing systems lines on the Apollo and the space shuttle orbiter depended on the braze alloy filling the capillary gap to ensure the inside of the tube would not have a crevice (see the discussion "Plumbing Lines" in this section). These stainless steel brazes were made under an inert gas shield, and no flux was required.

In brazing aluminum, however, it is necessary to use a flux to remove its tenacious oxide film and to ensure wetting of the faying surfaces. The brazing fluxes usually consist of mixtures of alkali and alkaline earth chlorides and fluorides, sometimes containing aluminum fluoride or cryolite (3NaF·AlF3). Braze fluxes will often become entrapped between the faying surfaces, and if the joint is completely sealed by a fillet, no problems arise. However, when incomplete fillets occur, the brazing fluxes, which are generally hygroscopic, will bleed out and cause corrosion. Cleaning of the part and verifying that surfaces are free of fluorides and chlorides do not always solve the problem, because weeks later the part can be covered with flocculent aluminum hydroxide corrosion products. Brazing of aluminum is generally avoided on the shuttle orbiter wherever possible. Aluminum cold plates are made by fluxless brazing in inert gas under temperatures and pressure sufficient to ensure braze alloy flow. In electronic boxes in which braze designs have been used, the joints are vacuum impregnated with a resin seal to avoid bleedout of entrapped fluxes. This appears to be quite satisfactory. On the Apollo program, aluminum-to-stainless steel tube brazes were used in discrete applications. The stainless tube was tin plated. Problems experienced with braze bleedout are shown in Fig. 16.

Fig. 16 Crevice corrosion of an Apollo aluminum-stainless steel brazed joint caused by bleedout of the brazing alloy. Upper portion is aluminum alloy 6061-T6, lower portion is tin-plated type 304L stainless steel. Brazing alloy was 718 aluminum. (a) Foaming of aluminum hydroxide corrosion products (arrows). Entrapped flux exposed to air caused corrosion. 1×. (b) Cross section through pockets of entrapped flux (arrows)

Fretting Corrosion Fretting is the abrasive wear of two touching surfaces subject to cyclic relative motions of extremely small amplitude. Fretting corrosion is an increased degree of deterioration that occurs because of repeated corrosion or oxidation of the freshly abraded surface and the accumulation of abrasive corrosion products between these surfaces. Although fretting is often limited to small, localized patches of wear, it can eventually provide a path for leakage (for example, valve seats) or an initiation site for fatigue. Fretting corrosion can be controlled by lubrication of the faying surfaces, restricting the degree of movement, or by the selection of materials and combinations that are less susceptible to fretting (see Table 4 in the article "Corrosion in the Aircraft Industry" in this Volume). Rudder Speed Brake Power Drive Unit Spacer. The mounting bolts of the rudder speed brake power drive unit of

the space shuttle orbiter are made of A-286 stainless steel heat treated to 965 MPa (140 ksi). Bolts are either 15.9 or 22.2 mm ( or in.) in diameter. The bolts are sleeved with a spacer that passes through a spherical bearing. The spacer is made of 17-4 PH steel (H1150M), and the bearing is Inconel alloy 718 heat treated to 1240 MPa (180 ksi). The surface finish on the spacer is 16 RHR (root height reading). When the power drive unit was removed from the Enterprise vehicle, fretting corrosion was discovered on the exterior of the spacer and the interior of the ball. The fretting corrosion of the 17-4PH was quite severe, as shown in Fig. 17. The corrective action consisted of changing the spacer to Inconel alloy 718 and applying a dry film (molybdenum disulfide) coating.

Fig. 17 Fretting corrosion of a steel spacer used to mount the rudder speed brake on the shuttle orbiter. The spacer is made of 17-4PH H1150M stainless steel. (a) Spacer on bolt shows contact area with an Inconel alloy 718 spherical bearing. Fretted area is between arrows. (b) Enlargement of fretting corrosion. 1×. (c) Mating Inconel alloy 718 bearing showing a similar pattern but only superficial marring of surface. 1.5×. (d) Cross section through fretting corrosion. 175×

Stress-Corrosion Cracking Stress corrosion requires the simultaneous occurrence of three conditions: a susceptible material or microstructure, a corrosive environment, and surface tensile stresses. Control of stress corrosion is achieved by avoiding any one of these conditions (see the discussion "Control of Stress Corrosion" in this section). Each metal family has its own unique environments in which it displays susceptibility to stress corrosion. Stress-corrosion cracking is one of the most insidious forms of corrosion because it often comes on without any warning and results in major, sometimes catastrophic, structural failures. In fluid systems, the presence or absence of a trace element can make the difference between no reaction and a major failure. Prior laboratory testing may have missed a potential problem that occurs in service, as illustrated in some of the failures below. It is mandatory that the cause of a stress-corrosion failure be demonstrated in the laboratory after a suspected stress corrodent has been identified and that testing be conducted as closely as possible to the chemical and metallurgical conditions experienced by the failed hardware. The injector tube of the space shuttle orbiter auxiliary power unit carries liquid hydrazine to a catalyst bed, where it is

heated and decomposes into nitrogen, hydrogen, and a trace of ammonia. The hot decomposed gases drive a turbine wheel to generate secondary power for spacecraft systems. Shortly after touchdown from the ninth launch of the orbiter, two of three APUs detonated (Ref 14). Extensive investigation determined that the cause of the detonations was the decomposition of hydrazine. While the orbiter was in orbit, hydrazine had leaked through cracks in the Hastelloy alloy B injector tube walls of two different APUs. In the space vacuum, evaporation withdrew heat from the leaking fluid, resulting in the formation of hydrazine snow balls. During reentry, the snow balls melted and ignited--somewhere below 12 km (40,000 ft) in altitude--and the heat resulted in decomposition and detonation of the hydrazine. The fractures on each tube were intergranular, started on the inside diameter, occurred in almost identical locations, and extended 220 to 240° around the periphery (Fig. 18). The cracks were determined to be caused by stress corrosion, as evidenced from their appearance and by extensive testing that eliminated all other failure mechanisms.

Fig. 18 Stress-corrosion cracking of a Hastelloy alloy B orbiter APU injector tube in a hydrazine environment. (a) Typical injector tube and catalyst bed. (b) Failed tube of APU #1. Arrow shows area of SCC. (c) Fracture faces of failed APU #2 injector. Dark areas (arrows) have been stress corroded. Bright areas were ductile fracture from tensile load used to separate parts. 6×. (d) Enlarged fracture face of (c) showing intergranular character. 940×. (e) Surface SEM of tube inside diameter showing etching of carbides in grain boundaries (arrow) and surface. 370×. (f) Sensitized surface layer on inside diameter due to carbide contamination (from electrical discharge machining process) entering braze cycle. 285×. (g) Enlargement of surface corrosion. 3450×

It was demonstrated by laboratory testing that ammonia or ammonium hydroxide was the only potential fluid that could cause stress corrosion of Hastelloy alloy B on the spacecraft. The ammonium vapors resulted from decomposition of hydrazine in the catalyst bed, probably as a result of hydrazine leaking into the injector tubes through the valve seat or from the equilibrium of the gas generator environment within the tube after shutdown. Moisture, resulting in the formation of ammonium hydroxide, was available from atmosphere migrating back into the exhaust duct. Misalignment and cocking of the injector tubes during installation resulted in high stresses in the failure area (up to yield stress). A sensitized microstructure of precipitated carbides along inside diameter grain boundaries provided a preferred path for stress corrosion to proceed. A carbide network found on the inside diameter of the injector tubes was the result of carbon deposited during the electrical discharge machining process to machine the injector tube bore. Subsequent injector tube brazing and cool-down cycles permitted grain-boundary diffusion and precipitation. The solution was to eliminate the preload stresses on the injector tube by instrumenting the installation and to eliminate the sensitized carbide network by reaming the tube inside diameter. The braze cycle was raised to 1185 °C (2165 °F) to ensure uniform diffusion of any carbides, followed by rapid cooling to ensure that any carbides inherent to the basic alloy

(0.05 max) would not precipitate at grain boundaries. Later designs of injector tubes were also internally coated with a thin chromized layer. The Reaction Control System injector of the Space Shuttle orbiter is made from uncoated niobium alloy C 103. This injector is bolted to a titanium mounting ring section. A niobium alloy C 103 chamber is then welded to the injector face. Cracking was observed in the sharp radius of the injector adjacent to the bolting ring. Examination of the fracture face showed that the attack was intergranular and had fine eruptions (popcorn balls) of niobium oxide on its surface.

A review of the manufacturing sequence isolated the problem to those steps performed between the bolting of the titanium flange section to the injector flange and the final operation in which the engine was baked at 315 °C (600 °F) for 20 h to remove resin from the insulation encasing the engine. The most probable cause of cracking was thought to be stress corrosion occurring from entrapment of an etchant (50% HNO3-50% HF) used to remove traces of iron and copper from the niobium prior to welding of the chamber to the injector. Repeated laboratory attempts to duplicate failure were unsuccessful until a method was devised to entrap the etchant between two pieces of niobium while tensile stresses were applied during sustained heating in the 290- to 315-°C (550- to 600-°F) range. The failure was duplicated and attributed to hot fluoride salt SCC (Fig. 19) (Ref 15). Performing the acid etching and rinsing prior to bolting the injector assembly prevented future occurrences.

Fig. 19 Fluoride hot salt SCC of niobium alloy C 103 injector used on the orbiter RCS chambers. (a) Schematic of C 103 injector and titanium bolting ring showing failure area. (b) Cross section through failure showing intergranular attack. 60×. (c) Fracture face showing grain boundaries and microscopic eruptions of niobium oxide. 1035×

Reaction Control System Oxidizer Pressure Vessels. Nitrogen tetroxide, a storable hypergolic oxidizer, was

used in the service propulsion system (SPS) and in the reaction control system on the Apollo program. The SPS provided the propulsion for orbit insertion, lunar flight, return from the moon, and deorbit for entry. The RCS provided for vehicle attitude control through roll, pitch, and yaw engines. Titanium alloy Ti-6Al-4V was chosen for pressure vessels in both systems as a result of laboratory tests on corrosion, stress corrosion, and impact ignition with N2O4. Qualification testing of the SPS pressure vessel for 46 days of exposure under a membrane stress of 690 MPa (100 ksi) was completed without problems in mid-1964. In January 1965, an RCS pressure vessel of the same alloy, protected from N2O4 by a teflon positive expulsion bladder, cracked in six adjacent locations (Ref 9, 10, 11, 12). The cracks were parallel to each other and perpendicular to the maximum stress. The fracture surface had a red stain and was flat and brittle in appearance. Because no N2O4 was supposedly in contact with the vessel, it was believed that misprocessing may have

caused the fractures. Subsequent testing of ten pressure vessels without bladders in June of that year was conducted to ascertain if misprocessing could be bracketed to certain manufactured pressure vessel lots over a period of time. Within 34 h after testing, one of the pressure vessels blew up (Fig. 20) and within a few days most of the others had failed.

Fig. 20 Stress-corrosion failure of an Apollo Ti-6Al-4V RCS pressure vessel due to nitrogen tetroxide. (a) Failed vessel after exposure to pressurized N2O2 for 34 h. (b) Cross section through typical stress-corrosion cracks. 250×. (c) Correlation between the number of cracks per square inch and stress level. (d) Cracking in cylindrical section where hoop stress predominates. (3) Cracking in biaxial area where stresses are approximately equal. (d) and (e) Both 35×

The cause of the failure was immediately suspected to be stress corrosion, but the aggressive fluid that contacted these surfaces was not immediately determined. Cracks occurred very close together on the inside of the pressure vessel, and the number of cracks per unit area was proportional to the local stress. Pressure vessels of the same design, when tested with N2O4 by Rockwell on the West Coast, did not fail. The RCS tank contractor. Bell Aerosystems Company, was soon able to demonstrate coupon failures in N2O4, but Rockwell could not, although over 300 specimens were cleaned and tested under more than 40 variables, including various contaminants.

Chemical testing of the propellants used at Bell and Rockwell revealed no differences or out-of-specification conditions. Existing chemical techniques were not capable of accurately quantifying all species present in the N2O4, especially compounds of nitrogen; however, cooling of the propellants revealed a color difference in N2O4 between the supplies at Bell and Rockwell. The N2O4 at Rockwell, when cooled to -18 °C (0 °F), turned green because of the presence of nitric oxide (NO), but the N2O4 at Bell was yellow. The green color resulted from a mixture of N2O4 (yellow) and dissolved NO as N2O3 (blue). Rockwell supplies of N2O4 had been purchased earlier than those of Bell Aerospace. Investigation revealed that a change in the military specification MIL-P-26539 had been made during this time period to improve the specific impulse of N2O4 by oxygenating the trace quantities of residual NO. This simple change had a devastating effect on its stress-corrosion behavior with titanium. Testing indicated as little as 0.2% NO was probably sufficient to inhibit the SCC of titanium. Specifications were changed thereafter to require a minimum NO content of 0.6%; present grades contain 1.5 to 3% NO. Addition of NO to existing supplies solved the problem. Early in the investigation, it was believed that no stress corrosion could occur in solutions as nonconductive as N2O4 (specific conductivity at 25 °C or 80 °F, is 3.1 × 10-13 - · cm-1). However, because of the low conductivity of N2O4, only closely spaced local cathodes and anodes could carry corrosion currents (Fig. 20). This resulted in a large number of cracks (up to 70 cracks/in.) rather than a single crack. This investigation also illustrated how past stress-corrosion test results or pressure vessel qualification can be voided by minor chemical changes in the corroding medium. This is further illustrated in the following discussion. Service Propulsion System Fuel Tanks. The storable hypergolic fuel used for the service propulsion system on the

Apollo Service Module was a blend of 50% hydrazine and 50% unsymmetrical dimethyl hydrazine. It was contained in two titanium Ti-6Al-4V pressure vessels approximately 1.2 m (4 ft) in diameter and 3 m (10 ft) long. Because hydrazine compounds are toxic and dangerous to handle, a "referee" fluid with similar density and flow characteristics was used in system checkout testing. Methanol was chosen as a safe fluid based on subcontracted studies conducted for the program. Methanol had been successfully used as a fluid in a tri-flush cleaning process for propellant systems at that point in time. Among the other advantages, it had a low explosive potential, was miscible with both fuel and water, and would leave surfaces residue free. During an acceptance test of the Apollo Spacecraft 101 service module prior to delivery, an SPS fuel pressure vessel (SN054) containing methanol developed cracks adjacent to the welds (Fig. 21). The test was stopped. This acceptance test had been run 38 times on similar pressure vessels without problems. Failure analysis could not reveal the cause of the cracking. The fractures had branching cracks characteristic of stress corrosion yet fracture faces exhibited a somewhat featureless quasi-cleavage appearance without typical stress-corrosion features. The adjacent material was ductile and within chemistry.

Fig. 21 Stress-corrosion cracking of a solution-treated and aged Ti-6Al-4V Apollo SPS fuel pressure vessel during a system checkout test. Fluid test medium was methanol. (a) Cross section adjacent to weld in cracked vessel. 65×. (b) Another crack near the same weld. 65×. (c) and (d) TEM fractographs of fracture surface showing no particular stress-corrosion features. Both 2500×

Misprocessing of pressure vessels during manufacturing was suspected, because the supplier had previously demonstrated that pickup of cleaning agents and contaminants on a Ti-6Al-4V pressure vessel prior to a heat-treat aging cycle would result in delayed stress-corrosion failure. These included such contaminants as finger prints, chlorinated kitchen cleansers, and liquid hand soap. A similar crack adjacent to the weld occurred on the first development pressure vessel for this program was thought to be caused by such contamination (Fig. 22).

Fig. 22 Stress-corrosion failure in an Apollo Ti-6Al-4V pressure vessel development test. (a) and (b) TEMs of fracture face of stress-corrosion crack in the vessel in a 24-h distilled water exposure after contamination of titanium with soap prior to heat treatment (aging). Fine hairlike wrinkles are characteristic of stress corrosion. (c) and (d) Stress-corrosion failure of first Apollo SPS development pressure vessel of Ti-6Al-4V. Cause unknown

The Spacecraft 017 service module was then put into test. An additional test was initiated to ensure that no marginal SPS pressure vessels would pass through system checkout. This additional test consisted of 25 pressure cycles, followed by a 24-h pressure hold. After only a few hours into the hold cycle, the replaced SPS pressure vessel failed catastrophically (Fig. 23).

Fig. 23 Stress-corrosion failure of a solution-treated and aged Ti-6Al-4V Apollo SPS fuel pressure vessel during a sustained pressure test in methanol. (a) Explosive failure of the tank occurred, resulting in severe ripping of the cylinder section. (b) Dome section fragmented by explosion. (c) Fracture origin and TEM fractographs of fracture face showing quasi-cleavage failure. (d) Surface of outside diameter showing crazing of oxide film under load and machining lines. 2500×. (e) Similar view of inside diameter surface. 2500×. (f) Titanium notched stress-corrosion specimens cracked at pin-loading areas in methanol. Pins were austenitic stainless steel.

The investigation conducted after this failure disclosed that titanium will undergo SCC in methanol (Ref 9, 16, 17, 18, 19, 20). Attack is promoted by crazing of the protective oxide film. It was learned that minor changes in the testing procedures could inhibit or accelerate the reaction. For example, the addition of 1% H2O inhibited the reaction completely. It could be restarted by a 5 ppm addition of chloride. Initial stress-corrosion testing in the laboratory was performed with available aluminum text fixtures, with 300-series stainless steel, and then with titanium. No failures occurred in aluminum fixtures (apparently it provided cathodic protection). Failures in stainless steel fixtures occurred at the specimen pin areas, not at the sharp notches used to initiate stress cracking. Failures occurred in only a few hours with stainless steel because it apparently accelerated the reaction by galvanic coupling. Specimens tested to the same stress levels in titanium fixtures often took several times as long to fail. The obvious solution to the problem was to replace the methanol with a suitable alternate fluid. Isopropyl alcohol was chosen after considerable testing. This incident further resulted in the imposition of a control specification (MF0004-018) for all fluids that contact titanium for existing and future space designs.

This failure further illustrates the importance of trace chemicals in accelerating or inhibiting a failure. It also points out that test fixture materials can affect test results. Although it is preferable to use fixtures of the same alloy, it may not always be critical in the screening of possible stress-corroding fluids, because it is often desirable to accelerate stresscorrosion testing during the screening phase. This acceleration can be carried out by significantly increasing the stress (using a notch), by raising the temperature (usually), or by making the specimen the anode in a fluid. It is notable that failures involving stainless steel fixtures did not occur when 1% H2O was added to methanol, nor did failures occur in isopropyl alcohol, benzene, Freon TF, Freon MF, and distilled water and MMH, even if HCl was added or bubbled through the solutions. Tower Leg Frangible Nut. The launch escape tower was designed to pull the Apollo command module free of the

Saturn V launch system in the event of an abort to ensure astronaut safety. Under normal launch conditions, the tower is jettisoned. It is separated from the command module by firing two pyrotechnic charges at opposite sides of each of the four attach nuts (frangible nuts). The detonation separates each of the frangible nuts into two pieces. The tower attachment to the command module lay inside the tower leg well. The frangible nut was made of 4340 steel heat treated to 1520 MPa (220 ksi). A few days before a scheduled launch, one of the nuts was reported to have fractured (Fig. 24).

Fig. 24 Stress-corrosion failure of an Apollo launch escape tower leg frangible nut. The nut was made of 4340 low-alloy steel heat treated to 1520 MPa (220 ksi). The corroding medium was ammonium fluoborate leached from the ablative on the solid rocket engine nozzle extensions of launch escape tower. Note salt accumulation.

(a) Frangible nut (arrow) located within launch escape tower leg. (b) Fracture face of failed nut in (a) showing extensive rusting. (c) Top view of nut. The fracture face is at top. (d) TEM fractographs. Views 6 and 7 are origin area. Note intergranular attack and corrosion of grain faces.

The part and fracture face were rusted and encrusted with salt. The access door to the tower leg well had been left open in a rain storm. Laboratory efforts to duplicate the stress-corrosion failure in salt (NaCl) water were unsuccessful after 30 days of exposure. The encrusted salt on the fractured nut was analyzed during this testing period; it was ammonium fluoborate, not salt from the sea coast. Stress-corrosion testing in this solution produced an immediate failure. Subsequent investigation disclosed that the ammonium fluoborate was a salt added to the ablative material on the launch escape motor nozzle skirts. Apparently, this salt had been leached out during a storm and ran down the tower leg in through the open access door. Obviously, this is not a scenario that would have been anticipated during design. The corrective action was to ensure that the doors were sealed in the rain and that the ablative was painted. Aluminum Structural Elements. In October 1967, a structural failure occurred on a Lunar Excursion Module (LM) test article (Ref 17). A crack occurred in a web splice plate made of aluminum alloy 7075-T651. Investigation revealed that shims had been omitted from the assembly and that high installation stresses resulted in a stress-corrosion failure in the atmospheric environment. The low stress-corrosion threshold of the alloy and temper ((a) >

0 to > 0 to < > 0

0 to > 0 to < V(e) 0

0(b) 0 0 to < HE(f)

0 0 0 HE

> < > 0

V




V

0

0

>

> >

>L, SCC(g) >L, SCC

> 0

0 to > 0 to >

>E(h) 0

> >

0 0, SCC if H 0 V 0 V

0 0, SCC if H 0 V 0 to < > 0 0, SCC if S(l)

> 0 0 > 0 > 0, SCC if H >

0 0 0 V 0 > 0 0, SCC if S

0 0, unless H(i) 0 0, unless H > 0(j) 0 >(k) 0, unless H 0

0 0 > 0 0 > 0 0

>SCC 0, unless H 0 >SCC 0 > > SCC >

>, generally increases attack. 0, generally has no effect on metal in question. 200 ppm NaCl). Thiosulfates also increase general corrosion rates of copper-base alloys. Sulfites. Sulfite additions are often made to control the pH of pulp stock. Bisulfite (the stable form of sulfite in white

water) is also produced during the decomposition of hydrosulfite brightening solutions. Sulfite is not considered to be aggressive; in fact, sulfites are often used as oxygen-scavenging corrosion inhibitors in other aqueous systems, such as boiler feedwater. Sulfites are readily oxidized to sulfate in the presence of dissolved oxygen. Sulfites may have inhibitive properties in paper machine white water. On the other hand, SO2 evolution can occur under conditions of high sulfite concentration and low pH, leading to severe atmospheric corrosion problems. Conductivity. White water conductivity depends on the concentrations of all dissolved ionic species, both inorganic and

organic. For carbon steel, cast iron, and copper-base alloys, higher conductivity is an indication of higher corrosivity. This is not the case for stainless steels as long as they are passive and are not undergoing localized corrosion attack. The corrosivity of carbon steel due to paper machine white water has been monitored by taking instantaneous linear polarization corrosion rate measurements together with simultaneous white water samples for analysis. Linear regression of the carbon steel corrosion rate with individual white water properties reveals that conductivity provides the best regression coefficient, r2:

Property Conductivity pH (on standing) Na2SO4 NaCl pH (on standing) pH (initial) Temperature

r2 84 74 72 66 64 18 2

Corrosion Mechanisms Compared with other pulp and paper mill process streams, paper machine white waters are not particularly aggressive. The conventional austenitic stainless steels used for the construction of wetted paper machine components are nonetheless subject to various forms of corrosion attack. The three major corrosion problems affecting stainless steels in paper machine white water systems are: • • •

Chloride pitting and crevice corrosion Thiosulfate pitting Microbiological attack

In recent years, there have been reports of rapid attack of type 304 stainless steel equipment in mills where this alloy had served well for decades. The recent unsuitability of type 304 stainless steel can be attributed to closure, hydrosulfite brightening, or both. Chloride Pitting and Crevice Corrosion. Stainless steels rely on the stable formation of a passive surface film for

immunity to corrosion in paper machine white waters. Oxidizing conditions must prevail for the passive film to form and be maintained. Dissolved oxygen from the air is sufficient to maintain stable passivity. Stainless steels can also withstand slightly reducing conditions in white waters without suffering serious attack. There is a certain safe range of oxidizing conditions within which the stainless steel corrosion potential can vary (these oxidizing conditions can be measured electrochemically as a range of oxidizing potentials). In the presence of aggressive anions such as chlorides, however, this safe potential range is narrowed. If a certain critical corrosion potential (called the breakdown potential) is exceeded, chloride ions can attack the stainless steel surface. The attack manifests itself as pits because passive film breakdown occurs only in isolated locations, such as weak spots in the film due to defects in the underlying metal. Pitting attack is favored by the following conditions: • • • • • •

Higher chloride concentration Higher temperature Highly oxidizing conditions Low pH Stagnant or low-velocity conditions Lower molybdenum content in the stainless steel

A new chloride pit tends to repassivate unless the favorable conditions for initiation are maintained until the pit has become established. The metal within the pit becomes anodic with respect to the surrounding metal outside, which becomes cathodic because of ready access to dissolved oxygen for the reduction half of the net corrosion reaction. Anodic dissolution and subsequent hydrolysis of metal ions inside the pit result in the generation of free hydrogen ions, which in turn promote the diffusion of additional anions into the pit to maintain charge neutrality. Because chloride anions are much more mobile than sulfate anions, chlorides preferentially migrate into the pit. Eventually, the solution inside a growing pit may come to resemble a solution of hydrochloric acid more than paper machine white water, and pit growth may continue almost independently of external conditions. Crevice corrosion initiates beneath deposits or in other areas shielded from direct contact with the white water environment (Fig. 8). Crevice corrosion initiates more readily than pitting attack, which it closely resembles, because the

conditions for an oxygen concentration cell already exist. The metal surface inside a crevice has difficulty maintaining passivity because of reduced access to dissolved oxygen; on the other hand, the surrounding metals has ready access to dissolved oxygen. Passive film breakdown within the crevice is thus facilitated by the natural tendency of the crevice to become anodic and the surrounding metal to become cathodic. Once crevice corrosion is initiated, its growth mechanisms is the same as that for pitting corrosion.

Fig. 8 Crevice corrosion of a type 316L checking plate located adjacent to a headbox apron. Corrosion developed under pulp pads that formed despite the highly polished surface.

The primary corrosion concern with white water system closure is that the critical chloride concentration and/or temperature for stainless steel breakdown will be exceeded. Under such conditions, pitting corrosion may occur spontaneously. This is a particular concern with type 304L stainless steel. Stainless steels are chosen for service in chloride pitting environments on the basis of molybdenum content. Because higher molybdenum alloys are also more expensive, it is common to select the alloy with the minimum molybdenum content required for resistance to pitting. A hierarchy of austenitic alloys, representing increasing resistance to chloride pitting and crevice corrosion, can be listed in increasing order or minimum molybdenum content:

Alloy

Molybdenum, %

Type 304L

0

Type 316L

2

SIS 2353

2.5

Type 317L

3

UNS N08904 4 UNS S31254 6 UNS N10276 15

Although type 316L stainless steel has thus far appeared to be resistant under conditions of high closure, it is often produced to have a molybdenum content barely above the minimum of 2% (rather than on the high end of the specified 2 to 3% range). A greater margin of safety for chloride pitting resistance can be ensured by instituting a supplementary requirement that the molybdenum content be no less than 2.5% or by purchasing according to Swedish specification SIS 2353. Thiosulfate Pitting. Thiosulfate is an aggressive pitting agent, especially for stainless steels that do not contain

molybdenum. Thiosulfate pitting, unlike chloride pitting, occurs below, rather than above, a certain critical potential--the thiosulfate reduction potential. Reduction of thiosulfate in the presence of hydrogen ions produces and adsorbed sulfur monolayer on the metal surface. The adsorbed sulfur activates the anodic dissolution of the metal and hinders repassivation. Excess hydrogen ions must be present for acidification of the pit; further, there must also be a larger amount of inert ions (sulfate and chloride) that can be transported into the pit to meet charge transfer requirements. The worst case for thiosulfate pitting occurs within the molar concentration ratio:

Above the range represented by the ratio, there is insufficient thiosulfate to reach the pit nucleus. Below this range, there is too much thiosulfate reduction, which prevents acidification of the pit. Thiosulfate corrosion particularly affects those grades of stainless steels that do not contain molybdenum. Once formed, pits are very stable and are not subject to spontaneous repassivation. Scratches encourage the initiation of pits. Few, large pits tend to form rather than many, small pits, as in chloride pitting. Sensitized type 304 stainless steel (weld heat-affected zones, HAZs) is particularly susceptible to thiosulfate pitting. Type 316L stainless steel is the minimum grade that should be used for white water service where high thiosulfate levels may exist. It is recommended that thiosulfate levels be controlled below 5 and 10 ppm for equipment made of type 304 and 316 stainless steel, respectively. Microbiological Corrosion. Paper machine white waters contain nutrients that can sustain bacterial growth.

Microbiological growth thrives in near-neutral pH environments. White water temperatures are also usually within the favorable range of 40 to 50 °C (105 to 120 °F). Although higher temperatures may prevent the growth of some forms of bacteria, increased temperatures can increase the metabolism of those bacteria that can adapt to heat. The result is the formation of slimes. Stock and white water flow systems are designed to minimize slime accumulations. Surfaces are polished and weld projections removed to prevent hang-ups. Wherever slime deposits can build up, however, microbiological corrosion can occur. Once a deposit has grown to a sufficient thickness to exclude oxygen, a colony of sulfate-reducing bacteria (Desulfovibrio desulfuricans) can become established. Enzymes produced by these anaerobic bacteria catalyze the reduction of sulfates to form free sulfide ions. Chemically reducing conditions quickly develop, resulting in the depassivation of the stainless steel surface beneath the deposit. Active corrosion in the form of pitting then proceeds. Corrosion due to Desulfovibrio desulfuricans is manisfested by the presence of large, shallow pits covered with a black crust (Fig. 9). Perforations through stainless steel equipment are usually small because the entry of oxygen at a leak will stop the activity of sulfate-reducing bacteria (Fig. 10).

Fig. 9 Section of type 304L stainless steel plate removed from a tapered header used to deliver stock to a headbox showing severe microbiological corrosion

Fig. 10 External view of the type 304L stainless steel tapered header in Fig. 9, showing leakage occurring at small perforations

Corrosion Control in Pulp Bleach Plants Andrew Garner, Pulp and Paper Research Institute of Canada

Pulp mill bleach plants have traditionally used austenitic stainless steels because of their combination of good corrosion resistance and weldability. Type 317L (18Cr-14Ni-3.5Mo) has been the typical bleach plant alloy for oxidizing acid chloride environments. However, bleach plants have become more corrosive over the past 20 years as mills have closed wash water systems and reduced effluent volumes. In modern closed bleach plants, type 317L is no longer adequate for long-term service (Ref 1), and many mills have turned to higher-alloy stainless steels, nickel-base alloys, and titanium for better corrosion resistance. Metals are chosen over nonmetals for moving equipment, such as washers. Metals are stronger, tougher, have better fatigue properties, and, if they have sufficient corrosion resistance, require virtually no maintenance. However, the more corrosion-resistant alloys are more costly, and the challenge is to choose an alloy with just enough resistance to avoid corrosion problems. A wide selection of alloys is available for bleach plant applications. The list includes three families of stainless steels (austenitic, ferritic, and duplex), whose differing merits are outlined in Table 1. Table 2 list most of the commercially available candidate alloys and their chemical compositions. Table 3 outlines the influence of each alloy component on bleach plant corrosion resistance.

Table 1 Characteristics of three families of stainless steels for bleach plant service Family Austenitics

Ferritics Lowinterstitial type

Ti- or Ti + Nb stabilized type

Duplex

Examples 316L 317L 904L 254SMO AL-6XN Nitronic 50

Characteristics Tough, ductile, readily welded without loss of corrosion resistance; corrosion resistance related to alloy content

Comments Bleach plant steels traditionally chosen from this group

As above, with better pitting resistance than type 317L

Manganese-substituted austenitic, may be better value than type 317L; not common at present

29-4-2 29-4

Not as tough as austenitics, particularly after welding thicker sections. Special precautions needed for welding to avoid N2 pickup. Corrosion resistance, related to alloy content, can be very good.

29-4C NYBY MONIT SEACURE 2205 Ferralium

0.02% C max versions

Higher alloys have remarkable corrosion resistance. Thin-section (0.015% S) in the alloy or contamination of the weld area. It is most commonly seen in the HAZ in the previous pass of a multiple-pass weld. Hot cracking rarely has a detrimental effect on the mechanical properties or structural integrity of a fabrication. However, it can be very detrimental to the corrosion properties of a weldment. Microfissures form crevice corrosion sites that are readily attacked. Recent laboratory tests have shown that other fillers can be used for 904L, such as Sanicro 27.31.4L CuR, Smitweld NiCro 31/27, and Thermanit 30/40 E. These fillers would be appropriate for microfissure-free shop construction. However, contaminated weldments are best repaired with Inconel alloy 112 or an equivalent filler. Another common problem with stainless steel weldments--sensitization in the HAZ--is avoided in bleach plants by the use of low-carbon steels (0.03% C max for austenitics). Similarly, fusion-line attack (sometimes called knife-line attack) due to precipitation of carbides at the fusion line in niobium- and titanium-stabilized austenitic steels is rarely seen, because these steels have been made obsolescent by new steelmaking technology. However, attack at the fusion line is possible when overalloyed fillers such as Inconel alloy 112 are used with high heat input welding. Such welding can create zones consisting of melted base metal that is not mixed with weld filler--called unmixed zones--at the fusion line. Cases of unmixed zone corrosion have occasionally been observed in the bleach plant. In practice, this can be minimized by the use of lower heat input on the final weld passes.

Electrochemical Protection The discussion thus far has centered on the selection of material or the control of the environment to minimize corrosion. A third approach has recently become available with the development of electrochemical protection for bleach plant washers (Ref 10, 15, 16). General Description. The life of a stainless steel washer can be greatly extended if the washer is cathodically

polarized from the oxidizing potentials imposed by residual oxidants such as chlorine (Fig. 17) to a more negative, passive potential by the use of a rectifier and a platinized anode mounted in the washer vat (Fig. 18). Electrical contact to the washer is made through a rotating mercury contactor; the washer potential is measured with a reference electrode and is automatically controlled with a feedback-controlled rectifier to a potential set point or window that minimizes corrosive attack.

Fig. 17 Effect of residual Cl2 concentration in the washer vat on the free-corrosion potential of a type 317L stainless steel bleach plant washer. Source: Ref 15

Fig. 18 Components of a washer electrochemical protection system

Principle of Protection. A detailed description of the principle of electrochemical protection is provided in Ref 15.

However, the essential feature of the technique is the use of the electron as an antichlor for corrosion protection. Electrons are fed from the rectifier to the washer in the form of electric current. At the washer surface, they react with chlorine (Ref 2):

Cl2 + 2e-

2Cl-

Therefore, like SO2, electrons react with the corrosive oxidant (chlorine) to form relatively harmless chloride ions. For corrosion protection, the electron has a clear advantage over SO2: It can be delivered to the cathodic reaction site. For this reason, comparatively few electrons are needed. The required current is low, and running costs are negligible. Because of the comparatively low cost of electrochemical protection and the ease of retrofitting to existing washers, commercialization of the technique has found wide acceptance. The first successful operation of one of these systems was in Nova Scotia, Canada, in 1978. By the end of 1986, there were about 90 installations worldwide. Monitoring Technique. The corrosion rate of each protected washer is monitored with coupons, using a technique

designed for comparison of protected and unprotected coupons (Ref 17). A mounting bolt is welded to the end face of the rotating washer as shown in Fig. 18, and two coupons, together with segmented crevicing disks, are mounted so that one is in electrical contact with the washer and the other is isolated, with all other mounting details being identical (Fig. 19).

The degree of protection is assessed by comparing the corrosion of protected and unprotected coupons after a 60-day exposure period (Fig. 20).

Fig. 19 Crevice corrosion monitor assembly

Fig. 20 Type 317L stainless steel monitor coupons after a 60-day exposure attached to a C-stage washer. The coupon on the left was in electrical contact with the washer and was therefore protected. The coupon on the right was isolated and unprotected. Substantial crevice corrosion occurred on the right-hand coupon at 7 of the 20 possible crevicing sites.

Results to Date. Both weight loss and depth of attack measurements have been made on coupons to assess the

comparative severity of pitting or crevice corrosion (Ref 10). The ratio of unprotected and protected coupon weight loss is used as a measure of protection. Table 6 lists protection ratios measured for washers that have been protected for up to 5 years (6 washers types/212 test coupons). These results show that protection lowers corrosion rates. On the average, unprotected coupons lose six times more weight than protected coupons.

Table 6 Type 317L test coupon data from the first six electrochemical protection systems Washer

C-stage D2-stage D2-stage D1-stage D2-stage D1-stage General average

Protection period, years 5.4 4.0 3.3 2.9 2.9 1.8 6.0

Average protection ratio(a) 4.9 2.6 12.3 8.2 1.9 6.2

Unprotected ÷ protected type 317L coupon weight loss (larger ratios indicate better performance).

(a)

Improved performance with protection can be compared with improved performance after alloy upgrading in the absence of protection. The results of an extensive bleach plant exposure program, which involved 880 coupons, 40 test racks, 10 mills, and no electrochemical protection (Ref 1), were reexamined to obtain the average weight loss ratios (using type 317L as a base case) for alloys 904L and 254SMO. These data are given in Table 7; greater ratios indicate better performance. Type 317L coupons lose 5.5 times as much weight as 254SMO coupons. Therefore, the improved performance achieved by the electrochemical protection of type 317L appears to be very close to that gained by upgrading type 317L to unprotected 254SMO.

Table 7 Test rack weight loss data for unprotected steels Steel

Average weight loss ratio(a)

Type 317L 1 904L

2.5

254SMO

5.5

(a)

This ratio is obtained, for example, by determining the average weight loss for type 317L coupons ÷ average weight loss for 904L coupons (larger ratios indicate better performance).

Economics. The installation and operating costs of electrochemical protection systems are small compared to the

resulting cost savings. Capital cost savings are such that if the life of a washer is extended from 5 to 10 years by protection, the protection system will have a payback period of about 1 year. Even this substantial saving can be overshadowed by savings in chemical costs in mills that add NaOH or SO2 to pulp before washing. Experience has shown that, in general, SO2 additions need not be made ahead of a protected washer to protect it from corrosion. If pulp souring is still required, then this can be done immediately after the washer, where much less SO2 will be needed. Additional savings can be realized by eliminating SO2 use in a closed bleach plant, because SO2 free recycled filtrate used for tower dilution consumes much less unreacted ClO2. Protection systems have been installed on washers made from types 316L, 317L, and 904L, and 254SMO stainless steels. A protected 254SMO washer probably represents the state-of-the-art for corrosion control in the most severe washer environments.

Corrosion by Sulfite Pulping Liquors C.B. Thompson and Andrew Garner, Pulp and Paper Research Institute of Canada

Sulfite pulping, one of the oldest methods of chemical pulping, dates back to the 1860s. For many years, sulfite pulping was the primary chemical method for making paper, although since the 1950s it has been overtaken in importance by the

kraft process to such an extent that the sulfite industry was perceived by many as dying out. In recent years, however, a growing appreciation of the versatility of the sulfite process and the appearance of combined sulfite and mechanical pulping methods (semichemical or chemi-mechanical) have combined to ensure the future of sulfite pulping. These changes in the industry are thoroughly documented in Ref 18). This section will discuss the various methods of sulfite pulping, the principal corrosion mechanisms occurring in sulfite environments, and the major corrosion problem areas found in sulfite mills. Information on corrosion in the disk refiners used in the later stages of semichemical and chemi-mechanical pulping can be found in the section "Corrosion of Mechanical Pulping Equipment" in this article.

Sulfite Pulping Sulfite Chemistry. Sulfite pulping liquors are prepared by dissolving SO2 in a solution of calcium, sodium,

magnesium, or ammonium hydroxide. The pH of the resultant cooking liquor depends on the base and the amount of SO2 dissolved, and it can range from 1 to above 13. The exact value is chosen according to the particular mill process and the desired pulp characteristics. Fresh sulfite liquor is essentially a mixture of sulfite and bisulfite ions in an aqueous solution of SO2. The ratios of these three components can vary widely, according to the liquor pH. The relative concentrations of bisulfite and sulfite ions a different pH levels are shown in Fig. 21; bisulfite dominates at pH values less than 6, while sulfite ions dominate at pH values above 7. The concentration of aqueous SO2 also increases at lower pHs (Ref 19).

Fig. 21 Relative concentrations of bisulfite ( ), sulfite ( ), and aqueous sulfur dioxide (SO2) as a function of liquor pH at 130 °C (265 °F). Aqueous sulfur dioxide and bisulfite dominate at acidic pHs, sulfite dominates at alkaline pHs. Source: Ref 19

Traditional Sulfite Pulping. For many years, most sulfite mills used a calcium hydroxide base cooking liquor, mainly because of the economy of the calcium carbonate feedstock. The calcium-base process is restricted to very low pH (typically 1.5) and low cooking temperatures (140 to 150 °C, or 295 to 300 °F) because of scaling and solubility problems. Furthermore, the process is difficult to adapt to high-yield pulping. These reasons, as well as increasingly stringent antipollution legislation, have led to the expanded use of soluble sodium, magnesium, and ammonium bases in recent years.

A typical sulfite operation is shown in Fig. 22. Sulfur (either in powder or liquid form) is burned to produce SO 2. The gas is cooled rapidly to minimize the formation of sulfur trioxide (SO3) and then passed into absorption towers, where it is

dissolved in the hydroxide base to form the raw pulping liquor. Pulping is carried out in either batch or continuous digesters; when the cook is complete the chips are released into a blow tank. The chips are then taken for washing, screening, and bleaching as required. The spent cooking liquor (red liquor) can be evaporated and burned in a furnace for steam generation and/or recovery of cooking chemicals, or it can be further treated to obtain chemical by-products. A typical operation might also include equipment for SO2 recovery from the cook and an acid accumulator.

Fig. 22 Typical layout for a traditional low-yield sulfite mill operation

Recent developments in sulfite mills have centered around the need to decrease pollution, to increase pulping efficiency (that is, increase the proportion of usable fiber obtained per cook), and to adapt to market demands for specific grades of pulp. These requirements have resulted in a variety of processes, such as high-yield sulfite pulping and chemithermomechanical pulping, in which sulfite chemical treatment of the chips is combined with additional treatment of the fiber in a disk refiner. The interrelationship of these process with pure chemical pulping and pure mechanical pulping is shown in Fig. 23. An important practical point to note is that high-yield and very high-yield sulfite pulping require the use of a digester, but chemi-mechanical and chemi-thermomechanical pulping use smaller impregnation vessels or steaming tubes for pretreatment of the chips.

Fig. 23 Interrelationship between chemical, semichemical, and mechanical pulping processes. The processes marked with an asterisk use impregnation vessels or steaming tubes rather than digesters. Source: Ref 18

Corrosion Mechanisms in Sulfite Liquors Stainless steels are used for most process equipment in the sulfite mill. As in any industrial plant, the steels in sulfite mills can experience many different types of corrosion. The most serious ones are usually connected in some way to the presence or absence of SO2. Chloride-induced localized corrosion and SCC may also be problems, particularly in coastal mills. The presence of SO2 in sulfite liquors can affect corrosion of stainless steels in three ways. It can: • • •

Maintain passivity against acidic cooking liquors Form H2SO4 by decomposition of the liquor Form H2SO4 by oxidation to SO3

These effects are discussed in more detail below. Maintaining Passivity. Sulfur dioxide in solution helps to maintain the passivity of stainless steels such as types 316L

and 317L that are commonly used in sulfite mills (Ref 20, 21). Therefore, a potential corrosion problem exists in any situation in which the concentration of dissolved SO2 becomes very low. This may happen, for example, in vacuum evaporators or in batch digesters during a cook. However, practical experience has shown that this type of corrosion is usually not of concern with batch digesters. Unless conditions are especially severe (that is, high temperature, low pH, high chlorides), austenitic stainless steels containing more than 2.7% Mo can withstand short periods without any dissolved SO2 present. Opening the digesters to the atmosphere between cooks also helps to maintain passivity. Sulfuric Acid Formation by Liquor Decomposition. Acid bisulfite liquors can spontaneously decompose to form

H2SO4 (Ref 22):

3SO2 + 2H2O 2H2SO4 + S 3NaHSO3 NaHSO4 + Na2SO4 + S + H2O

The decomposition reactions are accelerated by increasing temperature, the presence of thiosulfate, and low liquor pH (1000 ppm) chloride levels in the liquor at coastal mills. As a result, these digesters are usually constructed of type 317L stainless steel (Ref 26). More highly alloyed stainless steels, such as type 317LMN and 904L, and nickel-base alloys have also been used in particularly corrosive conditions. Digesters for use in neutral, alkaline, and chemi-mechanical sulfite pulping can be made from a range of materials, including carbon steel and austenitic stainless steels. In one case, accelerated corrosion in the vapor phase of a type 317L stainless steel continuous high-yield bisulfite digester was recently a problem because of SO3 formation (Ref 27).

Formation of SO3 was minimized by raising the liquor pH from 4 to 6, by using a plug screw feeder to reduce the entry of oxygen with the chips, by feeding chips in below the liquor level, and by preheating the water in the adjacent blow tank to lower its dissolved oxygen content. At the same time, the vapor space of the digester was clad in 904L stainless steel to provide increased corrosion resistance. Liquor Recovery Systems. Evaporations are commonly made from type 316L or 317L stainless steel, although type

304L or carbon steel has been used for the final effects in sodium-base systems. Pitting and crevice corrosion under scale deposits are the most commonly reported problems. Liquor decomposition to form H2SO4 may also occur. In general terms, magnesium- and calcium-base systems are considered to be most corrosive towards evaporators (Ref 26). Recovery boilers have been largely constructed from carbon steel, although increasing use has been made of composite stainless steel-carbon steel water tubing to combat corrosion. Flame-and plasma-sprayed stainless steel have also been used for the same purpose. Sulfite mill recovery boiler exhaust gases contain significantly more sulfur dioxide than those from kraft recovery boilers; they may also contain greater amounts of SO 3, hydrogen sulfide (H2S), or HCl (Ref 28). Type 316L and 317L stainless steels and fiber-reinforced plastics have been used for scrubber bodies and internals, but have sometimes been replaced by higher-alloyed stainless steels or nickel-base alloys because of severe pitting and crevice corrosion. Attack can be especially severe in the inlet quenching/chloride removal zone, where even titanium and nickelbase alloys have suffered extensive general dissolution. Special care must be taken in these areas to avoid scale buildup and the presence of wet-dry zones, where chloride concentration can occur. Corrosion in the Recovery Boiler

Recovery boilers are used in the wood pulp industry to recover pulping chemicals and to raise steam by burning the organic residues present in spent pulping liquors. By far, most of the recovery boilers in North America burn kraft process liquor (black liquor). When evaporated to about 65% solids and immediately before firing in the boiler, a heavy black liquor contains 50% organic solids and approximately 6% total sulfur, mostly in the form of Na2SO4 and Na2S2O3. Some coastal or closed cycle mills operate with high sodium chloride (NaCl) content (for example, 12% NaCl) black liquor. One typical boiler configuration is shown in Fig. 24. It differs from an electric utility boiler in that the underlying firing zone is a bed of molten salt (the smelt bed) that operates under reducing conditions and is contained in a water-walled furnace bottom. In addition, the combustion zone is extraordinarily tall so that chemical reaction products can condense and run down the water wall before reaching the superheater.

Fig. 24 Schematic of typical recovery boiler used in the wood pulp industry

In and above the smelt bed, chemical reactions take place that reduce sulfate to sulfide and form sodium carbonate (Na2CO3). The operating philosophy of the boiler places safety and a continuous supply of regenerated liquor (green liquor, made by dissolving smelt in water) ahead of steam supply and energy efficiency. Steam pressures are generally in the range 4135 to 10,340 kPa (600 to 1500 psi), and a typical boiler might generate 135,000 kg (300,000 lb) of steam per hour, burning 725 to 900 Mg/day (800 to 1000 tons/day) of solids.

Corrosion Mechanisms There are numerous mechanisms of corrosion or chemical degradation of the recovery boiler. Whether or not corrosion can take place and the rate at which corrosion proceeds depend on the composition and metallurgical condition of the

material of construction, on the environment to which the materials are exposed (for example, liquid or gaseous phase, concentrations of chemicals), and on the temperature of the material and environment. Composition of some alloys and coatings are listed in Table 8. The corrosion products formed on the surface of the metal depend on the metal composition and on environmental factors. These corrosion products often protect the underlying metal from exposure to the corrosive environment and therefore reduce or prevent further corrosion.

Table 8 Metal alloy and coating compositions for recovery boiler tubing Alloy/coating Tubing alloys A192

Composition, wt%(a) C Mn P

S

Si

Cr

Ni

Mo

Fe

Other

0.048

0.058

0.25

...

...

...

bal

...

0.048

0.058

...

...

...

bal

...

0.048

0.058

...

...

...

bal

...

0.045

0.045

...

...

bal

...

0.030

0.030

...

bal

...

0.030

0.030

8.0010.00 1.001.50 1.902.60 18.020.0 24.026.0 17.020.0 17.020.0 19.023.0 27-31

0.440.65 0.901.10 0.440.65 0.871.13 ...

bal

...

bal

...

bal

...

...

bal

Ti(b)

...

bal

Nb + Ta(c)

... ...

39.5 (min) 7-11

0.75 Cu; 0.15-0.60 Al, Ti 0.50 Cu

A210, grade A-1

0.060.18 0.27

0.270.63 0.93

A210, grade C

0.35

T-1 T-9

0.100.20 0.15

T-11

0.15

T-22

0.15

0.030

0.030

Type 304H

0.040.10 0.15

0.291.06 0.300.80 0.300.60 0.300.60 0.300.60 2.00

0.1 (min) 0.1 (min) 0.100.50 0.251.00 0.501.00 0.50

0.040

0.030

0.75

2.00

0.040

0.030

0.75

2.00

0.040

0.030

0.75

2.00

0.040

0.030

0.75

1.5

...

0.015

1.0

0.50

...

0.015

0.50

0.93

0.048

0.058

0.1 (min)

...

...

...

bal

...

2.00

0.040

0.030

0.75

18.020.0

8.0011.00

...

bal

...

... ... 1.00

... ... 0.040

... ... 0.030

... ... 0.08

9.0 27.5 14.516.5

bal ... bal

5.5 2.0 15.017.0

5.0 bal 4.007.00

7.0 Al 6.0 Al 3.00-4.50 W 2.50 Co

Type 310 Type 327H Type 347H Incoloy alloy 800H Inconel alloy 690

0.040.10 0.040.10 0.050.10 0.05

Composite boiler tubing Inside tube A210, grade 0.27 A-1 Outside tube 0.08 Type 304 Coatings Metco alloy 444 Metco alloy 465 Hastelloy alloy C276

(a) (b) (c)

... ... 0.02

... ... 8.0011.00 19.022.0 9.0013.00 9.0013.00 30.035.0 58.0 (min)

...

Maximum wt% unless otherwise specified. Type 321H shall have a titanium content of not less than four times the carbon content and not more than 0.60%. Type 347H shall have a niobium plus tantalum content of not less than eight times the carbon content and not more than 1.0%.

The environmental characteristics determine whether or not corrosion may be expected. Metals can be exposed to corrosive chemicals in solid, liquid, or gaseous phase, with the latter two being most harmful. The temperature of the metal in contact with the corrosion environment will determine the rate at which corrosion reactions will take place. Sometimes, a corrosion reaction proceeds slowly because of the temperature. Sometimes,

corrosion reactions take place only in a specific temperature range. Table 9 shows the typical locations where the different types of corrosion have been found in recovery boilers.

Table 9 Corrosion mechanisms and typical locations in recovery boilers Mechanisms

Locations

Sulfidation reaction

Water wall tubes, studs, and air ports in furnace hearth zone. Both furnace-side and back-side surface of tubes

Molten hydroxides

Around primary air ports on back-side surface of tubes

Sulfidation-oxidation reaction

Lower superheater section, or further downstream, when liquor/smelt carryover is excessive

Sulfur trioxide adsorption to form liquid Superheater and boiler bank pyrosulfates Sodium chloride deformation

enrichment

and Superheater and boiler bank

Sulfur trioxide adsorption to form liquid Economizer and back-end rows of boiler bank bisulfates Acid dew-point corrosion

Cold-side of economizer, direct contact evaporator, precipitator, ID fan, and all flue gas ducting connecting these sections

Metal wastage due to corrosion can occasionally be accelerated by and confused with erosion and overheating, although erosion and overheating fall outside the strict definition of corrosion. However, the combination of metal properties, composition of corrosive chemicals in contact with the metal, thermal cycling, and the temperature will determine the rate at which corrosion will take place. Corrosion in the Furnace Corrosion, in combination with erosion, can severely damage tube metal, pin studs, stack studs, flat studs, and protective coatings in the furnace hearth zone. The corrosion rate of carbon steel in recovery boilers rapidly increases above 315 °C (600 °F). Operating pressures above 6900 kPa (1000 psig) could result in external tube metal temperatures at or above this value. Corrosion by Flue Gas in the Absence of Sulfur Compounds. When the bare metal surface of carbon steel is

exposed to flue gases containing oxygen at elevated temperatures, corrosion may occur as follows:

3Fe + 2O2

Fe3O4

4Fe + 3O2

2Fe2O3

2Fe + O2

2FeO

(Eq 1) (Eq 2) (Eq 3)

The corrosion products formed, magnetite (Fe3O4), ferric oxide or hematite (Fe2O3), or ferrous oxide (FeO), depend on oxygen concentrations and temperatures. Of these products, only Fe3O4 is formed as a dense surface layer that prevents further attack of the underlying metal. The other iron oxides, Fe2O3 and FeO (formed at temperatures above 565 °C, or 1050 °F), do not form a dense protective layer, and corrosion proceeds in the underlying metal at the same rate. When the steel is alloyed with such metals as chromium, nickel, or molybdenum, the formation of Fe3O4 will prevail; in such cases, only minor amounts of Fe2O3 are formed, and no FeO is formed. These alloys are, therefore, more corrosion resistant. At higher alloy contents, other metal oxides will also be formed that protect against further oxygen attack. The corrosion resistance of higher-alloyed steels is primarily due to the formation of a protective layer of chromium oxide (Cr2O3). Corrosion by the Flue Gas in the Presence of Sulfur Compounds. In the presence of sulfur-bearing

compounds, such as in kraft recovery boilers, the dominating corrosion reaction of carbon steel is (sulfidation reaction):

Fe + S

FeS

(Eq 4)

Iron sulfide is formed as a porous layer on the metal surface and does not protect the underlying metal. Because of the chemical structure of iron sulfide, the underlying iron can migrate through the sulfide layer and react with flue gas. Thus, corrosion will proceed, and iron sulfide will continue to build up on the outer surface. At high sulfur concentrations, an outer layer of iron disulfide (FeS2) can also be formed:

Fe + 2S

FeS2

(Eq 5)

Elemental sulfur can be formed as a result of the following reactions in the boilers:

Na2S + 2CO2 2Na2S + 3SO2 2H2S + O2 Fe + S

FeS

Na2CO3 + CO + S

(Eq 6)

2Na2SO3 + 3S

(Eq 7)

2H2O + 2S

(Eq 8) (Eq 9)

Of these sulfur-producing reactions, the reaction of sodium sulfide (Na2S) with carbon dioxide (2CO2) as shown in Eq 21 appears to predominate. Some of this elemental sulfur then reacts with iron to form iron sulfides. Laboratory experiments have shown that iron monosulfide (FeS) will be formed at low reaction rates when oxygen is not present. The corrosion rate increases considerably when oxygen is also present, especially at temperatures above 150 °C (300 °F). At a hydrogen sulfide to oxygen ratio of approximately 1 (in the gas phase), the formation of FeS (Eq 23) is at a maximum rate. When the oxygen concentration is increased further, the corrosion reaction will increasingly lead to the formation of Fe3O4 (Eq 1) and FeS2 (Eq 5), which develop protective layers against continuing corrosion of underlying metal. Oxygen and gaseous sulfur concentrations can vary widely in time and place in the recovery boiler furnace, and little is known about actual concentration profiles and gradients. They are dependent on such variables as the sodium and sulfur content in the black liquor, bed temperature, bed size, primary air flow, and primary air pressure. Water wall tubes in the furnace hearth zone are usually covered with a layer of solidified smelt. This solid smelt layer decreases the corrosion rate because the tube metal is not exposed directly to molten smelt or corrosive gases and because the tube surface temperatures are lowered due to the insulating properties of the solidified smelt. One method of corrosion protection in the furnace hearth zone is to increase the thickness of the smelt layer by attaching pin studs to the tubes. The studs cool the smelt and serve as anchors to hold the solidified smelt layer onto the tubes. Corrosion from the smelt itself is negligible at normal tube metal temperatures (below 345 °C, or 650 °F). Stud wastage will occur and require repair before wastage of the underlying tube metal occurs. In nonwelded wall recovery units, furnace gases and smelt may penetrate to the outside or back-side surface of the tubes because of improper tube or stud alignment and/or flat (fin) stud wastage. This can lead to severe cold-side corrosion, as described in the following discussion "Corrosion by Molten Hydroxides" in this section. When liquor or smelt is carried into the superheater section, where tube metal temperatures are often above 425 to 510 °C (800 to 950 °F), corrosion may take place as the direct result of the smelt deposits on the metal (see the discussion "Corrosion in the Superheater" in this section). Some research work indicates that the solidified smelt layer is not always directly attached to the tube metal, but is separated from it by a thin layer of irregularly spaced porous material, probably dried black liquor. This provides a space between tube metal and smelt layer through which corrosive gases may attack the tube metal. The formation of this porous layer spacing will largely depend on operating variables.

Corrosion by Molten Hydroxides. Highly volatile hydroxides of sodium and/or potassium are released from the

furnace bed and may diffuse to the cold side of the furnace tubes, especially close to the primary air ports. Because temperatures are much lower here, condensation/sublimation of the hydroxides may occur. Liquid hydroxide causes corrosion of the tube metal. Sodium hydroxide has a melting point of 315 °C (600 °F) and may exist in the liquid state in high-pressure boilers. In combination with carbonates, sulfates, and potassium, deposit melting points have been found as low as 250 °C (480 °F). Laboratory experiments show even lower melting points (as low as 170 °C, or 340 °F), when substantial amounts of potassium hydroxide are also present. The condensation/sublimation deposits on lower-temperature metal surfaces are characterized by low sulfide content. In practice, molten hydroxide corrosion does not appear to take place in boilers with operating pressures below 6000 kPa (870 psig), because of lower saturation temperatures. The potassium content, which greatly affects the deposit melting point, can vary from mill to mill, depending on wood species and the source of chemical make-up. Molten hydroxides are much more corrosive toward stainless steel than carbon steel. There have been numerous reports of tube-side corrosion on the water walls of composite tube recovery boilers. Once cladding is wasted to expose underlying carbon steel, subsequent attack is much slower. This form of attack of stainless steel occurs at ports in the water walls, particularly the primary air ports. Corrosion in the Superheater Tube metal temperatures are highest in the superheater section. Superheated steam temperatures at the superheater outlet can be as high as 510 °C (950 °F). Outside tube metal temperatures make the superheater metals vulnerable to high rates of corrosion. At these temperatures, carbon steel is not corrosion resistant, even if the corrosive environment consists only of oxygen with no smelt or sulfur-bearing compounds present. Superheater tube materials in the high-temperature section consisting of structural steels, such as T-11 and T-22 (see Table 8 for compositions), and stainless steels have a higher corrosion resistance than carbon steel. Even these higher-alloy steels may also be vulnerable to corrosion. Corrosion by Flue Gas Constituents. Under normal conditions, ash and/or black liquor carryover is deposited on

the superheater tubes and dislodged by sootblower action. The porous structure of the ash deposits will allow bare tube metal to be exposed to the flue gases. The exposed surfaces are subject to corrosion by the flue gases, which contain the following:

Compound Oxygen (O2) Hydrogen sulfide (H2S) Sulfur dioxide (SO2) Sulfur trioxide (SO3) Carbon monoxide (CO) Water vapor (H2O) Carbon dioxide (CO2) Nitrogen (N2)

Concentration 0-5% 0-50 ppm 0-500 ppm 0-50 ppm 0-5% 30-35% 10-15% 50-55%

Oxygen. Unprotected metals at high superheater temperatures will react with excess oxygen in the flue gas, as described

in the discussion "Corrosion by Flue Gas in the Absence of Sulfur Compounds" in this section. Hydrogen Sulfide. Hydrogen sulfide and other total reduced sulfur (TRS) compounds are generated in the furnace

hearth zone and, with proper combustion, will be oxidized to SO2 by the time the flue gas reaches the superheater section. It is possible that the TRS concentration leaving the combustion zone is in the range of 0 to 1000 ppm, while O2 concentrations are normally 2 to 5%. In this case, the H2S/O2 ratio will be far below 1, which is the critical ratio for iron attack by dissociated H2S. Hydrogen sulfide dissociates as follows:

2H2S + O2 Fe + S

FeS

2H2O + 2S

(Eq 10) (Eq 11)

When TRS generation in the furnace is excessive and air supply to the tertiary and/or secondary combustion zones is insufficient to oxidize TRS, then the H2S/O2 ratio may increase to its critical ratio of 1, and iron attack by sulfur will be at a maximum rate. Therefore, excessive TRS and the use of low excess air may lead to this type of superheater corrosion. However, generation of TRS in the furnace is minimized by proper air distribution. Maintenance of proper air distribution usually avoids this type of corrosion. Further, proper air distribution is essential for compliance with air emission standards. Sulfur dioxide is formed in the combustion zone by the oxidation of sulfur bearing compounds. Sulfur dioxide causes

negligible corrosion in the superheater. Sulfur trioxide can be formed through the oxidation of SO2 at high temperatures in the gas phase:

2SO2 + O2

2SO3

(Eq 12)

Equation 12 is an equilibrium reaction, and under normal furnace conditions, only a small amount of the SO2 present will react to form SO3. Another mechanism for formation of SO3 is the catalytic action of metal surfaces along with the flue gases flow. Oxides of iron and oxides of vanadium are known catalysts. Excessive firing of oil with a high vanadium content (fuel oil may contain up to 500 ppm V and 3% S) may lead to deposits of vanadium pentoxide (V 2O5), which enhances the formation of SO3. Sulfur trioxide in the gas phase does not impose a corrosion problem. However, when it combines with water vapor and condenses, corrosive H2SO4 and/or sulfurous acid (H2SO3) are formed. This type of corrosion is known as dew-point corrosion and occurs when flue gases are cooled to the temperature at which condensation takes place (see the discussion "Corrossion in the Economizer" in this section). Another harmful effect of SO3 is adsorption on Na2SO4 deposits in the superheater and boiler bank areas; this lowers the pH of these deposits and causes sticky ash formation. Carbon Monoxide. Incomplete combustion leads to CO in the flue gas. Carbon monoxide alone does not cause

corrosion, but it slightly increases the corrosion process by H2S gases due to its reducing characteristics. Maintaining sufficient and proper air delivery, especially secondary/tertiary air, will minimize or eliminate the presence of CO in the flue gases. Water Vapor. The reaction of iron metal with water vapor:

3Fe + 4H2O

Fe3O4 + 4H2

(Eq 13)

is favored by thermodynamics to proceed to the right. The reaction rate, however, is slow, and corrosion inside or outside of the tubes is not usually due to this reaction. Carbon dioxide alone seldom corrosion. In combination with Na2S, however, it releases sulfur compounds from the

smelt (Eq 21), leading to the corrosion reaction of sulfur with iron. Nitrogen is inert and does not participate in any of the combustion or corrosion processes in recovery boilers.

Formation of Superheater Ash Deposits Deposits will, normally be present on superheater tubes. The source, structure, and composition of these deposits can vary. Recent research has established that two basically different mechanisms cause the formation of superheater deposits: • •

Carryover of black liquor particles Vaporization/condensation of sodium compounds

These two mechanisms act alone or in combination, depending on liquor composition, bed temperature, combustion air delivery system, and liquor spray pattern. The structure and composition of the deposits vary according to aging factors and the mechanism by which they were formed. The sodium to sulfur ratio in the flue gas is an important parameter of corrosion/deposition chemistry. This ratio is usually 1.4 or higher in kraft recovery flue gases. Deposits From Black Liquor Carryover. Black liquor particles can be entrained in the combustion gases. The organic will be burned off, leaving predominantly Na2CO3, Na2SO4, Na2S, and some NaOH. These deposits are found at lower elevations of the radiant superheater surfaces close to the furnace, especially at the leading side of the tubes. The carryover deposits may sinter and fuse, thus forming a hard, thick layer on the tube surface. This type of deposit formation may be found on units that are significantly overloaded. Deposits From Vaporization/Condensation of Sodium Compounds. Vaporization of sodium and sodium

compounds begins in the furnace bed. These compounds condense on the cooler sections of the superheater and the boiler bank tubes, forming a thin white deposit consisting mainly of sodium sulfate and some sodium carbonate. A minor amount of sodium chloride may also be present. Condensation of these compounds may occur on bare tubes or in the porous spaces between the previously laid down carryover deposits. This results in a mixture of carryover and condensation deposits. Deposits formed by the vaporization/condensation process are likely to be found at higher, cooler elevations away from the furnace and especially on the trailing side of the tubes where flue gas velocities are lower. Also, at lower elevations closer to the furnace, some condensation products may be found on the trailing side, where they are shielded from the gas stream. The type of process by which the deposits are laid down determines their chemical characteristics and composition. Carryover deposits will be richer in sodium carbonate. The pH of a saturated solution of this material is usually well above 7 (alkaline). Vaporization/condensation deposits usually do not contain much sodium carbonate, but consist mainly of sodium sulfate. The pH of a saturated solution is slightly below 7 (acid) or can be much lower because of SO3 adsorption on the surface. A typical deposit analysis is shown in Table 10.

Table 10 Typical tube deposit analysis Deposit

Na2CO3, % Na2SO4, % Other, % pH saturated solution

Superheater tube leading side; carryover 37.6 60.5 1.9 12.0

Boiler bank tube trailing side; vaporization/condensation 0 97.6 2.4 4.4

Deposit Enrichment With Chlorides. The two mechanisms of deposit formation--liquor carryover and vaporization/condensation--occur independently and may lead to a mixture of deposits. The porous spaces in the carryover deposits may be filled with deposits from condensation. These deposits are commonly found close to the tube metal where the temperature is lowest.

Sodium chloride (NaCl), potassium chloride (KCl), and potassium hydroxide (KOH) are among the most volatile chemicals present in the recovery furnace. They may also be found in high concentrations, together with deposits from condensation (Na2SO4), closest to the tube metal. Research work indicates Cl/Na and K/(K + Na) ratios increase when progressing from the outside of the deposit layer to the inside toward the tube metal surface. The increased amount of chlorides and potassium in the deposits is of particular concern with closed chemical recovery systems (that is, those with bleach plant effluent recovery) and in mills processing seawater-borne logs. IN other mills, chloride levels in the liquor cycle can also be excessive. Sodium chloride concentrations up to 40 g/L in the white liquor have been reported (the corrosive influence of chlorides is explained in the discussion "Effects of Chloride Compounds" in this section). A comparison of kraft recovery superheater deposit compositions from different types of mills is given in Table 11.

Table 11 Examples of kraft recovery superheater deposit compositions Deposit, %

Mill A(a) Mill B(b) Mill C(c) Mill D(d)

NaCl

2.2

16.0

22.5

3.7

Cl/Na (molar)

4.9

25.5

26.8

4.4

5.7

7.5

11.4

K/(K + Na) (molar) 1.7

Inland pulp mill with bleach plant (open cycle).

(a) (b) (c) (d)

Pulp mill with bleach plant (closed cycle). West Coast pulp mill using seawater-born logs. Pulp mill using hardwood chips

Corrosion Mechanisms in Superheater Ash Deposits The adsorption of SO3 on ash deposits and the resulting problem in pH impose a serious corrosion problem in that an

acidic environment is created. This type of corrosion may occur especially in sulfuric liquor systems with high sulfidities, but may also be found in kraft liquor systems. It has been found that Na2SO4 and SO3, in combination with water vapor, may form sodium pyrosulfate (Na2S2O7) and sodium bisulfate (NaHSO4) according to:

Na2SO4 + H2O + SO3 Na2S2O7 + H2O Na2S2O7 + H2O

2NaHSO4

(Eq 14) (Eq 15)

The adsorption of SO3 and the reaction with Na2SO4 occur only in deposits depleted of Na2CO3. Sodium trioxide will be neutralized when carbonate is present, such as in the deposits from black liquor carryover (that is, on the superheater and boiler bank tubes closes to the furnace), according to:

Na2CO3 + SO3

Na2SO4 + CO2

(Eq 16)

This reaction proceeds rapidly at temperatures above 400 °C (750 °F). When carbonate is no longer present in the deposits and SO3 is still present in the flue gas, Na2S2O7 may be formed. At lower temperatures, NaHSO4 may also be formed. The corrosive attack from Na2S2O7 is thought to occur only in the liquid state. The liquid state of pyrosulfate may occur in many areas of the recovery boiler because its melting point is about 400 °C (750 °F). Solid Na2S2O7 is not particularly corrosive. In recovery systems with high potassium content, sodium and potassium pyrosulfates (Na2S2O7/K2S2O7 eutectic) may be liquid at temperatures as low as 280 °C (540 °F). Sodium bisulfate is stable at temperatures below 280 °C (540 °F) and will therefore not be found in the highertemperature sections in the superheater. However, NaHSO4 can form on lower-temperature surfaces of the superheater, boiler bank, and economizer. Fouling and corrosion of these metal surfaces may result because of the presence of molten NaHSO4. This corrosion mechanism may be found in kraft recovery boilers having low sodium to sulfur ratios and high SO3 concentrations in the flue gas and in a narrow temperature range from approximately 185 to 280 °C (370 to 540 °F). In kraft pulp mills having normal white liquor sulfidities (25 to 30%) and normal recovery furnace operation, the concentration of SO3 in the flue gas will be low, and it all will likely be neutralized by Na2CO3. Some of the following conditions, however, may increase SO2 and therefore SO3 formation and lead to acidic deposits or sticky ash: • • • •

High white liquor sulfidity and low sodium to sulfur ratio in the liquor Low char bed temperature Excessive or uncontrolled spent acid additions Excessive firing of oil with high sulfur and vanadium content (catalytic formation of SO3)

•

High potassium content in the liquor

Any one of these conditions or a combination may be conductive to the formation of acidic sulfates and may require special attention and correction. Effects of Chloride Compounds. The role of chloride compounds in superheater corrosion is not clear. Little

research has been done on these corrosion mechanisms, and it is thought that chloride compounds may either directly attack tube material or decrease the melting point of deposits, thus causing fluxing of normal protective metal oxides. Sodium chloride vapor, when condensed in the tube surface deposits, may release HCl in accordance with the following reaction:

4NaCl + 2SO2 + O2 + 2H2O 4HCl + 2Na2SO4

(Eq 17)

Subsequent reactions of HCl fumes with tube material (either metal or metal oxides) may involve the formation of volatile iron chlorides and/or metal oxychlorides. From laboratory experiments, it is known that the deformation and melting points of a mixture of Na2S, Na2CO3, and Na2SO4 can be greatly affected by the NaCl content. The lowest first deformation point of a mixture of Na2S, Na2CO3, Na2SO4 and NaCl has been found below 595 °C (1100 °F). The deformation and melting points can be lowered by as much as 220 °C (400 °F) with NaCl concentrations above 10%. Potassium may decrease these melting points even further. In closed cycle mills with bleach plant effluent recovery or in mills using seawater-borne logs, the chloride input to the recovery boiler can be high. Because of the condensation/vaporization mechanism, the NaCl content of superheater deposits has been shown to be above 15%, and deposit deformation and melting may occur around 595 °C (1100 °F). Erosion-corrosion of tube metal may occur when 70% or more of the total deposit is in the liquid phase, thus making the entire deposit flux. In kraft pulp mills without bleach plant effluent recovery or seawater-borne logs, the deposit melting point is not significantly affected, and subsequent fluxing of deposits does not usually occur. However, some conditions may cause high chloride levels or other conditions conducive to deposit fluxing, such as: • • •

High chloride input to the system due to a high chloride content in the spent acid from ClO2 generation Excessive or uncontrolled spent acid additions producing temporarily high inorganics input to the liquor Local hot spots in the superheater due to flue gas channeling or improper design features

These conditions should be avoided or minimized. Effects of Miscellaneous Chemistry. Superheater metal wastage may also occur as a result of the reaction and/or

decomposition of sodium-sulfur compounds at elevated temperatures. The Na2S and Na2S2O3 contents in smelt or ash deposits between the furnace hearth zone and the superheater will generally decrease from their concentrations in the smelt to deplete the superheater tube deposits unless excessive liquor/smelt carryover takes place. Usually, a small amount of Na2S can be found in the deposits at the superheater bends. In laboratory experiments, it has been shown that sodium-sulfur compounds may react or decompose at elevated temperatures, resulting in the formation of free sulfur. These experiments (at 370 °C, or 700 °F) indicated the following reactions:

Na2SO4 + Na2S Na2SO3 + S + Na2O*

(Eq 18)

Na2S2O3 + Na2O*

Na2S + O2

(Eq 19)

Na2SO3 + S

Na2S2O3 Na2S + 2CO2

(Eq 20)

Na2CO3 + CO + S

2Na2S + 3SO2

(Eq 21)

2Na2SO3 + 3S

(Eq 22)

Equations 21, 22, and 18, 19, and 20 also occur in the furnace bed and around the water wall tubing where tube metal temperatures are much lower. The formation of free sulfur at the superheater lower bends may result in metal attack:

Fe + S

FeS

(Eq 23)

when bare tube metal is not protected. Corrosion in the Boiler Generating Bank Tube metal temperatures in the boiler bank area are lower than in the superheater. This makes boiler bank tubing less vulnerable to corrosion. Certain conditions, however, may cause elevated temperatures in the boiler bank riser tubes. This may lead to accelerated corrosion. In localized areas, conditions that cause overheating include: • • • • •

Internal scale Flue gas channeling Mechanical blockage of water flow Low water level Improper water distribution/circulation

At moderately elevated temperatures, accelerated corrosion may result when the tube metal is subjected to a corrosive environment. The corrosion mechanisms in the boiler bank are the same as for the superheater, that is, corrosion by flue gas constituents or by ash deposits. Lower temperatures in the boiler bank result in lower corrosion rates. In addition, there is less chance for the formation of chloride-containing deposits. Corrosion at the Drum Tube Seats. Tube metal wastage may occur at the drum tube seats. Corrosion is due to the

combined action of chemicals and water and may be accelerated in the high stress areas resulting from the tube rolling technique or vibration-induced stresses. The water present may be from tube seat leakage, water-washing remains, soot blowing, or absorption by deposits around the tube seat. Erosion by Sootblower Steam/Water Mixtures. Boiler bank tubes, superheater tubes, and economizer tubes may

be subject to steam/water mixture sootblower action, resulting in tube metal wastage. The damage usually appears as localized external wall thinning, and the affected area usually has a smooth, polished appearance. Proper sootblower spacing and care in piping design (which ensures that only dry steam is used for sootblowing) and proper maintenance of steam traps should be exercised. Leaking poppet valves cause steam to condense in the lance tube when the sootblower is idle. This water runs from the nozzle and may cause corrosion of wall tubes and casing. Leaking poppet valves and defective steam traps should be promptly repaired. Corrosion in the Economizer. External corrosion of economizer tubes may occur when economizer flue gas exit

temperatures and/or feedwater temperatures are low. The flue gas exit temperature and feedwater temperature entering the economizer determine the tube metal temperature. The tube metal temperature may drop below the acid dew point. The

dew-point temperature is highly affected by SO3 concentrations in the flue gas. An increase in SO3 in the flue gas elevates the dew-point temperature. Therefore, condensation is more likely to occur. It is generally good practice to maintain the flue gas temperature at the economizer outlet at or above 160 °C (325 °F) and to maintain the feedwater temperature entering the economizer at or above 135 °C (275 °F) in order to prevent condensation. If condensation occurs, the product is dilute H2SO4:

SO3 + H2O

H2SO4

(Eq 24)

The H2SO4 attacks metals and protective iron oxide coatings:

Fe + H2SO4

FeSO4 + H2

Fe + Fe3O4 + 4H2SO4 4FeSO4 + 4H2O

(Eq 25) (Eq 26)

Therefore, corrosion may take place, especially at localized areas where ambient air can leak through doors or holes in the casing. In this type of corrosion, the rate increases as the temperature decreases, not because the corrosion mechanism itself proceeds faster at lower temperatures, but because more of the corrosive H2SO4 is formed. Economizers are particularly susceptible to this type of corrosion because operating gas temperatures are near the acid dew point and any ambient air leakage or other air infiltration, such as at vents, will tend to lower gas temperatures in localized areas. It should be noted that economizers are particularly susceptible to high leakage/infiltration due to highdraft conditions. Acid dew-point corrosion can also occur in other downstream components of the system, such as electrostatic precipitators and flue gas scrubbers. Corrosion in the economizer can also be caused by the formation of molten bisulfates below 280 °C (540 °F), as explained in the discussion "Adsorption of SO3" in this section. Note cited in this section

*

Sodium oxide (Na2O) is very reactive and combines quickly with CO2 or H2O to form Na2CO3 or NaOH, respectively.

Suction Roll Corrosion Max D. Moskal, Stone Container Corporation

Suction rolls are used to remove water from paper at the wet end of the paper machine. One way to accomplish this is by passing the paper web through a roll nip, one roll of which is the suction roll. The suction roll is drilled to an open area of about 20%, and a vacuum is applied to the inside of the roll. A typical cross section through a suction roll is shown in Fig. 25. A rubber cover is often used on the suction press roll or mating roll to give a desired nip pressure relationship. Many machine designs have been used in the wet end of paper machines throughout the years. A typical paper machine may use three or more suction rolls designated as couch, wringer, pickup, or press rolls (Fig. 26).

Fig. 25 Cross section of suction roll configuration

Fig. 26 Schematic diagram of wet end of paper machine showing types of suction rolls

The critical component in the suction roll is the drilled shell, which is subject to corrosion and fatigue cracking. During the past 30 years, changes have evolved in design and machine environment that have increased the cracking and corrosion tendency of the suction shell. For example, demands for increased machine production and speed have been resolved by increasing roll nip pressures. Wider machines have also been designed. These new designs have required heavier and stronger shells to meet the increased stress that is imposed. The corrosive environment in the paper machine wet end has also become more severe as water systems have been closed, water temperatures increased, and chemicals used to clean paper machine felts. Numerous changes have also occurred in a search for materials and designs that better resist shell corrosion and corrosion-assisted cracking.

This section describe the history of suction roll and materials development, the operating environment, corrosion fatigue of suction rolls, shell manufacturing methods, and corrective measures.

History of Suction Roll Alloys Prior to about 1950, suction shells were cast from bronze having a composition of approximately 85% Cu, 5% Pb, 5% Zn, and 5% Sn. Bronze has many advantages, such as castability, machineability, corrosion resistance, and relatively low cost. However, bronze lacked the strength needed to resist the higher press loads that were introduced at that time, and it lacked the stiffness necessary to maintain shape and roundness on larger and wider paper machines. In 1964, it was noted that suction roll nip pressures had increased from the original levels of 26 kN/m (148 lb/in.) up to 79 kN/m (451 lb/in.) (Ref 29). Designs with nip pressures to 131 kN/m (750 lb/in.) were proposed in the 1960s and were actually achieved in the 1970s. Also, paper machine widths increased from typically 3.8 m (150 in.) in the 1950s to over 6.3 m (248 in.) in the 1970s, with maximum machine widths achieved in excess of 10 m (394 in.). The first stainless steel materials to be widely used were forged type 410 martensitic alloy and centrifugally cast CF-3M and CF-8M austenitic alloys. The austenitic alloys were deemed especially desirable for low-stress, high-corrosion applications, such as the couch position. From 1955 to 1965, the influence of corrosion on fatigue was not widely recognized. In one case (Ref 29), shell designs were based on the fatigue strength of these alloys in air (Fig. 27).

Fig. 27 Typical fatigue curves of suction shell alloys as tested in air. These data were used to establish early designs of suction rolls. Courtesy of the Technical Section, Canadian Pulp and Paper Association

In 1957, the first centrifugally cast CA-15 martensitic stainless steel shells were cast. Also during this early period, shells were cast from alloy A-63 stainless steel and aluminum bronze. Tables 12 and 13 list the nominal compositions of the alloys used for suction shells.

Table 12 Composition of copper-base suction shell alloys Material 1N bronze(a) GC-CuSn5ZnPb(b)

(a) (b)

Composition, % Cu Sn Pb 85.4 4.7 4.6 84-86 4-6 4-6 Typical composition. Composition range

Zn 4.6 4-6

Al ... ...

Ni 0.6 ...

Fe ... ...

Mn ... ...

Table 13 Nominal composition of stainless steel suction shell alloys Materials are categorized according to structure. Material Austenitic CF3M CF8M(b) PM-3-1811MN Martensitic C-169 CA-15 PM-4-1300 DSS-69(c)(d) A-704(d) Duplex A-63(d) A-170(d) A-171 A-271 3RE60(e) A-75 A-86 VK-A3780(f) A682(g) PM-3-1804M PM-2-2505

(a) (b) (c) (d) (e) (f) (g)

Composition, %(a) C Cr Ni

Mo

Cu

Mn

Si

0.02 0.05 0.015

17.7 17.7 16.5

13.8 13.8 13.5

2.3 2.3 2.1

... ... ...

1.3 1.3 1.6

0.8 0.8 0.5

0.07

12.4

0.6

0.5

...

0.5

0.6

0.12 0.04 0.03

12.5 12.4 11.9

0.4 4.0 4.0

0.4 0.7 1.5

... ... ...

0.7 0.7 0.8

0.5 0.6 0.5

0.05 0.07 0.07 0.06 0.02 0.02 0.02 0.06

21.8 23.3 22.2 24.6 18.5 26.0 26.0 20.0

9.4 10.7 8.3 4.3 4.7 6.8 6.0 5.0

2.7 2.1 1.2 0.7 2.8 ... ... 2.0

... ... ... ... ... ... 2.8 4.5

0.8 0.7 0.8 0.7 1.5 0.8 0.8 0.6

1.3 1.5 1.1 1.3 1.7 0.5 0.7 0.8

0.06 0.07

17.9 26.0

4.0 4.0

2.0 0.8

... ...

0.6 1.2

0.6 1.3

Remainder of compositions is iron. Alloy has been discontinued for suction shells in North America. Details of composition not published, but similar to ASTM A-296, alloy CA-6NM (discontinued specification). Obsolete alloy. Contains nitrogen. Contains nitrogen and tungsten. Alloy A-682 has a composition similar to that of VK-A378, only reduced chromium and nickel.

Between 1969 and 1977, a host of new centrifugally cast stainless steel alloys was introduced for suction shells. These were DSS69, C169, alloy 70 (martensitic), and alloy A170, A171, A271, and A-75 (duplex/ferritic-austenitic). Also during this period, rolled and welded stainless steel shells were produced in Sweden from type 316 and 3RE60 stainless steels. Continuous-cast bronzes GC-CuSn5ZnPb and GC-CuA19,5Ni also emerged during this period. In 1971, the reduced fatigue strength of bronzes and stainless steels in synthesized white water was reported (Ref 30). The importance of residual stresses on fatigue cracking was also more widely recognized at this time. Between 1978 and 1986, additional new stainless alloys were introduced. These included centrifugally cast precipitationhardening duplex alloys VK-A378 and KCR-A682 and duplex alloy A-86. In 1978, the argon-oxygen decarburization process was introduced in the United States for suction shells. This permitted more economical melting and casting of stainless alloys with a very low carbon content. Centrifugally cast CF-3M has since been produced instead of CF-8M. Several forged alloys were also introduced during this period: PM-4 1300 and PM-4-1300M (martensitic), PM-31811MN (austenitic), PM-3-1804M, and PM-2-2505 (duplex). During the 1980s, emphasis has been placed on stainless steel development with greater corrosion and fatigue resistance and low residual stresses.

Paper Machine Environment

A detailed discussion of the paper machine corrosion environment is given in the section "Paper Machine Corrosion" in this article. However, considerations specific to suction roll corrosion will be covered in this discussion. Suction rolls are subjected to a variety of corrosive environments. The severity of the environment depends primarily on the type of paper produced on the machine, the source of water, and the degree of closure of the mill. The corrosivity of the environment depends primarily on pH, temperature, dissolved solids content, chlorides, and the presence of sulfur compounds. The area around the suction rolls is subjected to splash from stock and white water, and deposits of paper fiber are present in crevices. Bacteriological growth can also occur, especially in areas that are difficult to treat with biocide chemicals (commonly used biocides are listed in the article “Corrosion Inhibitors in the Water Treatment Industry” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook). Suction rolls are also subject to corrosion from chemicals and waters used in showers around the rolls and from felt treatment chemicals. Laboratory studies have shown that Na2S2O3 in stock and white water systems results in severe pitting corrosion of the less resistant martensitic stainless steels, and thiosulfate is believed to be important in corrosion and cracking of suction roll shells. Sodium thiosulfate, residual sodium hydrosulfite (NaHSO2), and the sulfide ion are also highly corrosive to bronze. Biological corrosion has also been observed within the drilled holes of suction shells. The hole area is highly favorable to the growth of microbes because of the presence of dissolved organic materials, dissolved inorganic salts, and a favorable temperature range of 40 to 50 °C (105 to 120 °F). Anaerobic sulfate-reducing bacteria thrive in crevice locations and under salt and fiber deposits. Therefore, because of the concentration of ions and the localized pH reduction, the chemical environment beneath deposits can become far more aggressive than the bulk stock or white water chemistry. Additional information on biological corrosion can be found in the articles “Microbiologically Influenced Corrosion,” and “Evaluating Microbiologically Influenced Corrosion” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook. The showers and chemicals used on the paper machine also influence the environment around suction rolls. Showers are used to clean machine felts, to remove deposits from drilled holes in suction rolls, and to lubricate roll seal strips. A wide variety of shower cleaning chemicals are used for felts and rolls, including acids, aliphatic and aromatic solvents, or animal and vegetable oils. Lubricating and needle showers are often supplied with white water at temperature ranges of 40 to 50 °C (105 to 120 °F), this white water is more corrosive to the suction roll than freshwater. Attack from chemicals has also contributed to failure of elastomeric covers on suction rolls (Ref 31). Figure 28 shows damage to a CA-15 martensitic stainless steel suction shell from muriatic acid used for felt cleaning.

Fig. 28 Corrosion damage to a CA-15 stainless steel shell caused by muriatic acid felt cleaning

Off-line chemical cleaning has been used to remove deposits from drilled holes in suction rolls. There have been numerous cases of severe damage to both shells and rubber covers when corrosive cleaning chemicals were used (Ref 32).

Corrosion Fatigue Testing In the 1960s, several investigators recognized the need to obtain corrosion fatigue data for suction shell materials using a synthesized paper machine white water environment. Corrosion fatigue testing of suction materials was first described by R. Thompson, who used a conventional R.R. More rotating-beam machine equipped with a corrosion chamber around the specimen (Ref 33). Liquid was continuously dropped onto the stressed portion of the specimen. Corrosion fatigue tests were conducted using a reverse bending plate with a single hole drilled in the stressed regions (Ref 34). Crack growth rate tests were employed to evaluate the corrosion fatigue characteristics of suction roll materials (Ref 35, 36). These investigators used compact-tension specimens subjected to cyclic loading in an electrohydraulic fatigue machine. Corrosion fatigue test data have been published largely by shell producers and paper machine equipment manufacturers. A variety of test data have been shown for all of the alloys used for suction rolls. Because different test methods and a variety of environments have been used for tests, direct comparison of corrosion fatigue strength data is very difficult. Rotating-bending fatigue test data for several alloys in white water at pH 3.5 and varying chloride and sulfate contents are shown in Fig. 29. Rotating-bending fatigue test data by a different investigator for alloys in a highly aggressive white water are shown in Fig. 30.

Fig. 29 Fatigue curves in white water based on the data of material suppliers. 1500- to 1750-rpm rotatingbending test, pH = 3.5, Cl- = 20-400 ppm, Pulp and Paper Industry

= 250-1000 ppm. Courtesy of Technical Association of the

Fig. 30 Fatigue curves for suction shell alloys in an aggressive environment. Courtesy of Sandusky Foundry & Machine Company

In 1985, the TAPPI Corrosion and Materials Engineering Committee summarized survey results on the correlation among corrosion fatigue testing methods. The survey included data provided by ten investigators on the type of testing used, specimen configuration, test conditions, and environment (Ref 37). The TAPPI committee also published documents describing standard methods and test environments for suction shell alloys (Ref 38, 39). Recent investigations in near-threshold fatigue crack growth testing have indicated that mean stresses could have a significant effect on the threshold for crack growth in duplex stainless steel alloys (Ref 36). These data correlate with the concept that high residual tensile stresses in suction rolls decrease fatigue life in service.

Roll Failures Corrosion and cracking failures have been reported for couch, press, and felt wringer rolls. There has been no widespread problem with the lightly loaded suction pickup rolls. Failure by cracking is frequently unexpected, and is usually the most cause for concern. When cracks progress a significant distance around the shell, the suction roll begins to vibrate and must be immediately removed from the paper machine. Deterioration of shells by general corrosion and pitting is another important means of failure. Severe roughening of the outside diameter surface of bronze couch rolls is known to reduce the life of fourdrinier wires. An example of outside diameter surface corrosion in bronze is shown in Fig. 31. General corrosion in drilled holes results in hole enlargement, decreasing the land area between holes to a point at which corrosion fatigue cracking can occur (Fig. 32).

Fig. 31 Thiosulfate corrosion created surface roughening (dark areas) on this uncovered bronze roll. Courtesy of the Technical Section, Canadian Pulp and Paper Association

Fig. 32 General corrosion in drilled holes. (a) Inside surface of bronze shell adjacent to fracture. The hole enlargement varied periodically along the length of the shell and corresponded to shower nozzle spacing. (b) Close-up view of hole corrosion. Courtesy of MacMillan-Bloedel, Ltd.

Most suction rolls that are reported to have failed have been removed from service because of cracking. Cracked shells of bronze and martensitic stainless alloys almost always exhibit a visible condition of surface corrosion or pitting in the drilled holes. Higher-alloyed duplex and austenitic stainless steels seldom show visible corrosion, although fatigue failures in these alloys are known to be corrosion assisted. Failure Experiences. Two separate surveys covering roll failures have been published since 1979 (Ref 32, 40). The surveys indicated that most failures occured in bronze and martensitic stainless steel shells. However, it was also noted that there are far more bronze and CA-15 shells in service as compared to other alloys. Cracking was also more prevalent in the higher-speed and more corrosive newsprint machines than in fine paper, pulp, or board machines. Corrosion Fatigue Failures. Fatigue cracking is first observed in shells as single cracks in the land area between two

drilled holes or as cracks that begin at a hole and have not yet progressed to a neighboring hole. Cracking then proceeds across the land regions connecting several holes, eventually joining and proceeding circumferentially around the shell (Fig. 33). Cracking usually manifests on the roll inside diameter surface near the mid-span, where operating stresses are

the highest. Martensitic stainless steels show many fatigue initiation sites through the cross section of the fractured surface (Fig. 34). Bronze and duplex stainless steels seldom show well-defined fatigue initiation sites.

Fig. 33 Typical cracking between drilled holes on the outside surface in a stainless steel alloy shell. Courtesy of The Institute of Paper Chemistry

Fig. 34 Fracture surface of CA-15 stainless steel suction shell showing several fatigue nucleation sites (arrows). See also Fig. 35.

Longitudinal cracking is often observed near the mid-span of austenitic and duplex stainless steel shells, and this has been attributed to overloading or residual stresses from heat treatment. Longitudinal cracks eventually turn to the circumferential direction and proceed around the roll to ultimate failure. Rolls may develop only a single crack that progresses to failure, but more often, they will develop tens or hundreds of cracks before final failure occurs. Bronze and martensitic steel shells usually exhibit many small cracks before a larger crack develops. Duplex stainless steels develop fewer cracks or a single large crack. Microexamination of the cracked region frequently shows cracking associated with corrosion pitting (Fig. 35). Cracking has also been attributed to manufacturing defects, poor drilling quality (Fig. 36), excessive porosity (Fig. 37), welding repairs, and electrical discharge machining used to remove broken drill bits. Intergranular cracking has been observed in austenitic stainless steel that was sensitized during heat treatment and in bronze shells that were exposed to mercury and ammonium contamination.

Fig. 35 Photomicrograph of specimen in Fig. 34 showing a fatigue crack initiating at the base of a corrosion pit. Etched with Marbles reagent. 75×

Fig. 36 Fatigue cracking of CA-15 stainless steel suction roll due to poor drilling quality

Fig. 37 Casting porosity in duplex stainless steel shell

Rubber Cover Failures. A recent Canadian survey showed that the cost of premature rubber cover failures on suction

rolls was about as large as the cost of shell failures (Ref 40). In the survey, 38 mills were questioned about rubber cover

failures on suction and nonsuction rolls for the year 1983. A total of 70 cover failures were experienced in the 38 mills surveyed. A significant cause of rubber cover failures is deterioration from chemicals used on the suction rolls, both on-line and offline. Organic solvents have been found to be highly detrimental to rubber covers, especially when used in a continuous mode as opposed to batch treatment. High-pressure hydraulic cleaning can produce mechanical damage and detachment of rubber covers.

Shell Manufacturing Methods Most suction roll shells are produced by the centrifugal casting method, followed by continuous-casting (bronze), rolled and welded (stainless steel), and forging processes (stainless steel). Each method of manufacturing claims certain advantages. After production of the rough shell, it may be heat treated (depending on the alloy), machined to near-final shell dimensions, and drilled. The shell is then finished on the inside diameter surface and head seats machined. The outside diameter surface is finished by turning, belt sanding, or grinding. A smooth outside surface finish is not required if the shell is to be subsequently rubber covered. Castings. The process of centrifugal casting is the oldest and most versatile method of shell production. It is applicable

for both bronze and stainless steel alloys. Metal is poured into a spinning mold that is horizontally positioned, and the spinning motion is continued until the metal shell has completely solidified. In this process, the last metal to freeze is at the inside diameter surface of the casting. Nonmetallic and intermetallic impurities concentrate on this inner surface, which is later machined away. Continuous casting is a relatively new method of producing cast shells. It is applicable only to bronze alloys. In this method, the molten metal is poured through holes in a distributing vessel into a vertical collar-type mold. Metal flow, speed of extraction from the mold; and cooling are coordinated so that the casting process is continuous. An advantage of continuous casting is that the uniformity of metal properties through the wall can be readily controlled. Weldments. Welded stainless steel shells have been produced in Sweden since the early 1970s. Most production has

been limited to 3RE60 duplex stainless steel alloy, which is hot formed from plate into semicylinders. The semicylinders are subsequently welded longitudinally to form short cylinders, and several short cylinders are welded at the ends to form the suction shell. Forgings. Forged suction shells are produced by first upsetting and punching a hollow cylinder from a solid cylinder,

then enlarging and forging the shell on a pipe-forming mandrel. The process minimizes the porosity and segregation that may be found in castings, but can be subject to inclusions and other internal flaws. Heat treatment of suction shells has proved to be a critical operation affecting the corrosion and the residual stress

properties of shells. Current practice in heat treatment is limited to stainless, steel suction shell alloys. Although aluminum bronze is heat treatable, shells from this material are currently furnished in the as-cast condition. Each stainless steel alloy requires a specific heat treatment to provide the best combination of corrosion resistance and low residual stresses. Slow cooling from the final heat treatment produces lower residual stresses, but (depending on the alloy) may produce undesirable intermetallic phases and a low corrosion fatigue resistance (Ref 41, 42). Austenitic alloys may be given a single-solution anneal treatment, but martensitic alloys are quenched and tempered. Duplex alloys may be subjected to a single-solution treatment or, in the case of copper-bearing duplex alloys, may be given one or two additional subcritical aging treatments. Drilling. Suction shells are drilled using multiple-spindle machines. Couch rolls are frequently drilled with larger hole sizes of up to 6.3 mm (0.25 in.), but press roll drilling is typically 4.0 mm (0.16 in.). The drilling process is critical in that hole spacing must be maintained to within close dimensional tolerances.

Gun drilling was first developed in Europe for drilling stainless steels and is particularly applicable when drilling long, straight holes in thick-wall stainless steel shells. Twist drilling of these alloys sometimes results in converging of drilled holes at the inside diameter surface (Fig. 36). Gun drilling also has the advantage of producing a smooth surface finish within the drilled hole. This reduces the tendency toward the initiation of corrosion and deposit formation. Gun-drilled holes have been measured to have a surface finish of Ra = 0.5 m (20 in.) as compared to Ra = 5.0 m (200 in.) for

twist-drilled holes, where Ra is the arithmetic average roughness. A process of reaming after twist drilling has also been used to produce a good surface finish in the drilled holes. Shell Quality. Several types of manufacturing defects have been attributed to the failure of suction shells, as discussed

above. In an analysis of 52 suction roll failures, about 50% of the failures were attributed to defects in manufacture or design, 35% were due to improper material selection for the environment, and 15% could not be attributed to any cause (Ref 43).

Corrective Measures Three factors influence suction shell resistance to corrosion fatigue; stress, corrosion environment, and the corrosion fatigue strength of the alloy. A favorable change in any of these factors will produce an increase in shell life. Bronze shells are occasionally removed from service because of general or pitting corrosion. The factors that influence these failures are the corrosion environment and the resistance of the alloy to the environment. Good judgment in roll design, material selection, and control of the paper machine environment all contribute to the long life of suction rolls. Good quality control during the manufacture of the shell and timely in-service cleaning and roll inspection also contribute to a favorable service life. Operating Stresses. Both the applied stress and internal residual stresses inherent in the shell material contribute to the stress in operation of the suction roll. The responsibility for the level of applied stress rests chiefly with the roll designer who selects the limit of design stress for the material selected. Each shell material will have a limit of stress that will give long life in a particular environment. The magnitude of applied stress is a function of roll configuration and applied loads. Drilled hole patterns must also be favorably designed to reduce applied stress. The most important factor influencing the stress in the roll is shell thickness; of course, stress decreases as the shell thickness increases. A stress calculation procedure has been developed by TAPPI that can be used as a guide for determining stresses in suction rolls (Ref 44).

The residual stress that is inherent in the alloy and the heat treatment are also important factors contributing to roll life. Shell materials with known high residual stresses should be designed more conservatively or, preferably, not used at all. This is covered in the discussion "Material Selection" in this section. Gun-drilled or reamed holes for improved surface finish will reduce corrosion and improve the cleanliness of drilled holes in service. Gun drilling or reaming is usually not performed on bronze shells, but would be helpful in the maintenance of the shell in service. Corrosion Environment. The papermaker often has little control of some factors influencing the corrosion environment, yet there are many conditions that are within his control. For example, thiosulfate ion contamination and residual hydrosulfite (NaHSO2) in newsprint paper machine white water can occur in NaHSO2 brightening; practices for reducing this problem are addressed in Ref 40. Freshwater showers will provide more effective cleaning and reduce corrosive effects. Good maintenance of the roll is also an important factor in minimizing corrosion. Drilled holes must be kept clean of deposits by the use of needle showers and periodic off-machine cleaning. A careful program of biological control should be maintained. The papermaker should also avoid the use of cleaning chemicals (such as muriatic acid) that are known to be damaging to shell alloys and roll covers. The use of a corrosion inhibitor such as sodium mercaptobenzotriazole should be considered to reduce the corrosion of bronze rolls. The use of inhibitors for couch rolls is described in Ref 45. An example of severe pickup roll corrosion due to thiosulfate that was corrected with inhibitors is discussed in Ref 46. Materials Selection. When a new or replacement roll is contemplated, the question of material selection is of primary

importance. Past operating experience with existing suction rolls should be used as a guide in selecting the new shell material. Reference 47 describes conditions that contribute to roll failure and will aid in new shell materials selection. Duplex stainless steel has the highest resistance to fatigue cracking, followed by martensitic stainless steel and bronze. Austenitic stainless steel has good corrosion resistance to most environments, but has relatively low fatigue resistance and moderate-to-high residual stresses (Ref 41, 42). Duplex stainless steels have the highest fatigue strength and generally have moderate-to-low residual stresses; bronze has low residual stress, as does CA-15 martensitic stainless steel.

Selection of a new alloy should involve an evaluation of the condition of the existing roll and the factors contributing to its deterioration. The changes in the corrosion environment that have occurred on the paper machine should also be evaluated. One is tempted to select an identical alloy if the previous material served for 15 or 20 years. However, it is more likely that changes in the machine environment have occurred at a recent time and that the life of a new roll of the same alloy would be substantially reduced. This factor accounts for the recent popularity of the stronger corrosionresistant duplex stainless steel alloys. Suction shell material cost is also a consideration. In North America and Europe, bronze is the lowest cost shell alloy, followed by martensitic stainless steel, austenitic stainless, and the duplex alloys. In Japan, bronze is seldom used because of availability and cost. Manufacturing Quality. There is little published information relating to suction shell manufacturing quality. Shells should be inspected using liquid penetrant prior to drilling to detect such flaws as porosity, slag, and cracks. Skillful control of melting and casting procedures can minimize these defects. Weld repairs should be used with discretion, and only before shell heat treatment. Type 1N bronze should not be weld repaired, because of hot shortness, which develops in the alloy. Careful attention should be given to drilling and machining quality and dimensional control. Industrywide standards giving acceptance criteria for casting flaws and drilling quality have not yet been developed. In-Service Inspection. Periodic inspection of the suction roll is a critical step that will help to extend its life and

minimize unexpected shutdowns. Rolls should be removed from the machine, disassembled, and thoroughly cleaned and inspected at least annually. Inspection should include an evaluation of drilled hole cleanliness, fatigue cracking of the shell, and rubber cover deterioration or detachment. Water-washable liquid penetrant has been successfully used to detect cracks on the inside and outside surfaces. A careful visual examination should also be made for corrosion and cracks. When cracking is observed during an annual inspection, immediate consideration must be given to replacement because of the long lead time required for shell manufacture. Consideration should also be given to inspecting the roll more often than once per year to observe the rate of crack growth. Thus, a careful record must be made whenever cracks are observed. Corrosion by Kraft Pulping Liquors R.A. Yeske, The Institute of Paper Chemistry

The kraft process is the predominant pulping process used in North America to extract fibers from wood for use in the manufacture of paper, tissue, and board. The term kraft is derived from the German word for strong, which reflects the high strength of paper products derived from kraft pulp. This high strength, together with effective methods of recovering pulping chemicals, explains the popularity of kraft pulping. In 1984, more than three-fourths of the pulp produced in the United States--approximately 39,000,000 metric tons, or 43,000,000 short tons--was manufactured by the kraft process (Ref 48). In the kraft process, hot alkaline sulfide liquor is used to dissolve the lignin from wood chips and to separate individual wood fibers for use in paper-making (Ref 49). Wood chips are exposed to cooking liquor for several hours at elevated temperature and pressure in a process called digestion. Digestion may occur by repetitive batch processes in small batch digesters, or the process may occur continuously in larger continuous digesters. The contents of the digester are then discharged under pressure into a receiver called a blow tank. Finally, the fibers are separated from the spent liquor in a series of washing stages. Pulping chemicals are recovered from the spent liquor in a series of process steps shown schematically in Fig. 38. First, the spent liquor (known as black liquor) extracted for the pulp is concentrated to 60% solids content in multiple-effect evaporators. A soapy by-product called tall oil is usually extracted from the liquor at some stage of the concentration process. The heavy black liquor is burned as fuel in a chemical recovery boiler. The steam produced by the combustion of the organic constituents of black liquor is used for process use and for power generation. The inorganic constituents in the spent liquor fall to the bottom of the recovery boiler, where they are extracted as a molten salt called smelt.

Fig. 38 Schematic representation of kraft pulping and chemical recovery

The smelt recovered from the boiler is converted into cooking liquor in a process known as recausticizing. The molten smelt, which consists primarily of Na2CO3 and Na2S, is dissolved in water to make green liquor. The green liquor is first clarified to remove insoluble dregs and then causticized by exposure to slaked lime. In causticizing, Na2CO3 is converted to NaOH by the reaction:

Na2CO3 + Ca(OH)2 2NaOH + CaCO3

(Eq 27)

The CaCO3 precipitate is removed from the liquor by clarification or filtration, leaving a white liquor with approximately 100 g/L of NaOH, 30 g/L of Na2S, and lesser amounts of residual carbonates and sulfoxy compounds. The white liquor is stored for reuse as cooking liquor. Meanwhile, the CaCO3 precipitate (lime mud) is washed and converted to lime in a lime-burning kiln. This lime is later slaked and reused in the causticizing processes. Although some details vary from mill to mill, the process equipment used in kraft pulping is more or less the same throughout the industry. The principal items include pressure vessels for batch or continuous digestion, rotary drum pulp washers, multiple-effect liquor evaporators, storage tanks for various liquors, chemical recovery boilers, electrostatic precipitators, rotary lime kilns, slakers, green and white liquor clarifiers, and liquor storage tanks. Ancillary equipment includes pumps, valves, piping, heat exchangers, control instrumentation, and equipment for tall oil processing. To some extent, diffusion washers are replacing rotary drum washers and pressure filters are replacing clarifiers in the modern pulp mill. Much of the equipment used in the kraft pulp mill is fabricated from plain carbon steel, although carbon steel has limited resistance to corrosion and cracking when exposed to kraft process liquors. In critical locations (such as evaporator tubes in high-temperature evaporator effects), stainless steel is routinely used because of rapid attack of plain carbon steel. In other locations, stainless steel can be used to avoid the inconvenience of periodic replacement of carbon steel equipment. In general, the high temperatures and alkalinity of kraft liquors prevent the use of polymeric materials of construction in pulp mill equipment. However, brick and tile linings, as well as sprayed-on concrete linings, are often found in recausticizing equipment, particularly where abrasion is a concern. Periodically, pulp mill equipment is acid cleaned by

recirculating inhibited acid--usually HCl--to remove carbonate deposits that build up on screens and heat-exchanger surfaces.

Corrosion in Pulp Mill Equipment Batch digesters were initially thought to be immune to the severe corrosion affecting digesters used in the sulfite

pulping process. In 1930, a catastrophic digester failure killed several workers and prompted the industry to measure the wall thickness of kraft digesters (Ref 50). The digesters were inspected, and several were retired because of corrosioninduced wastage of the vessel wall. Thereafter, the industry was vigilant in monitoring the thickness of batch digester walls and retiring vessels that had corroded beyond safe limits. Corrosion damage appears in batch digesters in several forms, including uniform wastage, large gouges, pitting, and occasional cracking (Ref 51). In some cases, the gouges are clearly associated with erosion caused by steam impingement during direct steam heating of the digester, by recirculating liquors, or by impingement of high-velocity pulp slurries discharged at the end of a cook. Figure 39 shows an example of erosion-corrosion of a blow target plate used to break the momentum of the pulp mass discharged from the digester under pressure. Grooves are often found beneath blind nozzles and other areas where liquor-saturated pulp can lodge, allowing liquor to run down the walls between cooks. Corrosion is often more severe in the bottom cone area, where flow rates are higher during indirect heating and during blows. In some digesters, corrosion is most severe in the splash zone, where cooking liquor splashes on the hot sidewalls during digester charging. Welds are often sites of preferential attack in batch digesters, including accelerated corrosion and occasional episodes of SCC.

Fig. 39 Erosion-corrosion of a blow target plate by pulp discharged at high velocities from a batch digester

In the late 1940s, the industry became alarmed over an abrupt decrease in the service lives of batch digester vessels because of a corrosion-induced loss of wall thickness. At one mill with 20 digesters in operation, the service life of digesters installed after 1947 was only 3 to 5 years, while the lifetime for digesters installed before 1947 was 12 to 14 years (Ref 52). Several events apparently combined to increase the corrosion rate. The modern low-odor recovery boilers introduced about this time were more effective in retaining sulfur in the process, with resultant increases in the sulfidity of the cooking liquor. Batch digester production was increased by using more frequent batch cooks of greater severity (higher temperatures and more aggressive cooking liquors). Furthermore, semikilled steel was used to fabricate digesters, replacing the rimmed Bessemer steels used in earlier generations of digesters. Each of these factors played a role in the apparent increase in digester wall corrosion. Considerable progress was made in understanding and controlling batch digester corrosion as a result of the intensive investigations completed in the early 1950s. The effect of liquor composition was thoroughly investigated through surveys and laboratory corrosion studies. General wastage was greatly accelerated by thiosulfates and low concentrations of polysulfides in the cooking liquor (Ref 53, 54) and by the introduction of air into the digester during charging (Ref 55). Linear regression equations were developed to relate the corrosion rate to concentrations of various minor and major constituents in white liquor (Ref 55, 56). Laboratory and field studies showed that most digester corrosion occurred at the start of the batch cook (Ref 57, 58), and corrosion rates were described in the arcane unit, mils per two-thousand cooks

(MTTC). Digester walls were found to be passivated during later stages of the cook and reactivated when raw liquor was introduced into the vessel for the next cook. Particularly troublesome was a phenomenon called hot plate boiling, in which liquor that was charged into the vessel splashed off the chip pile onto sidewalls that were still hot from the previous cook (Ref 59). The evaporative concentration of dissolved chemicals and the damage to the passive film from the thermal shock combined to reactivate the surface of the vessel for further corrosion during the early stages of the next cook. The influence of the composition and the microstructure of digester steel on susceptibility to corrosion remains controversial. One researcher noted that vessels that were deoxygenated with silicon corroded more rapidly during contact with cooking liquor than steels with lower silicon and higher oxygen levels (Ref 60). This observation explains the preferential attack of high silicon welds that is frequently noted. Several laboratory studies confirmed the effect of silicon on the corrosion rate of steel (Ref 61, 62, 63), while other studies dismissed the effect as insignificant (Ref 64). A similar controversy regarding the benefits of normalizing heat treatments also developed (Ref 55, 65). The issue of silicon content has never been fully resolved, but many digesters have been subsequently fabricated with special digester steel (for example, ASTM A285 Grade C, containing no more than 0.03 wt% Si) and low-silicon weld filler metal (for example, E6010) in final passes on structural welds. Digester lifetimes were extended by a variety of measures involving mill operations, the use of linings, and the specification of low-silicon steels. Chip- and liquor-charging practices were modified to eliminate hot plate boiling. This was accomplished by introducing cooking liquor from the bottom of the digester, by using special liquor injection nozzles, and by simultaneous charging of chips and liquor. Graphite brick linings were once extensively used, but stainless weld overlays and thermal-sprayed coatings are now preferred. The durability of stainless weld overlays has often been inadequate in spite of a TAPPI standard for acceptable overlay practice (Ref 66). Pitting, cracking, and interpass corrosion are the most common complaints. Excessive dilution of the filler metal and production of partially martensitic overlays have been implicated in the poor performance of overlays. The choices of filler metal composition, heat input, and travel speed affect the uniformity of the overlay and the dilution of the filler metal. The weld overlay composition recommended by TAPPI for batch digester protection is as follows (Ref 66):

Alloying element Carbon Manganese Phosphorous Sulfur Silicon Chromium Nickel Molybdenum, when desired

Composition, wt% 0.15 max 2.50 max 0.045 max 0.030 max 1.00 max 18.0 min 8.0 min Cr:Mo 8:1

The schaeffler diagram predicts that overlays of the specified composition will be devoid of the ferrite required to prevent hot cracking. Furthermore, the high allowable carbon concentration may result in sensitizing of the overlay by the heat of subsequent weld passes. Consequently, mill operators often specify a higher-chromium equivalent and lower carbon in the as-deposited overlay. Blow Target Plates. Erosion-corrosion of the target plate in blow tanks is a continuing maintenance problem, but some advantage can be obtained by using more highly alloyed material. For example, Hastelloy alloy C-276 has endured for 4 years in target plate service where 18-8 austenitic stainless steels were lasting for only a few months. Piping and Ancillary Equipment. Carbon steel piping is subject to corrosion and erosion-corrosion in the kraft pulp

mill, but is often chosen for economy. Elbows lines and other areas of high abrasion are particularly susceptible to attack. Type 304L stainless steel is effective in controlling corrosion damage in piping. Pump impellers, pump casings, and valves made from cast CF-8M are usually resistant to attack by kraft liquors.

Continuous Digesters. In contrast to the cyclic environments encountered in batch digesters, wetted surfaces in

continuous digester experience the same environment for months at a time, although the exact environment depends on the elevation in the digester. Wood chips are added at the top of the vessel, and the pulp mass works its way downward while the chips are successively saturated with liquor, cooked, washed, cooled, and discharged from the bottom of the vessel. Cooking, extraction, and wash liquors are circulate through the pulp through a series of internal screens and downcomer pipes. Although the rate of uniform wastage in continuous digesters is usually quite low--of the order of 0.13 to 0.25 mm/yr (5 to 10 mils/yr)--higher rates of attack have been observed locally. In particular, erosion-corrosion can be a problem where fresh cooking liquor impinges on steel surfaces near the top of the digester. Pitting of welds and preferential corrosion of weld filler metal have also been observed in continuous digesters, but this attack may be result of acid cleaning practices, rather than attack by cooking liquor. A severe corrosion problem in continuous digesters literally burst upon the scene in 1980 with the catastrophic failure of a large continuous digester vessel during routine operation (Ref 67). The digester ruptured near the top of the vessel because of extensive SCC of shell girth welds. This cracking was restricted to the welds in the impregnation zone at the top of the digester. Caustic SCC was implicated because of the branched, intergranular appearance of the cracks and the presence of sodium hydroxide in the hot cooking liquor. Subsequent inspections of similar digester revealed that more than half of the 140 continuous digesters operating in North America exhibited cracking in structural welds (Ref 68). Cracking has been found in girth and vertical shell welds, attachment welds inside the digester, and nozzle welds. Both longitudinal and transverse cracking have been observed, but the deepest cracking has been found in longitudinal cracks in the weld HAZ. Examples of cracking in actual digester welds are shown in Fig. 40.

Fig. 40 Examples of digester weld cracking. (a) Macrograph showing cracking in a sample taken from a continuous digester weld. 10×. (b) Photomicrograph showing branched, intergranular nature of cracking in an actual continuous digester weld. 40×

Most of the cracking has been found in the impregnation zone, where the caustic concentration in the cooking liquor is highest. Furthermore, the residual stresses in impregnation zone welds were often not relieved by postweld heat treatment, because the shell is thinner in this zone and stress relief is not required under provisions of the ASME. Boiler and Pressure Vessel Code. An extensive survey of digester cracking statistic failed to reveal differences in digester design or operation that would account of differences in cracking susceptibility (Ref 68). As shown in Fig. 41 postweld heat treatment significantly reduced, but did not eliminate, susceptibility to severe cracking.

Fig. 41 Effect of stress relief on SCC susceptibility of continuous digesters

Cracking similar digester cracking was reproduced in the laboratory by using accelerated slow strain rate and fracture mechanics test (Ref 69). These tests indicated that the caustic concentration in cooking liquors at the impregnation zone was sufficient for caustic cracking of pressure vessel weldments. Furthermore, caustic cracking occured only when the potential of the digester steel was within a 100-mV range centered close to digester potentials. Figure 42 shows the dependence of cracking susceptibility on potential as determined in slow strain rate tests performed on welded specimen in simulated impregnation zone liquor (Ref 70). Potential measurements are made on an operating digester indicated that the digester rest potential remained above the cracking range, except for a few days following upset in operating routine (Ref 71).

Fig. 42 Plot showing the effect of potential on cracking severity in controlled-potential slow strain rate testing of digester steels exposed to a simulated impregnation zone liquor. SCE, saturated calomel electrode

Two measures have been successful in controlling digester weld cracking: high alloy barrier coatings placed over susceptible welds and anodic protection. Other remedial measures, such as shot peening, temper-bead weld repair, in situ stress relief, and unsealed thermal spray coatings, have not been uniformly successful in preventing recracking. Weld overlays and thermal sprayed coatings have both been successfully used to control cracking. Weld overlays-primarily Inconel alloy 82 and type 309L stainless steel--have been applied in bands over structural welds to isolate them from contact with cooking liquor. The overlay bands are themselves resistant to corrosion damage, but there have been cases of cracking of the carbon steel substrate in the HAZ at the edge of the overlay band. Experience with thermal sprayed coatings has been limited, but it appears that flame-sprayed and plasma-sprayed stainless steel coatings protect underlying welds if the coating is sealed with a silicone-modified furan after application (Ref 72). Additional information is available in the article “Thermal Spray Coatings” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook. Anodic protection has been effective in controlling both corrosion and caustic cracking in a number of continuous digesters (Ref 73). Protection is achieved by passing a controlled direct current (dc) through an electrolytic cell consisting of the digester wall, the cooking liquor, and a special cathode installed inside the digester. Currents as high as 1000 A (at 12 V dc) may be required to passivate the vessel initially, but only a few hundred watts of electrical power is required to maintain anodic protection. The digester potential is maintained approximately 100 mV above the upper limit of the potential range required for cracking. Anodic protection has suppressed further cracking in several digesters previously susceptible to severe and chronic cracking. Additional information is available in the article “Anodic Protection” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook. Liquor Reheater Tubing. Corrosion and cracking have been encountered in stainless steel tubes in shell and tube heat

exchangers used to heat recirculating liquors. An example of a cracked liquor reheater tube is shown in Fig. 43. In most cases, severe transgranular cracking occurs when liquor leaks onto the steam side of the tubes, where it is concentrated and heated by the steam. Caustic cracking and chloride SCC have both been implicated. Differential thermal expansion between the stainless steel tube bundle and the carbon steel shell apparently provides the stress required for cracking. Cracking has been eliminated in more than 100 kraft mills by the use of a duplex stainless steel (3RE60, see Table 13 for composition) in liquor heater tubes; Inconel heater tubes have also been successfuly used (Ref 74, 75).

Fig. 43 Transgranular SCC of type 304L stainless liquor reheater tubing

Corrosion of stainless steel liquor heater tubing is more puzzling, because stainless steel is generally resistant to attack by kraft liquors at liquor heater temperatures. Corrosion of liquor heater tubes may actually result from improper acid cleaning--for example, by using excessive temperatures uninhibited HCI.

Corrosion in Recausticizing Equipment The recausticizing plant provides a variety of corrosive media, including white and green liquor, lime mud slurries, lime mud wash waters, and lime kiln scrubber waters. White and green liquors attack storage tanks, clarifiers, and piping fabricated from plain carbon steel, but have little effect on stainless steel equipment (in the absence of abrasion). Erosioncorrosion is frequently seen on equipment in contact with abrasive lime mud slurries, such as clarifier rakes, lime mud slurry pump casings and impellers (Fig. 44), and lime mud discharge lines. Corrosion also occurs on lime kiln scrubber components.

Fig. 44 Erosion-corrosion of a lime mud pump impeller

Plain carbon steel storage tanks and clarifiers in white and green liquor service suffer severe corrosion attack, particularly in zones where liquor levels fluctuate frequently. Figure 45 shows the variation in corrosion rate at different elevations along the sidewall of a white liquor clarifier (Ref 76). In the zone immediately above the outlet bustle pipe, the corrosion rate approaches 0.75 mm/yr (30 mils/yr). Areas that are seldom immersed experience a much lower rate of corrosion, as do surfaces near the bottom, which are protected by a layer of lime mud sediment. In multitray clarifiers, the trays experience a similar form of attack. Little corrosion is seen on the top of the tray where lime mud settles, but severe attack occurs on the underside of the tray where air bubbles entrained in the liquor remain in contact with the tray. White liquor clarifiers and storage tanks often need extensive repair or replacement after only 10 years of service, and cases of severe corrosion in less than 2 years have been reported.

Fig. 45 Variation of corrosion rate with elevation at the wall of a white liquor clarifier

The corrosivity of white liquor is strongly influenced by the liquor composition. In simulated white liquors containing only the important pulping chemicals--NaOH and Na2S--the rate of corrosion of carbon steel is less than 0.25 mm/yr (10 mils/yr) (Ref 77). However, the presence of thiosulfates and low concentrations of polysulfide in white liquor can increase the corrosion rate to more than 1.3 mm/yr (50 mils/yr), as shown for thiosulfate in Table 14. The effect of thiosulfate and polysulfide is greater in liquors containing high concentrations of NaOH and Na2S. The other species typically found in white liquor--sulfates, sulfites, chlorides, and carbonates--are apparently innocuous (Ref 78). Because

the hydrosulfide in white liquor is rapidly converted to thiosulfate by air contact, the high rate of corrosion at air/liquor interfaces is probably related to the thiosulfate effect shown in Table 14.

Table 14 Effect of thiosulfate concentration on white liquor corrosivity Duration of exposure: 8 weeks Concentration, g/L NaOH Na2S Na2S2O3 20 0 80 20 2.5 80 20 10 80 20 25 80 20 50 80 40 2.5 80 40 5 80 40 10 80 40 25 80 40 50 80

Corrosion rate mm/yr mils/yr 0.18 7 0.18 7 0.28 11 0.71 28 1.35 53 0.15 6 0.51 20 0.53 21 1.09 43 2.21 87

Source: Ref 78

Liquor velocity relative to steel surfaces can also have a large effect on the rate of corrosion (Ref 79). As shown in Table 15, the rate of corrosion of carbon steel by white liquor is increased fivefold by a change from stagnant to laminar flow conditions. Under turbulent flow conditions, corrosion rates exceeding 3.8 mm/yr (150 mils/yr) are possible.

Table 15 Effect of liquor velocity on white liquor corrosivity of simulated liquor, 100 g/L NaOH + 33 g/L Na2S Velocity m/s ft/s 0.0 0.0 0.14 0.46 0.25 0.82 0.30 0.98 0.43 1.41 0.50 1.64 0.86 2.82 1.32 4.33 2.62 8.60

Corrosion rate mm/yr mils/yr 0.13 5 1.65 65 2.03 80 2.51 99 3.05 120 3.63 143 3.50 138 3.81 150 3.76 148

Several corrosion control measures have been effective in recausticizing applications, including stainless steel fabrication, specification of thickwall carbon steel at critical sites, controlling thiosulfate and polysulfide levels, the use of barrier linings, and anodic protection. Type 304L stainless steel is not affected by white or green liquor, and stainless steel recausticizing equipment will probably last for the life of the mill. The higher cost of stainless steel can be mitigated to some extent by eliminating the corrosion allowance on wall thickness. There is little justification for using more costly type 316L stainless steel in recausticizing applications; the high pH of the liquors inhibits chloride attack and precludes the need for molybdenum in the stainless steel. In vessels with corrosion at the liquid level line as shown in Fig. 45, stainless steel belly bands can be installed to prolong tank lifetime. Usually, formed stainless steel plates are seal welded to the inside wall. Careful installation is required to prevent liquor from leaking between the stainless cladding and the wall and accelerating wall corrosion. Experience with alkali-resistant concrete and brickwork linings in recausticizing vessels has been checkered. In some cases, the lining prevents further corrosion damage. In may cases, liquor penetrates behind the barrier linings, allowing renewed attack of the wall. Installation of a protective polymeric membrane between the vessel wall and the lining may eliminate damage if the lining is breached.

Thiosulfate levels can be reduced by minimizing air contact with liquor in order to lower corrosion rates in white and green liquor systems. Thiosulfates are formed wherever dissolved sulfides in the liquors come in contact with air. Likely sites for thiosulfate formation include shatter jets (where smelt is dispersed as it falls into the dissolving tank liquor), splitter box weirs, storage tank and clarifier liquid surfaces, and leaking pump seals. Steam shatter jets are preferred over compressed air jets for reducing thiosulfate levels. In some cases, thiosulfates are generated when weak wash liquor is used to scrub lime kiln flue gases before being used as dissolving tank makeup liquor. Scrubber waters should be completely oxidized to convert thiosulfates to innocuous sulfates before use as dissolving tank makeup liquor. Natural polysulfide levels are negligible in most white liquors. However, polysulfide-accelerated attack can occur if emulsified sulfur in added (for sulfur makeup) in the recausticizing plant, because elemental sulfur is rapidly converted to polysulfide. Makeup sulfur should instead be added to black liquor to minimize corrosion damage. Recently, anodic protection has been shown to be an effective and economical corrosion control measure for use in white liquor storage tanks and clarifiers (Ref 76). In a prototype installation on an operating clarifier with severe corrosion damage, corrosion rates were reduced by nearly an order of magnitude by holding the rest potential of the vessel in the passive range. The anodic protection system is shown schematically in Fig. 46. Although relatively large currents were required to achieve initial protection, little power is needed to maintain protective passivation.

Fig. 46 Schematic diagram of an anodic protection system for recausticizing storage tanks and unit clarifiers

Erosion-corrosion is a severe problem wherever abrasive lime mud is found. Stainless steel fabrication is appropriate wherever abrasive conditions prevail, as in rakes and rake arms in while and green liquor clarifiers. In severely abrasive conditions, such as lime mud pump impellers, hard-facing overlays such as cobalt-base Stellite alloys may be effective, but periodic replacement of components is usually necessary.

Corrosion in Black Liquor Processing Equipment Black liquor varies considerably, depending on the pulping process used, the pulping yield, and the stage of chemical recovery. Weak black liquor, the filtrate obtained from brown stock washing, contains approximately 20% solids consisting of organic compounds extracted from the wood (principally lignin compounds), inorganic salts not consumed in the pulping process, organic acids, and various hemicelluloses and sugars. The pH of weak black liquor is of the order of 12, and the residual active alkali (NaOH and Na2S) can range from 4 to 30 g/L. The solids content of heavy black liquor is increased by evaporation into the 60 to 75% range, and heavy black liquor at room temperature is viscous and

semisolid. Black liquor oxidation can be used to convert reduced sulfur compounds in black liquor to thiosulfates and sulfates in order to reduce odorous emissions from the recovery boiler and to conserve sulfur in the process. Corrosion of brown stock washing equipment is rarely a concern in the kraft pulp mil. Brown stock washing is usually carried out on rotary drum vacuum washers, although displacement washing is becoming more common. Washer drums are usually fabricated from carbon steel, while backing wires and face wires are usually stainless steel. The moderately high pH of weak wash filtrate, together with the low concentration of NaOH and the inhibitive effect of the dissolved organic species, generally prevent severe corrosion of brown stock washing equipment in spite of the dissimilar-metal construction. However, corrosion can become a problem if the pH of the weak wash filtrate is depressed by the use of acid bleach plant filtrate as shower water or by the addition of neutral sulfite semichemical pulping liquors. When acidic filtrates are used, it is common practice to raise the pH above 10 to reduce corrosion and to prevent odorous emission of hydrogen sulfide gas. The details of black liquor evaporation equipment vary from mill to mill. Multiple-effect evaporators are generally used to increase the black liquor solids to the 55% level, while concentrators or cascade evaporators complete the evaporation. The shell and tube heat exchanger (rising film, long tube vertical evaporators) is the most common evaporator design, but other configurations with plates and falling films are becoming more popular. Plain carbon steel and type 304L and 316L stainless steels are common materials of construction. Stainless steel tubes are used in the higher-temperature evaporator effects (>95 °C, or 200 °F); both stainless and carbon steel tubes are used at lower temperatures. Black liquor storage tanks and oxidation equipment are usually made of plain carbon steel. Stainless steel piping is commonly used for black liquor transport. In general, black liquor is much less corrosive toward carbon steel than the recausticizing liquors, because the concentrations of NaOH and Na2S in black liquor have been reduced by the pulping reactions. The residual alkali in black liquor maintains the pH near 12, which stabilizes a protective film on the steel surfaces. In spite of this film, carbon steel storage tanks wetted by black liquor may experience pitting and crevice corrosion, particularly at the liquor level line and underneath sediments and scales. Pitting attack, such as that shown in Fig. 47, is often encountered on roofs and other surfaces that are exposed to vapors emanating from the black liquor. This pitting is attributed to volatile organic acids that condense on cooler surfaces but are not neutralized by direct contact with the alkaline black liquor.

Fig. 47 Pitting of a carbon steel root section in a black liquor storage tank

Remedial measures for corrosion in black liquor storage tanks are limited. Generally, the moderate rate of attack does not justify the use of stainless steel for tanks and vessels. Although the high temperatures and organic sulfur compounds in black liquor make fiber-reinforced plastic generally unsuitable for immersion service, some mills have installed fiberreinforced polymeric roofs on black liquor storage tanks to minimize pitting attack. Thick-wall carbon steel construction and the use of stainless belly bands may also be beneficial when attack is restricted to one area, such as the liquor level line. Stainless steels are generally resistant to corrosion by black liquors at temperatures normally encountered in evaporation and concentration equipment. Over the years, experience has shown which evaporator components must be stainless to prevent excessive corrosion, therefore, severe evaporator corrosion is now relatively infrequent. Tubes, vapor deflectors,

and liquor boxes are generally stainless steel or stainless-chad carbon steel, particularly in the high-temperature effects. When localized corrosion does occur on carbon steel surfaces in modern evaporators, it can usually be traced to unanticipated contact between hot liquor and the steel due to foaming, excessive throughout, or other operating misadventures. In these cases, stainless steel weld overlays or thermal spray coating applied to the carbon steel surface usually forestalls continued corrosion. Although black liquor oxidation produces high levels of thiosulfate in the black liquor, the thiosulfates are much less injurious to carbon steel in black liquor service as compared to the recausticizing situation. Oxidized black liquors have low residual alkali levels because of consumption of NaOH in the digestion process and complete oxidation of Na2S. In the absence of significant concentrations of NaOH and Na2S, thiosulfates are relatively innocuous toward carbon steel. Consequently, it is desirable to oxidize weak black liquor, rather than strong black liquor, so that more of the black liquor process equipment benefits from the reduced corrosion that accompanies black liquor oxidation. Carbon steel is generally adequate for oxidation equipment and oxidized black liquor storage. Corrosion of Mechanical Pulping Equipment C.B. Thompson, Pulp and Paper Research Institute of Canada; D.A. Wensley, MacMillan Bloedel Research

In mechanical pulping, wood fibers are separated from each other by rolling, rubbing, or teasing, either against themselves or against a harder, specially designed surface. Chemical additions or the application of heat and pressure can be used to increase the efficiency of the process, but the predominant pulping forces are mechanical. Mechanical pulping can be divided into two main areas: groundwood and disk refining. in overall terms, mechanical pulping forms an important part of the pulp and paper industry and is notable for its technological and economic expansion in recent years (Ref 80).

Mechanical Pulping Systems Groundwood is one of the oldest commercial pulping processes and, until recently, was the predominant means of

mechanical pulping. The most common version is stone groundwood, in which pulp is formed by pressing debarked logs against a rotating grindstone, as shown in Fig. 48. Water sprays are directed onto the stone to prevent the fibers from burning, and the pulp is removed in the form of a slurry. Other variations include pressurized groundwood, in which pulping occurs at elevated temperature, and chemi-groundwood, in which a chemical pretreatment is used.

Fig. 48 Schematic sectional view of a stone groundwood machine. Cut, debarked logs are fed into magazines, from which they are pressed against a rotating grindstone to be pulped.

Corrosion is usually not of concern in these units, largely because most pulping is done by the stone groundwood method, in which water sprays are the only potentially corrosive environment. As a result, attack is limited to general wastage of exposed carbon steel. Disk refining has overtaken groundwood in importance since the 1970s, primarily because of its ability to produce

higher-strength pulps and to utilize chips or sawmill residues as feedstock. Modern disk refiners are typically large and powerful; disk diameters up to 1.8 m (70 in.) are common, with installed motor powers of 13,400 kW (18,000 HP) (Ref 80). A schematic sectional view of a refiner is shown in Fig. 49.

Fig. 49 Schematic sectional view of a disk refiner. Pulping takes place between rotating rings of refiner plates. In this example, the refiner plates are mounted on opposing rotor assemblies (shaded) that counterrotate at high speed. Such a unit is called a double-disk refiner. Single-disk refiners, in which only one set of refiner plates is rotated, are also common.

In operation, refiner pulp is produced by passing wood chips between two or more closely spaced serrated disks, at least one of which is being rotated at high speed by an electric motor. Each disk is made by bolting refiner plates onto a backing plate. A typical refiner plate is illustrated in Fig. 50, showing the characteristic series of raised bars. The space between two opposing sets of refiner plates is called the refining zone. As wood chips pass outward through the refining zone, they are gradually reduced to pulp by interaction with refiner plate bars. Although many plate bar patterns exist, it is common to have large bars near the inlet to cause rapid breakdown of the chips. As the pulp progresses through the outer refining zone, the bars become finer in order to impart maximum refining energy to the stock. In most cases, it is usual to have two or more refiners operating in series to obtain optimum refining efficiency.

Fig. 50 General view of a refiner plate. The large bars at the bottom are called breaker bars. The bars become progressively finer towards the top edge of the plate.

A large number of refiner pulping processes have been developed, primarily as a result of the need to produce tailor-made pulps and to decrease energy requirements. The simplest process is refiner mechanical pulping, in which the chips are fed directly to the refiner with no prior use of heat or chemical dissolution. For newsprint production, this process has now been largely superceded by thermomechanical pulping, in which the chips are softened in a steaming tube before entry into the primary refiner, and by chemi-thermomechanical pulping, in which both heat and chemical additives are used in order to accelerate the pulping. In both thermomechanical and chemi-thermomechanical systems, chip presteaming and primary refining take place at temperatures above 100 °C (212 °F); in the secondary refiner, the pulp is usually discharged to atmosphere. A schematic diagram of a thermomechanical pulping system is shown in Fig. 51.

Fig. 51 Schematic of a typical thermomechanical pulping system layout

In addition to the pure mechanical pulping described above, refiners are also used in a variety of chemi-mechanical pulping methods, in which chips are typically pretreated in a digester before entering the refiners. The principal difference between the two types of pulping is the pulp yield. Chemi-mechanical pulps are in the range 55 to 90% yield; the yield for pure mechanical pulping exceeds 90%. Corrosion is of more concern in disk refining than in stone groundwood due to the higher process temperatures and the presence of pulping chemicals. In addition, significant wear can occur because of the action of high-velocity jets of steam and water within the refiner and the presence of sand and grit entrained in the stock. Presently, the major problem areas for corrosion and wear are the refiner plates and the steaming tubes; these are discussed below in more detail.

Corrosion and Wear of Refiner Plates Most refiner plates are cast, although wrought plate sets are available for low-consistency units. Plates are produced by either refiner manufacturers or specialist refiner plate foundries. Alloys are often of proprietary composition, although most fall into the general classes of white cast iron or cast stainless steel (Ref 81, 82). Some typical compositions are shown in Table 16. Refiner plates cast in white cast iron generally have a large volume fraction of primary carbides set in a martensitic matrix, which sometimes contains large amounts of retained austenite. Cast stainless steel refiner plates usually possess a lower concentration of carbides set in a martensitic matrix. Some typical microstructures are shown in Fig. 52.

Table 16 Typical refiner plate alloy types and compositions Alloy type Martensitic white cast iron Martensitic high-chromium white cast irons Martensitic cast

Composition, wt% C Cr Ni Si 3.3 2.4 4.0 1.2 2.7 23.2 0.4 0.5 3.0 27.7 1.1 0.6 0.8 18.0 1.4 0.9

Mo 0.3 2.0 0.1 0.5

Mn 0.6 0.7 0.5 0.6

P 0.03 0.02 0.03 0.03

S 0.02 0.02 0.03 0.02

Fig. 52 Typical refiner plate alloy microstructures. (a) Low-chromium white cast iron; martensite in a eutectic matrix of martensite and iron carbide. Etched with 2% nital. (b) High-chromium white cast iron; iron-chromium primary carbides in a martensitic matrix. Depending on the heat treatment, a large amount of retained austenite may be present in the structure. Etched with 2% nital. (c) Cast stainless steel; iron-chromium primary carbides in a martensitic matrix containing secondary carbides. Etched with mixed acids. All 230×

Corrosion and Wear. Premature wastage of refiner plates by corrosion and wear adversely affects such critical pulp quality factors as burst, tear, and shives (Ref 83, 84). Worn plates can also decrease paper machine runnability; this can result in significant costs due to machine stoppages or lower speeds (Ref 85) and the need for chemical reinforcing pulps. The types of damage caused by the corrosion and wear of refiner plates are detailed below. It should be noted, however, that little systematic study has been reported in this area, and there is no consensus of opinion as to the actual mechanisms by which such wastage takes place. Bar rounding is usually found in the outer refining zone on the bar leading edges. Severe bar rounding leads to loss of

profile and a reduction in refining efficiency; a typical example is shown in Fig. 53. Bar rounding appears to take place by a combination of corrosion and low-stress abrasion due to sand and grit in the stock. Examination of rounded surfaces at high magnification shows numerous short abrasion furrows, often with associated lips or platelets, of deformed material. Typically, the primary carbides are left in relief on the surface.

Fig. 53 Severe bar rounding (arrows) on a cast stainless steel refiner plate. Approximately

×

Clashing. During operation, it is common for opposing sets of refiner plates to touch each other. This is called clashing,

and it can be due to loss of stockfeed or a temporary disruption of the pulp pad between the plates. Because of the high power applied during refining, clashing produces severe plate damage, commonly in the form of deep, circumferential grooves and serrations. A typical example is shown in Fig. 54. Some investigators have attributed clashing to electrochemical dissolution rather than mechanical contact, primarily because of the remarkably smooth surfaces that are often found on clashed plates (Ref 86). However, thermal cracking (Ref 87) and evidence of surface melting and smearing (Ref 82) have been observed on clashed plates, and it is probable that in most cases metal wastage due to clashing is primarily mechanical in nature. To reduce the frequency of clashing, various type of proximity sensors have been developed to monitor the gap between opposing plate sets (Ref 88).

Fig. 54 Clashed surface on a white cast iron refiner plate. The circumferential serrations are typical of this type of wear.

Cavitation and Liquid Droplet Impingement Erosion. Heavy pitting at the base of breaker bars and on the edges

and top surfaces of intermediate and outer zone refining bars is commonly attributed to cavitation erosion (Ref 89, 90). A typical example is shown in Fig. 55. In some cases, similar attack can be produced by liquid droplet impingement erosion, which is caused by water jets or condensing steam (Ref 91). Metal wastage due to these mechanisms is generally regarded as secondary in importance to bar rounding and clashing; plates are changed only if there is a danger of bar segments being undermined and breaking free during operation. In some cases, however, cavitation has caused such rapid loss of bar profile that plates have had to be changed after as little as 300 h of service. Such intense attack is probably due more to the design and operational characteristics of the particular refiner than to any deficiency in the cavitation resistance of the refiner plate.

Fig. 55 Heavy cavitation attack on the intermediate refining zone bars of a high-chromium white cast iron refiner plate

Corrosion in Pressurized Equipment Corrosion of pressurized equipment constructed from stainless alloys, such as thermomechanical or chemithermomechanical pulping steaming vessels and bisulfite chemi-mechanical pulping digesters, can cause serious materials

reliability problems if either chlorides or acidic sulfur compounds are present. Problems include chloride SCC, pitting, and crevice corrosion, as well as general wastage due to sulfuric acid condensation. Thermomechanical and Chemi-Thermomechanical Pulping Thermomechanical and chemi-thermomechanical steaming vessels may be of horizontal, vertical, or inclined configuration. Steaming is typically carried out at 130 to 150 °C (260 to 300 °F) and at 200 to 350 kPa (30 to 50 psig). Horizontal steaming tubes (Fig. 56) experience corrosion problems in areas not swept clean by the moving chip bed. Deposits can form on the vessel walls and on the lee sides of longitudinal rub bars. The rub bars are used to guide the chip mass along the vessel and to protect the vessel from accidental damage by the screw.

Fig. 56 View inside a horizontal chemi-thermomechanical pulping steaming tube with the screw removed for inspection. The vessel is of UNS NO8904 construction and has three longitudinal rub bars of the same composition stitch-welded to the shell. See also Fig. 59.

Corrosion by Chlorides. Chips from seaborne logs may introduce appreciable amounts of chloride. The liquid film

surrounding damp wood chips may contain 2500 ppm NaCl and may also have an acid pH (4 to 6) due to the release of wood acids. Chlorides can concentrate beneath deposits, perhaps by evaporation of areas intermittently wetted with chloride-containing water. Chemi-thermomechanical pulping steaming tube deposits have been analyzed to contain as much as 38,000 ppm NaCl. Chloride SCC has caused the rupture of a type 316 stainless steel steaming tube. Severe cracking was associated with the residual tensile stress fields around rub bar stitch welds (Fig. 57). Extensive cracking was also found beneath vapor space deposits in the same vessel--well removed from any welds (Fig. 58). Chloride SCC has also been observed in horizontal steaming vessels and screws constructed from UNS NO8904 stainless steel (Fig. 59 and 60).

Fig. 57 Cross section of a piece cut from a type 316 stainless steel steaming vessel that failed by chloride SCC. Cracking occurred under deposits on the lee-side of a 13-mm (0.5-in.) square rub bar and progressed through

the 6-mm (0.25-in.) thick shell.

Fig. 58 Photograph of the inside surface of a failed AISI 316 steaming tube showing chloride SCC that developed beneath deposits (the cracks have been opened by subsequent deformation). The lower part of the surface was swept clean by chip movement and did not crack.

Fig. 59 Micrograph of transgranular SCC observed in a UNS NO8904 rub bar removed from the steaming tube shown in Fig. 56

Fig. 60 Crows-foot SCC observed on the inside top surface of the shell of a UNS NO8904 steaming tube. Approximately 1×

Pitting or crevice corrosion has also been observed in steaming vessels, often in association with general attack due to condensation of acidic sulfur compounds. Chloride pitting attack tends to be particularly widespread beneath deposits that form on the screw shaft cladding and flights. Corrosion by Acidic Condensates. Additions of Na2SO3 can be made prior to the steaming tube. Sulfur dioxide

evolution occurs from sulfite solutions under the acidic conditions that exist naturally with damp wood chips. Sulfur dioxide is converted into the insidious sulfur trioxide in the presence of oxygen. Sulfur trioxide may condense or dissolve in condensed water films as sulfuric acid. Stainless steels are resistant to H2SO4 only in either very dilute or very concentrated solutions. Condensates of intermediate concentration are therefore liable to be quite corrosive. Indeed, condensate corrosion runs have been observed on a steaming tube wall (Fig. 61). Condensate drips that form directly overhead produce circular areas of corrosion. The progress of acidic condensate corrosion in a UNS NO8904 steaming tube has been monitored by photographs taken 16 months apart (Fig. 62 and 63).

Fig. 61 Acidic condensate corrosion runs observed across a longitudinal seam weld on the wall of a UNS NO8904 steaming tube (see Fig. 56 left side wall)

Fig. 62 Circular areas of corrosion observed directly overhead in a UNS NO8904 steaming tube due to formation of acidic condensate drips. Approximately

×

Fig. 63 Photograph 16 months later of the same area as in Fig. 62 showing the progress of corrosion from acidic condensation. Approximately

×

Bisulfite Chemi-Mechanical Pulping Digesters Bisulfite chemi-mechanical pulping involves cooking wood chips in continuous digesters with NaHSO3 liquors at 160 °C (320 °F) and at 620 kPa (90 psig). Depending on the amount of liquid in the digester, cooking may be in either the liquid or vapor phase. The conditions that promote the formation of vapor-phase condensates containing high concentrations of H2SO4 are the presence of air and low pH (Ref 92). If no effort is made to exclude air, the raw liquor pH must be 7.6 to prevent H2SO4 formation. At the optimum process pH of 6, conversely, extremely corrosive condensates containing 34.9% and having a contact pH of 0 can be formed. Corrosion-free operation at pH 6 has been obtained by strict exclusion of air by the following means: • • • •

Presteaming of the chips Plug-screw feeding into the digester, below the liquid level Preheating the cold blow water used to fill the digester prior to start-up Replacing the air in the vessel with steam prior to start-up

References 1. 2. 3. 4. 5. 6.

A. Garner, Pulp Paper Can., Vol 82 (No. 12), 1981, p T414 R.J. Brigham and E.W. Tozer, Corrosion, Vol 29, 1973, p 33 A. Garner, Corrosion, Vol 35, 1979, p 108 R.J. Brigham, Corrosion, Vol 30, 1974, p 396 A. Garner, Corrosion, Vol 37, 1981, p 178 B. Wallen, in Pulp and Paper Industry Corrosion Problems, Vol 2, National Association of Corrosion Engineers, 1977, p 43 7. A.H. Tuthill, J.D. Rushton, J.J. Geisler, R.H. Heasley, and L.L. Edwards, TAPPI J., Vol 62 (No. 11), 1979, p 49 8. A.H. Tuthill, in Pulp and Paper Industry Corrosion Problems, Vol 4, Swedish Corrosion Institute, 1983 9. J. Hill, in Pulp and Paper Industry Corrosion Problems, Vol 4, Swedish Corrosion Institute, 1983 10. L.H. Laliberté and A. Garner, TAPPI J., Vol 64 (No. 1), 1981, p 47 11. S. Henrikson and V. Kuchera, in Pulp and Paper Industry Corrosion Problems, Vol 3, National Association of Corrosion Engineers, 1983, p 137 12. A. Garner, How Stainless Steel Welds Corrode, Met. Prog., Vol 127 (No. 5), April 1985, p 31 13. American Welding Society Specification AWSA5.9-69 14. A. Garner, Weld. J., Vol 62 (No. 1), 1983, p 27

15. A. Garner, Mater Perform., Vol 21 (No. 5), 1982, p 43 16. A. Garner and L.H. Laliberté, U.S. Patent 4,285,787, 1981 17. A. Garner, U.S. Patent 4,285,232, 1981 18. O.V. Ingruber, M.J. Kocurek, and A. Wong, Ed., Pulp and Paper Manufacture, Vol 4, Sulfite Science and Technology, Joint Textbook Committee of the Paper Industry of the United States and Canada, 1985 19. S.A. Rydholm, Pulping Processes, 1st ed., John Wiley & Sons, 1965, p 469 20. P.-E. Ahlers, The Corrosion of Stainless Steels in Passive and Active States Under the Conditions of Sulfite Cooking, in Pulp and Paper Industry Corrosion Problems, Vol 3, National Association of Corrosion Engineers, 1982, p 66 21. H.-J. Rocha, Sulphurous Acid as Corrosion Inhibitor in Sulphite Mills, DEW-Technische Berichte, Vol 10 (No. 2), 1970, p 124 22. O.V. Ingruber, Sulfite Pulping Cooking Liquor and the Four Bases, in Pulp and Paper Manufacture, Vol 4, Sulfite Science and Technology, Joint Textbook Committee of the Paper Industry of the United States and Canada, 1985 23. D. McGovern, A Review of Corrosion in the Sulfite Pulping Industry, in Pulp and Paper Industry Corrosion Problems, Vol 3, National Association of Corrosion Engineers, 1982, p 60 24. S.K. Murarka and W.T.A. Dwars, Potential Formation of Acidic Condensates in Vapour Phase Sulphite Pulping, in Proceedings of the 71st Technical Section Annual Meeting, Canadian Pulp and Paper Association, 1985, p A109 25. G.J.C. Potter and R.B. Kesler, The Formation of Sulphur Trioxide and Calcium Sulphate in the Sulphite Process, TAPPI J., Vol 41 (No. 2), 1958, p 183A 26. K.-E. Jonsson, Stainless Steels in the Pulp and Paper Industry, in Handbook of Stainless Steels, D. Peckner and I.M. Bernstein, Ed., McGraw-Hill, 1977, p 43-47 27. S.K. Murarka, W.T.A. Dwars, and J.F. Langlois, Potential Formation of Corrosive Acidic Condensate and Its Monitoring in a Continuous Vapour Phase Bisulphite Chemi-Mechanical Pulping (BCMP) Process, in Pulp and Paper Industry Corrosion Problems, Vol 5, Canadian Pulp and Paper Association, 1986, p 187 28. B. Wallen, M. Liljas, and J. Olsson, Performance of a High Molybdenum Stainless Steel in Gas Cleaning Systems With Particular Reference to the Pulp and Paper Industry, Mater. Perform., Vol 21 (No. 6), 1982, p 40 29. J.B. Stirling, Modern Suction Roll Design and Construction, in CPPA Technical Section Proceedings, Canadian Pulp and Paper Association, 1964, p D74-D79 30. R.M. Vadas, The Suction Roll--Its Problems and Its Future, Pulp Paper Can., Vol 72 (No. 11), 1971, p 6670 31. J.P. McNamee, Influence of Suction Roll Covers on Shell Corrosion, in Pulp and Paper Industry Corrosion Problems, Vol 1, National Association of Corrosion Engineers, 1974, p 94-98 32. F.S. Wooster, M.D. Moskal, and J.P. McNamee, TAPPI J., Vol 62 (No. 9), 1979, p 71-74 33. R. Thompson, The Suction Roll--its Problems and Its Future. Part II, Pulp Paper Can., Vol 72 (No. 11), 1971, p 70-76 34. C.B. Dahl and E.P. Neubauer, Modern Practices Can Present Hazards to Stainless Steel Suction Rolls, Paper Trade J., Vol 155 (No. 34), 1971, p 46, 48, 50 35. C. Kelley, J. Vestola, V. Sailas, and R. Pelloux, Corrosion-Fatigue Crack Propagation of High-Strength Stainless Steels Used in Suction Rolls, TAPPI J., Vol 58 (No. 11), 1975, p 80-85 36. R.A. Yeske, "Prediction of Suction Roll Performance From Laboratory Testing," Paper 147, presented at Corrosion/86, National Association of Corrosion Engineers, 1986 37. D.F. Bowers and W.S. Butterfield, Corrosion Fatigue Data for Paper Machine Suction Roll Alloys--Its Significance, in Proceedings of the TAPPI Engineering Conference, Technical Association of the Pulp and Paper Industry, 1986 38. "Plate Bending Fatigue Testing in a Corrosive Environment," TIS 0402-08, Technical Association of the Pulp and Paper Industry, 1984

39. "Rotating Bar Bending Fatigue Testing in a Corrosive Environment," TIS 0402-09, Technical Association of the Pulp and Paper Industry, 1984 40. A. Garner, Suction Roll Failures in Canada, Pulp Paper Can., Vol 86 (No. 3), 1985, p 81-89 41. M.D. Moskal and G.W. Eggeman, The Effect of Residual Stresses on Corrosion Fatigue of Suction Rolls, in Pulp and Paper Industry Corrosion Problems, Vol 2, National Association of Corrosion Engineers, 1977, p 94-99 42. C.B. Dahl and C.W. Rainger, Stainless Steel Suction Roll Performance Design and Materials, in Pulp and Paper Industry Corrosion Problems, Vol 2, National Association of Corrosion Engineers, 1977, p 105-110 43. V. Sailas, Critical Components of the Paper Machine: Damage Caused by Metallic Corrosion, in Pulp and Paper Industry Corrosion Problems, Vol 2, National Association of Corrosion Engineers, 1977, p 100-104 44. "Guide for Evaluation of Paper Machine Suction Roll Shells," TIS 0402-10, Technical Association of the Pulp and Paper Industry, 1985 45. C.C. Nathan and A.J. Piluso, Wet End Corrosion Problems in Paper Mills, in Pulp and Paper Industry Corrosion Problems, Vol 2, National Association of Corrosion Engineers, 1977, p 126 46. W.S. Butterfield, private communication, 1986 47. "Suction Roll Shell Service Report," TIS 014-56, Technical Association of the Pulp and Paper Industry, 1982 48. Lockwood's Directory of the Paper and Allied Trades, Vance Publishing, 1985 49. R.G. MacDonald and J.N. Franklin, Ed., The Pulping of Wood, McGraw-Hill, 1969 50. W.D. Halsey, Paper Trade J., Vol 90, 1930, p 140-143 51. Inspection of Digesters, Monograph 12, Technical Association of the Pulp and Paper Industry, 1954 52. T.C. Johnson, TAPPI J., Vol 33 (No. 10), 1950, p 481-484 53. L. Ruus and L. Stockman, TAPPI J., Vol 38 (No. 3), 1995, p 156A-161A 54. W. Mueller, Corrosion, Vol 17 (No. 12), 1961, p 557 55. L. Stockman and L. Ruus, Svensk Papperstid., Vol 57 (No. 22), 1954, p 831-838 56. C.B. Christiansen and J.B. Lathrop, Pulp Paper Mag. Can., Vol 55, 1954, p 113 57. W. Mueller, Can. J. Technol., Vol 34, 1956, p 162-181 58. J. Hassler, TAPPI J., Vol 38 (No. 5), 1955, p 265-274 59. F. Flynn F. Richter, and F. Snyder, TAPPI J., Vol 36 (No. 10), 1953, p 433-444 60. C. von Essen, Svensk Papperstid., Vol 52, 1949, p 549 61. B. Roald, Norsk Skogind., Vol 11, 1957, p 446-450 62. E. Hopper, TAPPI J., Vol 36 (No. 8), 1953, p 345-352 63. L. Ruus and L. Stockman, Svensk Papperstid., Vol 56 (No. 22), 1953, p 857-865 64. N. Shoumatoff and H.O. Teeple, TAPPI J., Vol 38 (No. 4), 1955, p 202-207 65. R. Huseby and M. Scheil, TAPPI J., Vol 34 (No. 5), 1951, p 202-209 66. "Stainless Steel Weld Overlay in Sulfate and Soda Pulp Digester Vessels," TIS 405-5, Technical Association of the Pulp and Paper Industry, 1981 67. K. Smith, Pulp Paper, Vol 55 (No. 10), 1981, p 66-69 68. R. Yeske, Pulp Paper, Vol 57 (No. 10), 1983, p 75 69. D. Singbeil and A. Garner, in Proceedings of the Fifth International Symposium on Corrosion in the Pulp and Paper Industry, National Association of Corrosion Engineers, 1986, p 267-271 70. D. Singbeil and A. Garner, "Second Semi-Annual Progress Report on the Research Program to Investigate Cracking of Continuous Digesters," Pulp and Paper Research Institute of Canada, 1984 71. R. Yeske and C. Guzi, TAPPI J., Vol 69 (No. 5), 1986, p 104-108 72. C. Guzi, in Proceedings of the Fourth International Symposium on Corrosion in the Pulp and Paper Industry, National Association of Corrosion Engineers, 1983, p 37 73. I. Munro, in Proceedings of the 1985 TAPPI Engineering Conference, Technical Association of the Pulp

and Paper Industry, 1985, p 37 74. T. Lemke and D. Graver, TAPPI J., Vol 59 (No. 2), 1976, p 134-135 75. D. Jedlica and E. Montrone, in Proceedings of the 1977 TAPPI Engineering Conference, Technical Association of the Pulp and Paper Industry, 1977, p 327-329 76. R. Yeske, in Proceedings of the Fifth International Symposium on Corrosion in the Pulp and Paper Industry, National Association of Corrosion Engineers, 1986, p 219 77. R. Yeske, Paper 245, presented at Corrosion/86, National Association of Corrosion Engineers, 1986 78. D. Crowe, The Institute of Paper Chemistry, private communication 79. R. Yeske, Kraft Recovery Operations Seminar, Technical Association of the Pulp and Paper Industry, 1985, p 211-222. 80. G.A. Smook, Handbook for Pulp and Paper Technologists, Canadian Pulp and Paper Association, 1982, p 44-57 81. C.B. Thompson and A. Garner, The Metallurgy and Wear of Refiner Plate Alloys, in 1985 TAPPI Engineering Conference, Technical Association of the Pulp and Paper Industry, 1985, p 223-231 82. P. Clayton and D. Christensen, Metallurgical Aspects of Refiner Plates, in Pulp and Paper Industry Corrosion Problems, Vol 5, Canadian Pulp and Paper Association, 1986, p 175-179 83. W.G. Mihelich et al., Single-Stage Chip Refining--Some Major Operating Parameters and Their Effects on Pulp Quality, Pulp Paper Mag. Can., Vol 73 (No. 5), 1972, p 78-82 84. E.W. Nystrom and R.H. Okell, Sawdust Refining at Crofton, Pulp Paper Mag. Can., Vol 70 (No. 4), 1969, p 83-87 85. F.S. Giffin, Effect of TMP Primary Refiner Plate Rounding on Pulp Quality and Paper Machine Operation, in Proceedings of the 68th CPPA Technical Section Annual Meeting, Canadian Pulp and Paper Association, 1982, p B157-B162 86. D.J. Rideout, R.M. Hopkins, and J. Molgaard, Corrosion and Wear of Plates in the Production of Refiner Mechanical Pulp, Pulp Paper Can., Vol 83 (No. 5), 1982, p 54-58 87. C.B. Thompson and A. Garner, Wear Characteristics of Refiner Plates, in Proceedings of the 72nd CPPA Technical Section Annual Meeting, Canadian Pulp and Paper Association, 1986, p A117-A122 88. J.S. Jack, C. Mills, and P.H. Baas, A Device for Detection and Prevention of Plate Clash in Disc Refiners, Pulp Paper Can., Vol 82 (No. 9), 1981, p T311-314 89. P.O. Kettunen, Wear and Corrosion Resistant Materials for Refiner Blades, in Proceedings of the Fourth International Symposium of Corrosion in the Pulp and Paper Industry, National Association of Corrosion Engineers, 1983, p 98-102 90. R.N. Beaudry, Improving Refiner Plate Costs, in Proceedings of the 1984 TAPPI Pulping Conference, Technical Association of the Pulp and Paper Industry, 1981, p 275-279 91. W.C. Frazier, W.H. Arnold, T.C. Williams, and R.S. Charlton, Performance, Design and Metallurgy of Refiner Plates for TMP, Pulp Paper Can., Vol 82 (No. 3), 1981, p 43-55 92. S.K. Murarka, W.T.A. Dwars, and J.F. Langlois, Potential Formation of Corrosive Acidic Condensate and its Monitoring in a Continuous Vapour Phase Bisulfite Chemi-Mechanical Pulping (BCMP) Process, in Pulp and Paper Industry Corrosion Problems, Vol 5, Canadian Pulp and Paper Association, 1986, p 187 Selected References Paper Machine Corrosion • R.W. Barton, Thiosulfate Generated by Excess Application of Sodium Hydrosulfite, Pulp Paper, Vol 59 (No. 6), 1985, p 108 • D.C. Bennett and C.J. Federowicz, Prediction of Localized Corrosion of Stainless Steels in White Water, Mater. Perform., Vol 21 (No. 4), 1982, p 39 • B.M. Blakey and P.H. Thorpe, Failure and Redesign of Press Roll Ends in Paperboard Machines, APPITA, Vol 33, 1979, p 45

• D.F. Bowers, Changes in Water Properties and Corrosivity with Closure, TAPPI J., Vol 66 (No. 9), 1983, p 103 • W.S. Butterfield and P.E. Glogowski, Determination and Corrosivity of Thiosulfate Ion in Paper Mill Systems, in TAPPI Engineering Conference, Technical Association of the Pulp and Paper Industry, 1984, p 61 • E. Danielsson, Corrosion and Corrosion Control in the Paper Machine Wet End, in Pulp and Paper Industry Corrosion Problems, Vol 3, National Association of Corrosion Engineers, 1980, p 191 • J.L. Ewald and D.P. Hundley, Surface Finish Requirements of the Headbox, TAPPI J., Vol 63 (No. 11), 1980, p 121 • A. Garner, Thiosulfate Corrosion in Paper Machine White Water, Corrosion, Vol 41, 1985, p 587 • J.P. Gerhauser, Corrosion of Fourdrinier Wires, TAPPI J., Vol 43 (No. 4), 1960, p 207A • V.V. Gorelov and A.K. Talybly, Corrosion Protection of the Equipment in Closed and Reduced Water Recycling Systems, in Pulp and Paper Industry Corrosion Problems, Vol 4, Swedish Corrosion Institute, 1983, p 113 • M. Kurkela, N. Suutala and J. Kemppainen, On the Selection of Stainless Steels in Bleach Plants and White Water Systems, in Pulp and Paper Industry Corrosion Problems, Vol 5, Canadian Pulp and Paper Association, 1986, p 127 • W.A. Mueller and J.M. Muhonen, Pitting Corrosion of Stainless Steels in Six Paper Machine Headboxes: Mechanism and Prevention, TAPPI J., Vol 55 (No. 4), 1972, p 589 • J.M. Muhonen, Corrosion of Stainless Steels in Whitewater, in Pulp and Paper Industry Corrosion Problems, Vol 1, National Association of Corrosion Engineers, 1974, p 75 • C.C. Nathan and A.J. Piluso, Wet End Corrosion Problems in Paper Mills, in Pulp and Paper Industry Corrosion Problems, Vol 2, National Association of Corrosion Engineers, 1977, p 126 • R.C. Newman, Pitting of Stainless Alloys in Sulfate Solutions Containing Thiosulfate Ions. Corrosion, Vol 41, 1985, p 450 • A.J. Piluso and C.C. Nathan, Chemical Treatment to Control Corrosion in the Wet-End Operations of Pulp and Paper Mills, in Pulp and Paper Industry Corrosion Problems, Vol 1, National Association of Corrosion Engineers, 1974, p 14 • J.D. Rushton and P.A. Kelly, White Water Corrosion Data from an Integrated Newsprint Operation, in TAPPI Engineering Conference, Technical Association of the Pulp and Paper Industry, 1984, p 41 • M.K. Smith and D.A. Wensley, Effects of White Water System Closure on Paper Properties and Machine Corrosion, in 67th Annual Meeting, Technical Section, Canadian Pulp and Paper Association, 1981, p A159 • G. Sund and S. Strom, The Consequences of System Closure for Corrosion in Swedish Pulp and Paper Mills, in Pulp and Paper Industry Corrosion Problems, Vol 5, Canadian Pulp and Paper Association, 1986, p 51 • P.H. Thorpe, Corrosion in Paper Machines--An Overview, in Pulp and Paper Industry Corrosion Problems, Vol 3, National Association of Corrosion Engineers, 1980, p 184 • P.H. Thorpe, Microbiological Corrosion of Stainless Steel in Paper Machines and its Causes, in Pulp and Paper Industry Corrosion Problems, Vol 5, Canadian Pulp and Paper Association, 1986, p 169 Recovery Boiler Corrosion • P.E. Ahlers, Chloride in Kraft Mill Recovery Systems, in Pulp and Paper Industry Corrosion Problems, Vol 2, National Association of Corrosion Engineers, 1977, p 23 • R. Backman, M. Hupa, and P. Hyoty, Corrosion Related to Acidic Sulphates in Sulphate and Sodium Sulphite Recovery Boilers, TAPPI J., Vol 67 (No. 12), 1984, p 60 • R. Backman, M. Hupa, and E. Uppstu, Fouling and Corrosion Mechanisms in Recovery Boiler Superheater Area, in Pulp and Paper Industry Corrosion Problems, Vol 5, Canadian Pulp and Paper Association, 1986, p 243 • F. Bruno, Primary Air Register Corrosion in Kraft Recovery Boilers, in Pulp and Paper Industry Corrosion Problems, Vol 4, Swedish Corrosion Institute, 1983, p 68

• R. Cawein and C. Nin, Detection and Analysis of Generating Tube Thinning in Recovery Boilers, in TAPPI Engineering Conference Proceedings, Technical Association of the Pulp and Paper Industry, 1982, p 539 • D.G. Chakrapani and H. Czyzewski, Corrosion Fatigue in Steam Generating and Processing Systems of the Pulp & Paper Industry, in Pulp and Paper Industry Corrosion Problems, Vol 3, National Association of Corrosion Engineers, 1980, p 255 • J.A. Dickinson, M.E. Murphy, and W.C. Wolfe, Kraft Recovery Boiler Furnace Corrosion Protection, in Tappi Engineering Conference Proceedings, Technical Association of the Pulp and Paper Industry, 1981, p 607 • G. Hough, Ed., Chemical Recovery in the Alkaline Pulping Process, Technical Association of the Pulp and Paper Industry, 1985 • A. Jaakkola and T. Roos, Operational Experiences of Austenitic Superheater Tube Bends in Recovery Boilers, in Pulp and Paper Industry Corrosion Problems, Vol 4, Swedish Corrosion Institute, 1983, p 82 • K. Kuukkanen, S. Nikkanen, and P. Hyoty, "Operational Experiences in the Rauma-Repola High Sulfidity Recovery Boiler," Paper presented at the TAPPI International Sulfite Pulping Conference, Technical Association of the Pulp and Paper Industry, 1982 • M.K. Minty, An Emission Control System for Recovery Boilers and Materials to Handle the Corrosive Environment, in Pulp and Paper Industry Corrosion Problems, Vol 4, Swedish Corrosion Institute, 1983, p 91 • O. Moberg, Recovery Boiler Corrosion, in Pulp and Paper Industry Corrosion Problems, Vol 1, National Association of Corrosion Engineers, 1974, p 125 • C.R. Morin and J.E. Slater, Corrosion Problems in Power Generation Equipment, Individual Case Histories, in Pulp and Paper Industry Corrosion Problems, Vol 3, National Association of Corrosion Engineers, 1980, p 234 • K.W. Morris, A.L. Plumley, and W.R. Roczniak, The Effect of Chlorides on Recovery Unit Superheater Wastage: An R & D Progress Report, in Pulp and Paper Industry Corrosion Problems, Vol 3, National Association of Corrosion Engineers, 1980, p 47 • T. Odelstam, Performance of Composite Furnace Tubes in Recovery Boilers, in Pulp and Paper Industry Corrosion Problems, Vol 4, Swedish Corrosion Institute, 1983, p 64 • A.L. Plumley and W.R. Roczniak, Recovery Unit Waterwall Protection: a C-E Status Report, TAPPI J., Vol 58 (No. 9), 1975, p 118 • A.L. Plumley, W.R. Roczniak, and B.E. Lefebvre, "Recovery Unit Superheater Wastage and Control--Progress Report II," Paper presented at Black Liquor Recovery Boiler Symposium, Helsinki, Finland, 1982 • P. Pollak and R. Oesterholm, Corrosion Resistant, Safe and Economical Composite Tubes for Black Liquor Recovery Boilers, in Pulp and Paper Industry Corrosion Problems, Vol 2, National Association of Corrosion Engineers, 1977, p 147 • D.W. Reeve, D.C. Pryke, J.A. Lukes, D.A. Donovan, G. Valiquette, and E.M. Yemchuk, Chemical Recovery in the Closed Cycle Mill. Part 1: Superheater Corrosion, Pulp Paper Mag. Can., Vol 84 (No. 1), 1983, p 58 • D.W. Reeve, D.C. Pryke, and H.N. Tran, Corrosion in the Closed Cycle Mill, in Pulp and Paper Industry Corrosion Problems, Vol 4, Swedish Corrosion Institute, 1983, p 85 • W.R. Reeve, H.N. Tran, and D. Barham, Superheater Fireside Deposits and Corrosion in Kraft Recovery Boilers, TAPPI J., Vol 64 (No. 5), 1981, p 110 • W.B.A. Sharp, Composite Furnace Tubes for Recovery Boilers--a Problem Solved, TAPPI J., Vol 64 (No. 7), 1981, p 113 • O. Stelling and A. Vegeby, Corrosion on Tubes in Black Liquor Recovery Boilers, Pulp Paper Mag. Can., Vol 70 (No. 15), 1969, p T236 • R.G. Tallent and A.L. Plumley, Recent Research on External Corrosion of Waterwall Tubes in Kraft

Recovery Furnaces, TAPPI J., Vol 52 (No. 10), 1969, p 1955 • H.N. Tran, D.W. Reeve, and D. Barham, Formation of Kraft Recovery Boiler Superheater Fireside Deposits, Pulp Paper Mag. Can., Vol 84 (No. 1), 1983, p 36 • H.N. Tran, D. Barham, D.W. Reeve, P.H. Davis, and C.E. Guzi, Acidic Sulphate Corrosion in Kraft Recovery Boilers, in Pulp and Paper Industry Corrosion Problems, Vol 5, Canadian Pulp and Paper Association, 1986, p 201 • D.A. Wensley, Corrosion and Cracking of Composite Boiler Tubes, in 1986 Kraft Recovery Operations Seminar, Technical Association of the Pulp and Paper Industry, 1986, p 231

Corrosion in the Brewery Industry Edgar W. Dreyman, PCA Engineering, Inc.

Introduction BREWERIES are unique in many aspects in the food-processing industry. The product and its production, as well as cleaning procedures, storage, and bottling, all use great quantities of water. This requires a large amount of tankage and extensive hot water facilities; furthermore, the product is an acidic liquid that is aggressive to low-carbon steel. This is further complicated by the fact that the presence of iron ions in the product drastically affects its shelf life. Wet, damp, and high-humidity conditions all contribute to plant corrosion and premature equipment failure if not properly treated. These factors make the typical brewery a challenge to the corrosion engineer. Each brewery, whether old or modern, has a cross section of unique equipment ranging from standard power plant equipment, such as boilers, condensers, and oil-handling units, to highly sophisticated laboratory and pilot plant equipment for handling yeasts, malts, and hops essences. All types of materials are employed, from wood to glass, with many metals and special alloys used in large quantities. A brewery materials engineer must be familiar with the materials of construction for all of this equipment and must also be aware of diverse corrosion control techniques involving coatings, metals, cathodic protection, plastics, and even the use of inhibitors.

Corrosion Control Methods As in any other industry, there are always several ways to solve a given materials problem in the brewery. The task of the materials (corrosion) engineer is to sort out the available choices and to arrive at the most cost-effective solution to the problem. Traditionally, brewmasters, who at one time exercised complete control of all functions in a brewery, employed set conditions for plant equipment, such as wood with wax linings (tanks with slotted or perforated false bottoms used for filtering clear liquid from the grain mash) for fermentation and storage, copper for kettles and Lauter tubs, and lowcarbon steel for pasteurization. This was due in part to the training programs available in Germany and later in the United States. A series of brewing academies or schools provided most of the training for brewmasters and also acted as consultants to most breweries. Their influence dictated what type of equipment was considered suitable for brewing beer. This has changed dramatically in recent years, and many plants now have production superintendents in addition to or instead of brewmasters. With the consolidation of smaller plants into giant, multiplant brewing empires, the selection of equipment has become a staff engineering function. This group generally works out of the home offices and dictates and policy for all plants, because a major concern is that products made at any location taste alike. Some of the materials of construction used will be discussed below. Wood. Many smaller, older plants still use wood fermentors and storage tanks. These tanks must be internally lined with

an odorless wax or pitch that imparts no odor or taste to the product. Corrosion of steel hoops and rods with turnbuckles

had been a problem that required good coating systems, such as the wax or pitch used to line the barrel or food-grade enamels. The key to using wood was to keep it moist so that it remained swelled tight. In trying to achieve longer life of the linings, several breweries switched to epoxy linings. These were completely impervious to moisture; consequently, cracking of the lining occurred as the wood staves separated from each other. Some internally pitched, wooden beer barrels are still in use, but both of these areas have yielded to the use of metals. Wood was also susceptible to dry out, which imparted an unacceptable flavor to the product. Steel. Because iron affects the shelf life of beer, steel was not a major material of construction for tankage in contact

with beer until superior linings or coatings became available to the industry. Initially, pitches or wax-type coatings were employed. These coatings had to be applied hot, and this was difficult to accomplished in the cold cellars. Some of these coatings had poor abrasion resistance and would crack and chip; therefore, they gave way to the tasteless, pure, food and drug approved epoxies. This opened a new field for factory, epoxy-lined fermentation and storage tanks. Tank size was limited only by railroad rights-of-way; therefore, 3.7-m (12-ft) diam × 16.5-m (54-ft) long tanks became common in the industry. Concurrently, there was an increase in the use of glass-lined steel tanks and finally thin-wall stainless steel tanks. Because the cost of nickel has decreased, new construction is virtually all stainless steel tankage. Stainless Steel. Most of the brewery equipment currently being installed is fabricated from AISI type 304 stainless

steel and includes kettles, tanks, tubs, plate coolers, and even some pasteurizers. Stainless steel hot water tanks have been a problem; there has been some major stress-corrosion cracking of heavy-wall vessels holding up to 500 barrels. The problem has been one of chloride ion (Cl-) concentration due to evaporation under hot conditions and cracking of shaped plate sections at the bottom of the tanks. Remedial measures have consisted of treated water plus complete drainage and flushing. Stress-corrosion failures in stainless steel thin-wall piping have been common (Fig. 1). Stress-corrosion cracking has taken place where low-chloride insulation and concentrated at temperatures as low as 70 °C (160 °F). Remedial measures have consisted of replacing the piping, coating the exposed stainless steel with a chloride-free coating, and maintaining better valve maintenance.

Fig. 1 Stress-corrosion cracking of thin-wall stainless steel hot water piping

Stress-corrosion cracking has also been observed where heavy scaling has occurred because of hard water conditions, with chloride concentration in the hardness salt deposits. Remedial action has involved acid treating the water to remove carbonates. Another problem area has been contact knobs in plate-type coolers. Where deposits have collected, cracking has occurred because of vibration during operation. Remedial action has consisted of using special cleaning compounds to keep surfaces clean.

Special Alloys. Type 444 ferritic stainless steel has been successfully used where SCC problems have occurred because

of deposit formation. Monel has replaced type 316 stainless steel for beechwood-chip baskets in one proprietary application in which cracking had occurred around holes in the screening. Copper has been the traditional metal in breweries for centuries, but with the advent of new alkaline cleaners, some

corrosion problems have occurred. Large-diameter brewing kettles are joined by solder joints; these joints have been preferentially corroded and have required silver soldering repair. The use of in-place cleaning with jet sprays has resulted in metal thinning of the copper plates at points of impingement. Rotary cleaners appear to be superior to stationary jets. Kettle floor thinning has also been observed; cleaning compounds and mechanical abrasion from brushing have been responsible for this. Stainless steels are much more resistant to this type of action. On percolator refits using stainless steel units with stainless straps, cases of galvanic attack have been reported on bolting and strapping. Copper or brass fittings should not be used in contact with large areas of stainless steel. Beer is a very good conductor of electricity (low resistance of 200 /cm) due to high CO2 loading. Even hot water will promote galvanic effects. Thus, precipitation of scales on kettle surfaces can be protective, but when scales are removed with cleaners such as sulfamic acid, the surfaces can become very active. Experience has shown that type 304 or 316 stainless steel will perform well in this environment. Coatings. Breweries are large consumers of quality coatings, not only for tankage but also for structural steel, flooring,

and other working areas. The coatings used range from high-heat silicones for stacks to special superresistant grouts for floor pavers. With a special emphasis on sanitation, many coatings containing antibacterial and mold-control agents are employed throughout the industry. Bottling and canning operations have very severe conditions for most coatings. Broken bottles can spill product on the coating, and bottle washers use highly alkaline cleaners and label removers at high temperatures. High-build epoxies with inhibitive pigment epoxy primers perform well in many cases. The water-tolerant polyamide epoxies perform better in cellars and other damp areas. Special low-temperature curing epoxies have been most successful in cellars. Considerable evaluation has been done on water-base epoxies because odors can be readily picked up by the beer or other ingredients, such as hops or malts. Because quick maintenance repairs are critical in large, high-volume breweries, many products that are user friendly must be employed (for example, products that cure quickly on damp surfaces with low odor). Again, with the large amount of water vapor present in many areas, the use of polyurethane topcoats over epoxies is finding greater acceptance. The polyurethanes do not chalk, and they hold their gloss with good color retention. This is especially true of bright greens, blues, or reds. Floor Materials. Considerable glazed tile is used in breweries, and special epoxies with good adhesion to very smooth

surfaces have been employed to coal glazed ceramic tile in order to prevent crazing (cracking). Epoxy grouts (American Tile Council Specifications) have solved the problem of grout deterioration around kettles, syrup tanks, and other areas susceptible to acidic and microbiological attack. Finally, the problems associated with floors have been addressed by many coating producers with mixed results. Beer, as well as corn syrup, will attack concrete; therefore, many areas have tile pavers installed. However, as previously mentioned, proper grouts must be used. Where the concrete must be coated, epoxies with embedded grit (skidproofing) are being used in many areas. Chlorinated rubber has been used on cellar floors (Fig. 2), but it generally requires more maintenance because it does not have the abrasion resistance of the epoxies.

Fig. 2 Chlorinated rubber floor coating in a cellar

Proper surface preparation is most important in brewery floor work. On new floors, acid etching or a light sandblast will suffice. On old floors, removal of a minimum of 50 mm (2 in.) of concrete is mandatory; more extensive concrete removal is sometimes necessary, depending on the degree of contamination. Bacterial contamination deep in the pores of the concrete is a common occurrence. If floors are not properly sealed, corrosion of concrete rebars and structural steel can result, with eventual cracking and spalling of the concrete (Fig. 3).

Fig. 3 Cracking and spalling of a concrete floor

Glass Linings. At one point in the evolution of the brewing industry, glass-lined tanks were introduced for fermentors,

storage, and filling tanks. Glass is an excellent materials for preserving the purity of the product and is resistant to the mild acids and alkaline cleaners used to remove scales and deposits. Glass linings are applied by spraying a borosilicate glass frit on the sandblasted surface of the tank, and then firing the entire tank in a furnace until a smooth glass surface is achieved. Glass linings were installed at thicknesses of 0.4 to 0.8 mm (15 to 30 mils) and gave good service, but were prone to maintenance problems. For example, the lining could be fractured by workmen dropping tools or hose fittings on the glass. Initially, factory pinhole repairs were made by gold filling, similar to a dental inlay. This was followed by repairs with sprayed tin; even large repairs, such as the knuckle areas on tank heads, were repaired in this manner. Tank perforations were repaired with tantalum plugs. As epoxies became available, glass repairs were performed with these materials. In tanks with multiple pinholing, the glass lining was removed by gritblasting, and the tanks were recoated with epoxies. Cathodic protection was also employed in some cases; this will be discussed later in this article.

Because of the large capital investment required to produce glass linings and the major reduction in the number of breweries, glass linings are no longer produced commercially. Despite this, many tanks are still in use and require maintenance. Plastics. From a corrosion control point of view, plastic materials are very useful; therefore, they have found application

in breweries. Water treatment tanks, acid storage, roofing, and gutters are applications for plastics that are common to most industrial activity and as such are used by breweries. Fiberglass and polyvinyl chloride are among the plastics that have been employed. Small polypropylene tanks for yeast culture and other specialty service have some record of use. Inhibitors. Where large volumes of water are being used, constant attention must be paid to reducing the water demand,

not only because water is becoming expensive and scarce but also because wastewater must be handled in an ecologically sound manner. Brewery waste is high in bacteria and, if untreated, has a very high biological oxygen demand. Consequently, many closed cooling systems are being employed. This necessitates cooling water treatment, and even before the current environmental regulations, breweries could not use additives containing chromates. This means that zinc/phosphate-type inhibitors and proprietary, pure, food and drug approved systems must be employed. Even boiler water treatment must be monitored because live steam is used for heating water that goes into the product and would carry chemicals over into the product. Beer production is sensitive to most chemicals in terms of the organoleptic properties of the product and the fragility of yeast cultures. Cathodic protection is an electrochemical method of corrosion control that impresses a direct current onto the structure to be protected. This in turn overrides the many small dc corrosion cells existing on that structure and stops corrosion (see the article “Cathodic Protection” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook).

This method of corrosion control can overcome the effects of galvanic cells due to dissimilar metals and concentration cells due to differential oxygen concentration, temperatures, or stresses. Combined with good coatings systems, cathodic protection is a most cost-effective way of controlling the corrosion of immersed and underground structures. Breweries employ cathodic protection on waste storage tanks, fermentor tanks, beer storage tanks, hot water tanks, and underground pipelines for oil, sewerage, and, in some cases, product lines. Water treatment facilities also employ this type of corrosion control.

Equipment Problems Air Conditioning Equipment. Humidity control in cellars will reduce paint deterioration and mold growth and will

promote working safety with relatively dry floors. Lithium chloride units used for moisture control have had some deterioration problems that have necessitated the use of exotic alloys and superior coating systems. In dusty areas, such as malt plants, the poultice corrosion of aluminum fins on cooling units has been a problem (Fig. 4). However, as with many maintenance problems, proper cleaning on a scheduled basis can prevent the failure of such components.

Fig. 4 Corrosion on air conditioner cooling fins

Cooling systems require water treatment, and water tower pans should be properly coated with such products as coal tar epoxies. On occasion, sacrificial zinc anodes have been used for cathodic protection of bare areas in coatings to prevent corrosion. As in all industrial situations, cooling towers must not be placed where they become scrubbers for stack gases. Absorption refrigeration systems use sacrificial anodes in waterboxes and require periodic eddy current inspection of the tubes to minimize corrosion failures. Compressors. Many older breweries still use CO2 and ammonia (NH3) compressors, which may have intercoolers. Closed systems, if used, require water treatment and the use of sacrificial zinc anodes in cooler heads.

The brine systems associated with older breweries are gradually being phased out, but at one time, they were a major corrosion problem requiring proper inhibition. As mentioned, because chromate inhibitors are not permitted, mixed phosphate inhibitors are used for corrosion control. Pasteurizers. A major part of a brewery is the bottling and canning operation. One of the most expensive pieces of

equipment in the bottleshop is the pasteurizer. The old units were basket pasteurizers that dipped baskets of bottles into water at various temperatures to kill any residual yeast in the product. If this is not done, fermentation can occur in the bottle or can, and gushing takes place when the container is opened. Modern pasteurizers are continuous-belt operations in which water at different temperatures is sprayed over the bottles or cans. This highly oxygenated hot water is quite corrosive, and the use of inhibitors is limited by the potential for staining of the package. The normal pasteurizer is of stainless steel construction in the spray areas, with low-carbon steel tanks under each temperature zone. It is in these tanks that problems can be caused by broken beer bottles, which introduce trash and debris onto the tank floors. The lower-temperature zones of a pasteurizer encourage bacterial growth, and oxygen cells can form under the biological deposits. This corrosion is further enhanced by copper heating coils in the low-resistivity water and by concentration cells produced by hot wall effects in the various compartments. Coatings have not solved this problem, because of mechanical damage from broken glass and the effects of high temperatures on the coating. Cathodic protection has performed well. Originally, magnesium galvanic anodes were used, but the consumption rate was high; replacement was required every 2 years. High-purity zinc anodes have given 10-year service lives, but extreme care must be exercised so that water chemistry does not cause reversal problems where the steel becomes anodic to the zinc. This can be a problem in some very soft waters. Ultimately, impressed-current systems have proved to be the most efficient corrosion control method for pasteurizers. The anode materials used include resin-impregnated carbon, high-silicon cast iron, or platinized niobium or titanium.

Current densities of about 160 to 270 mA/m2 (15 to 25 m/A/ft2) of bare, immersed steel or copper are used. Anodes are mounted on epoxy-coated stand-off holders that hold the anode 150 to 300 mm (6 to 12 in.) from the floor, walls, or heating coils. A dc rectifier provides direct current through four circuits, each of which is resistorized to reflect the variance in resistivity due to water temperature. Normally, this requires anodes operating at water temperatures of 10 °C (50 °F), at 20 to 40 °C (70 to 100 °F) for preheat, at 50 to 60 °C (120 to 140 °F) for heat, at 75 °C (165 °F) (the pasteurizing temperature), and at 25 °C (80 °F) for cooling and final rinse. Anode lives are calculated from the following rates of consumption: • • •

Carbon: 0.45 kg/A·yr (1 lb/A·yr) consumption High-silicon cast iron: 0.34 kg/A·yr (0.75 lb/A·yr) consumption Platinum: 0.006 g/A·yr (2 × 10-4 oz/A·yr consumption

The platinum anodes can be much smaller because the allowable current density is much higher: • • •

Carbon: 10.8 A/m2 (1 A/ft2) anode surface High-silicon cast iron: 27 A/m2 (2.5 A/ft2) anode surface Platinum: 540 A/m2 (50 A/ft2) anode surface

These values have been compensated for use in freshwaters. In hard waters (high in carbonates), some scaling will occur on the cathodic surfaces, and overprotection in discouraged because it will cause stripping of coatings near anodes. Polarized potentials should not exceed -1.3 V versus a copper/copper sulfate reference electrode if epoxies are being used. Alkyd or zinc-rich coatings are not recommended. Although air-cooled rectifiers perform well as a power source, oilcooled, sealed units are recommended because of the extensive use of water during cleanup. Finally, some plants have pasteurizers dedicated to can pasteurization only. When tin-plated steel cans were used, the seam solder would fall to the tank floors, and galvanic pitting of the floor would take place. Most plants now use deepdrawn aluminum cans with rolled tops; therefore, this is no longer a problem. Beer Barrels. Barrels are generally no thought of as equipment, but they represent a large brewery investment and have

been a source of corrosion problems in the past. Except for the steel hoops, wooden barrels are not a problem; however, the lack of suitable wood, the few available coopers, and the heavy weight of wood has led to the use of aluminum and stainless steel barrels. Initially, aluminum appeared to be the best material for this purpose, but shortly after it was put into use, pitting appeared in the barrels. This ultimately led to perforation, and the rough, pitted surface became hard to clean. The cause of this corrosion was found to be the tin-plated, brass tap that was commonly used. This brass tube in close proximity to the aluminum became the cathode in beer and resulted in pitting of the aluminum. This required pitching the interior of the barrels, thus negating the cost various for aluminum. Currently, the preferred material of construction for barrels is stainless steel. Tanks. In addition to the use of coatings in tanks, some cladding with stainless steel has been employed. Lauter Tubs. The metal cladding of Lauter tubs, especially the bottom of the tub where a perforated or slotted stainless

or brass false bottom rests, has been successful. The stand-off legs of the drain plates would perforate the floor coating and cause galvanic cell corrosion action at these areas. Organic coatings were tried first, but the impact was too severe. Therefore, stainless pads or complete cladding with thin sheets of type 304 stainless steel was employed to correct the problem. Concrete Tanks. Several breweries conceived the idea of constructing buildings in a honeycomb fashion, with concrete

tanks being a part of the building support. This practice was employed in Europe and the United States.

Originally, these tanks were coated with mix, which would crack. Beer would be lost in the walls of the building, with the result being odor, corrosion of concrete rebars, and microbial contamination. Attempts were made to correct this problem with fiberglass (cloth) reinforced epoxy. This was very expensive, and failures in the lining occurred unless substantial layers of contaminated concrete were removed. A film in the United Kingdom worked out procedures to apply thin-sheet type 304 stainless steel to concrete without expansion buckling, and this technique was used successfully in Scotland and Holland. Similar techniques were then used in the United States with acceptable performance. The main problem in using thin type 304 stainless sheet lies in preventing movement, which can cause wrinkling and, ultimately, cracking. Proper welding techniques and support spacing are necessary to achieve a clad tank without movement of the cladding. Another problem area was the tank doors, which today are primarily stainless steel. Coated low-carbon steel and glasscoated doors chipped readily at the edges, which are in contact with the beer; heavy pitting in these areas resulted. The solid stainless door currently provides satisfactory service. The fermentation tank environment is more severe than that in storage tanks. Fermentation tanks fabricated of stainless steel provide good service in most plants; however, many low-carbon steel tanks still exist in working breweries and present maintenance problems. Fermentation tanks have copper heating and cooling coils inside the tank to control the rate of fermentation. The area ratio of bare steel (breaks in the coating or glass) to bare copper is very unfavorable and results in very rapid attack of the steel in the acidic CO2-saturated product.

Repairs made to fermentation tanks using epoxies have performed well, but care must be exercised in grinding out the pits to ensure absolute cleanliness. The resin must not only conform to pure food regulations but must also exhibit low shrinkage characteristics so that it does not pull away from the edge of the pit. The resin must not be feathered over the edge onto the good coating or glass surface; if this occurs, the thin coating may fail, and the product will get into the edges of the pit with the resin coming out as a plug.

Cathodic Protection Another method used to prevent corrosion in fermentation tanks is the use of cathodic protection to monitor the size of pinholes in coatings and linings. Galvanic anodes would introduce metal ions into the product; therefore, they cannot be used. Consequently, impressed-current systems must be employed, with anode selection becoming most important. High-silicon cast iron anodes cannot be used, because this would introduce iron ions into the beer. Platinum would appear to be the best choice because it is the most inert. Unfortunately, oxygen is liberated at the anodic surface and this would cause deterioration of the product. Carbon, therefore, is the most logical choice for anodes because the oxygen produced combines with the carbon and produces CO2, which occurs naturally in the beer. A second caution is that oil-impregnated carbon must not be used, because any uncarbonized oil would contaminate the product and inhibit foam formation, which is critical for a good head on the beer. Resin-impregnated carbon anodes 75 mm (3 in.) thick and 200 mm (8 in.) in diameter have been used in fermentors. These are mounted on stand-off insulators installed in the tank floor and penetrating the tank bottom so all wiring is on the outside of the tank. This hardware has a porcelain insulator with stainless fittings. These stand-off insulators isolate each anode and provide better throwing power of the current; therefore, fewer anodes are required. Current demand for such a tank is quite low, using 21.5 to 54 mA/m2 (2 to 5 mA/ft2) of bare area (including bare coil area). Typical current for an 1800-barrel tank is 0.3 to 0.5 A at 3.5 to 5 Vdc. In order to maintain a balanced direct current load to each tank, the doors are fitted with waterproof microswitches that transfer the load to a dummy resistor when the tank is open. This prevents any current flow to the tank when workers are inside. Generally, an entire cellar is equipped with anodes from one power supply with 6 to 12 circuits. Power supplies are air cooled and are installed outside of wet cellars. The bare steel must not be overprotected with excessive current, because coating stripping can occur. This is especially critical when phenolic coatings are used that are sensitive to alkalies. The result of the cathodic reaction is to increase the pH at the tank metal surface, which can soften phenolic resins.

In addition, hydrogen is liberated at the cathode and can be absorbed into the steel. Excessive hydrogen can cause lining failure in glass-lined tanks; therefore, excess current must be avoided. Potentials of less than -1.00 V versus a silver/silver chloride reference electrode have provided steel protection without coating or lining breakdown. Hot Water Tanks. Many breweries using low-carbon steel hot water tanks that are bare, coated, or lined with gunited

(sprayed concrete) linings have employed cathodic protection. Anodes have been resin-impregnated carbon; 75-mm (3in.) × 1.5-m (60-in.) hanging anodes have performed well, with service lives of 7 to 10 years. These systems have been most effective where flat copper heating coils are used. One system has been in service for 30 years in a tank that was originally scheduled for replacement. It must be remembered that only the submerged parts of a tank are protected; therefore, water tanks must be coated on the top, or dome, of the tank and down the walls to the waterline. In addition, complete coating should be considered, because this will substantially reduce the size of any cathodic protection system to be employed. The coating used must be compatible with impressed current. Inert coatings such as epoxies, chlorinated rubbers, or vinyls will function well. Alkyds and inorganic zinc-rich coatings or primers are not recommended. Cold Water Tanks. The same techniques employed to protect hot water tanks also are used for cold water storage and

fire protection tanks. Some fire protection storage tanks are used for dual service; these may contain 3.8 × 106 L (1 × 106 gal) of water, with 950,000 to 1.9 × 106 L (250,000 to 500,000 gal) of this being used for process water. This means that the water throughput in the upper third of the tank may be substantial. This can be highly oxygenated, aggressive, agitated water; consequently, current demand for a cathodic protection system may be very high in that area (323 to 540 mA/m2, or 30 to 50 mA/ft2) for proper corrosion control. These design factors must be considered for such systems.

Plant Structures Buildings. Although most structures are brick and concrete, which require good coatings and sealers (silicone),

corrugated siding also is employed. Tedlar (polyvinyl fluoride) coated siding has shown good service for the costs involved. Care should be exercised to employ proper fasteners as specified by the siding vendor. Galvanized and aluminized siding and roofing panels are commonly employed and should be specified with sufficient zinc or aluminum thickness to ensure reasonable life. If metal panels are to be recoated, special pretreatment is generally required, and coating suppliers should be consulted for the proper materials to use as primers with their top coats. Rebar Protection. The use of epoxy-coated concrete rebars for buildings and slab construction where water intrusion may be a problem is highly recommended to prevent concrete spalling. Where deicing salt contamination is a problem, as in heated driveways or parking garages, the small additional cost is well worthwhile.

Where concrete delamination is occurring in existing structures, cathodic protection of rebars has been successfully applied. Carbon strands or platinum wire anodes or conductive polymer anodes can be successfully used for rebar protection. Anodes for this purpose must be used with conductive coatings to distribute a low dc current density over the salt-impregnated areas. Buried Tanks and Piping. Buried steel fuel and product storage tanks, such as solvent tanks for can coatings, can be

readily protected with a combination of coatings, isolation from electrical contact with other structures, and cathodic protection. The Steel Tank Institute has specifications and designs for applying protection during tank manufacturing, with 30-year warranties against corrosion. Steel fuel storage tanks for large truck fleets, when these are a part of the brewery operation, can be readily protected. Fiber-glass tanks are an alternative choice, but require more stringent installation procedures than steel. Underground fire protection lines, gas lines, water lines, and waste lines are all readily protected with coatings and cathodic protection. If cast iron bell-and-spigot pipe (Fig. 5) is being employed, it must be bonded across the joints. Thermit bonds are preferred to brass wedges, which can loosen and fall out.

Fig. 5 Corrosion failure of an underground cast iron pipe

An alternative to cathodic protection is the use of polyethylene, plastic sleeve encasement. This technique was developed by the Cast Iron Pipe Research Institute, and it has performed well in the United Kingdom as well as in the United States. Miscellaneous Structures. Depending on the brewery location, the type of subsoils, and the proximity to rivers and lakes, other structures, such as steel bearing piles or sheet piling, must be protected. When cosmetic requirements are not a consideration, such coatings as black coal tar epoxies or polyurethanes can be used in conjunction with cathodic protection.

Water and waste treatment plants require protection for the steel structures, such as settling basins, flocculators, and clarifiers, and for the concrete. Superior concrete protection must be employed if soils are high in sulfates, with coal tar epoxies giving good performance. The selection of corrosion control materials and systems must be predicated upon the expected life of the structure, the expected difficulty of future maintenance, and the overall cost to the owner.

Selected References •

• • • • • •

W.D. Rigg, "Accelerated Galvanic Corrosion in Beer Tanks in the Presence of Dissimilar Metals," Paper presented at the MBAA 47th Annual Convention, New York, Master Brewers Association of America, 1954 H.B. Dwight, Calculation of Resistance to Ground, Electr. Eng., Dec 1936, p 1319-1329 Cathodic Protection Frequently Practical in Corrosion of Brewery Vessels, Corrosion, Vol 13, Jan 1957, p 134 J.D. Redmond, Solving Brewery Stress Corrosion Cracking Problems, MBAA Tech. Quart., Vol 21, Nov 1984 Solving Design Problems for Cathodic Protection of Glass-Lined Domestic Water Heaters, Corrosion, Vol 16, Sept 1960, p 9-17 E.W. Dreyman, Some Successful Applications of Cathodic Protection in Breweries, Mater. Prot., Feb 1962, p 58-62 C.W. Ambler and W.E. Allen, "Zinc Galvanic Anodes as a Supplement to Coatings in a Brewery Pasteurizer," Paper presented at the Northeast Regional Conference, National Association of Corrosion Engineers, Oct 1958

Corrosion Industry

in

the

Pharmaceutical

Ralph J. Valentine, VAL-CORR

Introduction THE PREVENTION AND MITIGATION of corrosion in the pharmaceutical industry presents a demanding challenge to materials engineers. In most cases, they are working with processes that require equipment and piping systems to be fabricated from material having extremely low corrosion rates when exposed to a wide variety of corrosive media and operating conditions. The substances produced by corrosion reactions contaminate the product being manufactured. This contamination must be removed in one of the subsequent process steps so that the product can pass the quality control tests required for compliance with the stringent purity and quality demands established by the applicable government regulatory agencies.

Materials of Construction The materials of construction found in pharmaceutical production facilities include: • • • •

Metals and alloys, both solid and clad Thermosetting plastics and thermoplastics, both solid and as linings Ceramics, both solid and as linings Impregnated carbon

Stainless Steels The austenitic stainless steels, especially the low-carbon and stabilized grades, have been the workhorse alloys in the pharmaceutical industry for many years. These alloys exhibit good corrosion resistance in many media, are readily fabricated, have excellent strength over a wide temperature range, offer good availability, and are relatively inexpensive. The surface condition of the austenitic stainless steels is critical where the pharmaceutical product must not be contaminated and where the stainless steel is required to resist an aggressive environment. The highly protective chromium oxide film that gives stainless steel its corrosion resistance is tenacious, durable, and self-healing in the presence of oxygen; however, this film can be damaged during equipment fabrication and postfabrication cleanup practices. Fortunately, serious problems can be minimized by following good procurement, handling, design, fabrication, and cleanup practices.

The austenitic stainless steels are widely used in oxidizing environments, high-purity water service, and in fine chemical and pharmaceutical production equipment and piping. They are not suitable for use in chloride-containing environments, particularly at high chloride concentrations and at high temperatures. In recent years, stainless steel producers have developed the so-called super austenitic stainless steels having excellent resistance to general corrosion and pitting/crevice attack in chloride-containing environments. These alloys are highly resistant to intergranular corrosion and stress-corrosion cracking (SCC); this makes them useful in oxidizing chloride solutions, oxidizing acids, and brines. The high molybdenum (2.5 to 6.5%) content and increased chromium and nitrogen give the super austenitics good resistance to pitting and crevice corrosion. The relatively high nickel content (18 to 31%) and the high chromium and molybdenum levels give the alloys excellent SCC resistance. The presence of copper in the alloys improves resistance to sulfuric (H2SO4), phosphoric (H3PO4), and acetic acids. The super stainless alloys are less costly than the nickel-base alloys and are readily available in a wide range of product forms, such as pipe and tubing, sheet, plate, and forgings, as well as a full range of welding consumables. In addition, the super austenitic stainless steels are more workable and have better weldability than high-alloy ferritic steels. When the

super austenitics are specified, the welding specifications of alloy producers should be followed explicitly so that the full chemical and cracking resistance of the alloy is maintained. The duplex stainless steels are, as the name implies, duplex in structure. At room temperature, their equilibrium structure is a mixture of austenite and ferrite phases. The compositions of these alloys are carefully controlled to maintain the proper balance of austenite to ferrite. Most of the physical properties of the duplex stainless steels are between those of the austenitic and ferritic stainless alloys. The thermal conductivity of the duplex stainless alloys is less than half that of carbon steel, but about 25% higher than that of the austenitic stainless steels.

The coefficient of expansion of carbon steel is similar to that of the stainless duplex alloys and is about 40% less than that of the austenitic stainless alloys. The duplexes have excellent toughness as well as high strength. Compared with the ferritic stainless steels, the ductile-to-brittle transition of the duplex alloys is more gradual and occurs at a lower temperature, thus allowing the production of a wide range of product forms. The duplex alloys are not suitable for cryogenic service the austenitic grades are preferred in this application. The duplex grades have good resistance of chloride SCC; however, the various alloys can show variability in pitting and crevice corrosion susceptibility because of the segregation of the ferrite and austenite phases. The high chromium content of the duplex stainless steels makes them strongly resistant to oxidation. However, prolonged exposure at elevated temperatures (above 345 to 370 °C, or 650 to 700 °F) can affect toughness and corrosion resistance in aqueous media and should be avoided. All commonly used duplex stainless steels are included in a variety of specifications for sheet, strip, plate, and seamless and welded tubing and pipe issued by the American Society for Testing and Materials (ASTM). Many duplex stainless steels are also included in Section VIII, Division 1, of the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code, in which they are identified by their UNS designations. The best corrosion resistance and mechanical properties in welded duplex stainless steels are achieved when the welding practice encourages the formation of a 50-50 austenite-ferrite phase in both the weld metal and the heat-affected zone (HAZ). Welding practices are required that emphasize cleanliness, the avoidance of carbon contamination, and the use of dry inert gas shielding. Because of the sensitivity of ferrite to hydrogen embrittlement, shielding gases containing hydrogen should not be used. Readily available duplex alloys include alloy 2205 (UNS S31803), 44LN (UNS S31200), and Ferralium 255 (S32550). High-Purity Ferritic Stainless Steels. The super ferritic stainless steels were introduced in the United States during

the past decade. The oldest and best known is E-Brite 26-1, an alloy containing 26% Cr and 1% Mo. E-Brite was electron beam refined to reduce carbon and nitrogen to very low levels. This process is no longer in use because of manufacturing problems; E-Brite is now made by vacuum melting. Other super ferritic alloys with approximately the same chromium and molybdenum contents with stabilizing additions of titanium have been introduced to the industry. The 26-1 alloy generally has corrosion resistance equal to or better than that of AISI types 304 and 316 stainless steel, the workhorse alloys of the pharmaceutical industry. In addition, the 26-1 alloy is resistant to SCC, a major shortcoming of the austenitic stainless steels. Because of the resistance of the high-purity ferritic alloys to a wide variety of aggressive environments, many compositions have been developed, such as 18Cr-2Mo, 29Cr-4Mo, 29Cr-4Mo-2Ni, and 27Cr-3.5Mo-2Ni. As-welded super ferritic stainless steels typically have poor weld zone ductility and are notch sensitive in the HAZ. If these alloys are heated to between 400 and 480 °C (750 and 900 °F) for prolonged times or are slowly cooled within this temperature range, notch toughness is further reduced, and the material becomes brittle. The use of the super ferritics at low temperatures is limited because of their high ductile-to-brittle transition temperatures. If the corrosion resistance of the super ferritic alloys is to be maintained, extreme care must be taken to avoid contamination with nitrogen or carbon during welding. Special welding procedures developed by the alloy producer should be used. Because of problems with welding and controlling the ductile-to-brittle transition temperatures, these alloys are generally used in sheet thicknesses under 3.2 mm ( in.) and for tubing. More information on the corrosion of all types of stainless steels is available in the article "Corrosion of Stainless Steels" in this Volume. Nickel and Nickel-Base Alloys Commercially pure nickel (Nickel 200) is highly resistant to many corrosive media. It is most useful in reducing

environments, and it can be used under oxidizing conditions that cause the development of a passive oxide film. Nickel has poor corrosion resistance in H2SO4, HCl, HNO3, and H3PO4. Pure nickel has outstanding resistance to alkalies, the

exception being ammonium hydroxide (NH4OH), which rapidly corrodes nickel. Oxidizing acid chlorides such as ferric, cupric and mercuric are very corrosive. Nickel is used for containing very reactive chlorides, such as phosphorus oxychloride, phosphorus trichloride, nitrosyl chloride, benzyl chloride, and benzoyl chloride. Pure nickel resists anhydrous chlorine, anhydrous hydrogen chloride, phenol, and bromine. Nickel-Copper Alloys. Monel alloy 400 is more resistant than nickel to corrosion under reducing conditions and more

resistant than copper to corrosion under oxidizing conditions. As a solid-solution alloy, Monel 400 is free from the corrosion that can result from local galvanic action between the phases of multiphase alloys. Monel is generally resistant to SCC; exceptions are mercury and solutions of its salts, fluorosilicates, concentrated caustic soda (NaOH), and potassium hydroxide (KOH). Monel 400 is resistant to all common dry gases at room temperatures. it is not resistant to chlorine, bromine, nitric oxides, ammonia, sulfur dioxide, and hydrogen sulfide in the presence of moisture. Monel is useful in handling H2SO4 under air-free reducing conditions. Aeration causes a sharp increase in the corrosion rate. Monel 400 can handle aerated HCl at 10% concentration at room temperature; above room temperature, applications are usually limited to 3 to 4% HCl. In unaerated hydrofluoric acid (HF), Monel 400 is resistant to all concentrations up to the boiling point. Monel 400 has very poor corrosion resistance in HNO3. Monel 400 has good resistance to HaOH up to about 50% concentration and to NH4OH up to 3% concentration. Nickel-Chromium Alloys. The high nickel content of these materials gives them considerable resistance to corrosion

under reducing conditions and in strong alkaline environments. Also, because of the high nickel content, the alloys are virtually immune to chloride SCC. However, at high temperature and in contact with concentrated alkalies, they are subject to SCC. Mercury will also cause SCC at elevated temperatures. Nickel-molybdenum alloys were developed to be resistant to HCl at all temperatures and concentrations. The nickel-

molybdenum alloys have good corrosion resistance to other nonoxidizing environments, including boiling 60% H2SO4, pure H3PO4 at most concentrations and temperatures, wet hydrogen chloride gas hydrogen chloride to 455 °C (850 °F), and wet halogenated organics. The presence of ferric or cupric salts or other oxidizing agents will cause rapid corrosion of these alloys. Nickel-Chromium-Molybdenum Alloys. The addition of chromium to the nickel-molybdenum alloys increases the resistance to oxidizing environments, giving them good resistance to HNO3 and H3PO4, as well as to most chloride salts. The alloys are very resistant to pitting and crevice corrosion because of the molybdenum content. The nickel-chromiummolybdenum alloys are the most versatile corrosion-resistant alloys available. They are some of the few alloys that are resistant to wet chlorine gas, hypochlorite, and chlorine dioxide solutions. They have good resistance to ferric and cupric chlorides and other oxidizing salts, alkalies, and acids. Nickel - chromium - molybdenum - copper alloys were developed to resist H2SO4 and HNO3 over a wide range

of concentrations and temperatures. They are resistant to H3PO4 even when the acid contains fluorides or oxidizing compounds. The article "Corrosion of Nickel-Base Alloys" in this Volume contains more information on corrosion of nickel and nickel-base alloys. Titanium Titanium forms a tight, adherent oxide film that makes it resistant to many oxidizing reagents, including HNO3 and chromic acid (H2CrO4). Titanium is attacked by reducing acids such as H2SO4 and H3PO4. It is useful in moist chlorine gas and hypochlorite at ambient and elevated temperatures; most organic acids; and water, seawater, and brine solutions at temperatures to the boiling point. At elevated temperatures, titanium is subject to pitting and crevice corrosion in a chloride environment. Titanium will not resist red fuming HNO3 or HF in any concentration or temperature and will ignite at very low temperatures in dry chlorine gas. More information on the corrosion of titanium and its alloys is available in the article "Corrosion of Titanium and Titanium Alloys" in this Volume. Zirconium The two zirconium-base alloys used most frequently in the pharmaceutical and chemical-processing industries contain some hafnium, which is metallurgically and chemically similar to zirconium and does not reduce its corrosion resistance. The alloy UNS R60702 contains a minimum of 99.2% Zr + Hf, with a maximum hafnium content of 4.5%. The second alloy, UNS R60705, contains a minimum of 95.5% Zr + HF, with a maximum of 4.5% HF and 2 to 3% Nb. Both are approved for use in the construction of pressure vessels according to the ASME Boiler and Pressure Vessel Code, Section VIII.

Zirconium has excellent resistance to HCl at all concentrations up to temperatures of 120 °C (250 °F). It resists HNO3 in all concentrations up to 90% and temperatures to 150 °C (300 °F); however, SCC may occur above 70% concentration if high tensile stresses are present. Zirconium is corrosion resistant in H3PO4 in concentrations up to 55% at temperatures of 175 °C (350 °F). Above 55% concentration, the corrosion rate increases with temperature, but even in 85% H3PO4 at 60 °C (140 °F) the corrosion rate is still less than 0.13 mm/yr (5 mils/yr). If there are fluoride ion impurities present at any concentration of H3PO4, zirconium may be subject to rapid corrosion attack. Zirconium alloys are resistant to H2SO4 up to 75% concentration and at temperatures to boiling. Ferric, cupric, and nitrate ion impurities cause corrosion of zirconium in H2SO4 concentrations above 65%. Fluoride ion concentrations as low as 1 ppm in 50% H2SO4 will cause corrosion of zirconium alloys. Zirconium has no corrosion resistance to HF and is rapidly attacked at concentrations as low as 0.001% Zirconium is resistant to virtually all alkaline solutions, either fused or in solution to boiling temperature. Zirconium has excellent resistance to corrosion in most organics and organic acids. It should be noted that SCC of zirconium occurs in ferric and cupric chloride solutions, concentrated HNO3, methanol-hydrochloric acid and methanol-iodine solutions, and liquid mercury or cesium. Detailed information on the corrosion of zirconium and its alloys is available in the article "Corrosion of Zirconium and Hafnium" in this Volume. Impervious Graphite Impervious graphite is made of impregnating, under pressure, raw graphite with phenolic, furan, or fluorocarbon resins. The resulting nonporous graphite is impermeable to gases and liquids and is highly corrosion resistant in acids and many solvents. Impervious graphite is dimensionally stable, does not fatigue, and will withstand thermal shock. The phenolic and furan impregnants leach out from the graphite when exposed to ammoniacal compounds. NaOH and KOH, wet halogens, hydrogen peroxide (H2O2), strong HNO3, hypochlorites, and some solvents. The introduction of the fluorocarbon impregnants to graphite has markedly increased corrosion resistance in solvents, acids, and alkalies. The maximum service temperature for the phenolic and furan impregnants is 170 °C (340 °F), for the Teflon impregnant, 205 °C (400 °F). Fluoropolymers All fluorocarbons have high molecular weights, high melting points, and excellent chemical resistance. They have found wide application in chemical and pharmaceutical plants as pipe liners, nozzle liners, gaskets, expansion joints, valve liners, diaphragms for valves and pumps, seals and seal components, and barrier linings for vessels. Polytetrafluoroethylene (PTFE) has a service temperature of 245 to 260 °C (475 to 500 °F) and is immune to most

corrosive environments. Among the materials that attack PTFE are molten alkali metals and free fluorine. This material is rapidly permeated by bromine and oxides of nitrogen, and low molecular weight amines tend to plasticize the polymer. It can also be used at cryogenic temperatures, giving it the widest temperature range of any polymer. Perfluoroalkoxytetrafluoroethylene (PFA) is a copolymer of tetrafluoroethylene and a perfluorovinyl ether. This

material has a service temperature of 195 to 230 °C (425 to 450 °F) and is similar in corrosion resistance to PTFE. It has a lower permeability to most chemicals than PTFE, which is always helpful for lining/barrier applications, and possesses higher tensile properties at elevated temperatures. Fluorinated Ethylene Propylene (FEP) is a copolymer of tetrafluoroethylene and hexafluoropropylene. It is a fully

fluorinated thermoplastic with a service temperature of 175 °C (350 °F). Like PTFE and PFA, this material is chemically inert, with a slightly lower permeability than that of PTFE. The uses for FEP are similar to those for PFA and PTFE. Ethylene-Chlorotrifluoroethylene (ECTFE) is the result of a 1:1 alternating copolymer of ethylene and

chlorotrifluoroethylene having a service temperature of 160 °C (320 °F). Because ECTFE is not completely fluorinated, there are sites along the polymer chain at which chemical attack may occur. This material is attacked by aromatic solvents above 120 °C (250 °F), chlorinated hydrocarbons, ethers, methanol, butanol, and ketones above 65 °C (150 °F), acetic acid above 95 °C (200 °F),and H2SO4 and aromatic amines above 65 °C (150 °F). It resists mineral acids up to 120 °C (250 °F), inorganic alkalies, inorganic salts, and oxidizing acids at room temperatures. Polyvinylidene fluoride (PVDF) is a crystalline, high molecular weight, partially fluorinated polymer of

vinylidenedifluoride having a service temperature of 135 °C (275 °F). If is attacked by hot alkalies, hot H2SO4, solvents, and warm organic acids. This material has good resistance to chlorine, bromine, and their compounds; however, it will

exhibit cracking when subjected to stress in the presence of nascent halogens. Virgin unplasticized PVDF has been successfully used for high-purity water piping as a replacement for electropolished type 316L stainless steel. Glass-Lined Steel This material offers the corrosion resistance of glass combined with the strength of steel, making it useful for process equipment operating at elevated pressure and temperature. Glass-lined steel has excellent resistance to corrosion over a wide range of pH and environments. Most glass-lined steel applications will not adversely affect product purity, flavor, or color. Glass-lined steel has an extremely smooth surface that resists fouling and is easily cleaned, which makes it attractive for use in pharmaceutical and fine chemical manufacture.

Material and Corrosion Failures Encountered The types of failures experienced in the pharmaceutical industry are similar in many ways to those seen in the chemicalprocessing industries (see the article "Corrosion in the Chemical Processing Industry" in this Volume). Three primary causes of failure in the manufacture of pharmaceuticals--embedded iron, failures of glass linings, and corrosion under thermal insulation--will be discussed in this section. Embedded Iron A common problem during the fabrication of stainless steel equipment is the embedding of iron in the stainless steel surface. The iron corrodes when exposed to moist air or when wetted, leaving rust streaks. Larger embedded iron particles can also initiate crevice corrosion attack in the stainless steel. Embedded iron cannot be tolerated in a fabrication destined for an application in the pharmaceutical industry in which the stainless steel is used to prevent contamination. Fabrication Practices. The following practices are recommended for minimizing embedded iron in fabrication. First,

sheet, strip, and pipe are usually purchased in a surface finish known as AISI 2B (a bright, cold-rolled finish; see the article "Surface Engineering of Stainless Steels" in Surface Engineering, Volume 5 of ASM Handbook). Plate is normally hot rolled, annealed, and pickled and is furnished with a 2B mill finish. If a plate having a better surface finish is required, it should be specified during the procurement stage. When cleanliness is very important, sheet and plate can be ordered with a protective adhesive paper that can be left in place during storage and fabrication. Pipe and tubing can be ordered with protective end covers, especially if the pipe and tubing is to be stored outdoors. Second, sheet and plates should be stored indoors and upright in racks, not horizontally on the floor. The dragging of sheets and plates over each other and worker foot traffic are often primary causes of embedded iron and deep surface scratches. Third, care should be exercised in handling the sheet and plate on layout tables, forming roll aprons, and benches. This will minimize iron contamination. Lastly, equipment design plays an important role in iron contamination. Equipment and piping should be free draining. If internal attachments are needed, they should not interfere with free drainage. Bottom connections should be completely free draining. Vessel bottoms used as a work area during construction collect debris, and the foot traffic grinds the debris into the surface. It is suggested that the vessel bottom be flushed down and drained completely at the end of the workday to remove collected dirt and debris. If the vessel is a large flat-bottom unit, a slatted wood floor should be installed to reduce the grinding of contaminants into the vessel bottom by foot traffic. Testing new fabrications for embedded iron is relatively easy. The surfaces should be washed which clean water,drained completely, and after a 24-h waiting period, inspected for rust streaks on the surface. Water testing of the fabrication as a minimum should definitely be part of the equipment purchase order. For items to be used in a pharmaceutical plant, a more sensitive test--the ferroxyl test (ASTM A 380) for free iron--should be called for in the purchase order. This test can be easily performed in the field as well as in the fabricator's shop. Removal of Embedded Iron. Pickling is the most effective method of removing embedded iron. The surfaces must

be cleaned of all surface oil, grease, and other organic materials so that the surface becomes wet by the pickling solution. The pickling solution is a mixture of HNO3 and HF at 50 °C (120 °F). The solution removes the embedded iron and other metallic contaminants and leaves the surface clean and in its most corrosion-resistant condition. It should be noted that HNO3 alone will remove only superficial iron contamination and will leave the deeply embedded particles. Small items

are usually pickled by immersion. Piping and vessels that are too large for immersion can be pickled by circulating the pickling solution through them. It is recommended that a competent chemical cleaning contractor be employed for the pickling operation. If the ferroxyl test shows only spotty patches of iron contamination, then it is recommended that these be removed by the use of an HNO3-HF pickling paste, rather than a complete pickling bath. Another method of cleaning the stainless steel surfaces is the use of glass bend blasting. The beads should be clean and of a proper size to abrade the surface slightly and remove the contamination. Gritblasting and sandblasting are not recommended; they leave a rough profile that makes the stainless steel prone to crevice corrosion. Organic contamination on stainless steel surfaces increases crevice corrosion. Contaminants include grease, oil,

marking crayons, paint, and adhesive tape. Removal of organic contaminants is best accomplished by the use of a nonchlorinated solvent. It is important that nonchlorinated solvents be used. If a proprietary degreasing solvent is used, it should be tested to ensure that it does not contain chlorides. Residual chlorides remain in crevices and cause chloride SCC of austenitic stainless steels. Weld Defects. Austenitic stainless steel surfaces can be affected by slag from coated welding electrodes, arc strikes,

welding stop points, grinding marks, and weld spatter. These factors have initiated corrosion in aggressive environments that normally do not attack stainless steels. Arc strikes damage the protective oxide film of the stainless steel and create crevicelike imperfections in or near the HAZ. Weld stops create pinpoint defects in the weld metal. Arc strikes and weld stop points are actually more damaging than embedded iron, because they occur where the protective film has been weakened by the heat of welding. Weld stop defects can be avoided by using runout tabs, by beginning the arc immediately ahead of the stop point, and by welding over each intermediate stop point. Arc strikes are more difficult to eliminate. Initially, the arc can be struck on a runout tab. It can also be struck on the weld metal when the filler metal will tolerate arc strikes. If the metal will not tolerate arc strikes, the arc must be struck alongside the filler metal in or adjacent to the HAZ. Weld spatter creates a small weld in which the molten glob of metal touches and adheres to the surface. The protective oxide film is penetrated, and small crevices or pits are formed where the film has been weakened. Heat tint formation also weakens the oxide film. The weakening is greater for some degrees of heating than for others, as indicated by the extent of color change. The necessity of removing heat tint in greatest where the environment is very aggressive and the stainless steel approaches the limit of its corrosion resistance. Pickling by immersion in the standard HNO3-HF solution is the simplest and preferred method of heat tint removal when size permits. Glass bead blasting, using beads the are clean and of proper size in order to prevent overroughening of the surface, can also be employed to remove the heat tint. Small slag particles from coated electrodes resist cleaning and tend to collect in slight undercuts or other irregularities. To remove slag from 300-series stainless steel, wire brushes fabricated from 300-series stainless steel should be used. For critical service, brushing should be followed by local pickling or glass bead blasting. Grinding is frequently used to remove slag, arc strikes, weld spatter, and other imperfections. Grinding wheels and continuous-belt grinders can overheat the surface and reduce corrosion resistance; therefore, they have limited usefulness. Abrasive disks and flapper wheels are not as harmful to the metal surface. Disks must be kept clean and replaced frequently. These procedures are good commercial fabrication practices and should be specified during the bidding and procurement stages in an effort to eliminate cost overruns and poor service performance. Failures of Glass-Lined Steel Equipment Most glass-lined equipment failures are not related to chemical deterioration of the lining, but rather to mechanical and thermal influences. The typical failures encountered in the use of this type of equipment in the pharmaceutical and fine chemical industries usually involve mechanical shock, corrosion, abrasion, thermal shock, and thermal stress. Mechanical shock is the most common cause of glass failure in pharmaceutical production facilities. It accounts for

approximately 70% of the failures in glass-lined steel process equipment and is frequently the result of human error. The most obvious cause of glass failure due to mechanical shock is objects falling on either the exterior or interior of the vessel. Care must be observed at all times when working near glass-lined equipment because a shock to the outside of a vessel may cause damage to the glass lining.

Lifting lugs are normally supplied on glass-lining equipment and should be used for lifting the equipment and setting it in place. The lugs are specifically designed for this purpose and should be used in accordance with manufacturer's recommended procedures for handling and rigging. Shortcuts in rigging, such as using a nozzle as a lifting lug, can easily subject the glass lining to undue stress and cause damage to the lining. If mechanical shock has occurred or is suspected, the equipment interior should be inspected immediately and, if necessary, repaired. Entering the interior of glass-lined steel equipment always creates a potential for mechanical damage. In addition to compliance with the routine safety precautions, the mechanic should wear clean, soft rubber soled shoes or sneakers to prevent scratching of the lining and should remove all loose objects from his pockets prior to entering the lined equipment. Even metal belt buckles should be removed to prevent accidental scratching of the glass on the entry manway. Tools can be lowered to the mechanic when he is safely inside. If it becomes necessary to remove product or by-products from the walls of glass-lined steel equipment, care must be taken to avoid scratching of the glazed surface. Metal tools should never be used. Plastic or wood scrapers can be used; high-pressure water jets are preferable. The introduction of nucleated glass linings by fabricators of glass-lined steel equipment has reduced the damage caused by impact and, to some extent, has lessened its occurrence. Nucleated or partially nucleated glass linings have higher tensile strength and fracture energy than conventional glassed steel linings. These properties have lessened the tendency toward spalling from mechanical shock, the releasing of internal stresses, or thermal shock. Nucleated glass linings cannot prevent cracking entirely, but they tend to limit the extent of the damage by restricting it to a small area, usually requiring only a tantalum plug for repair. Corrosion Failures of Glass Linings. Glass-lined steel is not completely inert and is constantly undergoing local chemical reactions at the glass surface. Glass-lined steel can be used with corrosive materials because of the low rate of reaction; the slower the rate, the longer the useful life of the glass lining. Acids. Except for HF, concentrated H3PO4, and phosphorous acid (H3PO3) above 85%, glass lined steel is resistant to

corrosion by acids. Generally, the corrosion rate decreases with concentration, but accelerates with increasing temperature. Acid attack is more severe in the vapor phase than in the liquid phase, especially in more dilute solutions, because of the water vapor in the vapor phase. Usually, acid attack will result in a gradual loss of the fire-polished surface, but the lining will generally retain a dull, smooth finish. Hydrofluoric acid will completely destroy glass-lined steel equipment. Even with concentrations as low as 20 ppm, fluorides in an acid environment corrode glass severely, especially in continuous reactions in which the fluorides are repeatedly replaced. Hydrofluoric acid reacts with silicon dioxide, the primary ingredient in glass, and destroys the silicon dioxide structure, forming silicon tetrafluoride and water vapor. The silicon tetrafluoride then hydrolyzes into silicon dioxide and hydrofluorine silicic acid, which are absorbed by the condensing water vapor. When this contaminated condensate reaches the heated vessel wall in the vapor space, it evaporates, depositing silicon dioxide and liberating the hydrofluorine silicic acid, which then disintegrates into silicon tetrafluoride and HF. The chain reaction keeps repeating itself with continuous replenishment of HF. The corrosion occurs both in the liquid phase and in the cooler areas of the vapor space in which the fluoride vapor can condense. Glass is not attacked by fluorine and its compounds in an alkaline environment, nor is it attacked by anhydrous hydrogen fluoride gas. The prerequisite for HF attack in the vapor space is the formation of water; thus, the corrosion rate will be greatly reduced if the vapor area is heated. In the liquid phase, fluorides will severely etch the glass and produce a roughened surface with a complete loss of the fire-polished surface. In the vapor phase, the attack is more localized and concentrated; chipping and pinholes will be seen, but with considerably less loss of the fire-polished glass surface. Reagents that contain fluoride impurities must be carefully analyzed to determine the fluoride level before they are used. Frequently, technical grade H3PO4 and its salts are often contaminated with fluoride, as are other mineral acids. It is important to realize this when using these chemicals in recovery operations. Alkaline attack of glass linings is much more severe than acid attack. The attack takes place only in the liquid phase in

the case of nonvolatile alkalies. The greater the concentration and pH of the alkali, the greater the amount of corrosion. Corrosion by alkalies is evidenced by pinholes, chipping, and a severe loss of the fire-polished surface.

Many glass-lined steel reactors have been lost prematurely in service by improper charging of reactants into the vessels. Caustic reactants charged into a vessel should always be fed directly into the liquid phase. If fed through a nozzle, the alkali will run down the side wall of the reactor in the vapor space and cause severe alkaline attack, especially if the reactor is being heated. Water can cause severe corrosion of glass-lined steel, and the severity increases with water purity and temperature,

becoming greatest above the boiling point. When water droplets condense on the relatively cool surface of the glass-lined equipment in the vapor space, they leach out alkali ions from the glass and form an alkaline solution that attacks the glass. A small amount of acid added to the water usually slows the corrosion caused by condensation in the vapor space. This addition of acid is frequently useful in steam distillations. Abrasion Failures. Failure of glass linings by abrasion alone is not very common. It is evidenced by a loss of fire polish of the glazed surface and results in a rough, sandpaperlike finish. Abrasion in conjunction with acid corrosion results in severe failure. The abrasive action weakens the silica network mechanically, allowing acid corrosion to accelerate rapidly. Thermal shock failure occurs because of abrupt changes in the temperature of the glass lining and results in relatively

small but thick pieces of glass spalling off in rigid fractures. There are four operations in which sudden temperature variations can cause thermal shock: • • • •

Sudden cooling of a glass-lined surface by subjecting a preheated surface to a cold liquid Sudden heating of a glass-lined steel wall by rapidly circulating a very hot fluid through the jacket of a cold vessel Sudden heating of the glass-lined steel surface by introducing a hot fluid into a cold vessel Sudden cooling of the vessel wall by rapid circulation of a cold fluid through the jacket of a preheated vessel

Thermal shock is strictly an operational problem and can be eliminated by adequate process controls. Unlike failure due to mechanical shock, thermal shock usually damages the glass lining so that repairs with tantalum plugs are not practical or possible. Reglassing of the vessel is then required. Failure due to thermal stress is caused by differential heating or cooling that is not instantaneous. Thermal stress

may occur on the vessel wall just below the area at the top jacket closure ring or in the area where the bottom jacket closure ring is welded to the vessel. In either case, the inside of the jacket can be heated or cooled, while the unjacketed area is not. At temperatures approaching 205 °C (400 °F), sufficient strain can be developed in the glass lining at these areas to cause the glass to crack. Thermal stress can be reduced by careful control of heating or cooling operations. Also, insulation of the unjacketed areas will help to reduce the extreme temperature variations. Water hammer has also been known to cause shock waves that add to the thermal stresses at the area of the jacket closure. If these stresses are allowed to persist for long periods of time, cracks can develop with or without chipping. Overstressing of Nozzles. Glass is strong in compression, but a concentrated point load can damage a glass-lined

nozzle. The convex radius at the top of the nozzle also makes it susceptible to damage due to overstressing. The two situations that can cause overstressing of the nozzle and lead to possible failure are overtorquing and overstressing by the attached piping. Overtorquing of bolts or clamps used to secure nozzles on glass-lined vessels to piping can cause glass lining to spall in large segments that may include a considerable area of the nozzle. Overstressing of nozzles through external piping can break the glass lining on the nozzle face. The design of external piping systems should include expansion joints or bellows as close as possible to the nozzles to minimize accentric loads and moments and to allow for thermal expansion of the piping and the vessel. Cold springing when connecting pipe to a nozzle cannot be tolerated. Supporting heavy gear drives on a nozzle can also create an unsafe condition with regard to the glass lining.

Failures of Repair Plugs. Tantalum is the most commonly used repair material for glass linings. Its corrosion

resistance is very close to that of glass, except in fuming H2SO4 above 65 °C (150 °F), H2SO4 above 98%, free sulfur trioxide (SO3) above 65 °C (150 °F), or nascent hydrogen. A repair made with tantalum will protect a glass-lined vessel for its useful life if the repair plug is installed in accordance with the recommendations of the manufacturer. However, the repair plug must be inspected periodically to ensure that it is secure. Galvanic corrosion can cause hydrogen embrittlement of tantalum repair plugs, because the tantalum is in contact with the steel substrate. If as second dissimilar metal is present in the vessel, such as a metallic dip tube fabricated from a material other than tantalum, a galvanic cell is produced. Tantalum, being more noble, will generally act as the cathode where hydrogen is liberated. Failure Due to Hydrogen Damage. Acids on the exterior steel surfaces of glass-lined steel equipment will, in time,

react with the steel, forming nascent hydrogen. This hydrogen diffuses through the steel behind the glass and causes the glass to spall because of pressure buildup. To avoid this type of failure, all acid spills on the outside of a glass-lined steel vessel should be washed off immediately to avoid acid attack. The so-called acid-resistant coatings give some degree of protection, but none is completely effective against acid attack on the steels. Jacket cleaning can create another source of nascent hydrogen. The cleaning solutions recommended by the manufacturer should be used for cleaning the inside jacket of a glass-lined vessel. Strong acid solutions, inhibited or otherwise, should never be used for descaling the jacket in order to avoid any possibility of nascent hydrogen formation. Corrosion Beneath Thermal Insulation Serious corrosion problems frequently occur under thermal insulation applied to vessels and piping components in pharmaceutical plants when the insulation becomes wet. Corrosion beneath insulation is an insidious that is also discussed in the article "Corrosion in the Chemical Processing Industry" in this Volume. The insulation usually conceals the corroding metal, and the situation can go undetected until metal failure occurs. The corrosion of metal components beneath insulation has led to very high maintenance costs and lost production time and has frequently required complete replacement of major components. In addition, operator and plant safety may be jeopardized. Thermal insulation received from manufacturers and distributors is dry, or nearly so. Therefore, if the insulation remains dry, there is no corrosion problem. Possible solutions to the problem of metallic corrosion beneath wet insulation include keeping the insulation dry of protecting the metal. Unfortunately, the application of this solution is not that simple. Insulation can become wet in storage and during field erection. Moisture or weather barriers are not always installed correctly, or they are not effective in preventing water entry. Weather barriers and protective coatings can become damaged and are often not maintained and repaired. The problem is further complicated by the fact that the degree of corrosion by wet insulation appears to be dependent on the insulation material as well as the atmospheric contaminants and moisture entering from external sources. Water extracts from calcium silicate base insulation, fiberglass, cellular glass, and ceramic fiber are generally neutral to alkaline, with pH values in the range of 7 to 11. Cellular glass is free of soluble chloride, while calcium silicate, fiberglass, and some ceramic fibers contain chlorides. Mineral wool gives a neutral environment when wet, usually a pH value of 6 to 7 with a low chloride content (2 to 3 ppm). Water extracts from organic foams can be quite acidic, with pH values of 2 or 3. In addition, where halogenated fire retardants have been added to the foam, water extracts show high levels of free halide, depending on the degree of hydrolysis achieved. Carbon and low-alloy steels are normally passive in alkaline environments and have minimal corrosion rates. However, chloride ions (Cl-), either from the insulation material itself or from airborne or waterborne contaminants, tend to break down the passivity locally and initiate pitting corrosion. If penetration by acidic airborne or waterborne contaminants of sulfur or nitrogen oxides is possible or if water extracts from the insulation are acidic, such as from organic foams, then general corrosion occurs. Occasionally, airborne or waterborne contaminants, notably the nitrate anion ( ), cause external SCC of nonstress-relieved carbon and low-alloy steel systems, especially if a cyclic wetting and drying

concentration mechanism is present. Generally, plant facilities operating continuously or intermittently between 65 and 205 °C (150 and 400 °F) are subject to corrosive attack. The most significant corrosion problem that occurs when insulated austenitic stainless steels are subjected to moisture is external SCC. The problem occurs because chlorides tend to concentrate under the insulation at the surface of the metal when the insulation becomes wet. The moisture can leach soluble chlorides out of the insulation, or the entering moisture may already contain chloride from the environment. At the warm metal surface, the moisture is vaporized, leaving behind an increasing concentration of chlorides. During operation, when the equipment or piping is in the susceptible temperature range, chloride SCC can then occur rapidly. The four factors necessary for SCC are an austenitic stainless steel, Cl- ions, tensile stress on the metal, and temperatures between 50 and 230 °C (120 and 450 °F). Unfortunately, thermal stress relief, which is usually effective in preventing SCC of carbon and low-alloy steels, is normally not practical for austenitic stainless steels. However, there are numerous controllable factors in the design, construction; and maintenance of insulated equipment that have a marked effect on the amount of damage caused by corrosion under thermal insulation. Equipment Design. The design of pressure vessels, tanks, and piping generally includes numerous details for support,

reinforcement, and connection to other equipment. These details may include stiffening rings, insulation support rings, gussets, brackets, reinforcing pads, ladder brackets, flanges, and hangers. The design of equipment, including these details, is the responsibility of engineers/designers utilizing construction codes to ensure reliable designs for both insulated and uninsulated equipment. Unfortunately, consideration of the problem of insulating these details and of leaving adequate room for the insulation is completely lacking in these codes and in the instructions to the engineers/designers. As a result, the items are designed as though they will not be insulated. Undesirable geometries and design features include: • • • • •

Flat horizontal surfaces (such as vacuum rings) Structural shapes that trap water (H-beams, channels) Shapes or configurations that are impossible to weatherproof properly (structural members, gussets) Shapes that lead moisture and contaminants into the insulation hanger rods (angle iron brackets) Inadequate spacing that causes interruption of the vapor or weather barrier (nozzle extensions, ladder brackets, deck or grating supports). The weather barrier on such designs is frequently broken because of inappropriate details for insulated equipment or the lack of space for the specified insulation thickness

The consequence of an incomplete moisture barrier is that more water and contaminants get into the insulation at each exposure cycle; this increases the time required for drying, cools the insulated equipment to temperatures at which corrosion is possible, and thus increases the damage. Some equipment details, such as gussets, brackets, and hangers, actually funnel water and contaminants into the insulation. Another consideration is the increased cost of insulating equipment that was not designed for insulation. In such cases, the insulation and jacketing must be cut and fitted by installers; thus, needless man-hours are spent insulating a complicated detail, with the result being an installation that is doomed to failure. The solution to such problems is to specify the type of insulation, thickness, and weather jacketing in the design stage. The service or operating temperature of the equipment is very important in corrosion beneath thermal insulation.

Higher temperatures make water more corrosive, and paints and caulking will fail prematurely. Generally, corrosion associated with equipment operating below freezing temperatures is corrosion outside of, not under, the insulation. Equipment operating between freezing and the atmospheric dew point is subject to continuous corrosion, and damage can occur as quickly as it does under warm insulation. However, corrosion beneath warm insulation is more difficult to control, because of the drying out of water entering the insulation and the concentration of contaminants carried in with the water drying out repeatedly in the same location. Insulation Materials. Corrosion is possible beneath all types of insulation. The insulation material is only a contributing factor. The insulation characteristics that are most influential in the corrosion of metal beneath insulation include water absorbency, chemical contributions to the water phase (not only from the insulation but also from external sources), and service temperature. Some of the more widely used insulation materials are discussed below.

Polyurethane foam is primarily used for cold and antisweat service. It does not absorb or wick water as long as the

cell structure remains intact. It is permeable to water vapor in cold service when the required vapor barrier fails. Vapor diffuses through cell walls to the temperature zone in which it condenses and diffuses further to the point at which it freezes. The maximum service temperature is 80 °C (180 °F). If used in continuously cold service, it does not corrode unprotected metal surfaces. If in intermittent service to its maximum service temperature, it can cause corrosion of unprotected wet metal surfaces from released chlorides in fire retardants and blowing agents. The ultraviolet rays from the sun will decompose this insulation. Polyisocyanurate foam is a fire-resistant organic foam having a low flame propagation rate. It does not absorb and

wick water as long as the cell structure remains intact. It is permeable to water vapor in cold service when the required vapor barrier fails. The maximum service temperature is 120 °C (250 °F). When polyisocyanurate is exposed to heat and moisture, the cell structure in the heated zone breaks down. The decomposition products may contain chlorides from the fire retardant and blowing agent and may aggressively corrode unprotected metal surfaces. Flexible foamed elastomer does not readily absorb or wick water, and it has a maximum service temperature of 80 °C (180 °F). Although not corrosive by itself, it will support the corrosion of unprotected metal surfaces when water is present, especially when the water contain chlorides from an external source. Cellular glass is a rigid glass foam whose blowing agent contains carbon dioxide and hydrogen sulfide. It does not

absorb and wick water. The maximum service temperature is 480 °C (900 °F). When water is present and the cell structure is damaged, release of the foam blowing agent may cause corrosion on unprotected carbon steel surfaces. Glass fiber for insulation is usually a pure glass fiber containing various types of binders. Fiberglass will absorb and

wick water; however, it drains excess moisture better than other types of insulation. The maximum service temperature is 230 °C (450 °F), with special formulations to 455 °C (850 °F). Mineral wool is a mineral or metal slag fiber that is basically an impure glass. It readily absorbs and wicks water.

Maximum service temperatures range from 650 to 980 °C (1200 to 1800 °F), depending on type and manufacturer. The fact that mineral wool will wick and hold water makes it a contributory factor in the corrosion on unprotected wet metal surfaces. Calcium silicate insulation is a cementitious mixture and readily absorbs and wicks water. Calcium silicate insulation

can hold up to 400% of its own weight in water without dripping. The maximum service temperature is 650 °C (1200 °F). Although its pH is initially high (10 average), it is aggressive in supporting corrosion on unprotected wet metal surfaces because of its moisture retention, particularly when the moisture contains chlorides from an external source. Protective coatings are extremely important in preventing the corrosion of metal surfaces beneath thermal insulation.

In the past, the attitude has been that a single coat of primer would be adequate based on the assumption that the weatherproofing never allowed water to penetrate into the insulation system. This was not the case. Basically, service under thermal insulation is virtually an immersion service. Once the weather-or vaporproofing barrier is broken, metal surfaces under isolation are wet longer than the surfaces of most uninsulated equipment. Under warm insulation, the coating is subject to higher temperatures than most coated, uninsulated equipment. The coatings beneath thermal insulation fail because of chemical degradation and the permeability of the coating. Highly permeable coatings allow corrosion to initiate behind the coating film, even in the absence of breaks or pinholes in the coating. In selecting a protective coating for use beneath insulation, consideration should be given to its abrasion resistance, temperature resistance, chemical resistance, and its ability to resist hot water vapors.Organic coating materials are discussed in details in the article “Organic Coatings and Linings” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook. In general, for carbon steel and stainless steel piping and equipment operating at -45 to 2120 °C (-50 to 250 °F) under insulation, it is recommended that the surface be abrasive blast cleaned to National Association of Corrosion Engineers (NACE) No. 2 or Steel Structures Painting Council (SSPC) SP10 Near White metal standards having a surface profile of 50 to 75 m (2 to 3 mils). The freshly cleaned surfaces should then be coated with 305 m (12 mils) (total dry-film

thickness) of an epoxy phenolic coating applied in two or more coats or with 305 m (12 mils) (total dry-film thickness) of a high melting point, amine-cured, coal tar epoxy applied in one or two coats. For carbon steel piping and equipment operating at temperatures of 120 to 260 °C (250 to 500 °F) with intermittent cycling service into the hot water range under insulation, it is recommended that the surface be abrasive blast cleaned to NACE No. 1 or SSPC SP10 Near White metal standards having a surface profile of 50 to 75 m (2 to 3 mils). The freshly cleaned surfaces should then be coated with 150 m (6 mils) (total dry-film thickness) of a coplymerized silicone resin applied in two or more coats. Carbon steel piping and equipment operating continuously above 260 °C (500 °F) under insulation should have the surface abrasive blast cleaned to NACE No. 3 or SSPC SP6 Commercial Blast standards having a surface profile of 25 to 50 m (1 to 2 mils). The freshly cleaned surfaces should then be coated with 100 m (4 mils) (total dry-film thickness) of a black zinc-free modified silicone coating applied in two coats. For stainless steel piping and equipment operating at temperatures from 20 to 370 °C (250 to 700 °F) under insulation, it is recommended that the surface be cleaned according to SSPC SP1 Solvent Cleaning standards. The freshly cleaned surfaces should be coated with 100 m (4 mils) (dry-film thickness) of a black zinc-free modified silicone coating applied in two coats. Inorganic and organic zinc-rich primers have given very poor performance under thermal insulation and should not be used. Possible reasons for this poor performance include: • • •

The likelihood of reversal of polarity of galvanic couples with increasing temperature Chemical salts carried in and deposited with the water that interfere with or destroy the effectiveness of the coating The environment beneath the insulation is not freely ventilated and may not have adequate oxygen or carbon dioxide for film forming reactions to occur

Weatherproofing and Vaporproofing. The outer covering of the insulation system is critical. It is the principal

barrier to the water necessary for the corrosion of metals beneath thermal insulation. Also, it is the only part of the insulation system that can be quickly inspected and economically repaired. The purpose of the weather barrier, which should be used on warm or hot equipment, is to keep liquid out but to permit the evaporation of any moisture that gets in. The purpose of a vapor barrier is to keep both liquid and vapor out of the insulation system. Vapor barriers should be used on all cold and dual-service insulation systems.

Selected References • • •

• • • • •

G.E. Moller Designing With Stainless Steels for Service in Stress Corrosion Environments, Mater. Perform., Vol 16 (No. 5), p 32-44 A. Garner, "Corrosion of High Alloy Austenitic Stainless Steel Weldments in Oxidizing Environments," Paper presented at Corrosion/82, National Association of Corrosion Engineers, 1982 Technical Bulletins S-1, 871-ENG; S-1, 872-ENG; S-1, 874-ENG; S-1, 880-ENG; S-1, 884-ENG; S1, 885-ENG; S-51-26-ENG; S-52-27-ENG; S-51-28 ENG; S-51-33-ENG; 2-52-72-ENG; and S-5418-ENG, Sandvik Steel AB Technical Bulletins H-2000A, H-2002A, and H-2010A, Haynes International, Inc. D.H. DeClerck and A.J. Patarcity, 32nd Report on Materials of Construction, Chem. Eng., Vol 93 (No. 22), p 46-63 Technical Bulletins T-5, T-7, T-15, T-37, T-40, T-42, and S-37, Inco Alloys International, Inc. W.I. Pollock and J.M. Barnhart, Ed., Corrosion of Metals Under Thermal Insulation, STP 880, American Society for Testing and Materials, 1985 S. Lederman, "Glass Lined Equipment--Typical and Atypical Failures Encountered in the Field" Paper presented at Corrosion/82, National Association of Corrosion Engineers, 1982

•

Technical Bulletins TWCA-8102ZR and TWCA 8103ZR, Teledyne Wah Chang Albany

Corrosion in Operations

Petroleum

Production

Chairman: James E. Donham, Consultant

Introduction THE PRODUCTION of oil and gas, its transportation and refining, and its subsequent use as fuel and raw materials for chemicals constitute a complex and demanding process. Various problems are encountered in this process, and corrosion is a major one. The costs of lost time, the replacement of materials of construction, and the constant personnel involvement in corrosion control are substantial and, if not controlled, can be catastrophic. The control of corrosion through the use of coatings, metallurgy, nonmetallic materials of construction, cathodic protection, inhibitors, and other methods has evolved into a science in its own right and has created industries devoted solely to corrosion control in specific areas. This article will discuss the particular corrosion problems encountered, the methods of control used in petroleum production, and the storage and transportation of oil and gas to the refinery. Refinery corrosion is discussed in the article "Corrosion in Petroleum Refining and Petrochemical Operations" in this Volume. Causes of Corrosion Arthur K. Dunlop, Corrosion Control Consultant

The fundamentals of corrosion and the forms of corrosion are discussed in detail in the Sections so named in this Volume. Therefore, this article will concentrate on those aspects that tend to be unique to corrosion as encountered in oil and gas production. Most unique, of course, are the environments encountered in actual production formations, which, in the absence of contamination, are devoid of oxygen. In situ corrosives are limited to carbon dioxide, hydrogen sulfide, polysulfides, organic acids, and elemental sulfur. Additional unique aspects are the extremes of temperature and, particularly, pressure encountered. In deep gas wells (6000 m, or 20,000 ft), temperatures approaching 230 °C (450 °F) have been measured, and partial pressures of CO2 and H2S of the order of 20.7 MPa (3000 psi) and 48 MPa (7000 psi), respectively, have been encountered. Convenient access to the most important literature on H2S corrosion (particularly with regard to sulfide stress cracking) and CO2 corrosion is available in Ref 1 and 2. Supplemental information is provided in Ref 3 and 4. To the initially oxygen-free geologic environment, a variety of oxygen-contaminated fluids may be introduced. Examples are drilling muds, which are used during drilling and maintenance of wells; dense brines; water and carbon dioxide injected for secondary oil recovery; and hydrochloric acid injected to aid formation permeability. Some of these fluids are inherently corrosive; others are potentially corrosive only when contaminated with oxygen. Oxygen is also responsible for the external corrosion of offshore platforms and drilling rigs. In oil and gas production, highly stressed structural members are directly exposed to a corrosive seawater environment. This makes corrosion fatigue a particular concern. Another unique aspect of oil and gas production operations, particularly in older fields, is the almost exclusive use of carbon and low-alloy steels. Only recently has any extensive use of corrosion-resistant alloys been justified.

Oxygen Although it is not normally present at depths more than approximately 100 m (330 ft) below the surface, oxygen is nevertheless responsible for a great deal of the corrosion encountered in oil and gas production. However, oxygeninduced internal corrosion problems tend to be greater in oil production, where much of the processing and handling occurs at near-ambient pressure. This makes oxygen contamination through leaking pump seals, casing and process vents, open hatches, and open handling (as in mud pits during drilling, trucking, and so on) highly likely. Also, failure of oxygen removal processes (gas stripping and chemical scavenging) is a relatively common occurrence in waterflood systems (see the discussion "Corrosion in Secondary Recovery Operations" in this article). A number of the properties of oxygen contribute to its uniqueness as a corrosive. Oxygen is a strong oxidant. This means that even trace concentrations can be harmful, and the corrosion potential of steel (almost 1.3 V) is high enough to overcome very substantial potential drops between anodic and cathodic sites. Also, the kinetics of oxygen reduction on a metal or conductive oxide surface are relatively fast. This, coupled with the low solubility of oxygen in water and brines, tends to produce conditions in which the mass transport of oxygen is the rate-limiting step in the corrosion of carbon and low-alloy steels in nonacidic environments. Mass transport is important in a number of aspects of oxygen corrosion and corrosion control. On newly installed bare steel offshore structures, mass transport of oxygen governs current requirements for cathodic protection. Poor mass transport under deposits and in crevices promotes localization of attack. In the final analysis, limiting the mass transport of oxygen plays a critical role in much of the corrosion control in oxygenated systems. The crucial role of mass transport can be illustrated as follows. At ambient conditions, water equilibrated with air will contain of the order of 7 to 8 ppm of oxygen. Under such conditions, mass transport limited rates of general corrosion of steel range from about 0.25 mm/yr (10 mils/yr) in a stagnant system to 15 mm/yr (600 mils/yr) in a highly turbulent one. However, by chemically scavenging the oxygen concentration down to the order of 7 to 8 ppb, the corresponding rates are reduced to less than about 0.01 mm/yr (0.4 mils/yr). Such rates are acceptable. However, under these conditions, magnetite forms as a stable protective corrosion product film and further lowers the corrosion rate by introducing a slower, anodically controlling step. An even more fundamental role of magnetite should be acknowledged. This is the protection of the steel surface from reaction with water or the hydrogen ions contained in the water. Therefore, if an excess of a chelating agent such as ethylenediamine tetraacetic acid (EDTA) dissolves a protective magnetite film, as would normally occur in regions of high turbulence, rapid corrosion ensues, despite the absence of oxygen.

Hydrogen Sulfide, Polysulfides, and Sulfur Hydrogen sulfide, when dissolved in water, is a weak acid and is therefore corrosive because it is a source of hydrogen ions. In the absence of buffering ions, water equilibrated with 1 atm of H2S has a pH of about 4. However, under highpressure formation conditions, pH values as low as 3 have been calculated. Hydrogen sulfide can also play other roles in corrosion in oil and gas production. It acts as a catalyst to promote absorption by steel of atomic hydrogen formed by the cathodic reduction of hydrogen ions. This accounts for its role in promoting sulfide stress cracking of high-strength steels (yield strength greater than approximately 690 MPa, or 100 ksi) (Ref 5). Hydrogen sulfide also reacts with elemental sulfur. In a gas phase with a high H2S partial pressure, sulfanes (free acid forms of a polysulfide) are formed so that elemental sulfur is rendered mobile and is produced along with the remaining gaseous constituents. However, as the pressure decreases traveling up the production tubing, the sulfanes dissociate and elemental sulfur precipitates. Various solvent treatments are used to avoid plugging by such sulfur. In the aqueous phase, under acidic conditions, sulfanes are also largely dissociated into H2S and elemental sulfur. However, enough strongly oxidizing species can remain either as polysulfide ions or as traces of sulfanes to play a significant role in corrosion reactions. Oxygen contamination of sour (H2S-containing) systems can also result in polysulfide formation.

Iron sulfide corrosion products can be important in corrosion control. Because of the low solubility, rapid precipitation, and mechanical properties of such corrosion products, velocity effects are generally not encountered in sour systems. Satisfactory inhibition at velocities up to 30 m/s (100 ft/s) has been proved. The great range of possible iron sulfide corrosion products and their possible effects on corrosion have been extensively studied (Ref 6, 7, 8, 9), but little of immediate practical value has resulted. At lower temperatures and H2S partial pressures, an adequately protective film often forms. The absence of chloride salts strongly promotes this condition, and the absence of oxygen is absolutely essential. At the high temperatures (150 to 230 °C, or 300 to 450 °F) and H2S partial pressures (thousands of pounds per square inch) encountered in deep sour gas wells, a so-called barnacle type of localized corrosion (Ref 10) can occur, resulting in corrosion rates of several hundred mils per year (Ref 11). This type of attack is strongly promoted by polysulfide-type species and requires the presence of some minimum chloride concentration. Although initially recognized in deep sour well environments, this same mechanism may operate at lower rates under much milder conditions. In the barnacle mechanism (Fig. 1), corrosion can be sustained beneath thick but porous iron sulfide deposits (primarily pyrrhotite, FeS1.15 because the FeS surface is an effective cathode. The anodic reaction beneath the FeS deposit is dependent on the presence of a thin layer of concentrated FeCl2 at the Fe/FeS interface. This intervening FeCl2 layer is acidic due to ferrous ion hydrolysis, thus preventing precipitation of FeS directly on the corroding steel surface and enabling the anodic reaction to be sustained by the cathodic reaction on the external FeS surface.

Fig. 1 Schematic showing the barnacle mechanism of sour pitting corrosion. Source: Ref 11

Carbon Dioxide Carbon dioxide, like H2S, is a weakly acidic gas and becomes corrosive when dissolved in water. However, CO2 must first hydrate to carbonic acid (H2CO3) (a relatively slow reaction) before it is acidic. There are other marked differences between the two systems. Velocity effects are very important in the CO2 system; corrosion rates can reach very high levels (thousands of mils per year), and the presence of salts is often unimportant.

Whether or not corrosion in a CO2 system is inherently controlled or uncontrolled depends critically on the factors governing the deposition and retention of a protective iron carbonate (siderite) scale. On the other hand, there are the factors that determine the rate of corrosion on bare steel. These latter factors govern the importance of maintaining corrosion control. Bare steel corrosion rates can be calculated from Eq 1, which was developed on the basis of electrochemical studies of the aqueous CO2/carbon steel system (Ref 12):

(Eq 1)

where R is the corrosion rate, t is temperature (°C), A is a constant, and is CO2 partial pressure. When R is calculated in millimeters per year and is in atmospheres, A = 7.96. When R is calculated in mils per year and is in pounds per square inch, A = 8.78. Corrosion rates calculated with Eq 1 reach 25 mm/yr (1000 mils/yr) at 65 °C (150 °F) and 1 MPa (150 psi) CO 2 pressure, and 250 mm/yr (10,000 mils/yr) at 82 °C (180 °F) and 16 MPa (2300 psi) CO2 pressure. Obviously, such rates are unacceptable. An alternative, idealized condition occurs when a protective carbonate scale is present and when the corrosion rate is limited by the need to replenish the film lost due to solubility in the aqueous phase. Under such conditions, the rates calculated for a hypothetical sweet (CO2-containing) gas well reached a maximum of about 0.15 mm/yr (6 mils/yr) as compared to calculated bare metal rates of 500 to 2000 mm/yr (20,000 to 80,000 mils/yr) (Ref 13). Conditions favoring the formation of a protective iron carbonate scale are: • • •

Elevated-temperature (decreased scale-solubility, decreased CO2 solubility, precipitation kinetics) Increased pH, as occurs in bicarbonate-containing waters (decreased solubility) Lack of turbulence

and

accelerated

Turbulence is often the critical factor in pushing a sweet system into a corrosive regime. Excessive degrees of turbulence prevent either the formation or retention of a protective iron carbonate film. The critical velocity equation has been used to estimate when excessive turbulence can be expected in a CO 2 system (Ref 14). There is no doubt that the velocity effect is real, but there is some question as to whether this exact form of the equation is the appropriate one:

(Eq 2)

where velocity is calculated in feet per second, K is a constant, and is the density of the produced fluid (liquid + gas combined). When is in kilograms per cubic meter, K = 7.6; for in pounds per cubic foot, K = 100. When both H2S and CO2 are present, simplified calculations indicate that iron sulfide may be the corrosion product scale when the H2S/CO2 ratio exceeds about 1/500 (Ref 13); sour system considerations would then be expected to apply. Even in a strictly CO2 system, iron carbonate may not always be the corrosion product. Magnetite may form instead. Figure 2 shows the stability fields expected for siderite and magnetite as a function of the redox potential (expressed here in terms of hydrogen fugacity) of the system (Ref 13). In actual experience, corrosion product scales are often found to consist of mixtures or layers of siderite and magnetite.

Fig. 2 Effect of temperature and hydrogen fugacity on the stability of FeCO3 and Fe3O4 in contact with aqueous CO2. Equilibrium calculations determine boundaries (indicated by the iso-hydrogen fugacity curves, with fugacities given in atmospheres) between FeCO3 and Fe3O4 stability fields in the produced fluid, which contained CO2, water, and traces of hydrogen. Curves 1 and 2 locate boundaries for locations 5180 m (17,000 ft) deep and at the wellhead, respectively, of a 170,000 m3/day (6,000,000-ft3/day) well corroding at a rate of 0.75 mm/yr (30 mils/yr). Source: Ref 13

Iron carbonate lacks conductivity and therefore does not provide an efficient cathode surface. Thus, the types of pitting mechanisms found in oxygenated and in H2S-containing systems do not occur. Rather, generalized corrosion occurs at any regions not covered by the protective scale. The result is that on any bare metal the anodic and cathodic regions are so microscopically dispersed that salt--to provide conductivity--is not needed to achieve the corrosion rates predicted by Eq 1.

Strong Acids Strong acids are often pumped into wells to stimulate production by increasing formation permeability. For limestone formations, 15 and 28% hydrochloric acids are commonly used. For sandstones, additions of 3% HF are necessary. In deep sour gas wells where HCl inhibitors lose effectiveness, 12% formic acid has been used. Corrosion control is normally achieved by a combination of inhibition and limiting exposure time to 2 to 12 h. If corrosion-resistant alloys are present (austenitic and duplex stainless steels, and so on), concern for stress-corrosion cracking (SCC) and inhibitor ineffectiveness (respectively) may rule out the use of HCl.

Concentrated Brines Dense halide brines of the cations of calcium, zinc, and, more rarely, magnesium are sometimes used to balance formation pressures during various production operations. All can be corrosive because of dissolved oxygen or entrained air. In addition, such brines may be corrosive because of the acidity generated by the hydrolysis of the metallic ions, as illustrated in Eq 3:

Zn2+ + H2O = ZnOH+ + H+

(Eq 3)

Corrosivity due to acidity is worst with dense zinc brines. More expensive calcium bromide brines are now often used at densities above about 1.7 g/cm3 (14 lb/gal) (attainable with CaBr2 brines) to avoid long-term exposure to ZnCl2 brines.

Stray-Current Corrosion If an extraneous direct current (dc) in the earth is traversed by a conductor, part of the current will transfer to the lowerresistance path thus provided. Direct currents are much more destructive than alternating currents (ac); an equivalent ac current causes only about 1% of the damage of a dc current (Ref 15). Regions of current arrival (where electrons depart) will become cathodic, and those regions where the current departs will become anodic. With corrodible metals such as carbon and low-alloy steels, corrosion in the anodic areas is the result. For example, 1 A · yr can corrode 9 kg (20 lb) of steel. Cathodic protection systems are the most likely present-day sources of stray dc currents in production operations. More detailed discussions are available in Ref 4 and 16. The article “Stray-Current Corrosion” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook also contains information on the causes and mechanisms of stray-current corrosion.

Under-Deposit (Crevice) Corrosion This is a form of localized corrosion found almost exclusively (if not exclusively) in oxygen-containing systems. Such corrosion is usually most intense in chloride-containing systems. It is essential to have some form of shielding of an area on a metal such that it is wet by an electrolyte solution but is not readily accessible to oxygen, the diffusing corrosive species. This type of attack is usually associated with small volumes of stagnant solution caused by surface deposits (sand, sludge, corrosion products, bacterial growth), crevices in joints, and gasket surfaces. Crevice corrosion is discussed in Ref 17, and a quantitative treatment of crevice corrosion (particularly of stainless steels) is provided in Ref 18, 19, and 20. The mechanism of crevice corrosion hinges upon the environmental conditions resulting from the loss of hydroxide production with cessation of the cathodic reaction when the initial oxygen in the shielded region is exhausted. Thus, in the shielded region, the anodic corrosion reaction continues because the corrosion potential is maintained by the reduction of oxygen on the outside surface. However, chloride or other anions now migrate into the developing anodic region to maintain electroneutrality. Thus, a relatively concentrated, essentially ferrous chloride solution accumulates in the shielded region. As a result of the hydrolysis of the ferrous ions, the pH drops to a value of 2 or 3. At this point, the crevice corrosion type of localized attack is fully established. The anodic reaction continues in the shielded region because in the low-pH environment the ferrous ions go readily into solution and have little tendency to precipitate as an oxide or hydroxide on the surface and thus stifle the anodic reaction. Outside, the cathodic reaction continues unperturbed. Because of the large ratio of cathodic-to-anodic surface area, high rates of localized corrosion can be maintained with very modest cathodic current densities. More information on the mechanisms of crevice corrosion is available in the article “Crevice Corrosion” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook.

Galvanic Corrosion When two dissimilar metals are electrically coupled--both electronically by a metal bond and ionically through an electrolyte--the more active (electronegative) metal tends to become a sacrificial anode and supply cathodic protection to the more noble metal. Such situations are often encountered in heat exchangers in which carbon steel tubesheets are used

with copper alloy tubes and at junctions between piping, fasteners, or corrosion-resistant sheeting with containers of another material. Problems with galvanic corrosion are the most acute when the cathode-to-anode area ratio is large. Such situations are often encountered inadvertently. This has happened when the normal polarity difference between zinc and steel in a galvanized pipe reversed in a bicarbonate/chloride brine so that the steel pipe walls at pinholes in the galvanizing perforated rapidly while trying to protect the extensive adjacent galvanized area. Another situation is when plastic-coated steel is coupled to more noble metal. At any pinholes in the coating, a very adverse area ratio will exist, and rapid corrosion rates can result. The article “Galvanic Corrosion” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook contains more information on this form of attack.

Biological Effects The most important biological effect on corrosion in oil and gas production is the generation of H2S by sulfate-reducing bacteria (Desulfovibrio Desulfuricans). These are anaerobic bacteria that metabolize sulfate ions (using an organic carbon source) and produce hydrogen sulfide. They can thus introduce H2S and all its corrosive ramifications into normally H2Sfree systems. Colonies of sulfate-reducing bacteria can also form deposits that are conducive to under-deposit corrosion. Contrary to previous beliefs, any resultant corrosion appears to be due to a mechanical shielding action, rather than any depolarizing action resulting from the metabolic processes of the SRB. However, this is not to deny that the introduction of H2S (whatever the source) into a crevice region could have an accelerating effect on corrosion, because H2S is known to be an anodic stimulant. More information on biological corrosion is available in the articles “Microbiologically Influenced Corrosion” and “Evaluating Microbiologically Influenced Corrosion” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook.

Mechanical and Mechanical/Corrosive Effects Cavitation. This metal removal--often grain by grain--is due to high-pressure shock wave impingement resulting from the rapid collapse of minute bubbles created under certain conditions in high-velocity fluid handling equipment. It is usually found on pump impellers operating with too low a suction pressure. Erosion. Most commonly, this is direct metal removal by the cutting action of high-velocity abrasive particles. Erosion failures (washouts) are seen in drill pipe when leaks (loose connections or a corrosion fatigue crack) allow drilling mud to flow through the wall under a high-pressure gradient. Erosion of flow lines at bends and joints by produced sand is probably the other most common occurrence of erosion in oil and gas production. Erosion-Corrosion. Strictly speaking, in erosion-corrosion, only the protective corrosion product film is removed by

erosive forces; however, with the protective film absent, corrosion can occur at a greatly accelerated rate. Erosioncorrosion may play a role in CO2 corrosion (Ref 21), and sand, under mild flow conditions, may also cause erosion corrosion. Erosion-corrosion has also been noted in heavy anchor chains where their use in an abrasive bottom mud allowed corrosion at contact regions to proceed at a rate of many hundreds of mils per year. Corrosion fatigue results from subjecting a metal to alternating stresses in a corrosive environment. At points of

greatest stress, the corrosion product film becomes damaged during cycling, thus allowing localized corrosion to take place. Eventually, this leads to crack initiation and crack growth by a combination of mechanical and corrosive action. Because of this combined action, damage per cycle is greater at low cycling rates, where corrosion can play a larger role. Also, in corrosion fatigue, a fatigue limit does not exist; rather than leveling out as in simple fatigue, the usable stress level continues to decrease with increasing cycles. The greatest concern for corrosion fatigue arises in connection with highly stressed, submerged, offshore structures. Welded connections on drill ships and on drilling and production platforms are particularly susceptible to this form of structural impairment. More information on attack resulting from combined corrosion and mechanical effects is available in the article "Mechanically Assisted Degradation" in this Volume.

Corrosion Control Methods* Arthur K. Dunlop, Corrosion Control Consultant; David H. Patrick, ARCO Resources Technology; Donald E. Drake, Mobil Corporation

Materials Selection Traditionally, carbon and low-alloy steels were virtually the only metals used in the production of oil and gas. This resulted from the fact that very large quantities of metal are required in petroleum production, and until a few years ago, crude oil and gas were relatively low-value products. In addition, insurmountable corrosion problems were not encountered. This situation changed when gas and oil prices increased dramatically and deeper wells were drilled that encountered corrosive environments of greatly increased severity. The final factor that made the current widespread use of corrosionresistant alloys possible was the development of high-strength forms of these alloys. This allowed thinner pipe and vessel walls and greatly reduced the amount of material required. The result of this situation is that essentially all of the high-tonnage uses of corrosion-resistant alloys in oil and gas production involve alloys in a high-strength form. Yield strengths typically span the range of 550 to 1250 MPa (80 to 180 ksi), but can reach nearly 1750 MPa (250 ksi) in wire lines. Metallurgical Considerations The great majority of corrosion-resistant alloys used in oil and gas production were originally developed for other applications not requiring high strength. Therefore, many of these have an austenitic microstructure and can be strengthened only by some form of cold working. This presents no problem for the production of tubulars used underground in the well, because they are joined by threaded connections. However, for other applications, such as welded flow lines and cast and forged valves, different techniques must be used. In addition to the problems of environmental compatibility, there are a number of difficulties of a metallurgical nature that will be briefly mentioned. For high-strength materials, particularly at the upper end of the range, adequate ductility is often difficult to achieve. Also, heat treatments developed for other applications may not be optimal for petroleum protection environments. Environmental Considerations There are several environmental factors that are more or less unique to oil and gas production. One factor is the general absence of oxygen in produced fluids. The dominant naturally occurring corrosives are carbon dioxide and hydrogen sulfide. However, in wells with high hydrogen sulfide concentrations, elemental sulfur--a relatively strong oxidizing agent--can also be present. Water, which must be present to make sulfur or the acid gases (CO2 and H2S) corrosive, can be assumed to be present in all productive geologic formations. Therefore, liquid water is present from the bottom of the well up to the point in the flow system at which it is removed. Corrosivity is aggravated by the presence of salt in formation waters. Such brines can range from a few percent sodium chloride up to saturated solutions containing as much as 300,000 ppm of total dissolved salts. Temperature also affects corrosivity. Deep well temperatures up to 205 °C (400 °F) are not uncommon. Therefore, temperatures from these levels down to ambient must be considered. Corrosion From other applications, enough is generally known to select alloys that are resistant to general corrosion. However, resistance to localized corrosion (pitting and crevice corrosion) often requires some experimental study.

Environmental Stress Cracking. In corrosion-resistant alloy selection, environmental stress cracking normally

requires the greatest portion of experimental effort. This follows from the use of high-strength alloys in environments in which there is relatively little experience with these alloys. The problem is complicated by the fact that the lowest alloy content and the maximum reliable strength level are needed to achieve an economically viable choice. Because this involves design close to the limits of the material, it is necessary to define these limits as accurately as possible. The term environmental stress cracking was selected because it includes two different mechanisms; cathodic and anodic. The cathodic mechanism is that of hydrogen embrittlement found in sulfide stress cracking of high-strength carbon and low-alloy steels in the presence of hydrogen sulfide. Here, H2S promotes entry of cathodically evolved hydrogen atoms into the metal. The anodic mechanism is that involved in the chloride SCC of austenitic stainless steels at temperatures generally above about 65 °C (150 °F). Nickel-base alloys are normally immune to this type of chloride cracking. However, the presence of H2S can induce susceptibility in normally resistant alloys by acting to promote anodic dissolution. In the most resistant alloys, such as Hastelloy alloy C-276 (Table 1), elemental sulfur and a temperature of about 230 °C (450 °F) must also be present to cause cracking. Cracking will occur only in high-strength, heavily cold worked samples.

Table 1 Compositions of corrosion-resistant alloys Alloys Hastelloy alloy C-276 Hastelloy alloy G Haynes No. 20 Mod Sanicro 28 Inconel alloy 625 Incoloy alloy 825 Monel alloy 400 MP35N Type 316 stainless steel Type 410 stainless steel Ferralium 255

(a)

UNS designation N10276 N06007 N08320 N08028 N06625 N08825 N04400 R30035 S31600 S41000 S32550

Composition, %(a) C Cr 0.02 14.5-16.5 0.05 21-23.5 0.05 21-23 0.03 26-28 0.10 20-23 0.05 19.5-23.5 0.3 ... 0.025 19-21 0.08 16-18 0.15 11.5-13.5 0.04 24-27

Fe 4-7 18-21 bal bal 5 bal 2.5 1.0 bal bal bal

Ni bal bal 25-27 29.5-32.5 bal 38-46 63-70 33-37 10-14 ... 4.5-6.5

Mo 15-17 5.5-7.5 4-6 3-4 8-10 2.5-3.5 ... 9-10.5 2-3 ... 2-4

Other 2.5Co, 1.0Mn, 4.5W, 0.35V 2.5Nb, 2.5Co, 2.5Cu, 2.0Mn, 1W, 1Si 2.5Mn, 1Si, Ti-4 × C min 1.4Cu, 2.5Mn, 1Si 0.4Al, 4.15Nb, 0.5Mn, 0.4Ti 0.2Al, 3Cu, 1Mn, 1.2Ti, 0.5Si bal Cu, 2Mn, 0.5Si bal Co, 0.15Mn, 1Ti 2Mn, 1Si 1Si, 1Mn 2.5Cu, 1.5Mn, 0.25N, 1Si

Maximum unless range is given or otherwise indicated.

Effects in the Cathodic Mechanism. In corrosion-resistant alloys, environmental effects on cathodic stress cracking

are generally similar to those encountered in the sulfide stress cracking of carbon and low-alloy steels. However, in extreme cases, the temperatures needed to avoid cracking susceptibility in the highest-strength materials can be as high as 120 °C (250 °F). An example is Hastelloy alloy C-276 at a hardness of 45 HRC. In contrast, the highest minimum temperature given in NACE standard MR-01-75 for high-strength low-alloy tubulars is 80 °C (175 °F). Also, on corrosion-resistant alloys, chloride ions can have a significant deleterious effect in cathodic cracking. This results from deterioration of passivity in the presence of chloride. The resulting higher anodic corrosion rate can then support a higher cathodic rate and thus gives rise to enhanced cathodic charging of the corrosion-resistant alloy with atomic hydrogen. Effects in the Anodic Mechanism. Determining regions of susceptibility to anodic cracking is much more

complicated than defining regions of susceptibility to cathodic cracking. This is due to the multidimensional nature of the problem. Therefore, in anodic cracking, susceptibility is a strong function of such environmental factors as temperature, chloride concentration, and hydrogen sulfide partial pressure; in addition, pH, oxidants (elemental sulfur or polysulfides), and galvanic coupling can be involved. Further, alloy composition (corrosion resistance), strength level, and stress all influence cracking susceptibility. The effect of H2S requires some elaboration. In cathodic cracking, there is a well-established threshold for the initiation of the sulfide stress cracking phenomenon; this is not the case in anodic cracking. The effect in anodic cracking is more monotonic. Therefore, there can almost be a different critical value for every combination of alloy, strength, applied

stress, and environmental condition. However, it is safe to assume that research in this area is leading to the development of some generalized patterns. Selection of Corrosion-Resistant Alloys Alloy selection, from a corrosion standpoint, can be considered to be a three-step process. First, resistance to general corrosion must be ensured. This is primarily a function of the chromium content of the alloy. Second, resistance to localized attack also must be ensured. This is primarily a function of molybdenum content. Finally, resistance to environmental stress cracking is sought at the highest feasible strength level. Nickel content plays a principal role in this instance, particularly in providing resistance to anodic cracking. The close correlation between pitting resistance and resistance to anodic cracking should be noted. This apparently results from the ease of crack initiation under the low-pH high-chloride conditions found in pits. Therefore, higher molybdenum can also increase resistance to anodic cracking. With the procedures given below, regions for alloy applicability can be shown schematically as a qualitative function of environmental severity. This has been attempted in Fig. 3, in which an aqueous, CO2-containing environment (hence low pH) has been assumed and the effects of temperature, chloride, and H2S concentration are illustrated. The effect of yield strength is not shown, but if environmental cracking is the limiting factor, reducing the yield strength should extend applicability to more severe environments.

Fig. 3 Schematic showing corrosion-resistant alloy selection for production environments containing aqueous CO2 and H2S

The reader should be cautioned that a diagram such as Fig. 3 is really more of a guide to alloy qualification than to direct selection for a particular application. Therefore, it may aid in developing a more efficient approach to alloy testing. Testing for Resistance to Environmental Stress Cracking The most directly applicable results are obtained by exposing samples of commercially produced alloys to an environment simulating as closely as possible that expected in actual production operations. Fortunately, an

understanding of the principles involved allows considerable simplification to be made without significantly altering the value of the results. Two simplifications can be readily made in the environmental parameters, as follows. First, only the CO2 and H2S partial pressures are usually reproduced. The overburden of methane pressure to create the actual total pressure of the environment is dispensed with. This can substantially lower the pressure ratings of test vessels. Second, only the chloride content of the brine is typically reproduced, rather than trying to simulate the total ionic spectrum of the produced fluids. Often, no reliable analysis is available. Assumption of a saturated sodium chloride solution is then a relatively conservation approach. A wide variety of test specimens are used, but for high-strength corrosion-resistant alloy tubulars, C-rings are particularly convenient. Double cantilever beam specimens are very attractive in that they can provide a quantitative measure of fracture toughness, which can then be used in mechanical design. Slow strain rate testing can also be very useful in assessing alloy limits. Another aspect of testing is the imposed stress level. Exposure at 100% of yield strength is the only conservative approach. However, some investigators place unwarranted reliance on design values and test at some lesser fraction of the yield strength. Detailed information on testing for resistance to environmental cracking is available in the articles “Evaluation of Hydrogen Embrittlement” and “Evaluation of Stress-Corrosion Cracking” in Corrosion: Fundamentals, Testing, and Protection, Volume 13A of ASM Handbook..

Note cited in this section

*

The section on materials selection was written by A.K. Dunlop; the section on coatings was written by D.H. Patrick; and the sections on cathodic protection, inhibitors, nonmetallic materials, and environmental control were written by D.E. Drake.

Coatings Internal protective coatings have been used to protect tubing, downhole equipment, wellhead components. Christmas trees (manifolds used to control the rate of production, receive the produced fluids under pressure, and direct the produced fluids to the gathering point), and various downstream flow lines and pressure vessels for more than 30 years. Because internal coatings are subject to damage, successful use is usually accompanied by chemical inhibition or cathodic protection as part of the entire protective program. Most of the coating use has been below 175 °C (345 °F). Tubing. The benefits derived from coating tubing depend on the coating remaining intact. Because no coating can be

applied and installed 100% holiday-free, inhibition programs are commonly employed to accommodate holidays and minor damage. The suitability of the service is dependent on specific testing and an effective quality control program. Inhibitor and volume will not change even though an operator decides to use coated tubing rather than bare tubing. The use of coated tubing improves the protection in shielded areas that are inaccessible to inhibitors. The two greatest dangers to coated tubing are wireline damage and improper joint selection. Wireline damage can be minimized by adjusting running procedures to include wireline guides and to slow wireline speed ( 22 Stainless alloys Aluminum alloys

Formation, directional or deviated hole

Observation

Erosion-corrosion

...

...

Underdeposit corrosion, uniform corrosion, pitting

...

...

...

...

Underdeposit corrosion, sulfide stress cracking, pitting Chloride SCC

...

...

Sulfide stress cracking, pitting

Underdeposit corrosion, galvanic corrosion, pitting

Control pH 9.5; remove oxygen, H2S, and/or CO2; use filming inhibitors.

Control pH

9.5; add biocides.

Select temperature-stable chemicals; use friction/torque reducers; cool mud; use oil mud. Use lubricants, torque reducers, and filming inhibitors; Control solids; use oil muds.

9.5; control oxygen, Control pH H2S, and CO2; limit stress. Control H2S to very low levels.

Limit salts and temperature exposure; control deposits. Limit pH to 9.5-10.5; use passivation.

Failure Analysis Analysis of used and failed equipment can provide a means of developing corrosion prevention methods. Identification of corrosion products (Ref 54) and corrosion forms (Ref 17, 55) can be developed into cause-effect mechanisms. Monitoring Drill pipe corrosion coupons are used to measure the rate, form, and cause of corrosion (Ref 56). Corrosion rates of 2.4 to 9.8 kg/m2/yr (0.5 to 2 lb/ft2/yr) free of pitting is an acceptable range. The lower rate should be used for deviated holes, deep drilling, and/or high-stress conditions. Special monitoring using linear polarization (Ref 57) or galvanic probes (Ref 58) will provide instant detection of corrosion and changes in rates that can be related to system conditions on a real time basis. On-site chemical and physical analyses of drilling fluid properties are conducted on a frequent basis (Ref 56). Test procedures should include oxygen, CO2, H2S, and bacterial analysis for comprehensive monitoring of corrosion problems. The Drilling Manual, published by the International Association of Drilling Contractors, presents equipment inspection methods as well as information on the care, handling, and specifications of tool joints, drill pipe, casing, and tubing. Reference is made to this manual to provide comprehensive information on drilling and production equipment. Oxygen Corrosion Control Oxygen causes crevice corrosion under deposits and is considered to be the most serious corrosion accelerator in drilling environments. Oxygen enters the drilling fluid system externally from the atmosphere, usually by way of solids control and mud mixing equipment (Ref 59). The operation of this equipment to reduce air entrapment into the circulating system is an effective technique for limiting oxygen levels. Foaming problems are characteristic of some mud systems and can result in high oxygen levels on the high-pressure side of the pump. Defoaming the fluid or maintaining properties to release gas quickly is required to overcome this problem. Oxygen scavengers, such as sodium sulfite or ammonium bisulfite, are used to remove oxygen from drilling fluid (see the discussion "Environmental Control" in this article). Treatment methods involve a continuous addition of chemical at the rate of 10 mg/L sulfite ion for each 1 mg/L of oxygen present in the fluid. A residual sulfite concentration of approximately 100 mg/L is maintained in the drilling fluid as a functional means of controlling oxygen in drilling systems. Oxygen scavenger catalysts are frequently required to overcome interfering side reactions that prevent the oxygen-sulfite reaction. Calcium in the fluid can combine with sodium sulfite and form calcium sulfite precipitate, thus

preventing the sulfite ion from scavenging oxygen. Aldehydes and chlorine dioxide used as biocides in drilling fluids react with sulfite ions and may prevent oxygen removal. The addition of cobalt or nickel catalysts overcomes many of these problems by increasing oxygen-sulfite reaction rates. Passivating compounds, such as sodium chromate or nitrite, are used to protect equipment during air, mist, or foam drilling operations. Treatment levels range from approximately 500 to 2000 mg/L of chromate or nitrite ion in fresh to slightly brackish fluid. Higher concentrations are required in high-brine solutions, and sodium nitrite is not recommended above approximately 25,000 mg/L of chloride ion concentration. A noteworthy disadvantage of using passivating agents is the tendency toward accelerated pitting attack if treatment levels are too low or if deposits exist under which the metal cannot be passivated. Zinc compounds are often combined with passivating agents to reduce pitting tendencies. Treatments for controlling deposits are recommended to mitigate under-deposit attack and are covered in the discussion "Scale and Deposit Control" in this section. Chromate or nitrite compounds are not compatible with sulfite-type oxygen scavengers. Care should be exercised in the use and disposal of chromate compounds. These materials are classified as carcinogens, and personnel safety should be ensured. Injection of drilling fluid within deep formations has been used to dispose of excess or waste fluid. A clear advantage is gained in the use of sodium chromate or nitrite chemicals when hydrogen sulfide is encountered in the well. These compounds oxidize and remove H2S (see the discussion "Hydrogen Sulfide Corrosion Control" in this Section). Atmospheric corrosion occurs on drilling equipment in urban, polluted, tropical, and marine environments. Protective coatings are commonly applied at the steel mill or storage yard and periodically between drilling operations. Many coating compositions are commercially available for both short-and long-term storage (2 or 3 years). The filming inhibitors that are typically used during drilling often provide good atmospheric protection for short periods between jobs. For long-term exposure, careful surface cleaning and selected atmospheric coating are recommended. Hydrogen Sulfide Corrosion Control Hydrogen sulfide causes two forms of corrosion surface attack--under-deposit (crevice) corrosion and sulfide stress cracking. Corrosion control methods include selecting resistant materials, removing the H2S from the fluid, and reducing stress. Hydrogen sulfide enters the drilling fluid primarily from the formation, but it can also come from thermally degraded mud products, sulfate-reducing bacteria, and makeup water. Alkaline pH control and sulfide scavengers are used to neutralize, precipitate, and/or oxidize H2S. Film-forming aminetype inhibitors are recommended for coating the drill string. Caustic soda or calcium hydroxide treatments are used to neutralize the acid gas. Alkaline pH above 9.5 results in the production of sodium bisulfide or sodium sulfide products that are almost totally water soluble. This treatment provides both personnel safety and corrosion protection. Compounds of iron oxide (Fe3O4), zinc carbonate, zinc oxide, and zinc chelates are used to precipitate sulfide ions from solution. Pretreatments of approximately 1 kg/barrel (2 lb/barrel) of one of the scavengers are commonly recommended as a precaution against a small influx of H2S entering the mud system and causing damage. Tests are used to monitor scavenger concentrations and treatment requirements. Sodium chromate, zinc chromate, or sodium nitrite compounds are used to oxidize H2S to sulfate or elemental sulfur. The oxidizing process is a fast and efficient method of removing H2S from the system. There is no compatibility problem with the sulfide scavengers listed above. Formaldehyde and chlorine dioxide are compounds that are frequently used as drilling fluid biocides. These products react with hydrogen sulfide, offsetting its corrosive action; however, their biocidal properties are diminished or eliminated in the process. Filming amine inhibitors provide protection from hydrogen sulfide surface attack and hydrogen embrittlement. Oilsoluble filming inhibitors applied directly on the drill pipe are recommended to offset corrosion fatigue and hydrogen embrittlement. Care should be taken with cationic filming inhibitors, which can damage mud properties by flocculating the anionic clays in drilling systems. Oil muds provide the most effective protection against all corrosion causes, including H2S. The oil phase provides a nonconductive film covering exposed equipment and thus preventing the corrosion process.

Stress reduction by mechanical changes, such as rotary speed and less weight on the bit, is effective in reducing sulfideinduced embrittlement failures. Torque-reducing agents, particularly in high-angle drilling, are effective in lowering stress. Material selection for drill pipe and casing can have a significant effect in controlling sulfide stress cracking. The brittle failures related to H2S are linked to the hardness and yield strength of the steel. Steels with hardness levels below 22 HRC or with maximum yield strengths of 620 MPa (90 ksi) have few sulfide stress cracking problems. Cold work, such as tong or slip marks, increases the hardness of steel, and sulfide stress cracking then becomes a problem. The service stresses in drilling frequently demand materials of great strength, requiring hardness and strength levels that are susceptible to sulfide stress cracking. Because of such requirements, the primary means of avoiding sulfide stress cracking is by control of the drilling fluid. A full discussion of the metals used in sulfide environments is provided in NACE MR-01-75. Higher temperatures (above 80 °C, or 175 °F) reduce sulfide stress cracking failures on high-strength steel. This factor can become advantageous in drilling and production operations if properly controlled. For example, an influx of H2S while drilling may not cause damage if the fluid temperature is above 80 °C (175 °F) in the hole. If H 2S is detected, scavenging should always be completed before operations are begun that would lower the metal temperature, such as pulling the drill pipe from the hole. Carbon Dioxide Corrosion Control Carbon dioxide causes pitting primarily by under-deposit corrosion cell action. Corrosion control methods involve controlling the pH in the higher alkaline ranges. An effective technique is to treat the mud with calcium hydroxide to neutralize this acid-forming gas and to precipitate carbonates, thus lowering CO2 levels. Film-forming inhibitors of the oil-soluble amine type applied by spraying the outside of the drill pipe and batch treatments for inside diameter filming are recommended to penetrate pits and deposits, stopping their corrosion action. Control of CO2 is quite similar to H2S corrosion control, and these two gases often enter the mud from the formation together. Scale and Deposit Control Mineral scale, corrosion by-products, and mud that form deposits on exposed metal are a major factor in setting up conditions that result in under-deposit pitting attack. The prevention and removal of these deposits with scale inhibitors is quite effective in offsetting this most serious drilling fluid corrosion problem. Inhibitors such as organic phosphonate, phosphate esters, and others of the acrylic, acrylamide, or maleic acid base structures have been effective. Products that exhibit threshold effect, temperature stability, and strong surface-active characteristics are useful. Treatments are variable because of environmental conditions, which differ greatly in drilling fluid compositions. As general rules apply, treatments of 15 to 75 mg/L are used on a daily basis for most mineral scale control. Treatments above this level are used to control deposits of metal corrosion by-products. Considerably higher treatment levels, up to 1000 mg/L, are used to provide corrosion protection. Care should be exercised in using the higher treatment levels, because these compounds may alter mud properties because of their dispersing characteristics.

Primary Production There are two main types of producing oil wells: artificial lift wells and flowing wells. Artificial lift wells can be further divided according to the method used to pump the hydrocarbon to the surface. These include rod-pumped wells, wells that use downhole hydraulic pumps, and gas-lift wells. Approximately 90% of the artificial lift wells in the United States are rod pumped. Artificial Lift Wells Rod-pumped wells. In a rod-pumped well, the potential for corrosion damage is aggravated by the sucker rods alternately being stretched and compressed and by the abrasion of the rod couplings on the inside of the tubing. It is common for a well to have continuing sucker rod failures. Pulling and replacing the rods is a quick fix, but the problem will continue to exist until the root cause of the failure is identified and corrected. Identifying the problem is the most important step, because corrective action cannot be taken if the cause is not clear. Rod breaks should be inspected immediately after the rod string is pulled to determine if corrosion is occurring and to determine the steps that can be taken immediately to prevent a recurrence.

Corrosion in rod-pumped wells can be caused by several mechanisms, as discussed below. Galvanic corrosion is

caused by dissimilar metals in contact or by the difference in metallurgy between two areas on a sucker rod. Most galvanic corrosion on rods is caused by differences in metal condition caused by hammer, wrench, or tong marks and the grooves left by rod-straightening machines. The impact area will be cathodic to the body of the rod, and corrosion will occur adjacent to the mark. Sucker rods have a soft decarburized layer or skin of low-carbon steel 0.13 to 0.2 mm (5 to 8 mils) thick. This layer can be broken by careless handling. Bent rods are sometimes straightened and used again. This is poor practice, because a bent rod is permanently damaged and should be discarded. The rod-straightening machine will put spiral grooves around the rod, and corrosion will occur directly adjacent to the groove. Any of these conditions will lead to pitting, and stress raisers will be set up. The cyclic stresses resulting from alternately stretching and compressing the rods during pumping operations will lead to rapid failure. Stray current from surface equipment or leakage from a cathodic protection system will cause severe corrosion where the current leaves the rod string. It is usually seen on couplings or the part of the rod that is close to the coupling. Damage from oxygen corrosion may take place when the rods are stored outdoors or when oxygen enters the wellbore through the annulus. Rusting of stored rods will often cause pitting, and rust deposits can set up concentration cells or under-deposit corrosion when the rods are run in the hole. Oxygen entry into the wellbore in wells that pump off or oxygen introduced during inhibitor treating operations will aggravate other forms of corrosion by depolarizing the cathodes on the metal surface during the corrosion reaction. Oxygen corrosion generally occurs in the lower part of the well: the casing, pump, tubing, and the lower part of the rod string. The effect lessens in the upper part of the well, because the oxygen is depleted by the corrosion reaction. Carbon dioxide or sweet corrosion is caused by CO2 from produced gases dissolving in water and forming carbonic acid. The carbonic acid ionizes to bicarbonate irons and hydrogen ions. A low pH results, and the bicarbonate and the carbonic acid will react directly with the steel rod and cause metal loss and pitting. The pits formed are usually round bottomed with sharp sides, and they may be connected in a line or will sometimes form a ring around the rod. Fatigue cracks will be initiated at the bottom of the pits. Carbon dioxide corrosion is aggravated by the presence of oxygen and organic acids. Oxygen depolarizes the cathodes, and organic acids deplete the bicarbonate ion concentration, which dissolves protective carbonate scale. The formation of iron carbonate scale is the major limiting factor in CO2 corrosion. Many pumping wells are in the temperature range (