ASM Handbook: Volume 7: Powder Metal Technologies and Applications (Asm Handbook) (Asm Handbook) (Asm Handbook)

ASM INTERNATIONAL

®

Publication Information and Contributors

Powder Metal Technologies and Applications was published in 1998 as Volume 7 of ASM Handbook. The Volume was prepared under the direction of the ASM Handbook Committee.

Editorial Advisory Board • • • • • • • • •

Peter W. Lee The Timken Co. Yves Trudel Quebec Metal Powders Limited Ronald Iacocca The Pennsylvania State University Randall M. German The Pennsylvania State University B. Lynn Ferguson Deformation Control Technology, Inc. William B. Eisen Crucible Research Kenneth Moyer Magna Tech P/M Labs Deepak Madan F.W. Winter Inc. Howard Sanderow Management and Engineering Technologies

Contributors and Reviewers • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Stanley Abkowitz Dynamet Technology Samuel Allen Massachusetts Institute of Technology Terry Allen David E. Alman U.S. Department of Energy Albany Research Center Sundar Atre Pennsylvania State University Christopher Avallone International Specialty Products Satyajit Banerjee Breed Technologies Inc. J. Banhart Fraunhofer Institute Daniel Banyash Dixon Ticonderoga Company Tim Bell DuPont Company David Berry OMG Americas Ram Bhagat Pennsylvania State University Pat Bhave Thermal Technology Inc. Sherri Bingert Los Alamos National Laboratory Jack Bonsky Advanced Manufacturing Center Cleveland State University Robert Burns Cincinnati Incorporated Donald Byrd Wyman Gordon Forgings John Carson Jenike and Johanson Inc. Francois Chagnon Quebec Metal Powders Tom Chirkot Patterson-Kelley Company Harsco Corporation Stephen Claeys Pyron Corporation John Conway Crucible Compaction Metals Kevin Couchman Sinter Metals Inc. Pennsylvania Pressed Metals Division F. Robert Dax Concurrent Technologies Corporation Amedeo deRege Domfer Metal Powders R. Doherty Drexel University Ian Donaldson Presmet Carl Dorsch Latrobe Steel Company John Dunkley Atomising Systems Ltd. William Eisen Crucible Research Mark Eisenmann Moft Metallurgical Corporation Victor Ettel Inco Technical Services Limited Daniel Eylon University of Dayton Zhigang Fang Smith International B. Lynn Ferguson Deformation Control Technology Inc.

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

Howard Ferguson Metal Powder Products Inc. Richard Fields National Institute of Standards and Technology Gavin Freeman Sherritt International Corporation Sam Froes University of Idaho Randall German Pennsylvania State University Herbert Giesche Alfred University New York State College of Ceramics Howard Glicksman DuPont Company Kinyon Gorton Caterpillar Inc. Mark Greenfield Kennametal Inc. Joanna Groza University of California at Davis E.Y. Gutmanas Technion--Israel Institute of Technology Richard Haber Rutgers University Jack A. Hammill, Jr. Hoeganaes Corporation Francis Hanejko Hoeganaes Corporation John Hebeisen Bodycotte, IMT Ralph Hershberger UltraFine Powder Technology Inc. Gregory Hildeman Alcoa Technical Center Craig Hudson Sinter Metals Inc. Ronald Iacocca Pennsylvania State University M.I. Jaffe W. Brian James Hoeganaes Corporation John Johnson Howmet Corporation Brian Kaye Laurentian University Pat Kenkel Burgess-Norton Manufacturing Company Mark Kirschner BOC Gases Erhard Klar Richard Knight Drexel University Walter Knopp P/M Engineering & Consulting John Kosco Keystone Powdered Metal Company Sriram Krishnaswami MARC Analysis Research Corporation David Krueger BASF Corporation Howard Kuhn Concurrent Technologies Corporation Chaman Lall Sinter Metals Inc. Larry Lane Brush Wellman Inc. Alan Lawley Drexel University Jai-Sung Lee Hanyang University Peter Lee The Timken Company Louis W. Lherbier Dynamet Inc. Deepak Madan F.W. Winter Inc. & Company Craig Madden Madden Studios Gary Maddock Carpenter Technology Corporation Dan Marantz Flame Spray Industries Inc. Alain Marcotte U.S. Bronze Powders James Marder Brush Wellman Millard S. Masteller Carpenter Technology Corporation Ian Masters Sherritt International Corporation Paul E. Matthews Brian J. McTiernan Crucible Research Center Steve Miller Nuclear Metals Inc. Wojciech Misiolek Lehigh University John Moll Crucible Research Center In-Hyung Moon Hanyang University Ronald Mowry C.I. Hayes Inc. Kenneth Moyer Magna Tech P/M Labs

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

Charles Muisener Loctite Corporation Alexander Sergeevich Mukasyan University of Notre Dame Zuhair Munir University of California at Davis Anil Nadkarni OMG Americas K.S. Narasimhan Hoeganaes Corporation Ralph Nelson DuPont Company Bernard North Kennametal Inc. W. Glen Northcutt Lockheed Martin Scott Nushart ATM Corporation James Oakes Teledyne Advanced Materials Barbara O'Neal AVS Inc. Lanny Pease III Powder Tech Associates Inc. Kenneth Pinnow Michael Pohl Horiba Laboratory Products John Porter Cincinnati Inc. Peter Price Tom Prucher Burgess-Norton Manufacturing Company David Pye Pye Metallurgical Consulting Inc. Thomas Reddoch Ametek Inc. John Reinshagen Ametek Inc. Melvin Renowden Air Liquide America Frank Rizzo Crucible Compaction Metals Prasan Samal OMG Americas Howard Sanderow Management and Engineering Technologies G. Sathyanarayanan Lehigh University Barbara Shaw Pennsylvania State University Haskell Sheinberg Los Alamos National Laboratory George Shturtz Carbon City Products John Simmons B.I. Thortex Ronald Smith Drexel University Richard Speaker Air Liquide America Robert Sprague Consultant Victor Straub Keystone Carbon Company C. Suryanarayana Colorado School of Mines Bruce Sutherland Westaim Corporation Rajiv Tandon Phillips Origen Powder Metallurgy Pierre Taubenblat Promet Associates Mark Thomason Sinterite Furnace Division Gasbarre Products Inc. Juan Trasorras Federal Mogul Yves Trudel Quebec Metal Powders Limited John Tundermann Inco Alloys International Inc. Christian Turner Hasbro Inc. William Ullrich AcuPowder Int. Arvind Varma University of Notre Dame Jack T. Webster Webster-Hoff Corporation Bruce Weiner Brookhaven Instruments Greg West National Sintered Alloys Donald White Metal Powder Industries Federation George White BOC Gases Eric Whitney Pennsylvania State University Jeff Wolfe Kennametal Inc. John Wood University of Nottingham C. Fred Yolton Crucible Materials Antonios Zavaliangos Drexel University

•

Robert Zimmerman

Arburg Inc.

Foreword In recognition of the ongoing development and growth of powder metallurgy (P/M) materials, methods, and applications, ASM International offers the new Volume 7 of ASM Handbook. Powder Metal Technologies and Applications is a completely revised and updated edition of Powder Metallurgy, Volume 7 of the 9th Edition Metals Handbook, published in 1984. This new volume provides comprehensive updates that reflect the continuing improvements in traditional P/M technologies as well as significant new coverage of emerging P/M materials and manufacturing methods. The ASM Handbook Committee, the editors, the authors, and the reviewers have collaborated to produce a book that meets the high technical standards of the ASM Handbook series. In addition to in-depth articles on production, testing and characterization, and consolidation of powders, the new volume expands coverage on the performance of P/M materials, part shaping methods, secondary operations, and advanced areas of engineering research such as process modeling. This extensive coverage is designed to foster increased awareness of the current status and potential of P/M technologies. To all who contributed toward the completion of this task, we extend our sincere thanks. Alton D. Romig, Jr. President, ASM International Michael J. DeHaemer Managing Director, ASM International

Preface On behalf of the ASM Handbook Committee, it is a pleasure to introduce this fully revised and updated edition of Volume 7, Powder Metal Technologies and Applications as part of the ASM Handbook series. Since the first publication of Volume 7 in 1984 as part of the 9th Edition Metals Handbook, substantial new methods, technologies, and applications have occurred in powder metallurgy. These developments reflect the continuing growth of powder metallurgy (P/M) as a technology for net-shape fabrication, new materials, and innovative manufacturing processes and engineering practices. Net-shape or near-net-shape fabrication is a key objective in many P/M applications. Many factors influence the economics and performance of P/M fabrication, and new methods and process improvements are constantly considered and developed. In this regard, the new Volume 7 provides completely updated information on several emerging technologies for powder shaping and consolidation. Examples include all new articles on powder injection molding, binder assisted extrusion, warm compaction, spray forming, powder extrusion, pneumatic isostatic forging, field activated sintering, cold sintering, and the consolidation of ultrafine and nanocrystalline materials. New articles also cover process modeling of injection molding, isostatic pressing, and rigid die compaction. Traditional press-and-sinter fabrication and high-density consolidation remain the major topic areas in the new Volume 7. This coverage includes new articles in several practical areas such as resin impregnation, dimensional control, machining, welding, heat treatment, and metallography of P/M materials. The traditional processes of rigid die compaction and sintering are also covered extensively with several updated articles on major production factors such as tooling, die design, compressibility and compaction, sintering practices, and atmosphere control. An overview article, "Powder Shaping and Consolidation Technologies," also compares and summarizes the alternatives and factors that can influence the selection of a P/M manufacturing method. Coverage is also expanded on high-density consolidation and highperformance P/M materials such as powder forged steels. Multiple articles on powder production and characterization methods have also been revised or updated in several key areas. The article on atomization is fully revised from the previous edition, and several new articles have been added to the Section "Metal Powder Production and Characterization." In particular, the new article by T. Allen, "Powder Sampling and Classification," is a key addition that provides essential information for accurate characterization of particle size distributions. The variability of sieve analysis is also covered in more detail in this new Volume.

The new Volume 7 also provides detailed performance and processing information on a wide range of advanced and conventional P/M materials. Ferrous P/M materials are covered in several separate articles, and more detailed information on corrosion, wear, fatigue, and mechanical properties are discussed in separate articles. New articles also provide information on several advanced materials such as aluminum-base composites and reactive-sintered intermetallics. This extensive volume would not have been possible without the guidance of the section editors and the dedicated efforts of the contributing authors. I would also like to thank Erhard Klar for organizing the previous edition, which formed the core for the structure of the new edition. Finally, special thanks are extended to ASM staff--particularly to project editor Steve Lampman--for their dedicated efforts in developing and producing this Volume. Peter W. Lee The Timken Company Member, ASM Handbook Committee

General Information Officers and Trustees of ASM International (1997-1998) Officers

• • • • •

Alton D. Romig, Jr. President and Trustee Sandia National Laboratories Hans H. Portisch Vice President and Trustee Krupp VDM Austria GmbH Michael J. DeHaemer Secretary and Managing Director ASM International W. Raymond Cribb Treasurer Brush Wellman Inc. George Krauss Immediate Past President Colorado School of Mines

Trustees

• • • • • • • • •

Nicholas F. Fiore Carpenter Technology Corporation Gerald G. Hoeft Caterpillar Inc. Jennie S. Hwang H-Technologies Group Inc. Thomas F. McCardle Kolene Corporation Bhakta B. Rath U.S. Naval Research Laboratory C. (Ravi) Ravindran Ryerson Polytechnic University Darrell W. Smith Michigan Technological University Leo G. Thompson Lindberg Corporation James C. Williams GE Aircraft Engines

Members of the ASM Handbook Committee (1997-1998) • • • • • • • • • •

Michelle M. Gauthier (Chair 1997-; Member 1990-) Raytheon Electronic Systems Craig V. Darragh (Vice Chair 1997-; Member 1989-) The Timken Company Bruce P. Bardes (1993-) Materials Technology Solutions Company Rodney R. Boyer (1982-1985; 1995-) Boeing Commercial Airplane Group Toni M. Brugger (1993-) Carpenter Technology Corporation R. Chattopadhyay (1996-) Consultant Rosalind P. Cheslock (1994-) Aicha Elshabini-Riad (1990-) Virginia Polytechnic Institute & State University Henry E. Fairman (1993-) MQS Inspection Inc. Michael T. Hahn (1995-) Northrop Grumman Corporation

• • • • • • • • • • • • • • • •

Larry D. Hanke (1994-) Materials Evaluation and Engineering Inc. Jeffrey A. Hawk (1997-) U.S. Department of Energy Dennis D. Huffman (1982-) The Timken Company S. Jim Ibarra, Jr. (1991-) Amoco Corporation Dwight Janoff (1995-) FMC Corporation Paul J. Kovach (1995-) Stress Engineering Services Inc. Peter W. Lee (1990-) The Timken Company William L. Mankins (1989-) Mahi Sahoo (1993-) CANMET Wilbur C. Simmons (1993-) Army Research Office Karl P. Staudhammer (1997-) Los Alamos National Laboratory Kenneth B. Tator (1991-) KTA-Tator Inc. Malcolm C. Thomas (1993-) Allison Engine Company George F. Vander Voort (1997-) Buehler Ltd. Jeffrey Waldman (1995-) Drexel University Dan Zhao (1996-) Essex Group Inc.

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

R.J. Austin (1992-1994) (Member 1984-) 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) D.D. Huffman (1986-1990) (Member 1982-) 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) W.L. Mankins (1994-1997) (Member 1989-) J.L. McCall (1982) (Member 1977-1982) W.J. Merten (1927-1930) (Member 1923-1933) D.L. Olson (1990-1992) (Member 1982-1988, 1989-1992) 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 Steven R. Lampman, Project Editor; Grace M. Davidson, Manager of Handbook Production; Bonnie R. Sanders, Copy Editing Manager; Alexandra B. Hoskins, Copy Editor; Randall L. Boring, Production Coordinator; and Kathleen S. Dragolich, Production Coordinator. Editorial assistance was provided by Amy E. Hammel and Anita D. Fill. The Volume was prepared under the direction of Scott D. Henry, Assistant Director of Reference Publications, and William W. Scott, Jr., Director of Technical Publications.

Conversion to Electronic Files ASM Handbook, Volume 7, Powder Metal Technologies and Applications was converted to electronic files in 1999. The conversion was based on the first printing (1998). 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, Marlene Seuffert, Gayle Kalman, Scott Henry, Robert Braddock, Alexandra Hoskins, and Erika Baxter. 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 © 1998 by ASM International All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the written permission of the copyright owner. First printing, December 1998 This book is a collective effort involving hundreds of technical specialists. It brings together 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, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, ARE GIVEN IN CONNECTION WITH THIS PUBLICATION. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this information. No claim of any kind, whether as to products or information in this publication, and whether or not based on negligence, shall be greater in amount than the purchase price of this product or publication in respect of which damages are claimed. THE REMEDY HEREBY PROVIDED SHALL BE THE EXCLUSIVE AND SOLE REMEDY OF BUYER, AND IN NO EVENT SHALL EITHER PARTY BE LIABLE FOR SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES WHETHER OR NOT CAUSED BY OR RESULTING FROM THE NEGLIGENCE OF SUCH PARTY. As with any material, evaluation of the material under enduse conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended. Nothing contained in this book 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 this book 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 handbook. Vols. 1-2 have title: Metals handbook. Includes bibliographical references and indexes. Contents: v. 1. Properties and selection--irons, steels, and high-performance alloys--v. 2. Properties and selection-nonferrous alloys and special-purpose materials--[etc.]--v. 7. Powder metal technologies and applications

1. Metals--Handbooks, manuals, etc. 2. Metal-work-- Handbooks, manuals, etc. I. ASM International. Handbook Committee. II. Metals Handbook. TA459.M43 1990 620.1'6 90-115 SAN 204-7586 ISBN 0-87170-387-4

History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Introduction POWDER METALLURGY has been called a lost art. Unlike clay and other ceramic materials, the art of molding and firing practical or decorative metallic objects was only occasionally applied during the early stages of recorded history. Sintering of metals was entirely forgotten during the succeeding centuries, only to be revived in Europe at the end of the 18th century, when various methods of platinum powder production were recorded (Table 1). Table 1 Major historical developments in powder metallurgy Date

Development

Origin

3000 B.C.

"Sponge iron" for making tools

Egypt, Africa, India

A.D. 1200

Cementing platinum grains

South America (Incas)

1781

Fusible platinum-arsenic alloy

France, Germany

1790

Production of platinum-arsenic chemical vessels commercially

France

1822

Platinum powder formed into solid ingot

France

1826

High-temperature sintering of platinum powder compacts on a commercial basis

Russia

1829

Wollaston method of producing compact platinum from platinum sponge (basis of modern P/M technique)

England

1830

Sintering compacts of various metals

Europe

1859

Platinum fusion process

1870

Patent for bearing materials made from metal powders (forerunner of self-lubricating bearings)

United States

1878-1900

Incandescent lamp filaments

United States

1915-1930

Cemented carbides

Germany

Early 1900s

Composite metals

United States

Porous metals and metallic filters

United States

1920s

Self-lubricating bearings (used commercially)

United States

1940s

Iron powder technology

Central Europe

1950s and 1960s

P/M wrought and dispersion-strengthened products, including P/M forgings

United States

1970s

Hot isostatic pressing, P/M tool steels, and superplastic superalloys

United States

1980s

Rapid solidification and powder injection molding technology

United States

1990s

Intermetallics, metal-matrix composites, spray forming, nanoscale powders, and warm compaction

United States, England

Metal powders such as gold, copper, and bronze, and many powdered oxides (particularly iron oxide and other oxides used as pigments), were used for decorative purposes in ceramics, as bases for paints and inks, and in cosmetics since the beginnings of recorded history. Powdered gold was used to illustrate some of the earliest manuscripts. It is not known how these powders were produced, but it is possible that some of the powders were obtained by granulation after the metal was melted. Low melting points and resistance to oxidation (tarnishing) favored such procedures, especially in the case of gold powder. The use of these powders for pigments and ornamental purposes is not true powder metallurgy, because the essential features of the modern art are the production of powder and its consolidation into a solid form by the application of pressure and heat at a temperature below the melting point of the major constituent. Early man learned by chance that particles of metal could be joined together by hammering, resulting in a solid metallic structure. In time, man learned how to build furnaces and develop temperatures high enough to melt and cast metals and to form lower melting alloys, such as copper and tin to make bronze. History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Earliest Developments Long before furnaces were developed that could approach the melting point of metal, P/M principles were used. About 3000 B.C., the Egyptians used a "sponge iron" for making tools. In this early process, iron oxide was heated in a charcoal and crushed shell fire, which was intensified by air blasts from bellows to reduce the oxide to a spongy metallic iron. The resulting hot sponge iron was then hammered to weld the particles together. Final shapes were obtained by simple forging procedures. Although the product often contained large amounts of nonmetallic impurities, some remarkably solid and sound structures have been discovered (Ref 1). W.D. Jones (Ref 2) wrote of a process modification developed by African tribes. After reduction, the sponge was broken into powder particles, washed, and sorted by hand to remove as much of the slag and gangue as possible. The powder was then either compacted or sintered into a porous material, which was subsequently forged. Another example of ancient reduction of iron oxide was carried out in the fabrication of the Delhi Pillar, which weighs 5.9 metric tons (6.5 tons). These crude forms of powder metallurgy ultimately led to the development of one of the commercial methods for producing iron powder. By grinding the sponge iron into fine particles, and heating in hydrogen to remove oxides and anneal or soften the particles, this process is today a viable technique for producing high-quality iron powder.

Powder metallurgy practices were used by the Incas and their predecessors in making platinum before Columbus made his voyage to the "New World" in 1492. The technique used was based on the cementing action of a lower melting binder, a technique similar to the present practice of making sintered carbides. The technique consisted of cementing platinum grains (separated from the ore by washing and selection) by the addition of an oxidation-resistant gold-silver alloy of a fairly low melting point to wet the grains, drawing them together by surface tension and forming a raw ingot suitable for further handling (Ref 3). A color change from the yellow of the sintered material to the whitish platinum of the final metal was caused by diffusion during heating prior to working. Heating is thought to have been accomplished by means of charcoal fires fanned by blowpipes. Analyses of these alloys vary considerably. The platinum content ranged from 26 to 72%, and the gold content ranged from 16 to 64%. Silver additions were found to vary from 3 to 15%, and amounts of copper up to 4% were traced.

References cited in this section

1. H.C.H. Carpenter and J.M. Robertson, The Metallography of Some Ancient Egyptian Implements, J. Iron Steel Inst., Vol 121, 1930, p 417-448 2. W.D. Jones, Fundamental Principles of Powder Metallurgy, London, 1960, p 593 3. P. Bergsöe, The Metallurgy and Technology of Gold and Platinum Among the Pre-Columbian Indians, Ing. Skrift. (A), Vol 44, 1937 History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Powder Metallurgy of Platinum The metallurgy of platinum, as practiced in the 18th and 19th centuries in Europe, is considered to be one of the most important stages of development for modern powder metallurgy. For the first time, complete records were available that provided insight into the various methods of powder production and the processing of these powders into solid, useful implements. Between 1750 and 1825, considerable attention was given to the manufacture of platinum. In 1755, Lewis (Ref 4) discovered that when a lead-platinum alloy was oxidized at high temperatures, a spongy, workable mass remained after lead oxide impurities had been volatilized. Scheffer (Ref 5) found that when platinum was heated with arsenic, the platinum showed signs of melting. This finding was confirmed in 1781 by Achard (Ref 6), who described the production of a fusible platinum-arsenic alloy, probably by forming the eutectic containing 87% Pt and melting at 600 °C (1110 °F). Achard formed solid platinum by hot hammering a sponge, welding the individual particles into a large solid. The sponge was obtained by high-temperature working of the platinum-arsenic alloy, which caused volatilization of the arsenic. This procedure formed the basis for a method of producing platinum that was first used in about 1790 in commercially manufactured chemical vessels by Jannetty in Paris. Mercury was used later in a similar process by von Mussin-Puschkin (Ref 7). Other metals worked in this way include palladium, by using sulfur instead of arsenic, and iridium (using phosphorus). Ridolfi (Ref 8) made malleable platinum for chemical vessels using lead and sulfur. In 1786, Rochon (Ref 9) successfully produced solid platinum without using arsenic by welding small pieces of scrap platinum. He produced malleable platinum by uniting purified platinum grains. Knight (Ref 10) found that if chemically precipitated platinum powder was heated at high temperatures in a clay crucible, it softened and could be compressed and forged. Tilloch (Ref 11) put platinum powder into tubes made of rolled platinum sheet, which were then heated and forged to produce a compact mass. In 1813, Leithner (Ref 12) reported production of thin, malleable platinum sheets by drying out successive layers of powder suspended in turpentine and heating the resulting films at high temperatures without pressure.

In 1882, a French process was reported by Baruel (Ref 13), in which 14 kg (30 lb) of platinum powder was made into a solid ingot by a series of operations. Platinum was precipitated in powdered form, slightly compressed in a crucible, and heated to white heat. The powder was then put in a steel matrix and put under pressure with a screw coining press. The compact platinum was repeatedly reheated and re-pressed until a solid ingot was formed. The final heat treatments were made in a charcoal fire at lower temperatures. Because the platinum powder was placed in the steel die while hot, this process was based on the hot pressing technique. In Russia in 1826, a high-temperature sintering operation was applied to previously compressed powder compacts on a commercial basis for the first time. This was in contrast to methods based on hot pressing. Sobolewskoy (Ref 14) described sifted platinum powder pressed into a cast iron cylinder that featured a steel punch actuated by a screw press. The resulting compacts were annealed for 1 days at high temperature in a porcelain firing kiln. The final product was highly workable, especially if the platinum powder had been well washed and was of high purity. Annealing, however, caused a decrease in volume; a cylinder 100 mm (4 in.) in diameter and 19 mm ( and 6 mm (

in.) in height shrank 19 mm (

in.)

in.) in these dimensions, respectively.

Another Russian method was reported by Marshall (Ref 15) in 1832. Platinum powder in a ring-shaped iron mold was pressed by a screw press, heated to a red heat, and re-pressed. After working in a rolling mill, the compacted discs were used as coins. The Wollaston process of producing compact platinum from platinum sponge powder is generally considered the

beginning of modern powder metallurgy. At least 16 years prior to his publication of 1829 (Ref 16), describing the manufacture of a product much superior to that of contemporary manufacturers, Wollaston devised the foundations for modern P/M technique. Wollaston was the first to realize all the difficulties connected with the production of solid platinum ingot from powdered metal, and thus concentrated on the preparation of the powder. He found that pressing the powder while wet into a hard cake (to be subsequently baked at red heat) was best done under considerable pressure. In addition, because available screw presses were not powerful enough, Wollaston developed a horizontal toggle press of the simple construction shown in Fig. 1. Wollaston used the following nine steps in the manufacture of compact platinum metal (Ref 17):

1. Precipitating ammonium-platinum-chloride from diluted solutions 2. Slowly decomposing the finely divided and carefully washed ammonium-platinum-chloride precipitate into loose sponge powder 3. Grinding this sponge powder without applying pressure to the powder particles, thus avoiding any burnishing of the particles and preserving all the surface energy of the particles 4. Sieving the sponge powder 5. Washing the sponge powder with water to remove all remnants of volatile salts 6. Separating fine particles from coarser particles through sedimentation (only the finest sponge particles were used) 7. Pressing the wet mass containing the finest platinum particles into a cylindrical cake 8. Drying the wet cake very slowly and then heating it to about 800 to 1000 °C (1475 to 1830 °F) 9. Forging the cake while it was still hot

Fig. 1 Simple toggle press used by Wollaston for making platinum powder compacts

By applying these steps, Wollaston succeeded in producing compact platinum, which when rolled into thin sheet was practically free of gas blisters. Crucibles made from this sheet were the best quality platinum implements of their time. Wollaston's process was used for more than a generation and became obsolete only with the advent of the platinum fusion procedure developed by Sainte-Claire Deville and Debray in 1859 (Ref 17). They succeeded in producing a powerful flame with illuminating gas and oxygen, the oxygen being manufactured from manganese dioxide. However, the fused metal which they produced was superior to Wollaston platinum in quality and homogeneity, and the fusion procedure was also less expensive and quicker than the Wollaston method. Fusion, therefore, was soon adopted by every platinum refinery. It is still considered the superior method for manufacturing standard-quality platinum.

References cited in this section

4. W. Lewis, Experimental Examination of a White Metallic Substance Said to Be Found in the Gold Mines of Spanish West Indies, Philos. Trans. R. Soc., Vol 48, 1755, p 638 5. H.T. Scheffer, Handlingar, Vol 13, 1752, p 269-275 6. K.F. Achard, Nouveaux Mem. Acad. R. Sci., Vol 12, 1781, p 103-119 7. A. von Mussiin-Puschkin, Allgem. J. Chem., Vol 4, 1800, p 411 8. C. Ridolfi, Quart. J. Sci. Lit. Arts, Vol 1, 1816, p 259-260 (From Giornale di scienza ed arti, Florence, 1816) 9. A. Rochon, J. Phys. Chem. Arts, Vol 47, 1798, p 3-15 (Rochon states that this was written in 1786 as part of his voyage to Madagascar) 10. R. Knight, A New and Expeditious Process for Rendering Platina Malleable, Philos. Mag., Vol 6, 1800, p 1-3 11. A. Tilloch, A New Process of Rendering Platina Malleable, Philos. Mag., Vol 21, 1805, p 175

12. Leithner, Letter quoted by A.F. Gehlen, J. Chem. Phys., Vol 7, 1813, p 309, 514 13. M. Baruel, Process for Procuring Pure Platinum, Palladium, Rhodium, Iridium, and Osmium from the Ores of Platinum, Quart. J. Sci. Lit. Arts, Vol 12, 1822, p 246-262 14. P. Sobolewskoy, Ann. Physik Chem., Vol 109, 1834, p 99 15. W. Marshall, An Account of the Russian Method of Rendering Platinum Malleable, Philos. Mag., Vol 11 (No. II), 1832, p 321-323 16. W.H. Wollaston, On a Method of Rendering Platina Malleable (Bakerian Lecture for 1828), Philos. Trans. R. Soc., Vol 119, 1829, p 1-8 17. J.S. Streicher, Powder Metallurgy, J. Wulff, Ed., American Society for Metals, 1942, p 16 History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Further Developments The use of P/M technology to form intricately shaped parts by pressing and sintering was introduced in the 19th century. In 1830, while determining the atomic weight of copper, Osann (Ref 18) found that the reduced metal could be sintered into a compact. Osann then developed a process for making impressions of coins from copper powder produced by the reduction of precipitated copper carbonate (Cu2CO3). Osann found that reduction was best done at the lowest possible temperatures that could be used to produce a metal powder of the fineness known in platinum manufacture. High reduction temperatures resulted in granular masses that did not sinter well. Contamination of the powder by the atmosphere was eliminated by using the powder immediately after reduction or storing it in closed glass bottles. The powder was separated into three grades, determined by particle size, before use. To make an impression of a coin, fine powder was sprayed on the surface, followed by layers of coarser grades. The powder and a die were placed in a ring-shaped mold and compressed by the pressure of hammer blows on a punch or use of a knuckle press. Volume of the copper powder was reduced to one-sixth of the original powder during compression. Sintering was done at temperatures close to the melting point of copper, after the compacts were placed in airtight copper packets sealed with clay. A nondistorted 20% shrinkage occurred, but the sintered copper was harder and stronger than cast copper. Osann also produced medals of silver, lead, and copper by the same procedure. Although he considered his process especially suitable as an alternative to the electrotype method of reproducing coins and medallions, Osann advocated its use as an initial production method for these articles. He believed powder metallurgy could be used for producing printing type and for making convex and concave mirrors by pressing on glass. Osann thought that measurement of the shrinkage of copper compacts could be used to calculate temperature, as the shrinkage of clay cylinders was used in the Wedgewood pyrometer. Among the advancements in the P/M industry during the second half of the 19th century were Gwynn's attempts to develop bearing materials from metal powders. Patents issued to Gwynn in 1870 (Ref 19) were the forerunners of a series of developments in the area of self-lubricating bearings. Gwynn employed a mixture of 99 parts of powdered tin, prepared by rasping or filing, and 1 part of petroleum-still residue. The two constituents were stirred while being heated. A solid form of desired shape was then produced by subjecting the mixture to extreme pressure while enclosing it in a mold. The patent specifically states that journal boxes made by this method or lined with material thus produced would permit shafts to run at high speeds without using any other lubrication.

References cited in this section

18. G. Osann, Ann. Physik Chem., Vol 128, 1841, p 406 19. U.S. Patents 101,863; 101,864; 101,866; and 101,867, 1870

History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Commercial Developments The first commercial application of powder metallurgy occurred when carbon, and later osmium, zirconium, vanadium, tantalum, and tungsten, was used for incandescent lamp filaments. Methods were developed from 1878 to 1898 for making carbon filaments by the extrusion and subsequent sintering of carbonaceous materials. Osmium filaments were used for a short time from 1898 to 1900. Auer von Welsbach (Ref 20) described the production of filaments of osmium by chemical precipitation of the powder and formation of a mixture with sugar syrup, which served both as binder and, if osmium oxide powder was used instead of the metal, as reducing agent as well. The mixture was squirted through fine dies, and the resulting fine threads were subsequently fired in protective atmospheres to carburize and volatilize the binder, reduce the oxide, and sinter the metal particles into a coherent metallic wire for use as an electrical conductor. The osmium electric lamp was soon succeeded by tantalum filament lights, which were used widely from 1903 to 1911. The general procedure (Ref 21) was similar to that used for osmium, with the exception that tantalum had to be purified by a vacuum treatment to become ductile. Similar techniques were used for the production of filaments from zirconium, vanadium, and tungsten; with tungsten, especially, extruded wires were bent into hairpin shapes before sintering to shape them for use as filaments. Because lack of ductility was the major shortcoming of these filaments, attempts were made to improve this property by the addition of a few percent of a lower-melting, ductile metal. Tungsten powder was mixed with 2 to 3% Ni, pressed into a compact, and sintered in hydrogen at a temperature slightly below the melting point of nickel. The resulting bars could be drawn, and nickel was removed from the final filaments by a vacuum heat treatment at a high temperature (Ref 22). Although this process was not commercially successful, it was an important step toward the industrial development of cemented carbides and composite materials. Tungsten was soon recognized as the best material for lamp filaments. The problem, however, was to devise an economical procedure for producing these filaments in large quantities. A number of procedures to produce powdered tungsten had been worked out earlier. In 1783, the D'Elhujar brothers (Ref 23) first produced tungsten powder by heating a mixture of tungstic acid and powdered charcoal, cooling the mixture, and removing the small cake, which crumbled to a powder of globular particles. The purification of tungsten powder by boiling, scrubbing, and skimming to remove soluble salts, iron oxide, clay, and compounds of calcium and magnesium was reported by Polte (Ref 24). Coolidge Process. At the beginning of the 20th century, Coolidge (Ref 25) made the important discovery that tungsten could be worked in a certain temperature range and would retain its ductility at room temperature. Few changes have been made over the years on the Coolidge procedure; it is still the standard method of producing incandescent lamp filaments. In this method, very fine tungsten oxide powder, WO3, is reduced by hydrogen. The powder is pressed into compacts, which are presintered at 1200 °C (2190 °F) to strengthen them so that they can be clamped into contacts. They receive a final sintering treatment near 3000 °C (5430 °F) by passing a low-voltage, high-current density current through the compacts. During sintering, the compacts shrink and reach a density near 90% that of solid tungsten. The sintered compacts can be worked only at temperatures near 2000 °C (3630 °F). When heated to this temperature, they can be swaged into rounds. With increasing amounts of warm work, tungsten becomes more ductile, the swaging temperature can be progressively lowered, and the swaged bars can be drawn into fine wire at relatively low temperatures. Other Refractory Metals. The procedures developed for the production of tungsten often were adaptable to the

manufacture of molybdenum. Lederer (Ref 26) developed a method of making molybdenum using powdered molybdenum sulfide. The sulfide, mixed with amorphous sulfur and kneaded into a paste, was formed into a filament. When exposed to air, the filaments became strong enough to be placed in a furnace. Heating in hydrogen resulted in formation of hydrogen sulfide and sintering of the metal into solid filaments. A similar process was patented by Oberländer (Ref 27), who used molybdenum chloride and other halides as starting materials. When the chloride was treated with a reducing agent such as ether, a paste was obtained. Tungsten, molybdenum, and tantalum are the three most important refractory metals used today in the lamp, aerospace, electronics, x-ray, and chemical industries. Other refractory metals of minor significance were developed by the P/M

method in the early 1900s, notably niobium, thorium, and titanium. However, at the same time another development, originating in refractory metal processing, took form and rapidly grew to such importance that it far overshadows the parent field. Cemented carbides have become one of the greatest industrial developments of the century. Cemented Carbides. Ordinary drawing dies were unsatisfactory for drawing tungsten wires and filaments. The need

for a harder material to withstand greater wear became urgent. Because it was known that tungsten granules combined readily with carbon at high temperatures to give an extremely hard compound, this material was used as the basis for a very hard, durable tool material known as cemented carbide. The tungsten carbide particles, present in the form of finely divided, hard, strong particles, are bonded into a solid body with the aid of a metallic cementing agent. Early experiments with a number of metals established that this cementing agent had to possess the following properties to permit solidification of the hard metal body: • • • •

Close chemical affinity for the carbide particles A relatively low melting point Limited ability to alloy with the carbide Great ductility (not to be impaired by the cementing operation)

Cobalt satisfied these requirements most closely. The early work was carried out mainly in Germany by Lohmann and Voigtländer (Ref 28) in 1914, by Liebmann and Laise (Ref 29) in 1917, and by Schröter (Ref 30) from 1923 to 1925. Krupp (Ref 31) perfected the process in 1927 and marketed the first product of commercial importance, "Widia." In 1928 this material was introduced to the United States, and the General Electric Company, which held the American patent rights, issued a number of licenses. The process entails carefully controlled powder manufacture, briquetting a mixture of carbide and metallic binder (usually 3 to 13% Co), and sintering in a protective atmosphere at a temperature high enough to allow fusion of the cobalt and partial alloying with the tungsten carbide. The molten matrix of cobalt and partly dissolved tungsten carbide forms a bond, holding the hard particles together and giving the metallic body sufficient toughness, ductility, and strength to permit its effective use as tool material. Composite Metals. The next development in powder metallurgy was the production of composite metals used for

heavy-duty contacts, electrodes, counterweights, and radium containers. All of these composite materials contain refractory metal particles, usually tungsten, and a cementing material with a lower melting point, present in various proportions. Copper, copper alloys, and silver are frequently used; cobalt, iron, and nickel are used less frequently. Some combinations also contain graphite. The first attempt to produce such materials was recorded in the patent of Viertel and Egly (Ref 32) issued shortly after 1900. The procedures used either were similar to those developed for the hard metals (Ref 33) or called for introduction of the binder in liquid form by dipping or infiltration. In 1916, Gebauer (Ref 34) developed such a procedure, which was developed further by Baumhauer (Ref 35) and Gillette (Ref 36) in 1924. Pfanstiehl (Ref 37) obtained patent protection in 1919 for a heavy metal, consisting of tungsten and a binder that contained copper and nickel. Porous Metal Bearings and Filters. In addition to the development of refractory metals and their carbides, another

important area of powder metallurgy that gained attention during the early 1900s was that of porous metal bearings. Special types of these porous bearings are referred to as self-lubricating. The modern types of bearings, usually made of copper, tin, and graphite powders and impregnated with oil, were first developed in processes patented by Loewendahl (Ref 38) and Gilson (Ref 39 and 40). Gilson's material was a bronze structure, in which finely divided graphite inclusions were uniformly distributed. It was produced by mixing powdered copper and tin oxides with graphite, compressing the mixture, and heating it to a temperature at which the oxides were reduced by the graphite and the copper and tin could diffuse sufficiently to give a bronzelike structure. Excess graphite (up to 40 vol%) was uniformly distributed through this structure. The porosity was sufficient to allow for the introduction of at least 2% oil. The process was later improved by Boegehold and Williams (Ref 41), Claus (Ref 42), and many others, primarily by utilization of elemental metal powders rather than oxides. Metallic filters were the next stage in the development of these porous metals, and patents date back as far as 1923 (Ref 43), when Claus patented a process and machine to mold porous bodies from granular powder.

References cited in this section

20. U.S. Patent 976,526, 1910 21. U.S. Patents 899,875, 1908 and 912,246, 1909 22. C.R. Smith, Powder Metallurgy, J. Wulff, Ed., American Society for Metals, 1942, p 4 23. A.W. Deller, Powder Metallurgy, J. Wulff, Ed., American Society for Metals, 1942, p 582 24. U.S. Patent 735,293, 1903 25. U.S. Patent 963,872, 1910 26. U.S. Patent 1,079,777, 1913 27. U.S. Patent 1,208,629, 1916 28. German Patents 289,066, 1915; 292,583, 1916; 295,656, 1916; 295,726, 1916. Swiss Patents 91,932 and 93,496, 1919 29. U.S. Patents 1,343,976 and 1,343,977, 1920 30. German Patent 420,689, 1925. U.S. Patent 1,549,615,1925 31. British Patents 278,955, 1927, and 279,376, 1928. Swiss Patent 129,647, 1929. U.S. Patent 1,757,846, 1930 32. U.S. Patent 842,730, 1907 33. U.S. Patents 1,418,081, 1922; 1,423,338, 1922; and 1,531,666, 1925 34. U.S. Patent 1,223,322, 1917 35. U.S. Patent 1,512,191, 1924 36. U.S. Patent 1,539,810, 1925 37. U.S. Patent 1,315,859, 1919 38. U.S. Patent 1,051,814, 1913 39. U.S. Patent 1,177,407, 1916 40. E.G. Gilson, General Electric Rev., Vol 24, 1921, p 949-951 41. U.S. Patents, 1,642,347, 1927; 1,642,348, 1927; 1,642,349, 1927; and 1,766,865, 1930 42. U.S. Patent 1,648,722, 1927 43. U.S. Patent 1,607,389, 1926 History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Post-War Developments Infiltration techniques, porous materials, iron powder cores for ratio tuning devices, P/M permanent magnets, and W-CuNi heavy metal compositions were developed during the periods between 1900, World War I, and the late 1920s. At the beginning of World War II in Europe, iron powder technology began its advance to commercial viability. The most spectacular development of iron parts made by powder metallurgy was during World War II in central Europe, where paraffin-impregnated sintered iron driving bands for military projectiles were extensively used. German powder metallurgists found this technique effective as a substitute for scarce gilding metal, a copper-zinc alloy containing 5 to 10% Zn. Production reached a peak of 3175 metric tons (3500 tons) per month for this application. The advent of mass production in the automotive industry made possible the use of iron and copper powders in large tonnages and spawned many of the technological advances of the modern P/M industry. The automobile has been the basis for most industrial applications of P/M, even in fields unrelated to the automotive industry. The first commercial application of a P/M product, the self-lubricating bearing, was used in an automobile in 1927. It was made from a combination of copper and tin powders to produce a porous bronze bearing capable of retaining oil within its pores by capillary attraction. At about the same time, self-lubricating bearings were introduced to the home appliance market as a refrigerator compressor component.

Through the 1940s and early 1950s, copper powder and the self-lubricating bearing were the principal products of powder metallurgy. Since then, iron powder and steel P/M mechanical components such as gears, cams, and other structural shapes have become dominant. While copper powder remains an important P/M material, consumed on the order of 21,000 metric tons (23,000 tons) per year, it is overshadowed by iron and iron-base powders with markets of 318,000 metric tons (350,000 tons) per year. Since the end of World War II, and especially with the advent of aerospace and nuclear technology, developments have been widespread with regard to the powder metallurgy of refractory and reactive metals such as tungsten, molybdenum, niobium, titanium, and tantalum and of nuclear metals such as beryllium, uranium, zirconium, and thorium. All of the refractory metals are recovered from their ores, processed, and formed using P/M techniques. With the reactive metals, powder metallurgy is often used to achieve higher purity or to combine them with other metals or nonmetallics to achieve special properties. Nuclear power plants use fuel elements often made by dispersing uranium oxide in a metal powder (aluminum, for example) matrix. The control rods and neutron shielding may use boron powder in a matrix of nickel, copper, iron, or aluminum. Tungsten combined with nickel and copper powders is used widely as a shielding component in applications where intricate configuration involving machining is required, such as in cobalt-60 containers. In aerospace, beryllium and titanium are used extensively. Rocket skirts, cones, and heat shields are often formed from niobium. Molybdenum is widely used in missile and rocket engine components. Nozzles for rockets used in orbiting space vehicles often are made from tungsten via the P/M process in order to maintain critical dimensional tolerances. The 1950s and 1960s witnessed the emergence of P/M wrought products. These are fully dense metal systems that began as powders. Hot isostatically pressed superalloys, P/M forgings, P/M tool steels, roll compacted strip, and dispersionstrengthened copper are all examples. Each of these processes and materials is covered in separate articles in this Volume. The commercialization of powder-based high-performance material emerged as a major breakthrough in metalworking technology in the 1970s by opening up new markets through superior performance, coupled with the cost effectiveness of material conservation and longer operational life. History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Recent Developments In the late 1970s, the experimental programs involving P/M wrought products began spilling over into the commercial industrial sector, principally in the form of P/M tool steels and P/M forgings. With the advent of P/M forgings, no longer were properties compromised by density. Fully dense components capable of combining the alloying flexibility and the net and near-net design features of powder metallurgy were very marketable. The later 1970s and early 1980s witnessed a significant metallurgical breakthrough in the recognition of P/M techniques for eliminating segregation and ensuring a fully homogeneous, fine-grained, pore-free, high-alloy structure. Categorized as P/M wrought metals, they led to the perfection of extremely high-purity metal powders and improved consolidation techniques such as hot isostatic pressing (HIP). The 1980s also saw the commercialization of ultrarapid solidification and injection molding technology. Both of these developments are also covered in separate articles in this Volume. Commercial powder metallurgy now spans the density spectrum from highly porous metal filters through self-lubricating bearings and P/M parts with controlled density to fully dense P/M wrought metal systems. The P/M parts and products industry in North America has estimated sales of more than $3 billion. It comprises 150 companies that make conventional P/M parts and products from iron- and copper-base powders and about 50 companies that make specialty P/M products such as superalloys, tool steels, porous products, friction materials, strip for electronic applications, highstrength permanent magnets, magnetic powder cores and ferrites, tungsten carbide cutting tools and wear parts, rapid solidification rate (RSR) products, and metal injection molded parts and tool steels. Powder metallurgy is international in scope with growing industries in all of the major industrialized countries. The value of U.S. metal powder shipments (including paste and flake) was $1.854 billion in 1995. Annual worldwide metal powder production exceeds 1 million tons. Trends and new developments include:

•

• • •

Improved manufacturing processes such as HIP, P/M forging, metal injection molding (MIM), and direct powder rolling through increased scientific investigation of P/M technology by government, academic, and industrial research and development programs Fully dense P/M products for improved strength properties and quality in automobiles, diesel and turbine engines, aircraft parts, and industrial cutting and forming tools Commercialization of technologies such as MIM, rapid solidification, P/M forging, spray forming, hightemperature vacuum sintering, warm compacting, and both cold and hot isostatic pressing The use of P/M hot-forged connecting rods in automobiles and a P/M camshaft for four- and eightcylinder automobile engines. The use of P/M composite camshafts in automotive engines and main bearing caps

A review of major historical developments in powder metallurgy is presented in Table 1. History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Powder Metallurgy Literature A number of literary works are worthy of mention in connection with the background of powder metallurgy. One of the earliest works of significance was Principles of Powder Metallurgy by W.D. Jones, published in 1937 in England (Ref 44). It was updated in 1960 and published as Fundamental Principles of Powder Metallurgy (Ref 45). The first Russian publication was by Bal'shin (Ref 46) and appeared in 1938; the first comprehensive text in German, Pulvermetallurgie und Sinterwerkstoffe, was published by R. Kieffer and W. Hotop in 1943 (Ref 47). In the United States, the first publication was by H.H. Hausner in 1947 (Ref 48), followed closely by P. Schwarzkopf (Ref 49). Two years later, the first of four volumes of a treatise on powder metallurgy, a major work by C.G. Goetzel (Ref 50), was published. Some current "Selected References" on powder metallurgy science and technology are listed at the end of this article.

References cited in this section

44. W.D. Jones, Principles of Powder Metallurgy, Arnold, London, 1937 45. W.D. Jones, Fundamental Principles of Powder Metallurgy, Arnold, London, 1960 46. M.Y.J. Bal'shin, Metal Ceramics, Gonti, 1938 (in Russian) 47. R. Kieffer and W. Hotop, Pulvermetallurgie und Sinterwerkstoffe, Springer, 1943; Re-issue Springer, 1948 48. H.H. Hausner, Powder Metallurgy, Chemical Publishing Co., 1947 49. P. Schwarzkopf, Powder Metallurgy, Macmillan, 1947 50. C.G. Goetzel, Treatise on Powder Metallurgy, Vol 1-4, Interscience, 1949 History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Powder Metallurgy Trade Associations The advancement of powder metallurgy from a laboratory curiosity to an industrial technology has been influenced greatly by various professional societies and the P/M trade association, whose annual technical conference proceedings chronicle the maturing of the technology. In 1944, an organization called the Metal Powder Association was founded by a

group of metal powder producers in the United States. It was reorganized in 1958 as the Metal Powder Industries Federation, a trade association whose representation embraced the commercial and technological interests of the total metal powder producing and consuming industries. International in scope, the Federation consists of the following autonomous associations, which together represent the primary elements of the P/M and particulate materials industries: • • • • •

•

Powder Metallurgy Parts Association: Members are companies that manufacture P/M parts for sale on the open market. Metal Powder Producers Association: Members are producers of metal powders in any form for any use. Powder Metallurgy Equipment Association: Members are manufacturers of P/M processing equipment and supplies, including compacting presses, sintering furnaces, belts, tools and dies, and atmospheres. Refractory Metals Association: Members are manufacturers of powders or products from tungsten, molybdenum, tantalum, niobium, and cobalt. Advanced Particulate Materials Association (APMA):Members are companies that use P/M or other related processes to produce any of a wide variety of materials not covered by the other MPIF associations as well as companies that have proprietary P/M parts manufacturing facilities. It also includes emerging technologies that use the powders as precursors in manufacturing processes. Metal Injection Molding Association (MIMA): Members are international companies that use the metal or ceramic injection molding process to form parts.

MPIF also has both Overseas and Affiliate/Consultant classes of membership. The Federation generates industry statistics, process and materials standards, industrial public relations and market development, government programs, research, and various educational programs and materials. The technology's "professional" society is APMI International. As distinguished from the Federation, APMI members are individuals, not companies. Members are kept informed of developments in P/M technology through local section activities, conferences, and publications, including the International Journal of Powder Metallurgy and Powder Technology. It is the only professional society organized specifically to serve the powder metallurgist and the P/M industry. Many of the major professional societies are also active in powder metallurgy, usually through committees working on standards, conferences, or publications. This includes the ASM International, the Metallurgical Society, SAE, the American Society for Testing and Materials, and the Society of Manufacturing Engineers. History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

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

H.C.H. Carpenter and J.M. Robertson, The Metallography of Some Ancient Egyptian Implements, J. Iron Steel Inst., Vol 121, 1930, p 417-448 W.D. Jones, Fundamental Principles of Powder Metallurgy, London, 1960, p 593 P. Bergsöe, The Metallurgy and Technology of Gold and Platinum Among the Pre-Columbian Indians, Ing. Skrift. (A), Vol 44, 1937 W. Lewis, Experimental Examination of a White Metallic Substance Said to Be Found in the Gold Mines of Spanish West Indies, Philos. Trans. R. Soc., Vol 48, 1755, p 638 H.T. Scheffer, Handlingar, Vol 13, 1752, p 269-275 K.F. Achard, Nouveaux Mem. Acad. R. Sci., Vol 12, 1781, p 103-119 A. von Mussiin-Puschkin, Allgem. J. Chem., Vol 4, 1800, p 411

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. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.

C. Ridolfi, Quart. J. Sci. Lit. Arts, Vol 1, 1816, p 259-260 (From Giornale di scienza ed arti, Florence, 1816) A. Rochon, J. Phys. Chem. Arts, Vol 47, 1798, p 3-15 (Rochon states that this was written in 1786 as part of his voyage to Madagascar) R. Knight, A New and Expeditious Process for Rendering Platina Malleable, Philos. Mag., Vol 6, 1800, p 1-3 A. Tilloch, A New Process of Rendering Platina Malleable, Philos. Mag., Vol 21, 1805, p 175 Leithner, Letter quoted by A.F. Gehlen, J. Chem. Phys., Vol 7, 1813, p 309, 514 M. Baruel, Process for Procuring Pure Platinum, Palladium, Rhodium, Iridium, and Osmium from the Ores of Platinum, Quart. J. Sci. Lit. Arts, Vol 12, 1822, p 246-262 P. Sobolewskoy, Ann. Physik Chem., Vol 109, 1834, p 99 W. Marshall, An Account of the Russian Method of Rendering Platinum Malleable, Philos. Mag., Vol 11 (No. II), 1832, p 321-323 W.H. Wollaston, On a Method of Rendering Platina Malleable (Bakerian Lecture for 1828), Philos. Trans. R. Soc., Vol 119, 1829, p 1-8 J.S. Streicher, Powder Metallurgy, J. Wulff, Ed., American Society for Metals, 1942, p 16 G. Osann, Ann. Physik Chem., Vol 128, 1841, p 406 U.S. Patents 101,863; 101,864; 101,866; and 101,867, 1870 U.S. Patent 976,526, 1910 U.S. Patents 899,875, 1908 and 912,246, 1909 C.R. Smith, Powder Metallurgy, J. Wulff, Ed., American Society for Metals, 1942, p 4 A.W. Deller, Powder Metallurgy, J. Wulff, Ed., American Society for Metals, 1942, p 582 U.S. Patent 735,293, 1903 U.S. Patent 963,872, 1910 U.S. Patent 1,079,777, 1913 U.S. Patent 1,208,629, 1916 German Patents 289,066, 1915; 292,583, 1916; 295,656, 1916; 295,726, 1916. Swiss Patents 91,932 and 93,496, 1919 U.S. Patents 1,343,976 and 1,343,977, 1920 German Patent 420,689, 1925. U.S. Patent 1,549,615,1925 British Patents 278,955, 1927, and 279,376, 1928. Swiss Patent 129,647, 1929. U.S. Patent 1,757,846, 1930 U.S. Patent 842,730, 1907 U.S. Patents 1,418,081, 1922; 1,423,338, 1922; and 1,531,666, 1925 U.S. Patent 1,223,322, 1917 U.S. Patent 1,512,191, 1924 U.S. Patent 1,539,810, 1925 U.S. Patent 1,315,859, 1919 U.S. Patent 1,051,814, 1913 U.S. Patent 1,177,407, 1916 E.G. Gilson, General Electric Rev., Vol 24, 1921, p 949-951 U.S. Patents, 1,642,347, 1927; 1,642,348, 1927; 1,642,349, 1927; and 1,766,865, 1930 U.S. Patent 1,648,722, 1927 U.S. Patent 1,607,389, 1926 W.D. Jones, Principles of Powder Metallurgy, Arnold, London, 1937 W.D. Jones, Fundamental Principles of Powder Metallurgy, Arnold, London, 1960

46. 47. 48. 49. 50.

M.Y.J. Bal'shin, Metal Ceramics, Gonti, 1938 (in Russian) R. Kieffer and W. Hotop, Pulvermetallurgie und Sinterwerkstoffe, Springer, 1943; Re-issue Springer, 1948 H.H. Hausner, Powder Metallurgy, Chemical Publishing Co., 1947 P. Schwarzkopf, Powder Metallurgy, Macmillan, 1947 C.G. Goetzel, Treatise on Powder Metallurgy, Vol 1-4, Interscience, 1949

History of Powder Metallurgy Revised by Donald G. White, Metal Powder Industries Federation and APMI International

Selected References • • • • •

M.E. Fayed and L. Oteen, Ed., Handbook of Science & Technology, Chapman & Hall, 1997 R.M. German, Powder Metallurgy Science, Metal Powder Industries Federation, 1994 A. Lawley, The Production of Metal Powders, Metal Powder Industries Federation, 1992 F. Thümmler and R. Oberacker, An Introduction to Powder Metallurgy, I. Jenkins and J.V. Wood, Ed., The Institute of Materials, 1993 A.J. Yule and J.J. Dunkley, Atomization of Melts for Powder Production and Spray Deposition, Oxford Science Publications, 1994

Powder Metallurgy Methods and Design* Howard I. Sanderow, Management & Engineering Technologies

Introduction THE POWDER METALLURGY (P/M) process is a near-net or net-shape manufacturing process that combines the features of shape-making technology for powder compaction with the development of final material and design properties (physical and mechanical) during subsequent densification or consolidation processes (e.g., sintering). It is critical to recognize this interrelationship at the outset of the design process because a subtle change in the manufacturing process can cause a significant change in material properties.

Note

* Adapted from article in Materials Selection and Design, Vol 20, ASM Handbook, 1997, p 745-753 Powder Metallurgy Methods and Design* Howard I. Sanderow, Management & Engineering Technologies

General P/M Design Considerations To begin a design using powder processing, six key design considerations must be recognized. With the variety of powder processing schemes available, the selection of the appropriate method depends to a great extent on these design constraints.

Size. Due to the physical nature of the processes and the physical limits of commercial manufacturing equipment,

product size has certain critical boundaries. For some powder processes, the product size is quite limited (such as metal injection molding, MIM), while for hot-isostatic pressing (HIP), size is not considered a serious constraint. Shape Complexity. Powder metallurgy is a flexible process capable of producing complex shapes. The ability to

develop complex shapes in powder processing is determined by the method used to consolidate the powders. Because a die or mold provides the container for the consolidation step, the ease of manufacture of the container and the ability to remove a green compact (unsintered) from the container, in most cases, determines the allowable shape complexity of a given part. Tolerances. Control of dimensional tolerances, a demanding feature of all near-net or net-shape manufacturing

processes, is a complex issue in powder processing. Tolerances are determined by such process parameters as powder characteristics, compaction parameters, and the sinter cycle. The amount of densification during sintering and the uniformity of that shrinkage controls dimensional tolerance in most P/M products. Due to the very small amount of size change during sintering of conventional press-and-sinter P/M parts, these products typically have the closest dimensional tolerances, as compared to HIP parts, which require the largest spread in tolerances. Material Systems. Powder shape, size, and purity are important factors in the application of a powder processing technique. For some consolidation processes or steps, powders must be smooth, spherical particles, but for other processes a much more irregular powder shape is required. Nearly every material and alloy system is available in powder form. For some materials such as cemented carbides, copper-tungsten composites and the refractory metals (tungsten, molybdenum, tantalum, etc.), powder processing is the only commercially viable manufacturing process.

As an example, for "press and sinter" processing, an irregular powder shape and distribution of particle sizes are desired for adequate green strength and sinter response. Hot isostatic pressing requires spherical powders (gas atomized) for lowest impurities and good particle packing. The MIM process also prefers spherical particles, but very small particle size (10 to 20 m) is needed to ensure proper rheology, homogeneous distribution in the plastic binder, and excellent sinter response. Properties. The functional response of any product is determined by its physical or mechanical properties. In powder processing these properties are influenced directly by the product density, the raw material (powder), and the processing conditions (most often the sintering cycle). As P/M materials deviate from full density, the properties decrease (as shown for the tensile properties--and electrical conductivity--of pure copper in Fig. 1). The mechanical response for 4% Ni steels is found in Fig. 2.

Fig. 1 Properties of pure copper. Source: Ref 1

Fig. 2 Effects of density on mechanical properties of as-sintered 4% Ni steel. Source: Ref 2

Quantity and Cost. The economic feasibility of P/M processing is typically a function of the number of pieces being

produced. For conventional press-and-sinter processing, production quantities of at least 1,000 to 10,000 pieces are desired in order to amortize the tooling investment. In contrast, isostatic processing can be feasible for much lower quantities, in some cases as small as 1 to 10 pieces. On a per pound basis, the approximate costs for steel P/M parts produced by various methods are roughly as follows:

Condition

Density range, g/cm3

1997 selling price(a), $/lb

Pressed and sintered

6.0-7.1

2.45-2.70

Pressed, sintered, sized

6.0-7.1

2.90-3.20

Copper infiltrated

7.3-7.5

3.50-3.55

Warm formed

7.2-7.4

3.10-3.30

Double pressed and sintered

7.2-7.4

4.00-4.10

Metal injection molded

7.5-7.6

45.0-70.0

Hot forged

7.8

5.00-5.50

Double press and sinter + HIP

7.87

6.00-7.00

(a) These numbers are only averages; smaller parts are more expensive and larger parts less expensive per pound.

References cited in this section

1. F.V. Lenel, Powder Metallurgy--Principles and Applications, Metal Powder Industries Federation, 1980, p 426 2. L.F. Pease III and V.C. Potter, Mechanical Properties of P/M Materials, Powder Metallurgy, Vol 7, ASM Handbook (formerly Metals Handbook, 9th ed.), American Society for Metals, 1984, p 467 Powder Metallurgy Methods and Design* Howard I. Sanderow, Management & Engineering Technologies

Powder Processing Techniques In order to understand the design restrictions of each powder processing method, it is best to review these processes individually. The P/M manufacturing methods can be divided into two main categories: conventional press-and-sinter methods and full-density processes. Conventional (Press-and-Sinter) Processes. The conventional press-and-sinter process technologies follow the

steps outlined in Fig. 3. The various powder ingredients are selected to satisfy the process constraints and still meet the requirements of the end product. For example, in cold compaction irregularly shaped powders are used to ensure adequate green strength and structural integrity of the as-pressed product. Special solid lubricants are added to the powder blend to reduce friction between the powder particles and the tooling. If these lubricants might contaminate the metal powder particles, then an alternate consolidation method would be needed.

Fig. 3 General steps in the P/M process

Because powder is compacted in hard tooling using a vertical compaction motion, the product size and shape are limited by the constraints of available press capacity, powder compressibility, and the density level required in the product. For most conventional P/M products these limitations have a maximum size of about 160 cm2 (25 in.2) compaction area, part thickness of about 75 mm (3 in.), and a weight of 2.2 kg (5 lb). However, parts as large as 200 mm (8 in.) diameter by 100 mm (4 in.) thick, weighing 14.5 kg (32 lb), have been produced on conventional equipment. Even parts 380 mm (15 in.) in diameter by 6 mm (

in.) thick have been produced by conventional P/M methods.

After compaction the green compact is sintered in a controlled-atmosphere furnace. Dimensional tolerance control is determined by the maximum temperature of the sintering cycle and the metallurgical changes that occur during sintering. If solid-state diffusion is the primary sintering mechanism, very little densification occurs, dimensional change is minimal, and tolerance control is very good. This practice is followed for most P/M steels where size change during sintering is held to less than 0.3%. In contrast, other alloy systems utilize liquid-phase formation as the primary sintering mechanism, causing a significant increase in density, large dimensional changes, and much lower tolerance control. Examples of these material systems include cemented carbides where dimensional changes of 6 to 8% are typical and tolerance control is in the range of ±0.25 mm (±0.010 in.). In addition to its effects on dimensional tolerance levels, the sintering step also plays a significant role in determining the final physical and mechanical properties of the product. Higher sintering temperatures and longer sintering times promote pore rounding and increase densification, thereby improving critical mechanical properties such as tensile strength, ductility, impact resistance, and fatigue limit (see Table 1). The sintering process is extremely important in determining the magnetic response of soft magnetic P/M alloys. As shown in Table 2 for the Fe-0.45 wt% P alloy, increasing the hydrogen content in the sintering atmosphere and raising the

sintering temperature improved the maximum permeability more than 100%, the tensile strength more than 15%, and the ductility more than 300%. In a similar manner, the mechanical properties and corrosion resistance of P/M stainless steels are strongly dependent on the sintering process parameters (Ref 5). Table 1 Effect of sintering conditions on the mechanical properties of two P/M nickel steels MPIF FN0205(b)

MPIF FN0208(c)

Belt(d)

Vacuum(e)

Belt(d)

Vacuum(e)

Tensile strength, MPa (ksi)

380 (55)

552 (80)

448 (65)

758 (110)

Yield strength, MPa (ksi)

193 (28)

414 (60)

331 (48)

586 (85)

Elongation, %

4

7

2

4

Impact energy, J (ft · lbf)

19 (14)

38 (28)

11 (8)

33 (24)

Hardness, HRB

64

80

80

90

Sintered density, g/cm3

(f)

7.32

(f)

7.30

(a) All samples pressed to a green density of 7.2 g/cm3.

(b) 1-3% Ni, 0.3-0.6% C.

(c) 1-3% Ni, 0.6-0.9% C.

(d) Belt, 30 min at 1125 °C (2060 °F) in nitrogen/endo atmosphere.

(e) Vacuum, 2 h at 1260 °C (2300 °F) with nitrogen backfill.

(f) Density not reported but estimable by MPIF Standard 35.

Table 2 Effect of sintering conditions on the properties of magnetic P/M iron (0.45 wt% P) Sintering conditions

Atmosphere(a)

Temperature, °C (°F)

Maximum magnetic induction (Bmax), kG

Coercive force (Hc), Oe

Maximum permeability

Tensile strength

MPa

ksi

Elongation, %

10% H2

1120 (2050)

13.2

2.3

2620

345

50

3

75% H2

1120 (2050)

13.3

2.0

3220

355

52

7

100% H2

1120 (2050)

13.4

1.7

3680

372

54

5

100% H2

1200 (2200)

13.7

1.3

5710

400

58

14

Source: Ref 4 (a) Balance N2.

Warm compaction is used to increase the green density and green strength of P/M steel parts. When combined with hightemperature sintering, this process can provide mechanical properties equivalent to double press-double sinter processing at a lower cost. Due to the much higher green strength, warm compacted parts can be machined in the green condition. This technique can also be used to produce insulated magnetic cores, a composite material suitable for high-frequency electromagnetic systems. Full-Density Processes. The second group of powder process technologies are formulated specifically to yield a

product as close to full density as possible. This contrasts significantly with the previous conventionally processed products where attainment of full density was not the primary goal. The full-density processes include powder forging (P/F), metal injection molding (MIM), hot isostatic pressing (HIP), roll compaction, hot pressing and extrusion. Powder Forging. In P/F a preform is manufactured using conventional P/M process techniques and then hot formed in

confined dies to cause sufficient material deformation so that nearly all the porosity is eliminated. Due to the high costs in developing the preform design and maintaining forging tools and automated production systems of the P/F process, it has been limited, in most commercial practices, to high-volume products such as automotive connecting rods and transmission components. The P/F process has been successful in developing mechanical properties in P/F steel comparable to wrought steels (see Table 3). This process successfully overcomes the mechanical property limitations imposed by the residual porosity in conventional P/M products. Table 3 Properties of powder forged steels Alloy

Hardness, HRC

Tensile strength

Yield strength

MPa

ksi

MPa

ksi

Elongation, %

Impact toughness

J

ft · lbf

10C60

23

793

115

690

100

11

2.7

2

11C60

28

895

130

620

90

11

4

3

4620

28

965

140

895

130

24

81

60

38

1310

190

1070

155

20

47

35

38

1310

190

1070

155

17

34

25

4640

4660

48

1585

230

1310

190

11

16

12

38

1310

190

1070

155

15

27

20

48

1585

230

1310

190

10

13.5

10

Source: Ref 6 Metal Injection Molding. The MIM process combines the structural benefits of metallic materials with the shape complexity of plastic injection molding technology. A uniform mixture of powder and binders is prepared and injected into a mold (see Fig. 4). The MIM powders are typically spherical in shape and much finer in particle size than those used for conventional cold-die compaction (MIM powder, 10 to 20 m; conventional die-compaction powders, 50 to 150 m). The binders are formulated specially to provide the proper rheological properties during injection molding as well as ease of binder removal after the molding step. Once the part is ejected from the mold, the binder material is removed using either solvent extraction or thermal processes (or both). After the debinding step the part is then sintered to complete the process. Due to the large amount of binder in the MIM starting material (up to 40% by volume), the MIM part undergoes a large reduction in size (as much as 20% linear shrinkage) during sintering. Dimensional tolerances, therefore, are not as good as in conventional die compaction and a straightening or coining step is sometimes needed.

Fig. 4 Metal injection molding process. Source: Ref 7

Hot Isostatic Pressing. This fully dense process method is the least constrained technique. However, due to its very

low production rate, costly equipment, and unique tool requirements, the HIP process is normally relegated to expensive materials such as tool steels, superalloys, titanium, and so forth. The process also requires high-purity powders (generally spherical in shape), and it is considered only a near-net-shape process. The powders are vibrated in place in a container, which is then evacuated and sealed. These metal or ceramic containers are placed in the HIP vessel, which applies an isostatic pressure (using a gaseous medium) and temperature to the container and the powder mass. This combination of heat and pressure on the container consolidates the powder to its final shape, as defined by the initial container configuration. The container must be removed from the HIP part after the process cycle, typically by machining or chemical etching. Other Full-Density Methods. The remaining full-density consolidation methods are used infrequently in commercial

practice and are limited to specialty materials. For example, roll compaction is used to form certain soft magnetic alloys, composite materials, and compositions unique to powder metallurgy. Hot pressing is used when the deformation

characteristics of the base powder require high temperatures to achieve plastic flow and adequate consolidation. The powder extrusion process requires a container, and it is similar to roll compaction and limited to specialty materials not suitable for conventional extrusion methods, such as composites, titanium, and nuclear materials.

References cited in this section

3. H.I. Sanderow, H. Rodrigues, and J.D. Ruhkamp, New High Strength 4100 Alloy P/M Steels, Prog. Powder Metall., Vol 41, Metal Powder Industries Federation, 1985, p 283 4. D. Gay and H. Sanderow, The Effect of Sintering Conditions on the Magnetic and Mechanical Properties of Warm Compacted Fe-P P/M Steels, Advances in Powder Metallurgy and Particulate Materials--1996, Vol 6, Metal Powder Industries Federation, 1996, p 20-127 5. "Material Standards for P/M Structural Parts," Standard 35 1994 edition, Metal Powder Industries Federation, 1994 6. "Standard Specification for Powder Forged (P/F) Ferrous Structural Parts," B 848-94, Annual Book of ASTM Standards, American Society for Testing and Materials 7. R.M. German, Powder Metallurgy Science, 2nd ed., Metal Powder Industries Federation, 1994, p 193 Powder Metallurgy Methods and Design* Howard I. Sanderow, Management & Engineering Technologies

Comparison of Powder Processing Methods Effective application of powder processing methods requires a general comparison of the major design features, focusing on the similarities, differences, advantages, and disadvantages of each method. Table 4 provides a qualitative comparison, while Table 5 offers more specific design information. Characteristics for each processing method are summarized below.

Conventional die compaction: • •

• • • •

Widest range of most frequently used engineering materials, including iron, steel, stainless steel, brass, bronze, copper, and aluminum Most applicable to medium-to-high production volumes; small- to medium-size parts such as gears, sprockets, pulleys, cams, levers, and pressure plates (automotive, appliances, power tools, sporting equipment, office machines, and garden tractors are typical markets) Greatest density range, including high-porosity filters, self-lubricating bearings, and high-performance structural parts Limited physical and mechanical properties caused by residual porosity Most cost-competitive of the powder processes Wide range of applications from low- to high-stress applications

Powder forging: • • • •

Potentially applicable to all engineering materials now hot forged, but actual applications currently limited to low-alloy steels Product applications limited to high-volume products such as automotive connecting rods and transmission components as well as power tool parts Mechanical properties equivalent to wrought steel Most cost-competitive of the full-density processes for medium-to-large parts

Metal injection molding: • • • • •

Limited range of materials, though most standard engineering alloys available as well as several specialty alloys Limited to relatively small, highly complex shaped products for medium-to-high production volumes Greatest range in shape complexity including high aspect ratios More costly than conventional die-compaction processes Superior physical and mechanical properties as compared to conventional process, due to higher density

Hot isostatic pressing: • • • • •

Materials limited only by the inherent cost of the process, therefore typically applied only to expensive materials Most suited for low-to-medium production volumes Competitive against large casting or forging products where substantial machining is needed to obtain the final product Much shape detail is machined after HIP processing; not normally a "net-shape" manufacturing process Physical and mechanical properties meet or exceed those of cast or wrought materials

Table 4 Comparison of powder processing methods Characteristic

Conventional

MIM

HIP

P/F

Size

Good

Fair

Excellent

Good

Shape complexity

Good

Excellent

Very good

Good

Density

Fair

Very good

Excellent

Excellent

Dimensional tolerance

Excellent

Good

Poor

Very good

Production rate

Excellent

Good

Poor

Excellent

Cost

Excellent

Good

Poor

Very good

Table 5 Application of powder processing methods Conventional die compaction

MIM

HIP

P/F

Material

Steel, stainless steel, brass, copper

Steel, stainless steel

Superalloys, titanium, stainless steel, tool steel

Steel

Production quantity

>5000

>5000

1-1000

>10,000

Size, lb

99.5 >99.5 >99.5 >99.5 >99.5

0.05 0.05 0.05 0.04 0.05

0.01 0.01 0.01 0.01 0.01

0.2 0.2 0.2 0.2 0.3

5.0 4.0 4.0 3.5 3.0

99.4 99.4 88

0.1 0.1 0.3-0.7

0.01 0.01 0.1

0.5 0.5 0.4-0.6

... ... ... ... 0.1 wt% coating ... ... 9-10 wt% P

SiO2

m m m m m

0.70 m 0.81 m 1.5 m

10.0 9.0 8.0 8.0 6.0

m m m m m

25.0 22.0 18.0 18.0 11.0

m m m m m

1.67 m 1.91 m 4.0 m

3.43 3.66 10.0

m m m

Maximum wt% unless a range is specified

(a)

Table 2 Carbonyl iron powders for powder metallurgy and injection molding Mean Iron, size, wt% m Reduced standard powders 7-8 >99.5 CL

BASF grade

Carbon (max), wt%

Characteristic properties

Oxygen (max), wt%

Nitrogen (max), wt%

0.05

0.01

0.2

Soft, spherical powder

CM

5-6

>99.5

0.05

0.01

0.2

Soft, spherical powder

CS

4-5

>99.5

0.05

0.01

0.2

Soft, spherical powder

5-6 >99.5 0.04 0.01 Unreduced standard powders for injection molding 4-5 >97.8 OM 0.9 0.9

0.2

Soft, spherical powder

0.4

Unreduced, hard powder; agglomerates broken up by grinding Unreduced, hard powder with low N content and higher O content Unreduced, hard powder; SiO2 coated

CN

ON

4-5

>97.5

1.2

0.1

1.2

OS

4-5

>97.3

0.9

0.9

0.7% SiO2

OX

3-4

>96.2

0.9

0.9

5%

OX

3-4

>94.7

0.9

0.9

10% Fe2O3

-Fe2O3 -

More stable form in debinding with improved sinter properties With lower or higher -Fe2O3 content on request

Table 3 Carbonyl iron powders for electronic and microwave applications BASF grade(a)

Mean size, m For electronic parts 4-5 EN 4-5 EW 3-5 EQ 3-4 ES 4-6 SP 4-6 SQ SL

7-8

Iron min, wt%

Carbon max, wt%

Nitrogen max, wt%

Oxygen max, wt%

Bulk, density, g/cm3

Characteristic properties

>97.4 >97.3 >97.2 >97.4 >99.5 >99.5

10 to 1). Because of its composition, MA 6000 has excellent resistance to oxidation and sulfidation. The characteristics of this alloy allow blade cooling to be reduced or eliminated as the metal temperature can be increased by 100 °C or more in engines where the stresses are medium or low. Iron-Base Alloys Table 3 lists the chemical compositions of mechanically alloyed iron-base superalloys. These alloys combine the hightemperature strength and stability of oxide-dispersion strengthening with excellent resistance to oxidation, carburization, and hot corrosion. These alloys are suitable for use in gas turbine combustion chambers. Incoloy alloy MA 956 is particularly well suited for use in heat processing applications. For example, vacuum furnace fixtures made of MA 956 have shown excellent durability and are able to compete with wrought molybdenum, which is also used in these applications. In comparison to molybdenum, MA 956 is 30% lower in density, providing weight savings and cost advantages. Further, since MA 956 has a lower vapor pressure than molybdenum, it will not coat the inside of the vacuum chamber or the parts being heat treated. Thus, MA 956 rods, flats, and sheets are used in numerous atmosphere and vacuum furnace applications including muffles, baskets, trays, and thermowells. Alloy MA 956 in tubing form has also been used for high temperature, severe service applications such as coke injection lance pipes in steel making. The alloy MA 956 is also being used in the glass processing industry because of its resistance to attack by molten glass. Because of this corrosion resistance, the alloy is being evaluated for applications such as firing-kiln rollers, muffle tubes, and furnace racks. Other applications include molten-glass resistance heaters, thermocouple protection tubes, glass-processing components used in nuclear waste disposal, and the bushings used to make single and multistrand fibers.

Table 3 Nominal compositions (wt%) of mechanically alloyed iron-base superalloys Alloy INCOLOY alloy MA 956 INCOLOY alloy MA 957

Fe bal bal

Cr 20 14

Al 4.5 ...

Ti 0.5 1.0

Mo ... 0.3

Y2O3 0.5 0.25

More recently, MA 957 has been evaluated for use as the fuel cladding in fast neutron, breeder reactors. Conventional austenitic alloys are unsuitable for this application due to the dimensional swelling phenomenon caused by the high neutron fluxes. The mechanically alloyed materials are also being evaluated for heat exchanger components in hightemperature gas-cooled reactors. Figure 12 shows some typical high-temperature applications of the alloy MA 956.

Fig. 12 (a) Furnace baskets made of INCOLOY alloy MA 956 operating in air at temperatures >1200 °C. (b) Removable hearth of vacuum heat treating furnace fabricated from 1.0 and 0.5 in. diam rod of INCOLOY alloy MA 956. The hearth legs are encased in alumina sleeves to prevent direct contact with the graphite heating elements. Courtesy of Inco Alloys International

Aluminum-Base Alloys The success of mechanically alloyed superalloys led to the development of dispersion-strengthened aluminum alloys. Table 4 lists the compositions of the mechanically alloyed dispersion-strengthened aluminum alloys. Because an aluminum oxide layer is always present either on the surface of the powder particles at the start of processing or during milling, its incorporation into the alloy contributes to significant improvements in the properties of the alloy. Further, since aluminum is a ductile metal, PCAs are added to assist in minimizing cold welding during processing. Aluminum carbides are formed during MA by the decomposition of the PCA. Both the oxide or carbide type dispersions are approximately 30 to 50 nm in size and stabilize the ultrafine grain size. This results in a 50% increase in strength, higher fracture toughness, and improved resistance to stress corrosion cracking and fatigue crack growth of the mechanically alloyed materials. IncoMAP alloy AL-9052 has a density 5% less than that of conventional age-hardenable aluminum alloy of comparable strengths such as 2024. With its combination of lightweight, high strength, and corrosion resistance, IncoMAP alloy AL-9052 is evaluated for aerospace applications where marine corrosion is also a factor.

Table 4 Nominal compositions (wt%) of mechanically alloyed dispersion-strengthened aluminum-base alloys Alloy IncoMAP alloy AL-9021 IncoMAP alloy AL-9052 IncoMAP alloy AL-905XL

Al bal bal bal

Mg 1.5 4.0 4.0

Li ... ... 1.3

C 1.1 1.1 1.1

O 0.8 0.6 0.6

Addition of lithium to mechanically alloyed aluminum alloys has produced an ultra lightweight alloy, IncoMAP AL905XL. Its density is 8% lower and stiffness 10% greater than the age-hardenable conventional alloy 7075-T73 of comparable strength. The excellent combination of the properties makes this alloy very attractive for airframe applications. In particular, the freedom from age-hardening treatments makes it possible to produce forgings and heavy sections with homogeneous metallurgical structures. Recently, high-strength aluminum-titanium alloys have been developed using MA by dispersing nanometer or submicronsized Al3Ti intermetallic particles (in addition to the Al2O3 and Al4C3 dispersoids from addition of PCAs) in an aluminum matrix. Similar approaches could be used to develop high strength alloys in other systems. Magnesium-base Alloys A useful application of the MA technique was in the production of supercorroding magnesium alloys that operate as short-circuited galvanic cells to corrode (react) rapidly and predictably with an electrolyte, such as seawater, to produce heat and hydrogen gas (Ref 20, 21). Such an alloy system is suitable as a heat source for warming deep-sea divers, as a gas generator to provide gas for buoyancy, or as a fuel in hydrogen engines or fuel cells. The corrosion rate of alloys can be maximized by providing (a) a short electrolyte path, (b) a large amount of exposed surface area, and (c) a strong bond (weld) between the cathode and the anode. It is also useful to provide a very low resistance path for external currents to flow through the corroding pairs. All these requirements can be met with MA processing. Consequently, magnesium-base alloys containing Fe, Cu, C, Cr, or Ti have been evaluated for such applications. The Mg-5 to 20at.% Fe alloy is ideal because of its extremely fast reaction rate, high power output, and the high percentage of theoretical completion of the actual reaction. For corrodable release links an alloy with a slower reaction rate, such as Mg-5at.% Ti is useful. There have also been a number of investigations in recent years to examine use of MA to produce metal hydrides. This is because metal hydrides are materials for safe storage of hydrogen, and they can store hydrogen with a higher volume density than liquid hydrogen. However, these are sensitive to surface oxidation and hence can be a limiting factor in their commercial utilization. Nanocrystalline hydrides have a high density of defects and interfaces that could enhance diffusion; therefore, nanocrystalline intermetallics would not require activation treatments at high temperatures and pressures after exposure to air (Ref 22). In comparison to coarse-grained materials, mechanically alloyed nanocrystalline intermetallics exhibit a narrower absorption plateau and a lower plateau pressure. Their hydrogen storage behavior is typical of amorphous systems. Several magnesium-base and iron-base intermetallics are being evaluated for this application.

References cited in this section

7. C.C. Koch, Mechanical Milling and Alloying, Materials Science and Technology--A Comprehensive Treatment, Vol 15, R.W. Cahn, Ed., VCH, 1991, p 193-245 8. C. Suryanarayana, Metals and Mater., Vol 2, 1996, p 195-209 9. C. Suryanarayana, Bibliography on Mechanical Alloying and Milling, Cambridge International Science Publishing, 1995 10. H. Bakker, G.F. Zhou, and H. Yang, Prog. Mater. Sci., Vol 39, 1995, p 159-241 19. C. Suryanarayana, Int. Mater. Rev., Vol 40, 1995, p 41-64 20. S.A. Black, Report of Civil Eng. Lab. (Navy), Port Hueneme, CA, 1979 21. S.S. Sergev, S.A. Black, and J.F. Jenkins, U.S. Patent 4, 264, 362, 13 Aug 1979 22. P. Tessier, L. Zaluski, A. Zaluska, J.O. Ström-Olsen, and R. Schulz, Mater. Sci. Forum, Vol 225-227, 1996,

p 869-874 Mechanical Alloying C. Suryanarayana, Department of Metallurgical and Materials Engineering, Colorado School of Mines

Displacement Reactions Mechanical activation of solids to induce chemical reactions has been used for over 30 years (Ref 5). A resurgence in this activity occurred in 1989 when it was recognized that MA can be used as the basis of a chemical refining process (Ref 23) by demonstrating that the pure metal copper can be synthesized when CuO and calcium were ball-milled together at room temperature. Simultaneous reduction of CuO and ZnO by calcium has also been shown to result in the formation of brass. Metal synthesis directly from oxides or chlorides has now been extended to the synthesis of a number of metals (Zn, Ti, Zr, Ta, Gd, Er, Sm, V, W and some intermetallics, e.g. SmCo5). A concise review on this subject is found in Ref 24. Most of the reactions studied to date have been displacement reactions:

MO + R

M + RO

where a metal oxide (MO) is reduced by a reductant (R) to the metal (M). Metal chlorides have also been reduced to metals this way. The displacement reactions studied by MA are characterized by a large negative free energy change at room temperature and are therefore thermodynamically feasible at room temperature. However, commercial operations by pyrometallurgical techniques are conducted at elevated temperatures to overcome the kinetic barriers and achieve sufficiently high reaction rates. Mechanical alloying can provide the means to substantially increase the reaction kinetics of the displacement reactions because the repeated fracturing and welding of powder particles increases the area of contact between the particles and allows fresh surfaces to come into contact repeatedly. This allows the reaction to proceed without diffusion through the product layer. Additionally, the high defect densities induced during MA accelerate the diffusion process. As a consequence, these reactions will now occur at room temperature. If a reaction cannot occur at room temperature, the particle refinement and consequent reduction in diffusion distances (due to microstructural refinement) can at least significantly reduce the reaction temperatures. Depending on the milling conditions, two entirely different reaction kinetics have been noted. The reaction can extend to a very small volume during each collision between the grinding media leading to a gradual transformation. Alternatively, if the reaction enthalpy is sufficiently high, a self-propagating combustion reaction can be initiated. Such reactions require a critical milling time for the combustion reaction to be ignited. If the temperature of the powder (or milling vial) is recorded during MA, a sudden increase in temperature at a critical time indicates the onset of the combustion event. Measurements of the ignition time provide a useful means of characterizing the structural and chemical evolution during MA. The product of the displacement reactions normally consists of two phases-the metal (or a compound) and the oxide or chloride associated with the reductant. The removal of the unwanted reaction by-product can be difficult due to the high reactivity of the metal phase associated with nanocrystalline grain sizes and intermixing of the phases induced by the MA process. The by-product phase can be easily removed if the metal particles are embedded in a continuous matrix. Removal of the by-product is achieved by leaching the product mixture in a dilute acid or hot water, or by vacuum distillation. The use of carbon or hydrogen as the reductant produces gaseous CO2 or water vapor as the byproduct and obviates the need for leaching/distillation. Process parameters such as milling temperature, ball diameter, ball-to-powder weight ratio, use of a process control agent, and relative proportion of the reactant phases seem to play an important role on the nature and amount of product phase obtained by the displacement reactions and these need to be optimized for the best yields.

The mechanically driven displacement reaction offers a number of advantages over conventional metal processing techniques. First, it enables the reduction of a number of oxides and halides to pure metals at room temperature, thus effecting energy savings. Second, if a number of components are reduced simultaneously, it is possible to produce an alloy without first having to convert the oxides to pure metals and then to the desired alloy. Third, for powder metallurgy applications, it allows the direct formation of powder product without first having to manufacture the bulk alloy and then convert it to powder form. Thus, a number of high-temperature processes can be combined into one single roomtemperature process with the potential for significant cost savings.

References cited in this section

5. G. Heinicke, Tribochemistry, Akademie Verlag, Berlin, Germany, 1984 23. G.B. Scahffer and P.G. McCormick, Appl. Phys. Lett., Vol 55, 1989, p 45-46 24. P.G. McCormick, Mater. Trans., JIM, Vol 36, 1995, p 161-169 Mechanical Alloying C. Suryanarayana, Department of Metallurgical and Materials Engineering, Colorado School of Mines

Powder Contamination A major concern in the MA process is the impurities that get into the powder and contaminate it. The small size of the powder particles, availability of large surface area, and formation of fresh surfaces during milling all contribute to the contamination of the powder. Thus, it appears as though powder contamination is an inherent drawback of the technique, unless special measures are taken to avoid/minimize it. As mentioned earlier, in the early stages of MA, the metal powder coats the surface of the grinding medium and the inner walls of the container. This was expected to have prevented contamination of the powder and so no attention was paid to the problem of powder contamination. However, when different results were reported by different groups of researchers on the same alloy system, it was recognized that contamination could be a problem. This problem appears to be ubiquitous and is now encountered in many investigations, especially when reactive metals such as titanium and zirconium are milled. The magnitude of contamination appears to depend on the time of milling, the intensity of milling, the atmosphere in which the powder is milled, and difference in strength/hardness of the powder, and the milling medium. Whereas 1 to 4 wt% iron has been found to be normally present in most of the powders milled with the steel grinding medium, amounts as large as 7 wt% (20 at.%) iron in a tungsten-carbon mixture milled for 310 h and 13 wt% (33 at.%) iron in pure tungsten milled for 50 h in a SPEX mill were also reported. These are very high levels of contamination. Similarly, large amounts of oxygen (up to 36.5 at.%) and nitrogen (up to 22.6 at.%) have also been reported to be present in aluminum-titanium powders milled for 400 h in a low-energy ball mill. Contamination of metal powders can be traced to chemical purity of the starting powders, milling atmosphere, milling equipment, and the process control agents added to the powders. Contamination from chemical purity can be either substitutional or interstitial in nature, while contamination from the milling atmosphere is essentially interstitial in nature and that from milling equipment is mainly substitutional, even though carbon from the steel equipment can be an interstitial impurity. Contamination from the PCA leads to interstitial contamination. The presence of interstitial impurities such as oxygen and nitrogen is deleterious to reactive metals like titanium and zirconium, and therefore maximum impurity levels are generally specified for acceptable microstructural and mechanical properties. Substantial amounts of nitrogen and oxygen (the amount of nitrogen is much higher than oxygen) are picked up during the milling of titanium and zirconium alloys and the presence of these impurities leads to a change in the constitution of the alloys (Ref 25, 26, 27). For example, the formation of an amorphous phase in titanium-aluminum alloys and a crystalline phase with a face-centered cubic (fcc) structure in powders milled beyond the formation of an amorphous phase have been attributed to the presence of large quantities of nitrogen in these alloys (Ref 12). Several attempts have been made in recent years to minimize the powder contamination during MA. One way of minimizing the contamination from the grinding medium and container is to use the same material for the container and grinding medium as the powder being milled. In this case also there will be wear of the grinding medium and this gets

incorporated into the powder. Thus, even though there is no contamination, the chemistry of the final powder will be quite different from the starting powder; the metallic content (of the container and balls) would be higher than in the initial powder. This can be compensated for if we know how much of the metallic content is increased. The above solution may be possible occasionally; but it is difficult in many cases due to the nonavailability of the special grinding medium and containers. The problem is becoming more and more complex because the technique of MA is being applied to a variety of materials, such as metals, alloys, ceramics, polymers, and composites, and it is impossible to get containers of all these types of materials. If a container of the same material to be milled is not available, then a thin adherent coating on the internal surface of the container with the material to be milled will minimize the contamination. If this is not possible, then a simple rule that could be followed to minimize contamination is that the container and grinding medium should be harder/stronger than the powder being milled. The problem of milling atmosphere is serious and has been found to be the major cause of contamination in many cases. In fact, it has been observed that if the container is not properly sealed, the atmosphere surrounding the container, usually air (containing nitrogen and oxygen), leaks into the container and contaminates the powder. Thus, when reactive metals like titanium and zirconium are milled in improperly sealed containers, the powders are contaminated with nitrogen and oxygen. It has been reported that flushing with argon gas will not remove oxygen and nitrogen absorbed on the internal surfaces. Pickup of impurities during milling would reduce the pressure within the container allowing outside atmosphere to continuously leak into the container through an ineffective seal. In practice, it has been noted that if it is difficult to open the container lid due to the vacuum present inside, it is an indication that contamination of the powder is minimum. Attempts have been made to improve the container seal integrity to prevent the outside atmosphere leaking inside. Use of high-purity argon (99.998%) atmosphere and improvements of the seal quality resulted in the processing of high-quality titanium alloy powder with as little as 100 ppm oxygen and 15 ppm nitrogen (Ref 26). This process, however, may not be economically viable and hence may not be feasible on an industrial scale. It is also important to remember that cross contamination could occur if a container that was used earlier to mill some powder is used again to mill another powder without properly cleaning it. The level of contamination can be different under different processing conditions and is dependent on the type of mill, intensity of milling, nature of the powder, nature of the grinding medium and container, ambient atmosphere, ball-topowder weight ratio, seal integrity, and others. Claims have been made in the literature about the superiority of certain mills and practices over others, but systematic investigations on milling the same powder under identical conditions in different mills and evaluating the contamination levels have not been undertaken.

References cited in this section

12. C. Suryanarayana, Intermetallics, Vol 3, 1995, p 153-160 25. G.H. Chen, C. Suryanarayana, and F.H. Froes, Metall. Mater. Trans. A, Vol 26, 1995, p 1379-1387 26. P.S. Goodwin and C.M. Ward-Close, Mater. Sci. Forum, Vol 179-181, 1995, p 411-418 27. T. Klassen, M. Oehring, and R. Bormann, J. Mater. Res., Vol 9, 1994, p 47-52 Mechanical Alloying C. Suryanarayana, Department of Metallurgical and Materials Engineering, Colorado School of Mines

Modeling From the previous description it is easy to realize that MA is a complex process. Like any other process, modeling of MA is carried out to identify the salient factors affecting the process and to establish process control instrumentation. By modeling the process effectively, it is possible to lower the number of actual experiments to be conducted to optimize the process and achieve a particular application.

The number of variables involved in the MA process is very large. For a particular alloy system, the variables include the type of mill, intensity of milling, type of milling media, ball-to-powder weight ratio, the atmosphere under which the powder is milled, purity of the powders, milling time, milling temperature, and nature and amount of the PCA used. These have a significant effect on the constitution of the powder. Even on a local scale, the nature of impacts between two balls, the frequency of impacts, and the amount of powder trapped between two balls during a collision could vary from point to point. Thus, modeling the MA process is a difficult task. In spite of this, some attempts have been made (Ref 28, 29, 30) and moderate success has been achieved in modeling the mechanics of the process. From the actual experiments conducted, attempts have been made to correlate the phases formed with the process parameters during milling. But, the ability to predict the final chemical constitution of the powder (type and description of phases formed) has not been achieved. It should be realized that due to the stochastic nature of the MA process, it is difficult to make absolute predictions.

References cited in this section

28. T.H. Courtney, Rev. Part. Mater., Vol 2, 1994, p 63-116 29. M. Magini and A. Iasonna, Mater. Trans., JIM, Vol 36, 1995, p 123-133 30. M. Abdellaoui and E. Gaffet, Acta Mater., Vol 44, 1996, p 725-734 Mechanical Alloying C. Suryanarayana, Department of Metallurgical and Materials Engineering, Colorado School of Mines

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

J.S. Benjamin, Sci. Amer., Vol 234 (No. 5), 1976, p 40-48 J.S. Benjamin, Met. Powder Rep., Vol 45, 1990, p 122-127 A.E. Yermakov, Y.Y. Yurchikov, and V.A. Barinov, Phys. Met. Metallogr., Vol 52 (No. 6), 1981, p 50-58 C.C. Koch, O.B. Cavin, C.G. McKamey, and J.O. Scarbrough, Appl. Phys. Lett., Vol 43, 1983, p 1017-1019 G. Heinicke, Tribochemistry, Akademie Verlag, Berlin, Germany, 1984 P.S. Gilman and J.S. Benjamin, Ann. Rev. Mater. Sci., Vol 13, 1983, p 279-300 C.C. Koch, Mechanical Milling and Alloying, Materials Science and Technology--A Comprehensive Treatment, Vol 15, R.W. Cahn, Ed., VCH, 1991, p 193-245 8. C. Suryanarayana, Metals and Mater., Vol 2, 1996, p 195-209 9. C. Suryanarayana, Bibliography on Mechanical Alloying and Milling, Cambridge International Science Publishing, 1995 10. H. Bakker, G.F. Zhou, and H. Yang, Prog. Mater. Sci., Vol 39, 1995, p 159-241 11. E. Ivanov, Mater. Sci. Forum, Vol 88-90, 1992, p 475-480 12. C. Suryanarayana, Intermetallics, Vol 3, 1995, p 153-160 13. E.G. Avvakumov, Mechanical Methods of Activation of Chemical Processes, Nauka, Novosibirsk, Russia, 1986 14. A. Calka, J.J. Nikolov, and J.S. Williams, Mater. Sci. Forum, Vol 225-227, 1996, p 527-532 15. G. Jangg, in New Materials by Mechanical Alloying Techniques, E. Arzt and L. Schultz, Ed., Deutsche Gesellschaft für Metallkunde, Oberursel, Germany, 1989, p 39-52 16. C.C. Koch, Nanostructured Mater., Vol 2, 1993, p 109-129; Vol 9, 1997, p 13-22 17. G.A.J. Hack, Metals and Mater., Vol 3, 1987, p 457-462 18. J.J. Fischer, J.J. deBarbadillo, and M.J. Shaw, Heat Treating, Vol 23 (No. 5), 1991, p 15-16 19. C. Suryanarayana, Int. Mater. Rev., Vol 40, 1995, p 41-64

20. S.A. Black, Report of Civil Eng. Lab. (Navy), Port Hueneme, CA, 1979 21. S.S. Sergev, S.A. Black, and J.F. Jenkins, U.S. Patent 4, 264, 362, 13 Aug 1979 22. P. Tessier, L. Zaluski, A. Zaluska, J.O. Ström-Olsen, and R. Schulz, Mater. Sci. Forum, Vol 225-227, 1996, p 869-874 23. G.B. Scahffer and P.G. McCormick, Appl. Phys. Lett., Vol 55, 1989, p 45-46 24. P.G. McCormick, Mater. Trans., JIM, Vol 36, 1995, p 161-169 25. G.H. Chen, C. Suryanarayana, and F.H. Froes, Metall. Mater. Trans. A, Vol 26, 1995, p 1379-1387 26. P.S. Goodwin and C.M. Ward-Close, Mater. Sci. Forum, Vol 179-181, 1995, p 411-418 27. T. Klassen, M. Oehring, and R. Bormann, J. Mater. Res., Vol 9, 1994, p 47-52 28. T.H. Courtney, Rev. Part. Mater., Vol 2, 1994, p 63-116 29. M. Magini and A. Iasonna, Mater. Trans., JIM, Vol 36, 1995, p 123-133 30. M. Abdellaoui and E. Gaffet, Acta Mater., Vol 44, 1996, p 725-734

Spray Drying and Granulation Introduction GRANULATION is a term used to describe two different types of processes. In one definition, granulation is the intentional agglomeration of fine particles into larger clusters to improve certain powder properties. For example, bulk powders typically have a low bulk density, do not readily flow, are dusty, and have low thermal conductivity. When properly granulated, the same powder pours easily, exhibits a high and uniform bulk density, does not experience dusting losses, and more efficiently transfers thermal energy. For example, in the case of a powder used as a feed material for high speed presses, the granules should typically be greater than 50 m (2 mils), but less than 1000 m (40 mils) in diameter, have a spherical shape, and should not be strong enough to retain their identity in the compacted part. Granulation is also defined as the production of metal particles by agitating molten metal. The most commonly used techniques are spray drying and spray granulation. Another method is water granulation. These three methods are discussed in this article. Other granulation methods include agitation and pressure technique for ceramics (Ref 1).

Acknowledgement This article was adapted from "Granulation and Spray Drying" by Stanley Lukasiewicz in Ceramic and Glasses, ASM International, 1991 and "Spray Drying of Metal Powders" by David Houck in Powder Metallurgy, Volume 7, ASM Handbook, 1984.

Reference

1. S. Lukasiewicz, Granulation and Spray Drying, Ceramics and Glasses, Vol 4, Engineered Materials Handbook, ASM International, 1991, p 100-108 Spray Drying and Granulation

Spray Granulation Spray granulation (Ref 2) forms granules by atomizing a liquid or a binder solution into a fluidized powder bed (Fig. 1). Fluidization is achieved by directing a heated gas, which is usually air, through a distributor at the bottom of the powder

bed. The gas imparts a vigorous motion to the particles, which prevents the formation of large lumps. The binding liquid is usually sprayed into the powder bed with a two-fluid nozzle. Spray granulators can be designed to operate in either a batch or a continuous mode.

Fig. 1 Fluidized bed spray granulator. Source: Ref 3

The formation of granules in a spray granulator occurs through the random nucleation of small seed agglomerates, followed by the growth of these seeds to the desired size. Growth occurs either by the layering of powder onto the seeds or by the agglomeration of seeds to form larger granules. Granule growth by seed agglomeration forms irregular shapes. Granule size increases as the fraction of the bed exposed to the binding liquid is reduced and as the spray nozzle is adjusted to give coarser droplets (Ref 3). Increasing the intensity of agitation of the bed (with a higher gas velocity) decreases the size of the granules. There is an upper limit on granule size because of the tendency of the powder bed to defluidize. However, spray granulation can form larger granules than is usually possible by spray drying because of longer residence times. Spray granulation is commonly used in the pharmaceutical industry to prepare feedstock for tablet presses. It is used less frequently in the metal powder industry. However, both spray drying and fluid bed conversion are scaleable technologies and together provide the means for producing bulk quantities of nanophase composite powders at low manufacturing cost. A few examples of fluidized bed granulation are described below. Metal

Powder Granulation in a Plasma-Spouted/Fluidized Bed. A direct current (DC) plasmaspouted/fluidized bed was applied to the granulation of spherical alloy grains from metal powder mixtures. From a mixture of iron powder (149 to 210 m) and aluminum powder (74 to 88 m and 125 to 149 m), alloy grains from 1 to 5 mm in diameter were obtained. The grains exhibited a dense homogeneous core and a porous nonhomogeneous shell structure. The grains were quite spherical; the average aspect ratio of the grains was >0.85. The mass fraction of iron in the core section of product grains depended on the Al/Fe ratio of initial powder mixtures but was insensitive to the size of product grains and to the size of initial Al particles. Selective growth of seed grains was observed when sufficient seed grains were added to the starting powder mixture (Ref 4).

Contamination-Free Processing of Pyrophoric Rare Earths and Abrasive Ceramic Powders. The

production of exactly limited particle fractions is an important process for rare earths (Sm-Co, Nd-Fe-B) and ceramic (oxides, carbides, nitrides, borides) products. Extremely fine products with tight upper particle size limits can be produced by a fluidized bed process (Ref 5). This allows practically wear- and contamination-free grinding of extremely hard products. The necessary grinding and classification can either be wet or dry, or a combination of both procedures (Ref 5).

References cited in this section

2. J.S. Reed, Introduction to the Principles of Ceramic Processing, John Wiley & Sons, 1988 3. C.E. Capes, Particle Size Enlargement, Handbook of Powder Technology, Vol 1, J.C. Williams and T. Allen, Ed., Elsevier Scientific, 1980 4. M. Tsukada, K. Goto, R.H. Yamamoto, and M. Horio, Metal Powder Granulation in a PlasmaSpouted/Fluidized Bed, Powder Technol., Vol 82 (No. 3), Mar 1995, p 347-353 5. H. Prem and D.J. Eddington, Contamination-Free Processing of Pyrophoric Rare Earths and Abrasive Ceramic Powders, Powder Handl. Process., Vol 1 (No. 1), Mar 1989, p 101-107 Spray Drying and Granulation

Water Granulation Water granulation is a process for producing coarse metal particles by pouring molten metals through a screen into water or by agitating molten metals into droplets with subsequent water quenching. Granulation of liquid metal offers a simple technique for solidification of metals, but pouring of liquid metal into water has always been regarded as hazardous. Nonetheless, several plants for granulation of steel and different kinds of ferroalloys have shown that granulation can be done in a safe and reliable way (Ref 6). For example, granulation of ferrosilicon and Si-metal has a risk of explosions when done at high flow rates (Ref 6). The explosion potential is considered to be due to the rapid formation of steam. Light metals like ferrosilicon and Si-metal have a comparatively low falling velocity in the water and thus result in a high power generation per unit volume of water. To decrease the power concentration, the heated water has to be efficiently replaced, or the heat has to be distributed to a big volume. It has been found that the replacement of the water from a small cooling zone in a big water tank is rather difficult, but metal can be efficiently distributed over a big water surface to decrease explosion potential (Ref 6). Another trial of water granulation of ferrosilicon was performed by impingement of molten alloy stream against stationary ceramic plate or rotary cone followed by water quenching of the drops. The test stand permitted variations of nozzle-plate/cone and plate/cone-water distances and adjustment of the inclination of the plate. Quality and size of the product was affected chiefly by the plate/cone-water distance. Better quality was obtained with impingement against rotary cone. Results provided the basis for development of an industrial-scale installation for granulation of ferrosilicon (Ref 7).

References cited in this section

6. P.-A. Lundstrom and A. West, Granulation of Ferroalloys and Silicon-Metal, 52nd Electric Furnace Conf., (Nashville, TN), Iron and Steel Society/AIME, 13-16 Nov 1994, p 309-315 7. L. Bulkowski et al, Trials of Water Granulation of Ferrosilicon, Pr. Inst. Metal. Zelaza, Vol 47 (No. 3), 1995, p 23-27

Spray Drying and Granulation

Spray Drying Spray drying is a powder-producing process in which a slurry of liquids and solids or a solution is atomized into droplets in a chamber through which heated gases, usually air, are passed. Figure 2 represents a typical spray dryer utilizing a disk atomizer with co-current air flow. Examples of commercially spray-dried powders are shown in Fig. 3. Spray drying is used widely in the pharmaceutical, chemical, and food industries. It is used to a lesser extent in the metals-related industries.

Fig. 2 Typical spray dryer arrangement

Fig. 3 Micrographs of commercially spray-dried granules. (a) Ferrite, 75×. (b) Zirconia, 30×

Spray drying offers several advantages over other powder-processing techniques, particularly in applications requiring agglomerates for subsequent pressing and sintering operations. Spray drying also is one of the most economical ways of drying slurries. Spray drying is a continuous rather than a batch process that allows close control of agglomerate size, bulk density, and moisture content. Agglomerate shape is spherical, facilitating excellent flowability. Particles have very short exposure time. Residence time within the dryer may range from 2 to 20 s, depending on dryer size. Therefore, heat-sensitive materials can be spray dried. Lubricating additives can be added easily for die pressing applications. Process Description During spray drying, atomized droplets dry rapidly because of their high surface-to-volume ratio. Coarse dried solids fall to the bottom of the chamber and are continuously collected, typically through a rotary airlock valve. The finer solids, which are entrained in the gases exiting the chamber, are removed by and collected from a cyclone separator. Baghouse collectors or wet scrubbers can be used to clean ultrafine particles from the air exhausted from the cyclone. Baghouse collectors and electrostatic precipitators also can be substituted for cyclone collectors. An alternate spray dryer design includes the introduction of the air through ductwork above the cyclone exit. Size of the drying chamber is dictated by (a) product characteristics, (b) product throughput, (c) atomization technique, (d) the properties of the slurry or solution, and (e) the evaporative capacity of the heating unit and associated fans. The evaporative capacity is a function of the allowable inlet and outlet temperatures of the dryer. These temperatures are, in turn, dictated by the heat sensitivity of the product. The interrelationship of these factors is shown in Fig. 4 (Ref 8). Generally, a spray drying facility is designed as an integral unit so that all component capacities are compatible.

Fig. 4 Relationship between chamber volume and evaporation capacity of spray dryers. Data based on retention time of 20 s. Source: Ref 8

Some special applications in the cemented carbide industry require the use of organic solvents, such as acetone, ethanol, hexane, and methanol. Because of the flammable nature of these materials, close-system drying technology (see Ref 9, 10, 11) requires that the medium be spray dried using a nonoxidizing gas, such as nitrogen. By recycling the nitrogen that contains organic vapors, organic emissions are eliminated, and the solvent can be recycled. Closed-cycle systems are gastight and operate slightly above atmospheric pressure. Most of the dried particles are collected from the dryer chamber; a small percentage of fines that pass through the chamber in the gas stream is collected in the cyclone. The drying gas subsequently passes through a scrubber/condenser to recover the solvent and to recondition the nitrogen. Typically, gas temperature is controlled by a recirculating scrubber and heat exchangers that are cooled by water or brine. For some spray drying applications, a cooling system may be required. Droplet-Air Mixing. The air flow pattern in the drying chamber controls the completeness of moisture removal from

the droplet, the maximum temperature that the granules will experience, and the formation of wall deposits. Droplet-air mixing in spray dryers is determined by the location of the air disperser and the atomizing device and is classified

according to the relative direction between the droplets and the air. These classifications are termed cocurrent, countercurrent, and mixed-flow mixing. Cocurrent conditions exist when the atomizing device is located near the air disperser at the top of the dryer, as in

Fig. 5(a). The droplets are exposed to the hottest air immediately after formation, but high evaporation rates maintain low product temperatures. As the moisture content of the droplets decreases, they come into contact with cooler air and surface temperature does not increase appreciably. Cocurrent air flow is common in dryers equipped with rotary atomizers.

Fig. 5 Types of droplet-air mixing in spray dryers. (a) Cocurrent. (b) Countercurrent. (c) Mixed-flow. Source: Ref 12

In cocurrent drying, the maximum temperature to which larger dried particles are exposed is approximately 10 °C (18 °F) less than the outlet temperature of the dryer. This is caused by the evaporative cooling effect that occurs when the liquid leaves the droplet while passing through the heated zone. Finer particles, which dry more quickly, experience temperatures approaching those of the inlet gas. Countercurrent conditions occur when the atomizing device is placed at the top of the chamber and the air disperser

is located at the bottom (Fig. 5b). Immediately after formation, the droplets contact cool humid air. However, as their moisture content decreases, they are exposed to increasingly hotter air. Because high internal temperatures can be realized in the granules, the organic binders should not be heat sensitive. When the slurry is introduced from the bottom and gas from the top of the dryer, droplets dry both in their upward and downward trajectories. Therefore, narrower, shorter dryers can be used. However, if partially dried products cannot be exposed to the high temperatures at the top of the dryer, this mode of spraying may not be applicable. Mixed-flow conditions are a combination of cocurrent and countercurrent air flow. This type is commonly utilized in

fountain-type dryers, where a nozzle atomizer is located at the base of the drying chamber and the air disperser is placed at the top of the chamber, as in Fig. 5(c). Mixed-flow conditions are frequently used in combination with pneumatic nozzles in small laboratory dryers because the fountain-like spray pattern increases droplet trajectory and provides sufficient airborne time to dry large droplets in small chambers. Control of Powder Properties Spray drying can be used to remove water from a slurry or to create agglomerates of fine particles contained within the slurry. Agglomeration is achieved by using a binder (discussed below in this section).

Most of the applications of spray drying in the metals industry require the formation of free-flowing agglomerates. Many of these powders are used for producing pressed parts. Therefore, the agglomerate size distribution and bulk density are the two most important properties of spray-dried powders. Agglomerate size distribution is a function of atomization conditions and the properties of the slurry. Generally, a

lower solids content yields a finer average agglomerate size. The maximum attainable solids content varies with material, but usually can be increased by using deflocculating or suspending agents. Bulk density is a function of the solids content of the slurry, inlet temperature, and slurry additives. Lower bulk

densities generally are achieved from slurries with low solids contents. Also, excessive inlet temperatures can cause lower bulk densities. Rapid evaporation of the liquids causes the partially dried droplets to expand rapidly, thus decreasing density. Introduction of frothing agents may entrap air in the slurry, which also leads to lower bulk densities. Typically, frothing agents are not added in metallurgical applications. Moisture content of a powder can be controlled by the inlet and outlet temperature of the spray dryer in conjunction

with the slurry feed rate. Moisture levels below 0.1% are possible. For a given airflow and inlet temperature, outlet temperature is controlled by the rate of slurry feed and the solids content. With higher percentages of solids, less water must be evaporated, which leads to higher throughput of dry product. If a product can withstand higher inlet temperatures, throughput can be increased (see Fig. 6).

Fig. 6 Relationship of evaporation capacity of spray dryers to production rate and slurry solids concentration. Source: Ref 3

Binders for Agglomeration. Suitable binder materials must be homogeneously dispersable (preferably soluble) in the liquid used to form the slurry. When dry, binders must form a coating and/or adhere to the material being agglomerated. They must impart the required strength and crush resistance to the particle for subsequent handling. In addition to the liquid, solids, and binders used to formulate a slurry, various other additives may be necessary. The following are typical components of spray drying slurries:

Organic binders • • • • • • • • • • • • • • • •

Polyvinyl alcohol Gum arabic Other natural gums Carboxy-methyl cellulose salts Polyvinyl acetate Methyl cellulose Ethyl cellulose Polyvinyl butyral dispersions Protein colloids Acrylic resin emulsions Ethylene oxide polymers Water-soluble phenolics Lignin sulfonates Propylene glycol alginates Flour Starches

Inorganic binders • • • • • • •

Sodium silicate Boric acid Borax Carbonates Nitrates Oxylates Oxychlorides

Plasticizers • • • • • • • • •

Glycerine Ethylene glycol Triethylene glycol Dibutyl phthalate Diglycerol Ethanolamines Propylene glycol Glycerol monochlorhydrin Polyoxyethylene aryl ether

Deflocculating agents • • • • • • • •

Sodium hexametaphosphate Sodium molybdenate Tetrasodium pyrophosphate Ammonium citrate Ammonium oxalate Ammonium tartrate Ammonium chloride Monoethylamine

Wetting agents • • • •

Synthetic detergents Alkylaryl sulfonates Alkylaryl sulfates Soaps

Suspending agents (high molecular weight) • • •

Sodium carboxymethyl cellulose Methyl cellulose Ethylene oxide polymers

Plasticizers may be used with binding materials that are hard or brittle and that tend to crack during drying. Suspending agents may be needed to prevent solids from settling within the slurry. Deflocculating agents aid in the formation of slurries by preventing the agglomeration of fine particles. Wetting agents also may be used to maintain solids in suspension. Some slurries have a tendency to foam during mixing. Antifoaming agents or defoamers may be used to control this action. Chemical activators also may be used as additives to aid in subsequent sintering or processing of powders. Atomization Techniques and Agglomerate Size Distribution. Three standard techniques are used to atomize

slurries for spray drying: • • •

Single-fluid nozzle atomization Centrifugal (rotating disk) atomization Two-fluid nozzle atomization

Table 1 gives the advantages and disadvantages of these techniques for atomization of slurries. Table 2 gives the relative agglomerate sizes produced by these techniques.

Table 1 Advantages and disadvantages of various spray drying atomization systems Atomization system Centrifugal (rotating disk) Single-fluid nozzle Two-fluid nozzle

Advantages

Disadvantages

High feed rates, less downtime, low-pressure pumps, never plugs Large agglomerate capability, smaller dryers can be used Less part wear

Larger diameter dryer required, coarse agglomerate size not obtainable Downtime due to part wear and plugging, high pressure pumps Broad agglomerate size distribution, compressed air needed

Table 2 Relative agglomerate sizes produced by various atomizing systems Atomizing system

Mean agglomerate size, m Centrifugal (rotating disk) 25-100 High speed 50-200 Medium speed 100-300 Low speed Single-fluid nozzle 25-100 High pressure Medium pressure 50-200 100-300 Low pressure

Very low pressure Two-fluid nozzle High pressure Medium pressure Low pressure

200-600 10-50 25-100 50-200

Source: Ref 12

As shown in Table 2, the largest agglomerate sizes (600 m) are achieved by the single-fluid nozzle. The centrifugal (rotating disk) atomizer yields agglomerate sizes up to 300 m, and the two-fluid nozzle produces agglomerates only up to about 200 m in size. Centrifugal atomization yields the narrowest agglomerate size distribution range, followed by single-fluid atomization and two-fluid atomization. Centrifugal atomization entails the introduction of the slurry into a horizontally rotating disk that is equipped with

vanes or holes through which the slurry exits. The slurry is atomized into fine droplets, the average size of which is a function of the design and peripheral speed of the wheel. Most wheel designs incorporate a wear-resistant material for extended life. Single-fluid atomization is capable of producing the largest diameter agglomerates. However, because of the high

pressures required to force the slurry through a single, small orifice to produce the desired droplet size, considerable downtime may be required to replace worn parts and to unplug the nozzle. Use of a nozzle in the fountain spray position extends the residence time of the droplets in the chamber. As a result, smaller dryers can be used for single-fluid atomization of materials that are not heat-sensitive. Two-fluid atomization, which uses a pressurized air blast to break up a slurry stream into droplets, produces the

widest range of agglomerate sizes of all three atomization techniques. However, it is a relatively easy technique for producing atomized droplets. Wear problems do not exist because high pressures are not employed. The orifices used are typically larger than those used for single-fluid atomization; therefore, plugging is not as serve a problem. This technique does, however, require the use and additional expense of pressurized air.

References cited in this section

3. C.E. Capes, Particle Size Enlargement, Handbook of Powder Technology, Vol 1, J.C. Williams and T. Allen, Ed., Elsevier Scientific, 1980 8. R.H. Perry and C.H Chilton, Spray Dryers, Chemical Engineers Handbook, 5th ed., McGraw Hill, 1973, p 58-63 9. A.O. Jensen and K. Masters, "Spray Dryer for Producing Tungsten Carbide Products," Bulletin F-125, Niro Atomizer, Inc., Columbia, MD 10. "Closed Circuit Spray Drying Systems," Bulletin 1342, Anhydro, Inc., Attleboro Falls, MA 11. "Closed Cycle Systems," Bowen Engineering Bulletin, Bowen Engineering, Inc., Somerville, NJ 12. K. Masters, Spray Drying, 4th ed., John Wiley & Sons, 1985 Spray Drying and Granulation

Applications Spray drying applications are most prevalent in the pharmaceutical, chemical, and food industries. There are, however, several areas in the metals industry that utilize spray drying techniques. These include production of cemented carbides, mineral processing, production of iron powders, production of oxide-dispersion-strengthened alloys, and production of powders used for thermal spraying applications.

Agglomeration by spray drying is a basic process that has unlimited possibilities in combining different materials to produce composite powders, which behave homogeneously in bulk. Investigations in producing custom-made powders have led to various multicomponent powders. The first step of this technique consists of the agglomeration of the starting powder, for example, metallic, oxide, and nonoxide hard materials. The second step includes plasma densification. Using the high energy of the plasma, materials with extremely high melting points can be melted. Cemented Carbides (Tungsten Carbide/Cobalt). Historically, cemented carbides have been pressed from

powders produced by various agglomeration techniques. With the advent of closed-cycle systems, production of cemented carbides by spray drying increased considerably in the 1980s. Closed-cycle spray drying is required for most cemented carbide powders because the binders that are used are soluble only in volatile organic fluids. The nitrogen drying gas that is used in the spray drying of cemented carbides is heated to 75 to 100 °C (170 to 210 °F), depending on the milling liquid used. The solids content of the slurry varies from 75 to 80%. Viscosity of milled slurries is sometimes modified with stabilizers, such as stearic acid (0.3 to 0.5 wt%). Pressures for single-fluid nozzle atomization range from 590 to 1470 kPa (85 to 213 psi). Pressure is a function of the particular type powder, slurry viscosity, and binder content. Recent work related to spray drying for cemented carbides is described in Ref 13, 14, 15, 16, 17, 18, and 19. Multicomponent Oxide Powders for Plasma Spraying. The concept of multicomponent powders consists of two

steps. The first step is the agglomeration of the starting powder by the spray drying process. The second step involves plasma densification. During spray drying, the starting powder combination is suspended in a liquid (water or solvent) and simultaneously mixed with binder and possible other needed auxiliary agents, for example, wetting agents, defoamers, etc. The suspension is sprayed through a nozzle into a heated chamber. Using rapid vaporizing, the formed droplets are dried briefly and micropellets are formed from the starting powder due to the adherence of the binder. This procedure and the properties of the resulting micropellets have certain advantages: • • • • • •

There are almost no limits to the combination of different materials. Particle sizes of the micropellets are reproducible, dependent upon the process parameters. The starting components are homogeneously distributed in the micropellets. Nearly ideal spherical shape produces powders with excellent flowability. Lower energy consumption compared to other melting or crushing methods Easier control on mean diameter and size range

However, a key disadvantage of many spray dried powders is the poor mechanical resistance of the coating due to particle breakage (Ref 20). Particle breakage results in poor transport during the plasma spray coating process, which thus makes spray dried powder difficult to use for thermal spray. Fused and crushed powders are still the most widely used powder types for plasma spraying because they give deposits with good thermophysical properties and mechanical resistance. To apply the advantages of spray dried powders for plasma spray coatings, a second step of plasma densification is required (Ref 20, 21, 22, and 23). In this step, the micropellets from spray drying are brought into a high energy plasma. Depending on the powder, one can choose whether the process takes place under atmospheric, inert atmospheric, vacuum, or underwater plasma spraying conditions. This additional step has the following advantages. During plasma densification, reactions between the components of the micropellets occur, which during further processing, such as plasma spraying, leads finally to a more stable microstructure in the product. Without the plasma treatment, the reaction time is not sufficient, so the resulting metastable microstructure can lead to certain disadvantages and to property changes during use of the material. Plasma densification enables high cooling rates so that disadvantages in microstructure--brittle phases or the solutioning of important phases--will not occur. This is an important advantage in contrast to conventional melting processes. The plasma densified powder obtained has a spherical, smooth surface, a high density, and a porosity approaching zero. Experience in this technique concerning the plasma densification of metals, metallic hard materials, and ceramics shows that the multicomponent powder conception is practicable for nearly all materials. To characterize the properties, plasma

sprayed coatings of plasma densified powders were investigated. Results revealed that coatings of plasma densified powders show better resistance to wear than coatings of agglomerated powders.

References cited in this section

13. L. Wu, Nanostructured Tungsten Carbide/Cobalt Alloys: Processing and Properties, Diss. Abstr. Int., Vol 54 (No. 9), Mar 1994, p 196 14. L.E. McCandlish, B.H. Kear, and S.J. Bhatia, Spray Conversion Process for the Production of Nanophase Composite Powders, U.S. Patent No. 5,352,269, 9 July 1991 15. S. Danzglock, Vacuum Plasma Spraying, Metalloberfläche, Vol 45 (No. 10), Oct 1991, p 455-458 16. B.-K. Kim, H.-S. Chung, L.E. McCandlish, and B.H. Kear, Fluid Bed Synthesis of Nanophase WC/Co Composite Powders, Novel Powder Processing, Vol 7, Advances in Powder Metallurgy and Particulate Materials, Metal Powder Industries Federation, 1992, p 51-61 17. B. Kim, "Synthesis, Processing and Characterization of WC/Co Nanophase Composites," Diss. Abstr. Int., Vol 52 (No. 3), Sept 1991, p 243 18. G. Nagarajan and K. Sadananda, Hard-Metal Powder Granulation for Spray Drying (Retroactive Coverage), PMAI Newsl., Vol 13 (No. 1), 12-15 Dec 1986 19. E. Lugscheider, H. Eschnauer, A. Nisch, and Z. Li, Plasma Treated Multi-Component Powders on the Basis of Metallic Hard Materials, 12th Int. Plansee Seminar '89, Vol 3, Metallwerk Plansee GmbH, Reutte, Austria, 1989, p 221-236 20. A. Denoirjean, A Vardelle, A. Grimaud, P. Fauchais, E. Lugscheider, I. Rass, H.L. Heijen, P. Chandler, R. McIntyre, and T. Cosack, Plasma Densification of Zirconia Powders: Optimization for Thermal Barrier Coatings in IC Engines and Gas Turbines (Brite Project P2280), Thermal Spray: Int. Advances in Coatings Technology, ASM International, 1992, p 975-982 21. Z. Bartnik, P. Bialucki, S. Kozerski, G. Clinton, K. Davies, P. Bork, B. Schrader, F. Guglielmi, and L. Pawlowski, Improvements in Manufacturing Technology of Wear Resistant Plasma Sprayed Cr 2O Coatings, Thermal Spray: Int. Advances in Coatings Technology, ASM International, 1992, p 983-993 22. P. Luo, P.R. Strutt, T.D. Xiao, Synthesis of Chromium Silicide--Silicon Carbide Composite Powders, Mater. Sci. Eng. B, Vol 17 (No. 1-3), 28 Feb 1993, p 126-130 23. E. Lugscheider, M. Loch, and H.G. Suk, Powder Technology--State of the Art, Thermal Spray: Int. Advances in Coatings Technology, ASM International, 1992, p 555-559 Spray Drying and Granulation

References 1. S. Lukasiewicz, Granulation and Spray Drying, Ceramics and Glasses, Vol 4, Engineered Materials Handbook, ASM International, 1991, p 100-108 2. J.S. Reed, Introduction to the Principles of Ceramic Processing, John Wiley & Sons, 1988 3. C.E. Capes, Particle Size Enlargement, Handbook of Powder Technology, Vol 1, J.C. Williams and T. Allen, Ed., Elsevier Scientific, 1980 4. M. Tsukada, K. Goto, R.H. Yamamoto, and M. Horio, Metal Powder Granulation in a PlasmaSpouted/Fluidized Bed, Powder Technol., Vol 82 (No. 3), Mar 1995, p 347-353 5. H. Prem and D.J. Eddington, Contamination-Free Processing of Pyrophoric Rare Earths and Abrasive Ceramic Powders, Powder Handl. Process., Vol 1 (No. 1), Mar 1989, p 101-107 6. P.-A. Lundstrom and A. West, Granulation of Ferroalloys and Silicon-Metal, 52nd Electric Furnace Conf., (Nashville, TN), Iron and Steel Society/AIME, 13-16 Nov 1994, p 309-315

7. L. Bulkowski et al, Trials of Water Granulation of Ferrosilicon, Pr. Inst. Metal. Zelaza, Vol 47 (No. 3), 1995, p 23-27 8. R.H. Perry and C.H Chilton, Spray Dryers, Chemical Engineers Handbook, 5th ed., McGraw Hill, 1973, p 58-63 9. A.O. Jensen and K. Masters, "Spray Dryer for Producing Tungsten Carbide Products," Bulletin F-125, Niro Atomizer, Inc., Columbia, MD 10. "Closed Circuit Spray Drying Systems," Bulletin 1342, Anhydro, Inc., Attleboro Falls, MA 11. "Closed Cycle Systems," Bowen Engineering Bulletin, Bowen Engineering, Inc., Somerville, NJ 12. K. Masters, Spray Drying, 4th ed., John Wiley & Sons, 1985 13. L. Wu, Nanostructured Tungsten Carbide/Cobalt Alloys: Processing and Properties, Diss. Abstr. Int., Vol 54 (No. 9), Mar 1994, p 196 14. L.E. McCandlish, B.H. Kear, and S.J. Bhatia, Spray Conversion Process for the Production of Nanophase Composite Powders, U.S. Patent No. 5,352,269, 9 July 1991 15. S. Danzglock, Vacuum Plasma Spraying, Metalloberfläche, Vol 45 (No. 10), Oct 1991, p 455-458 16. B.-K. Kim, H.-S. Chung, L.E. McCandlish, and B.H. Kear, Fluid Bed Synthesis of Nanophase WC/Co Composite Powders, Novel Powder Processing, Vol 7, Advances in Powder Metallurgy and Particulate Materials, Metal Powder Industries Federation, 1992, p 51-61 17. B. Kim, "Synthesis, Processing and Characterization of WC/Co Nanophase Composites," Diss. Abstr. Int., Vol 52 (No. 3), Sept 1991, p 243 18. G. Nagarajan and K. Sadananda, Hard-Metal Powder Granulation for Spray Drying (Retroactive Coverage), PMAI Newsl., Vol 13 (No. 1), 12-15 Dec 1986 19. E. Lugscheider, H. Eschnauer, A. Nisch, and Z. Li, Plasma Treated Multi-Component Powders on the Basis of Metallic Hard Materials, 12th Int. Plansee Seminar '89, Vol 3, Metallwerk Plansee GmbH, Reutte, Austria, 1989, p 221-236 20. A. Denoirjean, A Vardelle, A. Grimaud, P. Fauchais, E. Lugscheider, I. Rass, H.L. Heijen, P. Chandler, R. McIntyre, and T. Cosack, Plasma Densification of Zirconia Powders: Optimization for Thermal Barrier Coatings in IC Engines and Gas Turbines (Brite Project P2280), Thermal Spray: Int. Advances in Coatings Technology, ASM International, 1992, p 975-982 21. Z. Bartnik, P. Bialucki, S. Kozerski, G. Clinton, K. Davies, P. Bork, B. Schrader, F. Guglielmi, and L. Pawlowski, Improvements in Manufacturing Technology of Wear Resistant Plasma Sprayed Cr 2O Coatings, Thermal Spray: Int. Advances in Coatings Technology, ASM International, 1992, p 983-993 22. P. Luo, P.R. Strutt, T.D. Xiao, Synthesis of Chromium Silicide--Silicon Carbide Composite Powders, Mater. Sci. Eng. B, Vol 17 (No. 1-3), 28 Feb 1993, p 126-130 23. E. Lugscheider, M. Loch, and H.G. Suk, Powder Technology--State of the Art, Thermal Spray: Int. Advances in Coatings Technology, ASM International, 1992, p 555-559

Rotating Electrode Process Steven A. Miller, Starmet Corporation, and Peter R. Roberts, American Superconductor

Introduction THE ROTATING ELECTRODE process is a method for producing metal powders where the end of a metal bar is melted while it is rotated about its longitudinal axis. Molten metal is centrifugally ejected and forms droplets that solidify to spherical powder particles. The basic process is shown in Fig. 1. The electrode can be melted by any energy source, but usually an electric arc or plasma is used and the process is then identified by the acronyms REP or PREP, respectively. In this article REP is used generically unless a distinction needs to be made between the two processes. The REP concept was first invented by Starmet Corporation (then Nuclear Metals, Inc.), and resulted in the granting of patents (Ref 1, 2). While Starmet remains the sole source of REP/PREP (REP and PREP are registered trademarks of the Starmet Corporation) powder, powders are also made in Russia, China, and Japan using similar rotating electrode technology.

Fig. 1 Schematic of REP

The rotating electrode process has several inherent characteristics that make it uniquely suitable for the fabrication of specific alloy powders to provide manufacturing and product advantages. First, REP is a means of contactless melting and atomization to make powder with the highest level of cleanliness possible. This is a critical feature for reactive, highmelting-temperature alloys that are aggressively corrosive in their molten state and attack conventional ceramic crucibles. Such alloys are routinely atomized by REP without incurring contamination. Examples are titanium, zirconium, molybdenum, and vanadium alloys. Second, REP powder is almost perfectly spherical and practically satellite free. Because the atomized droplets are dispersed and move radially away from each other there is little opportunity for collisions between droplets and particles and the resulting coalescence of the two into irregularly shaped clusters. This single-particle nature of the powder spheres results in REP powder being very free flowing and having a high packing density, approximately 65%. Also, REP powder has both a tighter size distribution and a larger median size than can be produced by gas atomization. Finally, because REP atomization is produced by centrifugal forces rather than by aerodynamic drag, the powder is essentially porosity free when compared to gas atomized powder (Ref 3). The rotating electrode process has evolved to include equipment that consists of a vacuum/controlled-atmosphere tank 2440 mm (96 in.) in diameter by 300 mm long in which powder is produced. The tank dimensions are determined by the trajectory and solidification distance of the largest droplets of the molten spray. This is mounted with its circular plane section in the vertical position. Figure 2 shows a typical production setup. The consumable rotating electrode is introduced through a seal-and-bearing assembly; the long axis of the electrode is horizontal and centrally located in the tank and is made the anode of a direct-current power circuit. The permanent cathode may be a simple tungsten-tipped

device provided with adequate cooling (REP) or a transferred arc plasma torch (PREP). This equipment projects through the other face of the tank to oppose the rotating electrode. Usually, melting is conducted under inert gas; the preferred medium is helium, which offers improved heat transfer properties and electric arc characteristics.

Fig. 2 Rotating electrode/plasma rotating electrode process machine for generation of specialty spherically shaped powders

References

1. A.R. Kaufman, Method and Apparatus for Making Powder, U.S. Patent 3,099,041 2. A.R. Kaufman, Production of Pure, Spherical Powders, U.S. Patent 3,802,816 3. P.R. Roberts, The Production of PREP Titanium Powder, 1989 Advances in Powder Metallurgy, Vol 3, Metal Powder Industries Federation, 1989, p 427-438 Rotating Electrode Process Steven A. Miller, Starmet Corporation, and Peter R. Roberts, American Superconductor

Equipment There are two standard types of machines used in REP: short bar and long bar. Short-Bar Apparatus. The short-bar machine (Fig. 3) accepts consumable anodes up to 89 mm (3 in.) in diameter by 250 mm (10 in.) long. The anode is held in a collet in a precision spindle, the head of which projects into the tank through a rotating seal mechanism.

Fig. 3 Short-bar PREP

Usually at least 80% of the length of a short bar is converted to powder. Electrode stub removal and the introduction of new electrodes to the collet is performed manually through a glove port that is located in the front of the machine adjacent to the cathode or plasma torch. Short-bar methodology is appropriate for converting experimental quantities of material, alloys that are inherently brittle, and materials that have a low specific stiffness where an electrode of long aspect ratio is not practical. Long-Bar Apparatus. Increased productivity and conversion efficiency are realized in the long-bar operation (Fig. 4),

which is designed to consume 63.5 mm (2 in.) diam electrodes that are up to 1830 mm (72 in.) long. Typically the assembly includes a precision spindle similar to that used in the short-bar machine. Instead of being fixed in position relative to the tank, the spindle is mounted on a table that also carries the drive motor and the electrical transfer brush mechanism. The table moves toward the tank from an extended starting position, feeding the electrode through the special seal-and-bearing assembly.

Fig. 4 Long-bar PREP

When the head of the spindle approaches the seal housing, the process is interrupted while the stub of the consumed electrode, now typically 230 to 250 mm (9 to 10 in.) long, is pushed through the seal by another long bar that is mounted in the spindle head when it is retracted to the starting position. Stubs produced in long-bar machines may be joined to new bars to obtain effectively 100% conversion to powder. The rotational speed used is determined by the desired particle size. Standard speeds range between 314 rad/s (3000 rpm) and 1570 rad/s (15,000 rpm). Smaller diameter short-bars have been rotated up to 2620 rad/s (25,000 rpm). Because highspeed rotation rates are employed, electrodes must have precise dimensions; they must also be straight to keep mechanical out-of-balance forces to a minimum. Rotating Electrode Process Steven A. Miller, Starmet Corporation, and Peter R. Roberts, American Superconductor

Particle Size Distribution Accurately controlled and maintained rotation speed of the anode is necessary to obtain a desired range of particle size distribution. Molten droplet diameter of a given material is determined by the properties of the liquid metal, the centrifugal ejecting forces (related to rotation speed), and to a limited extent the aerodynamics of the droplet trajectory through the inert cover gas. The following equation predicting median droplet size is obtained from a force balance of the centrifugal forces acting on the molten surface tending to cause atomization with the surface tension forces resisting atomization:

(Eq 1)

where d is the median droplet diameter (microns); is rotation rate (rad/s); and are the surface tension (dynes/cm) and density (g/cm3) of the alloy being atomized, respectively; D is the electrode diameter (cm); and k is an empirical constant principally determined by the method of droplet formation which is in turn controlled by the melting rate (Ref 4). For any particular alloy, the material properties are fixed and Eq 1 can be further simplified to:

(Eq 2) where K is a constant (over a limited melting rate range) that has been determined for many alloy systems (Ref 5). Figure 5 shows the powder size distributions of 1018 steel and Ti-6Al-4V produced at different rotational speeds. Within the steel data there is a clear reduction in particle size produced with increasing rotational speeds as predicted by Eq 1 and 2. Equation 1 also predicts that Ti-6Al-4V, due to its decreased density (approximately 57%) and increased surface tension (approximately 107%) compared to 1018 steel, would have greater inherent resistance to centrifugal atomization. Again, the data of Fig. 5 confirm the trend of the prediction, although the magnitude of the shift in sizes due to the change in melt properties is actually greater than that predicted by Eq 1.

Fig. 5 Effect of rotational speed and alloy properties on particle size

High-speed video images of the commercial REP process have shown the powder is formed by direct droplet production; that is, the liquid metal forms discrete droplets, with little or no ligament formation, as soon as it detaches from the bar surface (Fig. 6). Spheroidization of the liquid is nearly instantaneous; droplets tend to come off in clusters of stringers, and the droplet flight path is nearly tangential to the bar surface (Ref 6).

Fig. 6 High speed video of the droplet formation process. Instabilities formed in the liquid metal foil grow into ligaments, which ultimately separate into primary and secondary molten droplets

Champagne and Angers (Ref 7, 8) have demonstrated that a bimodal number distribution of particle size occurs due to this mechanism in which liquid drops separating from the electrode rim remain connected momentarily by a necked-down column of liquid metal that breaks into smaller drops when the major spheroid is freed. Bimodal particle size distribution is shown in Fig. 7. Although smaller particles are not present in large quantities by weight, their absolute numbers may be significant. This behavior is superimposed on distributions and is shown as a function of anode melting rate (Fig. 8) and rotation rate (Fig. 9).

Fig. 7 Bimodal particle size distribution attributed to satellite drop formation

Fig. 8 Particle size distribution as a function of melting rate

Fig. 9 Particle size distribution as a function of rotation rate

The ordinate values of these curves are expressed in terms of the percentage of retained particles on a given sieve size divided by the size range between sieves, which eliminates the effect of this range on the shape of the curve. Additional information can be found in Ref 7 and 8.

References cited in this section

4. R. Angers et al., Inverted Disk Centrifugal Atomization of 2024, Int. J. Powder Metall., Vol 30 (No. 4), 1994, p 429-434 5. P.R. Roberts, "Powders Made by the Rotating Electrode Process," Starmet Int. Report, 1986 6. S.A. Miller, Gas Enhanced Rotating Electrode Atomization, Advanced Particulate Materials and Processes, F.H. Froes and J.C. Hebeisen, Ed., Metal Powder Industries Federation, p 457-454 7. B. Champagne and R. Angers, Fabrication of Powders by the Rotating Electrode Process, Int. J. Powder Metall. Powder Technol., Vol 16 (No. 4), 1980, p 359-367 8. B. Champagne and R. Angers, Size Distribution of Powders Atomized by the Rotating Electrode Process, Modern Development in Powder Metallurgy, Proc. 1980 International Powder Metallurgy Conference (Washington, D.C.), H. Hausner, H. Antes, and G. Smith, Ed., Metal Powder Industries Federation, 1981, p 83-104

Rotating Electrode Process Steven A. Miller, Starmet Corporation, and Peter R. Roberts, American Superconductor

Aerospace Applications Spherical metal powders made by REP or by gas atomization are not well suited for cold pressing into green compacts. Therefore, spherical powders tend to be used in specialized applications where consolidation is achieved by hot isostatic pressing (HIP), or other high-temperature processing in which interparticle voids are more readily closed. These particles flow well into complex mold shapes and can be tapped to a reproducible density of packing to provide fully dense parts that closely approach the dimensions of the finished component (Ref 9). Much research has been performed on near-net shape compaction for military airframe parts in titanium alloys. An example of a complex shape that can be made from titanium alloy powders pressed to near-net shape is shown in Fig. 10. Overseas research has resulted in as hot isostatically pressed PREP superalloy powder being used in both land- and airbased gas turbines (Ref 10, 11). This use of powders can offer both economic and metallurgical advantages. Near-netshape technology ensures more efficient material utilization, which is important for high-cost materials such as titanium. Additionally, the metallurgical benefits include improved homogeneity and control of microstructure to achieve enhanced mechanical properties (Ref 12).

Fig. 10 Gas turbine engine compressor rotor made from hot isostatically pressed plasma rotating electrode processed Ti-6Al-4V powder

Because REP prevents contact of the melted alloy with any container material, it provides a decided advantage over other methods of making contamination-free spherical particles. This process generates powder with cleanliness and composition approximating that of the precursor electrode. Molten titanium is extremely aggressive and reacts with all container materials; consequently, REP is ideally suited for production of clean titanium alloy powders in commercial quantities. Typical characteristics for Ti-6Al-4V powder made by REP are shown in Table 1.

Table 1 Typical Ti-6Al-4V REP powder compositions and properties Element Aluminum Vanadium Oxygen(a) Oxygen(b) Iron Carbon Nitrogen Hydrogen Tungsten Other

Composition, wt% 5.50-6.75 3.50-4.50 0.13-0.20 0.05-0.13 0.30 max 0.10 max 0.05 max 0.0125 max 99.9%). The power requirements for electrowinning of copper are approximately ten times as large as those for electrorefining of copper with soluble anodes. Copper may be concentrated from low-content leach solutions by solvent extraction, followed by stripping with dilute sulfuric acid into an aqueous solution and electrowinning. Carboxylic acid and hydroxylamine-based compounds have been found to be selective solvents of low water solubility, to have good stability, and to be compatible with inexpensive diluents. Direct powder precipitation with hydrogen or ammonia is an alternative to stripping the metal from the organic solvent into an aqueous solution. Metals can be precipitated from their acid or basic solutions by reduction with hydrogen. Sulfuric acid, ammoniacal ammonium carbonate, and ammoniacal ammonium sulfate solutions have been used to produce copper powder by this method. Sulfuric acid leaching of a cement copper and hydrogen reduction of the filtered solution in an autoclave at 120 to 140 °C (250 to 280 °F) and 3 MPa (425 psi) is reported to produce a precipitate with a purity of about 100% Cu. Drying and furnace processing in a reducing atmosphere at 540 to 790 °C (1000 to 1450 °F) increases particle size, due to agglomeration of the very fine powder. Production of Copper Powders David F. Berry, OMG AMERICAS; Erhard Klar, Consultant

References 1. E. Klar and D. Berry, Copper P/M Products, Vol 2, Properties and Selection: Nonferrous Alloys and Pure Metals, Vol 2, ASM Handbook, 9th ed., ASM International, 1979, p 392 2. S. Harper and A.A. Marks, Electrodeposition of Copper Powder with the Aid of Surfactants, Copper Development Association/American Society for Metals Conf. on Copper, Vol 3, Conference Paper No. 059/2, Copper Development Association, 1972 3. L.I. Gurevich and A.V. Pomosov, The Effect of Chloride on Electrodeposition of Powdered Copper Precipitates, Sov. Powder Metall. Met. Ceram., Jan 1969, p 10-15 4. V.P. Artamonov and A.V. Pomosov, Effect of Foreign Electrolytes on the Production of Copper Powder by Contact Deposition, Izv. V.U.Z. Tsvetn. Metall., No. 2, 1976, p 30-34 (in Russian); Metall. Abstracts, No. 54-0503, Nov 1976

5. S.L. Kotovskaya et al., Manufacture of Coarse Copper Powder from Sulfamate Electrolytes, Sov. Powder Metall. Met. Ceram., Feb 1973, p 93-96 6. S.K. Singh and D.D. Akerkar, A Continuous Self-Regulating Method of Making Copper Powder by Electrolysis, NML Tech. J., Vol 17, 1975, p 23-26 7. F. Willis and E.J. Clugston, Production of Electrolytic Copper Powder, J. Electrochem. Soc., Vol 106, 1959, p 362-366 8. A.V. Pomosov, M.I. Numberg, and E.G. Krymakove, Protection of Copper Powder Against Corrosion during Manufacture and Storage, Sov. Powder Metall. Met. Ceram., March 1976, p 175-177 9. Technical data from AMAX Metal Powders, AMAX Copper, Inc., 1968 10. D.L. Adamson and W.M. Toddenham, "Production of High Quality Electrolytic Copper Powder," American Institute of Mining, Metallurgical and Petroleum Engineers Annual Meeting (New York), 1-4 March 1971 11. E. Peissker, Metal Powders, Norddeutsche Affinerie, Aug 1974 12. P.W. Taubenblat, W.E. Smith, and C.E. Evans, Production of P/M Parts from Copper Powders, Precis. Met., April 1972, p 41 13. D.E. Tyler and W.T. Black, Introduction to Copper and Copper Alloys, Vol 2, Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, ASM Handbook, ASM International, 1990, p 216-240 14. D. Kumar and A.K. Gaur, Electrochemical Studies on Production of Electrolytic Copper Powders, J. Electrochem. Soc. India, July 1973, p 211-216

Production of Copper Alloy Powders Thomas W. Pelletiers, II, United States Bronze Powders, Inc.; David F. Berry, OMG Americas

Introduction COMMERCIAL COPPER ALLOY POWDERS, including brasses, bronzes, and nickel silvers, are manufactured in a similar manner. Usually, the same integrated manufacturing facilities are used to complete the melting process, atomization, final screening, and blending of a lot or batch. The powder-producing process is similar to a foundry operation in which high-purity virgin metals are charged to a melting furnace in preweighed batches to be processed at predetermined heating rates and times. High-purity raw materials are required, because melting produces minimal refinement (only a partial transfer of contaminants from the melt to the formed slag occurs). To ensure continuity, homogeneity, and uninterrupted atomization, the molten batch of alloy is transferred to a second furnace with a greater holding capacity than the primary furnace melting rate. Induction heating is preferred for at least one furnace to ensure the constant induced metal movement required for alloy homogeneity and to ensure uniform lead dispersion in lead-bearing alloys. Atomization is achieved by particulation of a controlled, constant-flowing, molten stream emitted from the secondary furnace by medium-pressure dry air. A typical melting and atomizing sequence is shown in Fig. 1. Subsequent reduction of oxides is not required for standard P/M grades.

Fig. 1 Flowchart for copper alloy powder air atomization. SQC, statistical quality control; SPC, statistical process control

Air-cooled, atomized powder is collected and passed over a primary control screen (sieve) to remove oversize particles. Usually, these are remelted concurrently during the processing of each alloy. Finally, the screened alloy powder may be blended with dry organic lubricants, such as lithium and zinc stearate, for use in the manufacture of P/M structural components. Adjustment of powder properties (particle size distribution, apparent density, green strength, etc.) of each alloy is accomplished by controlled manipulation of the atomizing process conditions (e.g., atomizing air flow rate, molten metal temperature, nozzle configuration, etc.). Powder properties are maintained by periodically recording the atomizing parameters and by evaluating the properties of representative samples from the in-process atomized product. A variety of copper-base alloys can be manufactured by the atomizing process; however, commercial P/M applications normally are confined to a rather narrow range of specific, single-phase ( ) compositions. Production of Copper Alloy Powders Thomas W. Pelletiers, II, United States Bronze Powders, Inc.; David F. Berry, OMG Americas

Brasses Brasses comprise the major portion of copper-base alloy powders used for parts fabrication, in comparison with prealloyed bronze and nickel silver materials. Typical copper-zinc brass powders contain zinc contents ranging from 10 to 30%. Lead may be added in small amounts (1 to 2%) to improve machinability of the sintered compact. A photomicrograph of an 80%Cu-18%Zn-2%Pb alloy powder is shown in Fig. 2. The melting temperatures of these alloys range from 1045 °C (1910 °F) for 90%Cu-10%Zn to 960 °C (1760 °F) for 70%Cu-30%Zn. As zinc content increases, melting temperature is lowered.

Fig. 2 Scanning electron micrograph of prealloyed, air-atomized brass (80%Cu-18%Zn-2%Pb). 165×

Additional superheat, or the temperature in excess of the alloy melting temperature, depends on heat losses within the manufacturing system and the physical property requirements of the atomized powders. Typical physical properties of brass alloy compositions are given in Table 1.

Table 1 Physical properties of typical brass, bronze, and nickel silver alloy compositions Property Sieve analysis, % +100 mesh -100 + 200 -200 + 325 -325 Physical properties Apparent density Flow rate, s/50 g Mechanical properties Compressibility(c) at 414 MPa (300 tsi), g/cm3 Green strength(c) at 414 MPa (300 tsi), MPa (psi)

(a) (b) (c)

Brass(a)

Bronze(a)

Nickel Silver(a)(b)

2.0 max 15-35 15-35 60 max

2.0 max 15-35 15-35 60 max

2.0 max 15-35 15-35 60 max

3.0-3.2 24-26

3.3-3.5 ...

3.0-3.2 ...

7.6 10-12 (1500-1700)

7.4 10-12 (1500-1700)

7.6 9.6-11 (1400-1600)

Nominal mesh sizes; brass, -60 mesh; bronze, -60 mesh; nickel silver, -100 mesh. Contains no lead. Compressibility and green strength data of powders lubricated with 0.5% lithium stearate

Production of Copper Alloy Powders Thomas W. Pelletiers, II, United States Bronze Powders, Inc.; David F. Berry, OMG Americas

Bronzes Prealloyed atomized bronze compositions are not used extensively as base powders for compacted parts fabrication because of their nodular particle form and high apparent density, both of which contribute to poor compacted green strength. Common prealloyed compositions are 90%Cu-10%Sn and 85%Cu-15%Sn, prepared in the same manner as brass powder, except that high-purity elemental copper and tin are used. A scanning electron micrograph of an 89%Cu9%Sn-2%Zn alloy powder is shown in Fig. 3. Typical physical properties of a bronze alloy composition are given in Table 1.

Fig. 3 Scanning electron micrograph of prealloyed, air-atomized bronze (89%Cu-9%Sn-2%Zn). 165×

Prealloyed bronze powders are also made commercially by water atomizing. Application is more extensive in Europe, where 90/10 prealloyed bronze powders are incorporated in bronze premixes for bearing manufacture. Low green strength due to high apparent density (3.2 to 3.6 g/cm3) is overcome by incorporating lower apparent density copper powders and choice of lubricants that have a less deleterious effect on green strength. Physical properties are similar to air-atomized powders, but particle morphology is different (Fig. 4). Powders contain 0.1 to 0.2% P to aid sintering.

Fig. 4 Scanning electron micrograph of a typical prealloyed water-atomized bronze powder (90%Cu-10%Sn); apparent density 3.4 g/cm3. 200×

Spherical 89/11 bronze powders are used to make filters. These are made by horizontal air atomizing and dry

collection. The spherical shape is achieved by addition of small amounts of phosphorus, 0.2 to 0.45% (in the form of a Cu/15% P alloy) to the molten bronze prior to atomizing. During air atomizing, surface oxidation of atomized molten particles of bronze and brass, which cause them to solidify in an irregular shape (see Fig. 3) is prevented. The oxygen in the air preferentially reacts with phosphorus to form phosphorus pentoxide (P2O5), which is volatile at atomizing temperatures. The spherical powders are screened to produce a number of grades, each with a narrow particle size range. An assortment of filters made from bronze powders and properties of four grades of filters are shown in Fig. 5 and Table 2.

Table 2 Properties of four grades of filter materials produced by loose powder sintering spherical powders Particle size of spherical powder particles

Tensile strength

Mesh range 20-30 30-40 40-60 80-120

MPa 20-22 25-28 33-35 33-35

Range in m 850-600 600-425 425-250 180-125

Largest dimensions of particles retained, m

Viscous permeability coefficient, m2

ksi

Recommended minimum filter thickness mm in.

2.9-3.2 3.6-4.1 4.8-5.1 4.8-5.1

3.2 2.4 1.6 1.6

50-250 25-50 12-25 2.5-12

2.5 × 10-4 1 × 10-4 2.7 × 10-5 9 × 10-6

0.125 0.095 0.063 0.063

Source: Ref 1

Fig. 5 Assorted filters made from P/M bronze. Courtesy of Arrow Pneumatics, Inc.

Microbearings. A more recent development has been prealloyed bronze powders for microbearings. These are very

small bearings, often weighing less than 1 g, used in electronic equipment such as computers, audiocassette players, and videocassette recorders. Most powders used in this application are made by diffusion alloying tin into copper powder to produce a substantially alloyed powder with particles of uniform composition and a particle shape that gives high green strength (apparent density of 2.3 to 2.7 g/cm3). As previously stated, water-atomized bronze powders have relatively high apparent density (3.2 to 3.6 g/cm3) and are usually limited to application in high-density structural parts. Figure 6 shows the particle morphology of a typical diffusion-alloyed bronze powder.

Fig. 6 Scanning electron micrograph of a typical diffusion-alloyed bronze powder (90%Cu-10%Sn); apparent

density 2.6 g/cm3. 200×

Reference cited in this section

1. F.R. Lenel, Powder Metallurgy Principles and Applications, Metal Powder Industries, 1980 Production of Copper Alloy Powders Thomas W. Pelletiers, II, United States Bronze Powders, Inc.; David F. Berry, OMG Americas

Nickel Silvers Only one nickel-silver-base alloy composition, 65%Cu-18%Ni-17%Zn, is commonly used in the P/M industry. This alloy can be modified by the addition of lead to improve machinability. Foundry practices are similar to those employed for brasses, except the melting temperature is in excess of 1093 °C (2000 °F). A micrograph of a 63%Cu-18%Ni-17%Zn2%Pb alloy powder is shown in Fig. 7. Typical physical properties of a nickel silver alloy composition are given in Table 1.

Fig. 7 Scanning electron micrograph of prealloyed, air-atomized nickel silver (63%Cu-18%Ni-17%Zn-2%Pb). 165×

Production of Copper Alloy Powders Thomas W. Pelletiers, II, United States Bronze Powders, Inc.; David F. Berry, OMG Americas

Reference 1. F.R. Lenel, Powder Metallurgy Principles and Applications, Metal Powder Industries, 1980

Production of Tin Powders William J. Ullrich, ACuPowder International, LLC

Introduction TIN POWDERS are used extensively in the production of porous, self-lubricating bronze bearings and as constituents in soldering and brazing pastes and powders. Low-melting-point tin powders normally are produced by air atomization, although other methods of production such as chemical precipitation and electrodeposition have been used (Ref. 1). Tin powders also are used in powder metallurgy (P/M) structural parts, friction disks, clutches, brake linings, metal graphite brushes, diamond abrasive grinding wheels, bronze filters, peen plating, plasma arc spraying, chemical formulations, additives for rubber and plastics, chemical manufacturing, smokeless powder for pyrotechnics, and tin flake.

Reference

1. C.G. Goetzel, Treatise on Powder Metallurgy, Vol 1, Interscience, 1949, p 199-200 Production of Tin Powders William J. Ullrich, ACuPowder International, LLC

Melting High-purity pig tin is melted in a gas-fired or electrically heated crucible (cast iron, clay, graphite, or ceramic). Melt temperature is controlled to maintain the desired degree of superheat above the low melting point of tin (232 °C, or 449 °F). Production of Tin Powders William J. Ullrich, ACuPowder International, LLC

Atomization Atomization results in a fine powder whose average particle size can be regulated over a wide range. Resultant powders are of high purity. Excessive oxidation, characteristic of other atomized products, does not occur because of the rapid chilling effect of the expanding gases released through the nozzle Thus, the oxygen content of atomized tin is normally below 0.2%. The thin film of oxide produced during atomization with steam or air is sufficient to inhibit further oxidation of the particles. The molten tin can be atomized vertically (up or down) or horizontally, depending on the nozzle direction, to produce a powder (Fig. 1). There are two atomization techniques: annular nozzle and cross-jet atomization. In an annular nozzle, the gas stream aspirates liquid tin into the nozzle, where it is disintegrated into tiny droplets by the high-velocity gas stream. In cross-jet atomization, the gas is at right angles to the molten tin stream. This method usually produces coarser particles than annular nozzles.

Fig. 1 Schematic of apparatus for atomizing tin powder

To atomize uniformly fine tin powders, temperature, stream diameter, and flow rate of the molten tin and temperature, pressure, velocity, and angle of impingement of the atomizing gas must be well controlled. Pressures commonly employed to atomize tin powder range from 345 to 1725 kPa (50 to 250 psi). Generally, finer powders require the higher pressure; however, the exact pressure is related to the nozzle design used. Compressed air is usually used as an atomizing medium. It is frequently preheated in a gas-fired heat exchanger to prevent solidification of tin within or around the orifice caused by the chilling effect of expanding air as it is released through the nozzle. Nozzles are designed to facilitate the atomization of several different particle size ranges, usually by changing orifice diameter and air pressure. A blower (fan) at the end of the system pulls the atomized tin powder from the atomization chamber into a cyclone, where it is collected. The finest particles, which have not settled in the cyclone, are then retained in the cyclone filters. The tin powder collected in the cyclone is sieved to remove the oversized particles, most commonly +100, +200, or +325 mesh. A uniform mixture of tin powder of the desired lot size is produced by tumbling in a blender. A sample of the blended tin is then analyzed for physical and chemical properties. The tin powder is packed in steel or waterproof fiber containers weighing up to 320 kg (700 lb). Production of Tin Powders William J. Ullrich, ACuPowder International, LLC

Atomization Atomization results in a fine powder whose average particle size can be regulated over a wide range. Resultant powders are of high purity. Excessive oxidation, characteristic of other atomized products, does not occur because of the rapid chilling effect of the expanding gases released through the nozzle Thus, the oxygen content of atomized tin is normally below 0.2%. The thin film of oxide produced during atomization with steam or air is sufficient to inhibit further oxidation of the particles. The molten tin can be atomized vertically (up or down) or horizontally, depending on the nozzle direction, to produce a powder (Fig. 1). There are two atomization techniques: annular nozzle and cross-jet atomization. In an annular nozzle, the gas stream aspirates liquid tin into the nozzle, where it is disintegrated into tiny droplets by the high-velocity gas stream. In cross-jet atomization, the gas is at right angles to the molten tin stream. This method usually produces coarser particles than annular nozzles.

Fig. 1 Schematic of apparatus for atomizing tin powder

To atomize uniformly fine tin powders, temperature, stream diameter, and flow rate of the molten tin and temperature, pressure, velocity, and angle of impingement of the atomizing gas must be well controlled. Pressures commonly employed to atomize tin powder range from 345 to 1725 kPa (50 to 250 psi). Generally, finer powders require the higher pressure; however, the exact pressure is related to the nozzle design used. Compressed air is usually used as an atomizing medium. It is frequently preheated in a gas-fired heat exchanger to prevent solidification of tin within or around the orifice caused by the chilling effect of expanding air as it is released through the nozzle. Nozzles are designed to facilitate the atomization of several different particle size ranges, usually by changing orifice diameter and air pressure. A blower (fan) at the end of the system pulls the atomized tin powder from the atomization chamber into a cyclone, where it is collected. The finest particles, which have not settled in the cyclone, are then retained in the cyclone filters. The tin powder collected in the cyclone is sieved to remove the oversized particles, most commonly +100, +200, or +325 mesh. A uniform mixture of tin powder of the desired lot size is produced by tumbling in a blender. A sample of the blended tin is then analyzed for physical and chemical properties. The tin powder is packed in steel or waterproof fiber containers weighing up to 320 kg (700 lb). Production of Tin Powders William J. Ullrich, ACuPowder International, LLC

Reference 1. C.G. Goetzel, Treatise on Powder Metallurgy, Vol 1, Interscience, 1949, p 199-200

Production of Aluminum Aluminum-Alloy Powder

and

A. Ünal, D.D. Leon, T.B. Gurganus, and G.J. Hildeman, Aluminum Company of America

Introduction ALUMINUM POWDERS and ALUMINUM alloys are produced almost exclusively by gas atomization. For most applications, the atomizing gas is compressed air; however, in some special cases, inert gases (helium, nitrogen, and argon) are used. Between 25 and 30 countries in the world are known to have production facilities for aluminum powder at an estimated total capacity of 200,000 tons per year (Ref 1). A large portion of this capacity is in North America where annual shipments are about 40,000 tons (Ref 2). Atomized aluminum powders are used in a variety of applications that include pyrotechnics, explosives, rocket fuel, thermite welding, aluminothermic reduction, chemical processes (as catalyst or reagent), additives for lightweight concrete, pharmaceuticals, and pigments for paints and printing inks. With few exceptions, the demand is for unalloyed powder of various standards of purity. Aluminum alloy P/M, historically a small percentage (about 1%) of the total powder market, has been enjoying a revival in recent years in automotive applications because of the need to reduce weight, lower emissions, and boost fuel economy. Viable P/M applications include engine cam caps and air conditioning compressor parts. Powder metallurgy offers competition to wrought aluminum castings and stampings because of its net shape advantage.

References

1. P.D. Liddiard, Aluminium Powder Metallurgy in Perspective, Powder Metall., Vol 27 (No. 4), 1984, p 193200 2. A.J. Yule and J.J. Dunkley, Atomization of Melts for Powder Production and Spray Deposition, Clarendon Press, Oxford, 1994

Production of Aluminum and Aluminum-Alloy Powder A. Ünal, D.D. Leon, T.B. Gurganus, and G.J. Hildeman, Aluminum Company of America

Historical Background Aluminum powder was first used commercially as flakelike pigment products called aluminum bronze powder in the United States around 1900. It was made from aluminum sheet by the Bessemer dry stamping process developed fifty years earlier for the manufacture of gold powders. Stamping mills were both inefficient and dangerous to operate, because, unlike gold powders, aluminum powder forms an explosive mixture with air over a wide range of metal-to-air ratios (Ref 3). Several fatal fires and explosions marred the early days of the aluminum powder industry. Two major breakthroughs developed by Hall (Ref 4) occurred in aluminum powder production in the late 1920s. First, the development of safer ball mill production processes that are still employed today for milling powders into flakes. Second, the introduction of the atomization process, which allowed the manufacture of the forerunners of today's atomized powders. Atomized aluminum powder was used initially only as raw material to produce aluminum flake pigments by ball milling. After World War II and the development of aluminized high explosives, a major market for atomized aluminum was created. The application of atomized powders broadened in postwar years from military explosives to commercial markets listed above.

To meet the individual needs of each of these markets, manufacturers now produce many grades of aluminum powders in several different general categories: granules, regular and coated atomized powders, spherical powders, high-purity powders, alloy powders, blended powders, and dedusted atomized powders. All are produced by the same basic gas atomization technique, but starting materials may differ and, as with flake powders, further processing may be required. Powders atomized in inert gas (nitrogen, argon, and helium), which have a spherical shape, are preferred for some applications. Water atomization of aluminum is not practiced because it is dangerous, due to the creation of hydrogen and the potential presence of rust in the equipment. The latter can lead to explosions through thermite reaction with the dry powder. Some manufacturers still use the dry stamping process for converting foil scrap into coarse particulate (granules), although it is now a very small percentage.

References cited in this section

3. K.L. Cashdollar, Flammability of Metals and Other Elemental Dust Clouds, Proc. Safety Progress, Vol 13 (No. 3), 1994, p 1139-1145 4. E.J. Hall, Process for Disintegrating Metal, U.S. Patent 1,659,291, 14 February 1928 Production of Aluminum and Aluminum-Alloy Powder A. Ünal, D.D. Leon, T.B. Gurganus, and G.J. Hildeman, Aluminum Company of America

Gas Atomization There are many methods that have been used to produce aluminum and aluminum alloy powders including gas and centrifugal atomization, ultrasonic or pulsed atomization, melt spinning with attrition, and mechanically alloying. Gas atomization is used for several applications. In gas atomization, a high-velocity gas jet disintegrates molten aluminum into droplets that solidify to form the powder. In the Alcoa process (Fig. 1), the material flow during atomization is vertically upward (also known as "updraught"). Molten metal of the appropriate composition is supplied from a holding or melting furnace at the required temperature to the atomizing bay. The liquid is drawn from the bay through a liquid delivery tube into the atomizing nozzle. This is achieved by the aspirating effect (suction) caused at the nozzle end of the delivery tube by the flow of the high-pressure atomizing gas in the nozzle. When the liquid metal meets the high-velocity gas, it is broken up into droplets and sprayed as a jet. The droplets are quenched by the gaseous atmosphere in the chamber to solidify as powder particles. These particles, together with a substantial volume of cooling air, are then drawn through a chiller chamber into the collection system consisting of two sets of cyclones. After the cyclones, the powder is transported in an atmosphere of inert gas to the screens and pack-outs where they are packed under inert gas into drums and bins or onto railroad cars or trucks.

Fig. 1 Schematic of the Alcoa process for atomizing aluminum powder. Entire operation is under computer control. Powder is packed in drums or bins or is loaded for bulk shipment in trucks or railroad cars.

Because aluminum powder forms an explosive mixture with air over a wide range of metal-to-air ratios, aluminum powder plants have explosion hazards associated with them. The Alcoa process (Fig. 1) incorporates several safety features. The bottom of the chiller chamber is closed rather than open to prevent ingress of combustible dust and foreign matter into the atomizing zone. Cooling air entering the chiller chamber is filtered in order to remove foreign matter and ignition sources, such as rust particles. The chiller chamber walls are reinforced. Explosion vents are located in the chiller chamber, along the exhaust duct, and in other critical locations, such as the lids of the cyclone abatement chambers. These

vents open up and minimize pressure buildup in the event of an explosion. To reduce explosion hazards between the atomizing nozzle and the cyclones, the airborne concentration of aluminum powder is kept to a level below which forms an explosive mixture (lower explosive limit, LEL). Airflow in the system is created by means of an eductor, so that powder does not have to travel through a rotating fan, which eliminates a significant static electricity ignition source. The whole plant is built in stainless steel and fully grounded to eliminate the danger of sparking and static buildup. The transport of powder after the cyclones is done in an atmosphere of inert gas, as noted above, to substantially reduce the explosion hazard. A certain level of oxygen is maintained in the inert gas atmosphere to ensure that the powder does not become pyrophoric. Workers are removed entirely from areas containing the greatest hazards by substitution of computerized controls, remote TV surveillance, and automated operations. Conductive shoes are provided to workers in powder handling areas. Good housekeeping and training of operators are essential elements of safe operation in an aluminum powder plant. Regular audits of plant equipment and operating practices are carried out to ensure compliance. Areas of improvement identified in such audits and by risk assessment procedures are implemented on a timely basis. The industry has recognized the hazards of aluminum powder. Manufacturers and users of aluminum powders cooperate in reporting and analyzing incidents through the auspices of The Aluminum Association. The Association holds safety workshops and supports research relating to safety funded by the members. Control of particulate emissions in industrial plants is an important criterion. In the United States, these requirements vary from state to state. Pennsylvania has one of the strictest specifications limiting particulate emissions rate to 0.09 g/Nm3 (0.04 grains/ft3) of dry process air for aluminum powder plants. To meet these stringent requirements, manufacturers of aluminum powders often collect the powder in two stages of cycloning followed by a baghouse. Opacity of the plume from the plant is monitored and recorded. The upper limit for opacity is often specified at 10% max. These two specifications of emissions do not in general correspond to each other for aluminum powder plants (Fig. 2). This may be attributed to fine particles (5 m. For sedimentation in liquids, particle sizes down to 0.1 m can be determined. Particles in the suspension must be perfectly dispersed, and the suspension must be diluted enough to guarantee independent motion, which translates to a maximum concentration of 1 vol% of particles in the suspending medium.

Finally, wall effects should be minimized; consequently, the inside diameter of the sedimentation chamber should be sufficiently large to overcome this phenomenon.

References

1. L. Svarovsky, Solid-Liquid Separation, Butterworths, 1977, p 15-23 2. T. Allen, Particle Size Measurements, 3rd ed., Chapman and Hall, 1981 Sedimentation Methods for Classifying Metal Powders

Micromerographs This equipment is a sedimentation balance occasionally used for determining the particle size distribution of subsieve metal powders. The powder is suspended in air by projecting the sample with a burst of nitrogen through a deagglomerating device consisting of a conical annulus into the settling chamber. The chamber consists of a thermally insulated vertical aluminum tube with a 10 cm (3.94 in.) inside diameter that is 2.5 m (8.20 ft) high. The pan of an automatic balance that weighs the amount of powder settling on it is located at the bottom of the chamber. A recorder calculates the cumulative weight of powder settled as a function of time. Particle size distribution is calculated from this value on the basis of Stokes's law. Particle size distribution ranging from 2 to 100 m can be determined with this instrument. One limitation of the method is the tendency of powder particles of various sizes to cling to the wall of the column. Sedimentation Methods for Classifying Metal Powders

Light and X-Ray Turbidimetry Turbidimetry methods are widely used to determine the particle size distribution of refractory metal powders, such as tungsten and molybdenum, and of refractory metal compound powders, such as tungsten carbide. The turbidimeter is standardized in ASTM B 430. Because turbidimetry is used for refractory metals and compounds, several additional factors must be considered to obtain reproducible size and size distribution data. The particles of fine refractory metal powders are often agglomerated by sinter bonds. In ASTM B 430, determination of particle size distribution of the powder is described in the as-supplied condition and after deagglomeration by rod milling (laboratory milling). To obtain consistent, reproducible results by turbidimetry, not only the turbidimetric procedure, but also the deagglomeration procedure, must be standardized. Dispersion of the powder in the liquid before turbidimetric analysis must also be standardized. Figure 1 shows a typical turbidimeter. The procedure for determining particle size distribution of refractory metal powders with this instrument is included in ASTM B 430. A sample of powder dispersed in a liquid is poured into a glass cell and allowed to settle. A collimated beam of light is passed through the cell at a level having a known vertical distance, h, from the liquid level. The intensity of the light beam is determined by the current generated in a photocell. The current is passed through a potentiometer, where the voltage drop across it is measured by a recording millivoltmeter.

Fig. 1 Schematic of turbidimeter

The reading of the millivolt recorder for the intensity of the light beam shining through the clear solution is adjusted to 100%, while the concentration of the suspension is adjusted so that the reading for the intensity of the light beam through the suspension before any settling has occurred ranges from 20 to 40% of that through the clear solution. As the suspension settles, the projected area of the particles in the suspension decreases, and the intensity of the light beam increases. At the beginning of the settling, all particle sizes are uniformly distributed through the volume of the sedimentation cell. As settling proceeds, large particles settle faster than small particles. After a given time (tx), all particles larger in diameter than x have settled below the level of the light beam. The concentration of particles at the light beam level is now equal to the original concentration of particles, minus all particles with diameters equal to or larger than x. The projected surface of the particles at time, tx, is therefore smaller than that of the particles in the original suspension, and the intensity of the transmitted light is greater. Determination of Particle Size Distribution. To obtain information on particle size distribution from the plot of

light intensity versus time, use is made of the relationship between the total weight of n particles of size x, which is proportional to nx3, and the projected surface area of the n particles, which is proportional to nx2. Therefore, the cumulative weight of particles up to a given particle size xlim, which is:

is proportional to

which is the integral of the product of particle size and projected surface area integrated from 0 to size xlim. Turbidimetric measurements using white light are relatively inexpensive and have proven quite reproducible and useful. They are used in research and in routine analyses for comparing different lots of refractory metal powders. These methods generally are used for comparative work on the same type of material. When x-rays instead of white light are used to determine particle size distribution of a subsieve particle suspension, the attenuation of the x-ray beam intensity is proportional to the mass of the powder particles rather than their projected area. Sedimentation Methods for Classifying Metal Powders

Roller Air Analyzers

This apparatus determined the particle size distribution of metal powders by suspension in a stream of air. Using this apparatus, a powder is classified into particle size fractions that range from 5 to 40 m. A stream of high-velocity air flowing through a nozzle of suitable size impinges the powder sample contained in a U-tube, so that the powder becomes dispersed within the stream. The velocity (v) of the air stream through the chamber balances the settling velocity of particles with diameter and density determined by a Stokes law calculation:

v = 29.9 × 10-4 x2 where v is the velocity of the air stream, cm/s; is the particle density, g/cm3; and x is the particle diameter, m. Given this velocity, particles with a size smaller than x are carried through the settling chamber into the collecting system, which consists of an extraction thimble. Larger particles fall back into a U-tube. By using a series of vertical settling chambers with diameters in the ratio 1 to 2 to 4 to 8 and a constant volumetric rate of flow, the powder could be classified into particle size fractions with the maximum sizes in a ratio of 1 to 2 to 4 to 8 (for example, 5, 10, 20, and 40 m). As previously noted, this sedimentation is no longer in use. Sedimentation Methods for Classifying Metal Powders

Other Sedimentation Methods Besides gravity sedimentation, other methods of particle size measurement also employ the Stokes principle. Andreasen's Pipette. In the pipette method of size analysis, the concentration changes that occur within a settling suspension are followed by using a pipette to draw off definite volumes. Stokes's law is used to calculate the size distribution.

In photosedimentation, a narrow horizontal beam of parallel light is projected through a suspension at a depth, h, onto a photocell. In a homogeneous suspension, if the particles are allowed to settle, the number of particles that leave the light beam will be balanced by the number entering it from above. However, after the largest particle, Dm, in the suspension has fallen from the surface to the measurement zone, the emergent light flux will begin to increase, because no more particles of this size will enter the measurement zone from above. Hence, the concentration of particles in the light beam at any time, t, will be the concentration of particles smaller than Dst, where Dst is given by the Stokes equation (Ref 3). The particle size distribution is calculated from the relationship between the attenuation of the light beam and the projected surface area of the particles. X-Ray Sedimentation. Similar to the use of a light beam, x-rays can be used in combination with the gravity settling

of particles. In this case, x-ray density is proportional to the weight of the powder in the beam:

I = Io exp (-BC) where B is a constant and C is the concentration of powder in the beam. The x-ray density, Xd, is defined as:

The sedigraph is a well-known instrument based on the gravitational settling and x-ray absorption analysis. This instrument automatically presents the cumulative percentage frequency, and the sedimentation cell is driven in such a way that the concentration is recorded directly as a function of the Stokes' diameter. The x-rays from a 22.6 keV source are collimated into a narrow beam and pass through a fixed thickness of suspension. The sedimentation cell (Fig. 2) is filled and emptied with the suspension by a built-in pump. The transmitted radiation is detected as pulses by a xenon-filled scintillation detector. The counting electronics give a voltage that is proportional to the powder concentration. Particle size distribution is calculated by the application of Stokes' law to these data.

Fig. 2 Sedigraph schematic

Centrifugal Techniques. Because the centrifuging process speeds up the rate of settling of particles, it overcomes one

of the serious disadvantages of the gravitational sedimentation techniques, which take an unduly long time for particles 500

or

(Dm)max where dm is diameter, m, and and

Density, g/cm3 dm (max), m

2 500

5 62

10 19

0

500| -

0|

-1.5

are densities of particle and liquid, respectively. Typical values are:

17 8

Below a value of 100 for this function, sieving becomes increasingly effective. Thus, some samples may require both sieve and electrozone analyses, with blending of the two data sets in the overlap region of 100 to 500 for this size-density function. Metal Conductivity/Magnetism. The oxide surface layer present on most metal particles causes them to appear to be nonconductive, in effect creating a barrier impedance. Electrozone currents high enough to produce 1 to 3 V across the surface layer overcome this barrier impedance, thus causing conduction for only part of the pulse cycle (prior to the occurrence of sulfur polarization and ion depletion). Consequently, the particle pulse will erroneously appear to be smaller.

This effect can be detected by test observation of the total relative volume of particles per unit volume of suspension with increased current value. Samples may be treated with coating reagents or surfactants to increase surface layer impedance. Orifice current settings then must be kept below the level where this effect appears. Similarly, porous particles appear to be too small, due to the conductance of the pores aligned with the electric field, unless they are plugged with nonconductive material. Magnetic powders may be dispersed for a sufficient time to permit electrozone measurement if they are degaussed for 20 min at temperatures as high as 750 °C (1380 °F) (but below the sintering point, however) and/or spatulated with viscous honey to inhibit reagglomeration. A microscope slide covered with the honey suspension is then dipped into the electrolyte in the stirred sample beaker. In this manner, data acquisition is initiated instantly, while the suspension is being formed from the dissolving honey. Advantages. The electrozone method rapidly measures particle size distribution with excellent low-end sensitivity and

with good resolution (volume response) and precision (minimal side effects). By contrast, photozone (photicsensing zone) methods, which include light-beam scattering, blockage, and diffraction, have much lower resolution and precision. This is due to their area response and several significant side effects, including photic properties of particles, high coincidence levels, and beam/sensor instabilities. Sieving and sedimentation methods have similar limitations to response and side effects and are much slower in producing data. Electric and Optical Sensing Zone Analysis of Powders

Optical Sensing Zone The light blockage principle (obscuration) was used by Carver in 1958 to develop a method for counting and sizing particulate contamination in hydraulic fluids (Ref 1, 2). To date, this principle has found application in other particle-fluid systems and has been used for the measurement of metal powder particles.

In an apparatus using the light-obscuration principles, particles are suspended in a fluid with a refractive index that differs from the refractive index of the particles. The suspension is then passed through a restricting orifice--the sensing zone-across which a collimated white light beam passes. Figure 2 illustrates the sensing zone. The beam of light falls on a photodetector, which measures the intensity of the light passing through the zone.

Fig. 2 Sensing zone for (1 to 45

m) particle sensor based on light obscuration

Large particles block the light beam, thus lowering the intensity with which the beam strikes the sensor. Small particle sizes (80 psi in a short distance provides the desired force. Usually this is sufficient energy to disperse the samples.

Fig. 11 Powder jet systems

Comparison of Wet and Dry Analysis. The increasing popularity of dry analysis has caused comparisons of wet

and dry analysis. In countless cases, these two analyses have produced comparable results. In cases where differences have persisted, issues have been identified to explain the differences. In some cases, powders have flowed so poorly that they could not be fed properly. The addition of a small amount of silica enabled proper feeding of the material. In the case of strongly agglomerated particles, the particles actually behaved that way in the dry process. As a result, the dry, agglomerated results were actually more correct. Despite problems with dry dispersion, dry analysis is a firmly established technique. Automated Features. Most of the automation that has been developed to date has been in the area of wet analyses.

This can be seen in areas such as the simplification of the analysis through personal computer (PC) programs or the use of sophisticated programs to allow one button analysis. Automatic alignment of the optical bench of the instrument is also a feature. Although these are important developments, they typically make operation more convenient but not necessarily faster. The major improvement in this area has been in the form of automated feeding of the sample or the liquid used to dilute the sample. These systems can take on a variety of configurations depending on the type of sample that is being analyzed. Liquid feeding can be achieved through a variety of mechanisms. In the simplest configuration, a container of fluid is

placed at a high elevation and allowed to flow into the instrument. In most cases, a source of fluid at a positive pressure fluid is available to introduce the liquid at an appropriate rate and in a correct volume. Reservoir systems also afford the opportunity to select a variety of liquid and surfactants to be fed into the instrument. This versatility increases the speed of sample preparation, sample analysis, and sample clean up. Sample feeding is a major innovation for automation. A variety of configurations are available and perform quite

differently. The best systems are those that are specifically designed for particle size analyzers rather than those that are "borrowed" from a gas chromatograph or similar analyzer. In specially designed systems, different samples are fed to the instrument, and a scheduler program allows for many types of experiments to be run. These include dispersion curves,

kinetic studies, stability experiments, reproducibility tests, etc. All of these can be completed in an unattended fashion with increased time savings and sample throughput. On-Line Analysis. Typically, the sample is in a jar or beaker for analysis, which means an operator mechanically withdrew the sample to be analyzed. This is an extra step requiring additional time and effort and is a possible source of additional error. To overcome these issues, laser light-scattering instruments also are used for on-line process analysis (Ref 12). While this marks the evolution of all analytical technologies, it is a very recent development in the particle sizing arena.

Historically the acceptance of this technology for plant use has been limited by many problems. They include the lack of ability to obtain a representative sample, lack of ability to keep the sample dispersed, lack of adequate cleaning for the instrument, and lack of ruggedness in the instrument design. Recently these problems have been solved for some of the well understood and better behaved processes. Many discussions of this analysis are now encountered in the literature (Ref 13). On-line analysis of dry powders is becoming more common. Examples of on-line measurement of wet systems also are becoming quite common (Ref 15). Future Developments. In the past 10 to 12 years, an explosive growth has occurred in the popularity of laser light

scattering as a technology for performing particle-size analysis. This popularity has led to the development of many new features and capabilities. The range of analysis has been greatly expanded as have the types of samples that can be analyzed. The analysis has been automated, and the reproducibility and repeatability have been improved. Even the algorithms used to determine the particle size have been improved. There are some indications that it can be used routinely in an on-line application. All of these trends are expected to continue well into the future. As new types of particles are developed and new particle systems are introduced, adaptations will develop to feed them to a laser lightscattering instrument to measure the particle size.

References cited in this section

8. "Determination of Particle Size Distributions--Laser Diffraction Methods," ISO TC 24/SC4 Draft 2, to be published 12. V.A. Hackley, Measuring Particle Size of Slurries with Sound Waves, Ceram. Ind., June 1997, p 52 13. A.P. Malcolmson and M.P. Bonin, Moving Laser Diffraction Particle Size Analysis from Lab to Line, Ceram. Ind., April 1996, p 106 15. M. Bumiller and T. Oja, Measuring Particle Size at High Concentration, Ceram. Ind., July 1997, p 24 Light Scattering Measurement of Metal Powders Michael C. Pohl, Horiba Instruments, Inc.

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

H.N. Frock, in Powder Metallurgy, Vol 2, ASM Handbook, American Society for Metals, 1984, p 216 E. Orr, New Pigment Control Techniques: Part 2, Am. Paint Coat. J., 14 Aug 1995, p 43 G. Mie, Ann. Phys., Vol 25, 1908, p 377 M. Kerker, The Scattering of Light and Other Electromagnetic Radiation, Academic Press, 1969 T. Allen, Particle Size Measurement, 4th ed., Chapman & Hall, 1990, p 484 G. Row, Particle Sizing Technologies, Int. Ceram., in Issue 1, March 1997, p 57 CRC Handbook of Chemistry and Physics, CRC Press "Determination of Particle Size Distributions--Laser Diffraction Methods," ISO TC 24/SC4 Draft 2, to be published 9. W.C. McCrone, The Particle Atlas, Vol 1, Instrumentation and Techniques, Ann Arbor Science, 1973, p 72

10. D. Pugh, Am. Ceram. Soc. Bull., Vol 76 (No. 7), July 1997, p 44 11. S. Twomey, On the Numerical Solution of Fredholm Integral Equations, J. Assoc. Comp., 10 March 1963, p 97 12. V.A. Hackley, Measuring Particle Size of Slurries with Sound Waves, Ceram. Ind., June 1997, p 52 13. A.P. Malcolmson and M.P. Bonin, Moving Laser Diffraction Particle Size Analysis from Lab to Line, Ceram. Ind., April 1996, p 106 14. T.L. Harvill, J.H. Hoog, and D.J. Holve, In-Process Particle Size Distribution Measurements for Pharmaceutical Applications, Powder Handl. Process., Vol 7 (No. 2), April 1995, p 139 15. M. Bumiller and T. Oja, Measuring Particle Size at High Concentration, Ceram. Ind., July 1997, p 24

Time of Flight Measurement of Metal Powders Brian H. Kaye, Department of Physics and Astronomy, Laurentian University

Introduction THE GALAI INSTRUMENT is one of the first time-of-flight instruments used to characterize the size of powder grains (Galai Production Ltd., Ha Hameak, Israel). For a time this instrument was widely sold in the United States by the Brinkmann Company, so it is sometimes referred to as the Brinkmann size analyzer. Figure 1 shows the basic principles of the Galai instrument. Light from a laser is focused to a fine beam, and the beam is rotated by the rotating optical wedge. As the laser tracks a circular path in the cell containing the particles to be characterized, the size of the profile is measured by the time required for the laser beam to track across the profile of the powder grain. The instrument determines when a particle is intercepted at its maximum width, and it is calibrated using standard fine particles. The instrument also incorporates a video camera to view the particles being characterized at right angles to the rotating laser (Ref 1).

Fig. 1 Galai particle size analyzer. (a) Basic layout of the Galai instrument. (b) Laser beam tracing a circular path within the cell and the logic of the instrument rejecting any particles that are off center or out of focus. (c) Laser beam blocked from reaching the photodetector

Figure 2 demonstrates the accuracy of the time-of-flight, or time-of-transit method, by the agreement in the measurement of the powder-size distribution measured by the Galai instrument, with the known size distribution of a standard powder prepared and distributed by the European Commission on Standard Materials, Brussels, Belgium. The sample studied in Fig. 2 is BCR-67.

Fig. 2 Data for BCR-67, standard quartz test powder, obtained form the Galai particle size analyzed compared with known size distribution

Reference

1. O.B. Weiner, W. Tscharnuter, and N. Karasikov, Improvements in Accuracy and Speed Using the Time-ofTransition Method and Dynamic Image Analysis for Particle Sizing; Some Real World Examples, Proc. of the Symposium on Particle Size (Orlando, FL), American Chemical Society, 1996 Time of Flight Measurement of Metal Powders Brian H. Kaye, Department of Physics and Astronomy, Laurentian University

Other Time-of-Flight Measurement Systems Another system that uses time-of-transit measurements to obtain particle size information is manufactured by Lasentec (Ref 2). This system is often used in on-line monitoring of material. Figure 3 shows the basic system used in their equipment. In technical literature (Ref 2) Lasentec states: "The focused beam will cross the particle structure on a straight line between any two points on the edge of the structure. The distance between two points is a chord length. Thousands of chords are typically counted per second. The number of chords measured over a specific time period are sorted by chord length into a 38 channel distribution. The resulting chord length distribution tracks changes in particle geometry, a function of the shape and dimension of the particles, and particle structures as they naturally exist in process."

Fig. 3 Lasentec focused beam reflectance measurement system

Two time-of-flight instruments that study the size of fine particles in the aerosol format are available from TSI Inc., St. Paul, MN, and Amherst Process instruments Inc., Amherst, MA. Figure 4(a) shows the basic system of the instrument manufactured by TSI, known as the aerodynamic particle sizer (APS) analyzer. After the powder to be characterized is made into an aerosol, the aerosol is fed into the equipment. As shown in the Fig. 4(a), some of the aerosol under study is filtered to create a clean air sheath, which confines the aerosol being created to the center of the measurement zone. The two laser beams are 125 m apart and 200 m downstream from the nozzle orifice. The air feed to the interrogation zone is used to accelerate the particles, which respond to the accelerating forces at varying rates depending on their mass. The time-of-flight between the two laser beams is measured and used to calculate the particle velocity. The instrument is able to measure particles in the region of 0.5 to 30 m. The instrument is calibrated using standard polystyrene spheres. Figure 4(b) shows typical results for the instrument.

Fig. 4 The TSI aerodynamic particle sizer. (a) Basic layout. (b) Results from a mixture of three sizes of polystyrene latex spheres. (c) Results for a therapeutic aerosol

The parameter of the powder grains measured by this instrument is known as the aerodynamic diameter of the particles. The aerodynamic diameter of a particle is defined as the size of a sphere of unit density having the same velocity as the particle under the same flow conditions. This aerodynamic diameter differs from diameters measured in other methods of particle size, and data need to be interpreted in a knowledgeable manner. Several publications deal with the relationship between particle sizes measured by time-of-flight and other well-known methods of size analysis, such as microscope image analysis and diffractometers (Ref 3). The TSI Inc. system works at subsonic velocities and uses one photocell. In 1994, TSI Inc. announced an advanced version of their instrument, the Model 3320 aerodynamic particle size spectrometer, using double crested optics. TSI Inc. claims that the new model sizes particles from 0.5 to 20 m aerodynamic size and can operate at 1000 particles/cm3. Figure 5 shows the system used in the Aerosizer, manufactured by Amherst Process Instruments Inc. Note that one of the problems in using instruments such as a time-of-flight aerosol spectrometer is ensuring that the powder under study is efficiently dispersed as an aerosol (Ref 4). The Amherst instrument has an auxiliary piece of equipment, the AeroDisperser, for dispersing powders that enable the investigator to vary the shear rate applied to the powder being aerosolized (Fig. 6). The equipment is used to carry out investigations of various shear rates applied to powder as it is aerosolized, until further increase in the shear rate does not result in a difference in distribution measurements. The aerosol being interrogated is surrounded by sheath air as the particles under study are accelerated into the interrogation zone. As shown in Fig. 5, the interrogation zone differs from that of the APS equipment in that two photocells are used to measure the optical energy scattered by the particles as they move through the two beams. A sophisticated electronic editor distinguishes which signal arriving at the photodetectors is associated with the initial signal from a given particle as it enters the first laser beam. The Aerosizer operates at sonic velocities and can count fine particles at a very high rate. The equipment measures the aerodynamic diameter and is calibrated using particles of known size. The output can be either a measure of the number of particles of a given aerodynamic size (Fig. 7a) or as a transformation into geometric size by volume (Fig. 7b). The geometric size by volume is a transformation of the aerodynamic diameter and takes into account the density of the particles being measured.

Fig. 5 Schematic diagram of the inspection zone of the Amherst Process Instruments Inc. Aerosizer time-offlight particle sizer

Fig. 6 Aero-Disperser used with the Aerosizer for aerosol preparation using controlled shear rates to disperse the sample and ensure that there is no agglomeration

Fig. 7 Output of measurements for the size distribution of an irregular iron powder. (a) Number size distribution of the iron powder. (b) Volume size distribution of the iron powder

References cited in this section

2. Technical literature, Lasentec, Redmond, WA 3. F.M. Etzler and M.S. Sanderson, Particle Size Analysis: A Comparative Study of Various Methods, Part. Part. Syst. Charact., Vol 12, 1995, p 217-224 4. B.H. Kaye, Generating Aerosols, KONA, (No. 15), 1997, p 68-80 Time of Flight Measurement of Metal Powders Brian H. Kaye, Department of Physics and Astronomy, Laurentian University

References 1. O.B. Weiner, W. Tscharnuter, and N. Karasikov, Improvements in Accuracy and Speed Using the Time-ofTransition Method and Dynamic Image Analysis for Particle Sizing; Some Real World Examples, Proc. of the Symposium on Particle Size (Orlando, FL), American Chemical Society, 1996 2. Technical literature, Lasentec, Redmond, WA 3. F.M. Etzler and M.S. Sanderson, Particle Size Analysis: A Comparative Study of Various Methods, Part. Part. Syst. Charact., Vol 12, 1995, p 217-224 4. B.H. Kaye, Generating Aerosols, KONA, (No. 15), 1997, p 68-80

Particle Image Analysis Introduction MICROSCOPY is the most definite method of particle size analysis, because individual particles are observed and measured. However, optical and electron microscopy also afford the opportunity for particle shape analysis. Particle shape, like particle size, is a primary powder characteristic that must be considered when determining the optimum utilization of a P/M material. Behavioral characteristics of a metal powder, such as flow rate, apparent density, compressibility, and sinterability, are all influenced by particle shape and size. This article briefly reviews the common methods of optical and electron microscopy for image analysis of particle size distributions and particle shape characterization. Image analysis of particle size distribution is a very precise quantitative practice, while particle shape analysis is more qualitative in nature. Advancements in qualitative shape analysis are based on the use of the scanning electron microscope for determination of particle shape. In addition to its high magnification capability, this device produces an image with a three-dimensional appearance. Because of its advantages, scanning electron microscopy (SEM) has replaced optical microscopy to a large extent in the area of shape and topographical analyses of metal powders. Quantitative shape analysis has posed numerous difficulties. However, with the advent of advanced computers and graphics systems, interest in quantitative analysis has increased. Through the use of optical and scanning electron microscopy, qualitative descriptors have been developed to label different particle shapes. As computational technology advanced, mathematical algorithms were written to describe particle shape. Fourier analysis and fractal geometry are two mathematical disciplines that are being applied to describe particle shape. Because of the computational intensity of these disciplines, most of the work has been confined to a two-dimensional space. The current challenge is to develop threedimensional modeling programs that can accurately describe the shape of irregularly shaped particles.

Particle Image Analysis

Image Analysis of Powder Size

Microscopy is a very precise method of particle size analysis, but some of the techniques used are more art than science. However, if the basic principles of sampling, preparation, and counting are followed, a precise count can be made with a thorough understanding of the nature of the particles being studied. ASTM standard E 20 details the use of microscopy for particle sizing. Particle Image Analysis

Size Measurements Various techniques are used to measure the size of irregularly shaped particles when viewed through a microscope. This has resulted in different measurements, which are used to classify these two-dimensional particle images in terms of an equivalent spherical particle. Accepted measurements (Fig. 1) include: • • • • • •

Feret's diameter (F): The maximum length of the particle measured in a fixed direction Martin's diameter (M): The length of a line that bisects the area of the particle image; all particles measured in the same direction Project area diameter (da): The diameter of a circle with the same area as the two-dimensional image of the particle Longest dimension: The maximum Feret's diameter for each particle: no set direction Perimeter diameter (dp): The diameter of a circle having the same circumference as the perimeter of the particle Maximum horizontal intercept: The length of the longest line that can be drawn through the particle in a fixed direction

Fig. 1 Techniques for measuring sizes of irregularly shaped particles. See text for identification of variables.

Feret's diameter is the easiest to measure manually. The average Feret's diameter is related to the perimeter of the convex hull of the particle, as shown in Fig. 2, by the relationship:

where PCH is the perimeter of the convex hull. Martin's diameter is related to the specific surface Sv of the particle by:

The projected area diameter gives the best estimate of the true cross-sectional area of the particle.

Fig. 2 Feret's diameter

Generally, the three diameters above are related to one another by the expression M < da < F. Their ratios remain fairly constant for a given material. The expression F/M has been used as a shape function, which is equal to 1 for spherical particles and increases in magnitude as particle shape becomes more acicular. Perimeter diameter and maximum horizontal intercept (or longest chord) are measured easily by some of the automatic particle sizing analyzers. Particle Image Analysis

Sampling Techniques The most important step in any particle size analysis, but especially by microscopic techniques, is sampling from the bulk. Because an extremely small quantity of material is used to determine the particulate size, an accurate analysis cannot be obtained if the bulk material is not properly sampled. Particles tend to segregate according to size. If handling has caused vibration of the sample, coarse material tends to collect near the surface. When free-flowing material is poured into a pile, the coarse portion collects near the outside of the pile, and the fines concentrate in the center. The following sampling techniques are recommended for obtaining a valid representation of particle sizes in a laboratory sample. Chute rifflers (Fig. 3) consist of V-shaped hoppers that feed a series of chutes, which alternately feed into trays on opposite sides of the hopper. A sample that is properly poured into the hopper is halved, and the process is repeated until a suitable sample size is obtained. Sampling efficiency depends on an even feed of the sample into the hopper and on the width of the chutes. Numerous narrow chutes provide more accurate sample statistics than a few wide chutes, provided the powder can flow easily through them.

Fig. 3 Chute riffler

Spinning rifflers (Fig. 4) divide a sample by feeding it into a number of containers that rotate on a table under the feed.

This method is efficient, provided the table rotates at least 100 times during the flow of the sample.

Fig. 4 Spinning riffler

Oscillating Sampler. The feed hopper oscillates rapidly between two sampling cups, thereby halving the sample. Feed

times must be long compared to oscillation times for optimum efficiency. Cone and Cup Sampler. In this device, the sample is fed onto a cone that disperses the powder in all directions along

the cone surface. A sample cup is rotated around the base of the cone, gathering portions of the sample at every position. The width of the cup and length of the cone periphery determine the sample reduction. Sampling tables (Fig. 5) are used to reduce larger quantities of the sample. Sampling tables consist of an inclined

plane with a series of holes that halve the sample in each of the four stages, leaving sample.

of the original volume as a final

Fig. 5 Sampling table diagram

Cone and quartering involves pouring the sample into a cone-shaped pile on a glass plate, flattening the top, and

dividing it into quarters with a thin metal blade. One of the quarters can then be mixed, repoured, and requartered until a suitably small sample is obtained. This method requires some technique to ensure symmetry of particle sizes in the pile and accurate quartering. Scoop sampling is the least precise, but perhaps the most frequently used, sampling technique. The sample container is

shaken, and a small portion is extracted with a scoop or spatula. However, segregation can be imparted to the sample by the shaking method. The scoop sampling technique that provides the least variance is shown in Fig. 6 and consists of shaking the container back and forth while rotating. Because particle sampling is the initial and most important step in particle size analysis, miniature sampling devices should be used whenever possible.

Fig. 6 Scoop sampling apparatus showing direction of shaking

Particle Image Analysis

Optical Microscopy The optical microscope is used to count and size particles ranging from 100 to 0.5 m diameter. The resolution of an optical microscope is optimally about 0.25 m, but diffraction effects at the edges of small particles result in significant errors in measurements within this size range. Resolution down to 0.1 m can be improved by using ultraviolet light and quartz optics. A severe limitation of the optical microscope is its small depth of field, which is about 10 m at 100 and 0.5 m at 1000×. This requires that the specimen powders be mounted carefully in one plane and eliminates the use of automatic counting devices on samples with wide particle size ranges. Another limitation is that due to the small number of particles usually counted, only the field of view can be characterized, not the entire lot from which the specimen was obtained. Preparation of slides for optical microscopy is an art, and the quality of a dispersion depends more on the skill of the preparer than on the particular technique used. Care must be taken to prevent segregation of particle sizes and to prevent the formation of agglomerates. Preparation techniques vary with the type of powder and depend on particle size, particle size range, and particle composition. Typical dispersion techniques are described below. Dry Mounting. When a permanent slide is not required, and the powder exhibits good flow characteristics, slides can be

prepared by "puffing" dry powder onto the slide or by applying it with a brush. Dispersing Fluids. In applications that require shearing action to disperse the powder, a dispersing fluid such as cedar

oil or glycerol can be used. The simplest technique involves placing a drop of dispersing fluid on a glass slide and placing a small amount of powder onto the drop. A cover slip is then placed carefully over the mounting fluid and pressed evenly with an eraser, while gently rotating the cover slip back and forth. Another method involves placing some of the powder on a glass slide, adding a few drops of dispersing fluid, and working the mixture with a flexible spatula or soft brush. More fluid is added until the proper concentration is obtained. A drop of this mixture is then transferred to a new slide and covered with a cover slip. Permanent Slides. A small amount of powder is placed in a beaker to which 2 or 3 mL (0.07 or 0.10 fluid oz) of a

solution of 1 to 2% collodion in butyl acetate is added. The mixture is stirred vigorously, and a drop of the suspension is placed on the surface of distilled water. The drop spreads and forms a thin film as the solvent evaporates. The film is then picked up on a microscope slide and dried.

Other films can be produced by dissolving parlodion in amyl acetate, Canada balsam in xylene, and polystyrene in xylene. Films are formed by casting a 1% solution on water or a % solution directly on a glass slide. After the films dry, the powder is applied by spraying or by letting a drop of aqueous suspension dry on the surface. The particles sink into the medium, forming a permanent slide. Melted Canada balsam or glycerol jelly also can be used to form permanent samples. Other Methods. Certain applications may require samples prepared by specialized techniques, such as heating

magnetic particles above their Curie points for dispersion. These examples generally serve as guidelines for technique development. Particle sizing usually is performed indirectly in the optical microscope, because the small depth of field frequently requires refocusing on individual particles. If the size range is narrow and the preparation is carefully made in a single plane, photographs or automatic particle counting devices can be used. Projection microscopes are also available that reduce operator fatigue for analyses that require numerous counts. Particle Image Analysis

Transmission Electron Microscopy The transmission electron microscope is used for counting particles that range from 0.001 to 5 m in diameter. This instrument has a large depth of field; consequently, all particles in the field of view are in focus regardless of size. Particles are usually not counted directly from the viewing screen of the transmission electron microscope. Photographs are normally taken, and counts are made from prints or projected images using the negatives. The electron beam is easily absorbed, and films greater than 100 to 200 nm (1000 to 2000 ) are completely opaque. It is therefore necessary to produce very thin support films on which powders can be dispersed for counting. These films usually are made of carbon. One of the best techniques for producing strong, flat carbon support films involves cleaving high-quality mica and placing it immediately in a vacuum evaporation unit. After a hard vacuum has been obtained, carbon is evaporated onto the mica surface to a thickness of about 10 nm (100 ). Film thickness can be calculated by including a white porcelain slide containing a drop of vacuum oil in the vacuum chamber. When the porcelain that is not covered by the oil appears as a light chocolate shade, the carbon film is about 10 nm (100 ) thick. The mica is then removed from the vacuum and slowly lowered at an angle into a beaker of distilled water. The carbon film floats off the mica onto the water surface, where it can be picked up on an electron microscope grid. Another technique involves casting a thin film of parlodion onto the surface of distilled water by depositing two or three drops of 1% parlodion dissolved in amyl acetate onto a clean glass slide. The slide is allowed to drain and dry, and the film is floated from the slide onto distilled water. The plastic film is then picked up on transmission electron microscopy (TEM) grids, placed in a vacuum evaporator, and coated with carbon. The parlodion can be carefully dissolved in amyl acetate by placing the grids on a fine stainless steel mesh bridge suspended in the center of a petri dish. Amyl acetate is added to the petri dish until the level of the liquid reaches the bottom of the mesh. The dish is then covered and left for a few hours, after which the grids are carefully removed. If the parlodion film is thin enough, it can be left on the grids. However, this requires fresh, water-free solutions and care in handling, or holes will form in the thin plastic film as the solvent evaporates. Prepared carbon substrates on grids are available. Powders are dispersed on the surface of support films by puffing from an aspirator or by allowing a drop of aqueous suspension to dry. Carbon films are hydrophobic; use of an aqueous suspension requires that they be rendered hydrophilic by either treatment with a thin solution of albumen or exposure to reactive oxygen in a low-temperature oxygen asher. Exposure to reactive oxygen must be done at a very low setting (4 W) for several seconds, or the carbon film will be destroyed.

Powders that require additional dispersion can be prepared by placing a small amount of powder on a clean glass slide and adding several drops of 1% parlodion in amyl acetate. This mixture is sandwiched between a second glass slide. As the amyl acetate evaporates, the mixture becomes more viscous. Immediately before the parlodion dries, the slides are sheared apart. This preparation can be viewed in the optical microscope to find the area of best dispersion. This area is then scribed into 3.2 mm ( in.) squares and floated from the glass slide onto distilled water. The squares are picked up on electron microscope grids and coated with carbon to add stability to the electron beam. For some samples that are difficult to disperse, ethylene glycol may be a suitable dispersing agent. A small amount of powder is added to a few drops of this viscous liquid on a glass slide and mulled until a good dispersion is obtained. A carbon substrate grid is then touched to the thin liquid film, which sticks to the grid. The grid is placed in a vacuum chamber and pumped for a few minutes, during which time the ethylene glycol sublimates, leaving the particles on the carbon substrate. Particle thicknesses can be measured in the transmission electron microscope by "shadowing" particle dispersions in which the particles sit on a substrate and are not embedded in a plastic film. This is done by evaporating a small amount of metal placed at an angle to the substrate surface in a vacuum chamber. The metal coats the surface and particles in a line-of-sight fashion, leaving a "shadow" cast behind the particles. Commercially available precision-sized latex spheres can be included with the powder sample so that the shadow length to particle height ratio can be calculated. Of the pure metals, platinum is best suited for shadowing because of its fine structure in an evaporated film. About 12 mm (0.5 in.) of 250 m (10 mil) platinum wire, evaporated from a point source at an angle of 30° and a distance of 100 mm (4 in.), casts a good shadow. Many other preparation techniques are used to observe powders in TEM, and most analysis methods are modified to suit the application and the operator's ability. Particle Image Analysis

Scanning Electron Microscopy The scanning electron microscope has a resolution of about 10 nm (100 ) and is capable of very low magnification (about 10×) up to about 50,000×. It therefore can be used to count particles ranging in size from 1 mm to 0.1 m. Particles smaller than 0.1 m usually have too low a contrast with the background to be counted efficiently. The scanning electron microscope has about 300× the depth of field of an optical microscope. The image in the scanning electron microscope usually is obtained by using the secondary electron output of the sample as it is scanned by a very narrow electron beam. The contrast of the image depends more on the topography of the sample than on differences in atomic number. Therefore, prepared powders must not be embedded in films, but dispersed on a smooth substrate. Any smooth surface can be used as a substrate. However, if energy dispersive x-ray analysis (EDXA) is to be performed for particle identification, a carbon or polystyrene surface is preferred. An excellent substrate can be made by placing a polystyrene pellet on a glass slide and heating it on a hot plate until it softens. A second glass slide is then placed over this slide and pressed until the pellet forms a thin disk. The slides are removed from the hot plate and pressed together until the polystyrene sets. The disk thus formed is as smooth as the glass and contains no elements that may hinder EDXA. For sample preparations using aqueous suspensions, polystyrene surfaces can be rendered hydrophilic by a brief treatment in an oxygen asher at low power (5 to 10 W for 5 s). While the substrates for SEM do not have to be as thin as those used for TEM, they must be conductive. Consequently, if glass or plastic surfaces are to be used, they must be coated with an evaporated metal (or carbon, for EDXA) film. This coating is usually applied after the particles have been dispersed on the surface. Many of the dispersing techniques used for TEM can be applied to SEM. Particle dusting, drying from liquid suspensions, and mulling in liquids that can be sublimed in a vacuum are suitable dispersing methods, depending on the powder. If the technique of mulling in parlodion and amyl acetate is used, parlodion can be removed in an oxygen asher, thus leaving the particles on the substrate. Suitable substrates include glass or metal, because they are not affected by the ashing.

The prepared sample should always be placed in the scanning electron microscope with the surfaces normal to the electron beam so that the magnification, which changes with working distance, will be the same on all areas of the viewing screen. Particle counting can be done directly from the viewing screen, from photographs, or by using an automatic image analyzer. Particle Image Analysis

Measurement Techniques Direct Measurements. Particle diameters can be measured directly in an optical microscope with the use of a filar

micrometer eyepiece. This eyepiece contains a scale and a movable cross hair that is operated by a calibrated knob on the side of the eyepiece. A particle is moved so that one side touches one of the fixed scale markings, and the cross hair is moved to touch the other side of the particle. The difference between the two readings is the Feret's diameter of the particle. The eyepiece is calibrated with an optical stage micrometer. This technique is time consuming, and because absolute measurements are not required for accurate size analysis, eyepiece graticules are usually used for direct measurements. Figure 7 shows typical examples of eyepiece graticules. These patterns are etched on glass disks that are positioned in the back focal plane of the microscope ocular and are therefore in the same focal plane as the particle images. Using one of these gauges, the size range, into which the actual diameter of a particle falls, can be measured easily. Feret's diameter, projected area diameter, or perimeter diameter are measured conveniently with graticules. Where measurements are made from photographs, projected images, or cathode ray tube screens, simple scales or plastic overlay graticules can be used to measure any of the accepted diameters.

Fig. 7 Eyepiece graticules

Image shearing eyepieces are used directly within the optical microscope. This device divides the optical beam into two parts by using mirrors and/or prisms. The distance between the two formed images is adjusted by shearing of the prisms, which is controlled by a micrometer dial on the side of the eyepiece. The maximum horizontal intercept is measured when the two images barely touch. With this device, a particle does not have to be moved to a particular position within the field of view to be measured.

The micrometer can be set to a given diameter, and the number of particles larger or smaller than this measurement can be counted, as determined by whether or not the images are touching. Red and green filters facilitate this by coloring the two images. Semiautomatic Techniques. Many devices are available to shorten the time required to complete a particle size

analysis. Most of these instruments record the number of particles in a size range, as judged by the analyst. Adjustable light spot analyzers use a circle of light projected onto a photograph. The diameter of the circle can be adjusted by a diaphragm; a foot switch causes one of a bank of registers to record the diameter according to the diaphragm setting. Each particle on the photograph is passed under the light, which is adjusted to measure the projected area diameter or the perimeter diameter. A transparent electronic graticule is a plastic sheet on which various size circles have been drawn. Under each circle, an electrical contact is connected to one of a bank of registers. When a contact is touched by the analyst with a pencil contact, the register advances one number. This device can be used with photographs, back-projected images, and cathode ray tube screens. The analyst decides which circle best represents the diameter of a particle and touches the corresponding contact. Recording calipers work in the same manner, with a bank of registers connected to the caliper spread adjuster. Any particles of the prescribed diameter can be recorded with this device. Recording micrometer eyepieces use a similar recording mechanism, but work directly within the optical microscope. Two movable cross hairs that move toward or away from each other with the turn of a micrometer screw are used to bracket a particle; a button is subsequently pushed to record the diameter. This technique is particularly useful when the depth of field prevents photography of the particles, because each particle can be sharply focused before measuring. Sensitive surfaces connected to computers facilitate counting from photographs. The computer can be programmed to record diameters when the opposite sides of particles are touched with a pointer, or specific areas of the particle are touched. The computer consequently compiles data to suit user requirements. Automatic techniques are used for counting large numbers of particles directly within the microscope. The main limitations of these techniques are the inability to distinguish particles that are touching and the fastidious sample preparation that is required.

Spot scanning devices use the moving spot from a cathode ray tube, which is projected through the microscope onto the specimen. When the spot passes over a particle, the light beam is interrupted, and a photocell records the particle. Particles are sized by scanning with different spot sizes. This technique also has been used on photographs, so electron micrographs can be used. The slit scanning method projects the microscope image onto a slit. The microscope slide is mechanically scanned, and the signal produced as the particle images pass the slit is recorded. The width of the slit can be varied to eliminate coincidence and overlap. Quantitative image analyzers are computer-controlled devices that use television cameras for direct analysis within an optical microscope. Analysis of photographs can also be accomplished. A quantitative image analyzer can be connected directly to a scanning electron microscope to control the scanning system. Particles can be counted according to their maximum horizontal chords, vertical chords, perimeters, or areas. Data Presentation. Table 1 shows the most precise form of presenting particle size data. However, graphical

representations are more concise and visually show the mean and deviation from the mean. Some graphical presentations can yield specific values for descriptive constants such as the mean, median, and the standard deviation.

Table 1 Particle size data Particle size range, m 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-20 20-30

No. of particles

Particles, %

0 3 8 15 79 163 121 64 28 13 2

0.0 0.6 1.6 3.0 15.9 32.9 24.4 12.9 5.7 2.6 0.4

Cumulative percent less than size 0 0.6 2.2 5.2 21.1 54.0 78.4 91.3 97.0 99.6 100

A histogram is a bar graph that illustrates the frequency of occurrence as a function of the size range. Figure 8 shows a typical histogram for a log-normal size distribution. The smooth curve drawn through the histogram is a valid sizefrequency curve if sufficient particles are counted and the number of size intervals is at least ten.

Fig. 8 Histogram and size frequency curve for log-normal size distribution

In cumulative plots, such as Fig. 9, the median particle size can be noted easily. These plots are made by plotting the percentage of particles less than (or greater than) a given particle size against the particle size.

Fig. 9 Cumulative plot used in determining median particle size

A special grid for plotting size-frequency data was developed so that the resulting curve is a straight line. This grid consists of a coordinate based on probability. Cumulative percentages are plotted on the probability scale against particle size on either an arithmetic or a log scale. Symmetrical size distributions, characterized by a bell-shaped size-frequency curve symmetrical to a vertical line through the mode (apex) point, plot as a straight line on arithmetic-probability paper. Most natural particle distributions do not follow this normal probability, but follow a skewed or log-normal distribution. These distributions provide a symmetrical size-frequency plot if the particle-size scale is logarithmic and forms a straight line on log-probability paper. The median particle size is found at the 50% point on these probability plots, and the standard deviation for arithmetic-probability plots can be calculated as:

= 84.13% size - 50% size = 50% size -15.87% size and for log-probability plots can be calculated as:

Particle Image Analysis

Particle Shape Characterization Brian H. Kaye, Department of Physics and Astronomy, Laurentian University

Figure 10 shows a typical scanning electron micrograph of an array of iron particles for the type of powder grains encountered by the powder metallurgist in various areas of technology. Many of the techniques for characterizing the shape of powder grains make use of image analysis using pictures such as that shown in Fig. 10(a) (Ref 1). When an array of fine particles to be characterized is inspected, the first thing to be done is to decide what constitutes a separate powder

grain. The economics of taking photographs using an expensive scanning electron microscope require working with a relatively dense field of view, which creates pseudoagglomerates from random juxtaposition. Thus, for the field of view shown in Fig. 10(a), the decision was made that the dashed lines represent separations between individual grains, which are close to each other.

Fig. 10 Iron powder grains illustrate the type of shape characterization problems facing powder metallurgists. (a) Electron micrograph of an iron powder. Source: D. Alliet, Xerox Corporation. (b) Erosion of an image can be used to separate particles.

Several commercial instruments are available for automated image analysis. Some automated image analysis systems have built-in logic procedures known as erosion logic. This procedure enables the device to strip ribbons of pixels away from an image to separate contiguous profiles to allow juxtapositioned particles to be evaluated as separate entities. Figure 10(b) shows the operation of erosion logic to separate profiles for subsequent image analysis. The software used in the work reported here was a program in which a television camera was used to obtain images through a light microscope. The simplest method for characterizing the shape of profiles such as those shown in Fig. 10(a) is to evaluate what is known as geometric shape factors. In many cases, studies are carried out using fundamental measures on a set of profiles,

such as those of Fig. 10(a), which are assumed to be in a position of maximum stability on the support system when they are photographed. It should be noted that three-dimensional measurements on powder grain profiles can be used if the image is shadow cast with either evaporated metal or light illumination used in an appropriate manner. Thus, Fig. 11 shows a set of pieces of material as they might have been viewed by an optical microscope. The profiles are illuminated at a given angle to create shadows shown in the diagram (Ref 2).

Fig. 11 Three-dimensional shape information determined by creating an image of a set of fine particle using shadow casting techniques.

The British Standards Institute has prepared a standard glossary of terms for use in the description of the appearance of powder grains (Ref 3): • •

• • • • • • • • •

Acicular: Needle shaped Angular: Sharp edged or roughly polyhedral shaped. (Polyhedral derives from poly, meaning many, and hedra, meaning a base; therefore, polyhedral is understood to be a geometric shape having many faces, each of which can act as a base.) Crystalline: A geometric shape freely developed in a liquid Dendritic: A branched crystalline shape Fibrous: Regularly or irregularly threadlike Flaky: No formal definition in the British Standards. (It is assumed that a flaky fine particle is recognizable.) Lamellar: Platelike Granular: Approximately equidimensional but irregularly shaped Irregular: Lacking any symmetry Modular: Rounded, irregularly shaped Spherical: Globular shaped

The pioneer work on the development of shape factors was carried out by Heywood (Ref 4, 5, 6) and Hausner (Ref 7). Figure 12 shows the length, width, and thickness of a powder grain, as defined by Heywood. Using these measures, Heywood defined the elongation ratio as the length divided by breadth and the flakiness ratio as the breadth divided by thickness. Extensive reviews of procedures for deriving shape indices from geometric measurements have been prepared by Phal et al. (Ref 8) and Loveland (Ref 9).

Fig. 12 Geometric shape indices defined on the basis of the magnitude of measurements made, either directly on the powder grain, or on images of the powder grains

Three-dimensional measurements of sugar grains have been used to develop three-dimensional graphs of a shape population variation within a powder (Ref 10). The original shape factors described by Heywood are not currently in widespread use. They have, to some extent, been displaced by the aspect ratio and chunkiness, both of which are defined in Fig. 12. Figure 13(a) shows the simple silhouettes of the profiles shown in Fig. 11. Figure 13(b) shows the aspect ratio distribution for this population of profiles. When the population of aspect ratios are plotted in this way, it appears that there are two Gaussian populations of shapes in the powder. When plotted as a distribution function, the link of shape with size is lost in the calculation of the distribution of aspect ratios. In populations containing a wide range of shapes, the aspect ratio is not always useful, because in theory, it can vary from 1 to , whereas the chunkiness can vary only between 0 and 1. In the graphical display known as the chunkiness-size domain, the shape of each profile can be plotted, as shown in Fig. 13(c). This type of plot retains the information on shape and size for each profile. Thus from the data display of Fig. 13, the technologist can see immediately that the shape increases with fineness of the powder, information that is not retrievable from Fig. 13(b).

Fig. 13 Shape of fine particles can be characterized by aspect ratio or its reciprocal, chunkiness. (a) Silhouettes of the fine particles shown in Fig. 11. (b) Aspect ratio distribution of a larger set of a fine particles like those of (a). (c) Chunkiness-size domain for the profiles describe in (b)

Figure 14 shows the chunkiness size domain for the powders shown in Fig. 10(a), as determined by image analysis. This data display shows that powder is bimodal in size distribution. This probably arises from the fact that because the powder is magnetic, some fine powders were clinging to the coarser grains when the powder was made by a sieve fractionation process.

Fig. 14 Chunkiness-size domain summary of data generated by image analysis of the iron powder of Fig. 10 demonstrates a bimodal distribution indicating that the power was probably sieved and that the smaller fine particles were most likely clinging to the larger fine particles.

If measurements are carried out using the dimensional information obtained by shadow casting the images, a triangular grid can be used to plot the three-dimensional information on the profiles. Figure 15 illustrates the way measurements are made on the three-dimensional structure of a grain and then turned into a graphical display. To numerically quantify the scatter of data points on such a diagram, the points in space can be treated as unit masses. The centroid and the moment of inertia can then be calculated as a measure of the range of shapes present in the powder. Although several descriptions of this technique exist, no publications describe how to use such a population of shapes and relate it to the functional behavior of a powder (Ref 11, 12).

Fig. 15 Three-dimensional shape of a fine particle summarized using triaxial graph paper. (a) Demonstration of the use of triaxial graph paper to plot the dimensions of simple shapes. (b) Direct measurement of the three dimensions of the particles of Fig. 11. (c) Dimensions of the particles of Fig. 11 using shadow casting to determine the third dimension

The gross shape information about a particle grain embodied in a geometric shape factor will govern the packing properties of a powder and is an important parameter to be considered because it contributes to the flow properties of a powder. When the surface reactivity of a powder and its sinterability is considered, the surface structure of the powder grains needs to be quantified.

Fourier Analysis. One method of characterizing the shape and texture of a powder grain is to generate and analyze

what is known as the geometric signature waveform of a powder grain (Ref 13, 14, 15). A two-dimensional profile of the particle grain is needed to generate a geometric signature waveform. In one of the more frequently used techniques, the centroid of the profile viewed as a laminar shape is chosen as a reference point. The magnitude of the vector of the series of angles as shown in Fig. 16(a) is then recorded. The high resolution recording of the vector shows that the pattern generated by the rotated vector has the appearance of a complex wave, known as the signature waveform of the profile shown in Fig. 16(b). If the signature waveform is treated by Fourier analysis, the power spectrum of the waveform contains information on the shape and the texture of the profile. Flook (Ref 15) has analyzed the profile in Fig. 16 and shown that the first 5 harmonics of the Fourier analysis describe the gross shape of the profile, whereas the strength of the harmonics from 5 to 25 describe the texture of the profile as shown in Fig. 16(c) and 16(d). The use of signature waveforms and Fourier analysis were pioneered by Beddow et al. in the study of metal powder grains (Ref 16). Difficulties arise in choosing this straightforward geometric signature waveform if the profile of the particle has deep indentations, because in this situation the vector does not have a unique value crossing over the indenture in the profile. This can be overcome in modern data processing by creating the Fourier transform of the actual profile directly on computer logic, but this has not been widely used in powder metallurgy.

Fig. 16 Geometric signature waveform of a profile subjected to Fourier analysis to extract information on structure and texture. (a) Method used to obtain a simple geometric signature waveform. (b) More detailed signature waveform for the profile of (a) along with its harmonic spectrum. (c) First 5 harmonics used to reconstruct the profile show the gross morphology of the profile. (d) First 25 harmonics used to reconstruct the profile and add textural information to the structure

Fractal Analysis. Another technique for characterizing the texture for a metal profile is to measure the fractal

dimension. Again current techniques available for measuring the fractal dimension are restricted to the use of twodimensional images. The concept of a fractal dimension for measuring the ruggedness of a boundary was pioneered by Mandelbrot (Ref 17). Mandelbrot proposed that one way of describing the ruggedness of a boundary in space is to add a fractional number to the topological dimension of the system being studied to describe space-filling ability. In the

description of the boundaries of powder grains, the topological dimension of a boundary in a two-dimensional space is one. Figure 17(a) shows basic concepts of systems put forward by Mandelbrot. There are several ways to characterize the fractal dimension of the profiles; one convenient technique is to explore the perimeter structure of various inspection resolutions and to construct a Richardson plot of the data generated.

Fig. 17 Properties of the mathematical construction known as the Koch Triadic Island (a) Construction algorithm of the Koch Triadic Island. Each straight-line segment is divided by three, and the center segment is replaced with two segments of the same length at 60° to each other. (b) Triadic Island illustrating the property of self similarity. The magnification of an image of a boundary cannot be known because it looks the same at any magnification.

The physical significance of these statements can be appreciated from the diagram summarized in Fig. 18. The first step in the evaluation of the fractal dimension of a profile is to digitize the profile as shown. A polygon from the perimeter of the profile is then constructed by counting off a number of steps around the profile. A polygon constructed in this way is an estimate of the perimeter at the resolution fixed by the number of steps taken around the perimeter. It can be shown that for an irregular perimeter that can be described by a fractal number, the plot of the estimates of the perimeter versus the size of the inspection stride (the term used to describe the step created by going a certain number of steps around the profile) on log-log graph paper generates a linear-data relationship. Furthermore, the absolute value of the slope is the fractional number to be used in the fractal dimension of the profile. This plot is known as the Richardson plot, named in honor of a pioneer scientist who studied the structure of irregular boundaries. Figure 18(b) and 18(c) show such a series of measurements and the data plot. A very rugged profile can have more than one data line on the graph. For example, in the case of powder formed by a fuming process such as a carbon black, a profile such as that of Fig. 19 can have two different data lines on the Richardson plot. The data line at high resolution represents the texture of the profile, whereas the data line of the coarser resolution is a measure of the overall gross structure of the profile. Figure 20 shows the shape, size, and fractal dimension of the population of powder grains from Fig. 10(a). Again the change of values of the three parameters can be quantified by treating the points representing the information as unit masses in space. This type of graph is used to look at the range of shapes in an abrasive powder.

Fig. 18 Boundary fractal dimension of a profile--a dimensionless descriptor of the ruggedness of a fine particle boundary. (a) First step of the equipaced method of determining the boundary fractal. Profile is converted to a series of equally spaced (x,y) coordinates. (b) Estimates of the perimeter of the profile made at a series of decreasing step sizes. (c) Data normalized by dividing by the Feter's diameter and plotting perimeter estimate versus resolution of inspection on log-log graph paper. Fractal dimension ( ) = 1 + |m| where m is the slope of the best fit line through the data.

Fig. 19 Structural boundary fractal dimension ( S) relates to the gross morphology of the fine particle. Textural boundary fractal ( T) relates to the smaller features of the profile.

Fig. 20 Individual powder grain of Fig. 10(a) characterized by its fractal dimension and combined with size and chunkiness to form a single point in three-dimensional space. When a group of powder grains are treated in this way, they produce a cloud of points characteristic of the powder.

References cited in this section

1. B.H. Kaye, Particle Shape Characterization, Chap. 2, Handbook of Powder Science and Technology, 2nd ed., M.E. Fayed and L. Otten, Ed., Chapman & Hall, 1997, p 35-52 2. T. Allen, Particle Size Analysis, 4th ed., Chapman & Hall, 1996 3. British Standards 2955 Glossary of Terms Relating to Powders, British Standards Institute, 1965 4. H. Heywood, Size and Shape Distribution of Lunar Fines Sample 12057, 72, Proc. of Second Lunar Science Conf., Vol 13, 1971, p 1989-2001 5. H. Heywood, Numerical Definitions of Particle Size and Shape, Chem. Ind., Vol 15, 1937, p 149-154 6. H. Heywood, Particle Shape Coefficients, J. Imp. Coll. Eng. Soc., Vol 8, 1954, p 25-33 7. H.H. Hausner, Characterization of the Powder Particle Shape, Proc. of the Symp. on Particle Size Analysis (Loughborough, England), The Society for Analytical Chemistry, 1967, p 20-77 8. M.H. Phal, G. Schädel, and H. Rumpf, Zusammenstellung von Teilchenform Beschreilbungsmethoden, Aufbereit.-Tech., Vol 5, 1973, p 7-11 9. R.P. Loveland, ASTM Committee E29.02 on Fineparticle Shape, Report, ASTM, 1975 10. B.H. Kaye, Chap. 10, Direct Characterization of Fineparticles, John Wiley & Sons, 1981 11. B.H. Kaye, G.G. Clark, and Y. Liu, Characterizing the Structure of Abrasive Fineparticles., Part. Part. Syst. Charact., Vol 9 (No. 1), 1992, p 1-8 12. R. Davies, A Simple Feature Based Representation of Particle Shape, Powder Technol., Vol 12, 1975, p 111-124 13. H.P. Schwartz and K.C. Shane, Measurement of Particle Shape by Fourier Analysis, Sedimentology, Vol 13, 1969, p 213-231 14. R. Ehrlich and B. Weinberg, An Exact Method for Characterization of Grain Shape, J. Sediment. Petrol., Vol 40 (No. 1), March 1970, p 205-212 15. A.G. Flook, A Comparison of Quantitative Methods of Shape Characterization, Acta Stereol., Vol 3, 1984, p 159-164 16. J.K. Beddow, M.D. Nasta, and G.C. Philip, Characteristics of Particle Signatures, Proc. of the Int. Conf. on Powders and Bulk Handling Systems, May 1977 (Chicago, IL), International Powder Institute, 1977

17. B.B. Mandelbrot, Fractals; Form, Chance and Dimension, Freeman, 1977 Particle Image Analysis

Particle Shape Factors*

The most common approach to describe and differentiate particle shapes has been the use of qualitative concepts. Two fundamental concepts have been used: (a) the dimensionality of the particle and (b) the surface contour of the particle. By the use of these concepts, a model system of shape characterization is presented in Fig. 21. Photomicrographs of several types of loose powders described in the International Standards Organization standard ISO 3252 are shown in Fig. 22. Basic shapes are: •

•

•

One-dimensional particles (Fig. 22a and 22d). Two different types of one-dimensional particles can be considered on the basis of their surface contour. One particle is smooth and the other particle is roughened with an irregular type of surface. Two-dimensional particles. These are very flat in nature and the surface contour of such particles is usually irregular. The dendritic type (Fig. 22c) is characterized by a tree-like shape and is often associated with electrolytic powders. However, secondary mechanical treatments often destroy such a shape. Flake particles (Fig. 22e) are also considered two-dimensional particles. Three-dimensional particles. Most powders are three dimensional in nature. These powders can be equiaxed and nodular. The simplest type of particle in this category is the spherical type (Fig. 22i). By departing from this perfect shape and contour, irregular particles (Fig. 22g) and nodular types (Fig. 22h) are obtained.

Fig. 21 Various shapes of powder particles and their methods of manufacture

Fig. 22 Common particle shapes as depicted in ISO 3252. (a) Acicular powder particles. (b) Angular powder particles. (c) Dendritic powder particles. (d) Fibrous powder particles. (e) Flaky powder particles. (f) Granular powder particles. (g) Irregular powder particles. (h) Nodular powder particles. (j) Spheroidal powder particles

Porous particles are different from irregular particles because of the presence of porosity, which can also be very irregular in both size and shape. Internal porosity may be isolated or interconnected. Large amounts of porosity make shape characterization difficult. It is best observed by examination of cross sections of mounted powder.

Note cited in this section

*

Adapted from ASM Handbook, Volume 7, 1985, p 233-241

Particle Image Analysis

Electron Microscopy Qualitative SEM Shape Analysis. The most significant feature of the scanning electron microscope, in addition to its high magnification capability (useful magnifications beyond 10,000×), is its ability to produce images with a threedimensional appearance. This ability derives from the fact that the depth of field for the scanning electron microscope is over 100× that of the optical microscope. This increased depth of field (ranging from 1 m at 10,000× to 2 mm at 10×) also accounts for the extensive particulate analysis performed on the scanning electron microscope at magnifications that are within the capabilities of optical microscopy.

The imaging capabilities of the scanning electron microscope make it a useful tool for P/M applications concerned with all phases of powder production and processing. Scanning electron microscopy has been used to study all aspects of

particle morphology, including size, shape, surface topography, surface structure (crystal, grain, and dendrite), coating or thin film characteristics (oxides), inclusion, void, and agglomeration characteristics, and satellite formation. The scanning electron microscope has also been used to study surface topography, effect of oxides or other coatings, porosity, inclusions, and other contaminants on P/M materials. One of the more useful applications of the scanning electron microscope in powder metallurgy is qualitative particle characterization in terms of visual appearance. Despite the development of automated instruments for quantitative particle characterization, no substitute has been found for the interpretive capability of man. The use of the scanning electron microscope extends this capability into the microscopic domain. Inspection of powder particles to be used in P/M processing is essential, because individual and agglomerate characteristics can have significant effects on final material properties. Figure 23 shows representative alloy powder particles used in the direct rolling of aluminum P/M strip. The irregular shape promotes interlocking of particles, which provides sufficient green strength for strip processing.

Fig. 23 Scanning electron micrographs of 7091 aluminum alloy particles used in direct powder rolling of strip. (a) 70×. (b) 700×

Quantitative Analysis of Projected Images. In general, microscopists encounter two types of projected images. The first type of image results from a transmitted beam through the specimen, representing the features located within the three-dimensional space (such as by thin-foil transmission electron microscopy). In the second type, the projected image is generated by a reflected beam from the external surface of the specimen (such as by scanning electron microscopy).

Only the most rudimentary quantitative calculations can be made on images projected by the reflection techniques (Ref 18). In rough surfaces, the intensity levels depend on topography, and some features may be masked by others. Threedimensional characterization is based on the photogrammetric analysis of stereopairs, for which automatic imageanalyzing techniques are not yet available (Ref 19). Quantitative statistical treatment of transmitted-beam images, however, has matured to a much greater extent. These analyses (Ref 20, 21) are too lengthy and complex to be treated here, but are described in the literature (see Ref 5). One final topic will be included, because of its importance to the analysis of particulate systems. Figure 24 provides interrelated general equations of convex particles that express the important spatial parameters in terms of measurements made on the polish and projection planes. Application of the equations to specific particles is summarized in Table 2 for the sphere, the truncated octahedron (or tetrakaidecahedron), and convex particles in general. Tabulations of the type presented in Table 2 permit the microscopist to approximate microstructures with particles of known shape when other techniques are not feasible.

Table 2 Properties of a sphere, truncated octahedron, and convex particles Property V S A' H' A L3 r, a, Nv

(a)

Sphere, D = 2r 4 r3/3 4 r2 r2 2r 2 r2/3 4r/3 r = 2NL/ NA

Truncated octahedron, edge length = a 11.314a3 26.785a2 6.696a2 3.0a 3.77a2 1.69a a = 0.45NL/NA

General equations for convex particles V = A'L3 = AH' S = 4A' = 4V/L3 A' = S/4 = V/L3 H' = V/A = A'L3/A A = V/H' = A'L3/H' L3 = 4V/S = ANA/N1 = H'/2 = NAA/2NL = A'/2L2 NV = NA/H' = NL/A'

General equation: VV = NVV + NAA = NLL3

Source: Ref 23

(a)

= half of mean tangent distance.

Fig. 24 Relationships among convex particles in space, their sections, and their projections. Projected quantities are indicated by primes. Source: ASM Handbook, Volume 9, p 134

References cited in this section

5. H. Heywood, Numerical Definitions of Particle Size and Shape, Chem. Ind., Vol 15, 1937, p 149-154 18. J.E. Hilliard, J. Microsc., Vol 95 (Part 1), Feb 1972, p 45-58 19. T.O. Johari, Res. Develop., Vol 22 (No. 7), 1971, p 12 20. E.E. Underwood, The Stereology of Projected Images, J. Micros., Vol 95 (Part 1), Feb 1972, p 25-44 21. E.E. Underwood, The Mathematical Foundations of Quantitative Stereology, Stereology and Quantitative Metallography, STP 504, ASTM, Philadelphia, 1972, p 3-38 23. H. Heywood, Symposium on Particle Size Analysis, Trans. Inst. Chem. Eng., Vol 25, 1947, p 14-24 Particle Image Analysis

Conventional Shape Factors Particle shape is a fundamental characteristic of powder particles and thus influences the properties of particulate systems. Various shape terms have been proposed to quantitatively represent particle shape. Early systems tended to measure one specific feature of a particle. Table 3 (Ref 23, 24, 25, 26, 27, 28, 29, 30) lists some of the most frequently used shape terms. The applicability and/or limitations of various shape factors, also referred to as shape parameters, are discussed in the following section of this article on stereological characterization of shape. For discussion purposes, the following method developed by Hausner (Ref 28) is used to demonstrate the applicability of shape factors in quantitative analysis of particle shape.

Table 3 Shape terms and their definitions Volume specific surface (Sv) Sv = S/V

V: Particle volume S: Particle surface area

Heywood ratios (Ref 24) Elongation ratio (n) n = Lh/Bh Flakiness ratio (m) m = Bh/Th Heywood shape factor ( h

Bh: Breadth--the minimum distance between two parallel planes that are perpendicular to planes defining Th Th: Thickness--minimum distance between two parallel planes tangential to the particle maximum stability plane Lh: Length--distance between two parallel planes that are perpendicular to both Th and Bh planes h)

= f/k where f =

(Ref 25) da: Diameter of a circle having the same projected area S: Particle surface area V: Particle volume

k= Wadell Sphericity ( s) (Ref 26) 2 s = (dv/ds) 2/3 4.84 (k/f) Roundness ( r) (Ref 27)

Krumbein (Ref 28) Sphericity ( k)

h

=6

dv: Diameter of the sphere that has the same volume of the particle f, k: Diameter of the sphere having the same surface area as the particle (defined in Heywood ratios) ri: ith radius of curvature along the particle profile which is smaller than the radius of the largest inscribed circle N: Total number of radii of curvature smaller than the inscribed circle R: Radius of the largest inscribed circle

Lk: Longest dimension of the particle Bk: Breadth--measured perpendicular to L Tk: Particle thickness

Hausner (Ref 29) Elongation ratio (x) x = a/b Bulkiness factor (y) y = A/ab Surface factor (z) z = c2/12.6A Church shape factor (

For spheres: n = m = 1,

a: Length of the enveloping rectangle that has the minimum area b: Width of the rectangle A: Projected area of the particle c: Perimeter of the projected profile

c)

(Ref 30)

dM: Martin's diameter--length of the chord which divides the profile into two equal areas with respect to a fixed direction E(dM): Expectation of Martin's diameter dF: Feret's diameter, the distance between a pair of parallel tangents of the particle profile with respect to a fixed direction E(dF): Expectation of Feret's diameter Centroid aspect ratio (CAR) (Ref 31) dm: Longest chord passing through the centroid CAR = dm/dp c

= E(dM)/E(dF)

dp: Length of the chord passing through the centroid and perpendicular to the one defining dm

A rectangle of minimum area is drawn around the cross section of a particle (particle projection) as it is observed under the microscope (Fig. 25). The ratio of the rectangle side lengths permits calculation of particle elongation:

Fig. 25 Determination of particle size characterization. (a) Length of the enveloping rectangle, which has a minimum area. (b) Width of the rectangle. (c) Circumference of the projected particle. A, surface area of the projected particle

The ratio of the area (A) of the projected particle to the area of the enveloping rectangle of minimum area (a×b) indicates the bulkiness of the particle:

In this way, the cross-sectional area of the particle is correlated with some of its linear dimensions. To characterize the surface configuration (or surface area), which is an essential factor of shape, the surface of the respective particle should be compared with the surface of a sphere of identical volume, or a cross section should be compared with that of the particle:

where c is the perimeter of the projection profile of the particles. For a spherical particle, z = 1, and for particles of any other shape, z > 1. Although the three ratios (x, y, and z) do not permit exact characterization of particle shape, they are nevertheless descriptive. It is possible, therefore, to correlate particle behavior with the three ratios--elongation factor, bulkiness factor, and surface factor. As indicated, however, there have been other shape factors proposed to characterize particle shape (see Table 3 and the next section of this article).

References cited in this section

23. H. Heywood, Symposium on Particle Size Analysis, Trans. Inst. Chem. Eng., Vol 25, 1947, p 14-24 24. H. Heywood, J. Imperial Coll. Eng. Soc., Vol 8, 1954

25. H. Wadel, J. Geol., Vol 40, 1932, p 343-357 26. H. Wadel, J. Geol., Vol 43, 1935, p 250-280 27. W.C. Krumbein, J. Sedimen. Petrol., Vol 11 (No. 2), 1941, p 64-72 28. H.H. Hausner, Planseeber, Vol 14 (No. 2), 1966, p 75-84 29. T. Church, Powder Technol., Vol 2, 1968-1969, p 27-31 30. J.K. Beddow, Particulate Science and Technology, Chemical Publishers, 1980 31. E.E. Underwood, Stereological Analysis of Particle Characteristics, Testing and Characterization of Powders and Fine Particles, J.K. Beddow and T.P. Meloy, Ed., 1980, p 77 Particle Image Analysis

Stereological Characterization Stereology refers to the study of the three-dimensional structure of materials from two-dimensional sections or projections. This discipline is primarily concerned with the geometrical evaluation of microstructural features, based primarily on geometrical probabilities. However, statistics, topology, and projection geometry also contribute to the overall quantitative characterization (Ref 20, 31). Quantitative characterization of the shape of microstructural features (e.g., pore space, precipitates, phase regions, and grains) and of particles on a substrate or embedded in a resin by means of data derived from planar images (sections or projections) is the prime consideration of stereology. Stereology usually does not apply to the geometric nature of single particles, but rather to the geometric characteristics of a large number of particles or to the geometric characteristics of the "average particle" in a powder mass. Detailed information on the fundamentals, instrumentation, and applications of quantitative stereology can be found in Ref 18, 19, 20, 21, 22, 32, and 33. Numerous shape parameters have been proposed, but only a few are practical. A useful stereological shape parameter (factor) must fulfill several requirements: • • •

Shape sensitivity: The value of a particular parameter must vary systematically with changes in shape--it should be sensitive to specific aspects such as elongation, bulkiness, and symmetry. Independence of other geometric properties: Size, size distribution, volume fraction, or other nonshape-related geometric characteristics should not influence the value for a given shape parameter. Accessibility: Quantities from which a stereological shape parameter is calculated must be available by simple measurements performed on planar images.

Other less significant requirements are that shape parameters must be dimensionless and independent of rotation or translation of the objects. Furthermore, it should be easy to visualize the significance of a shape parameter in terms of interpretable microstructural shape changes. The range for typical microstructural changes should be wide compared to statistical fluctuations and measuring errors. From the parameters listed in standard textbooks (Ref 22, 32) and from parameters proposed in literature (Ref 34, 35, 36, 37), none conforms to these requirements completely. In the following sections, some of the more useful stereological shape parameters, their limitations, and their practical alternatives (fingerprinting by means of simple two-dimensional parameters) are discussed. Image Data for Shape Characterization With manual or semiautomatic data acquisition (in which counting and measurements are performed by the operator), only average quantities for the measured fields of view (field quantities), such as area fraction, total and mean intercept length, total and mean perimeter length, mean curvature, and number of features per unit area or unit length, can be obtained in a timely manner. Automatic instruments facilitate measurement of parameters describing the geometric

properties of each individual planar feature, as seen in the fields of view at a high speed. From those parameters that can be used for shape description, the following have been shown to be of practical use: individual intersect areas, intercept length, tangent (Feret) diameters, as a function of direction and their extreme (maximum and minimum) for each closed planar feature; individual curvature; and moments of inertia. Concise characterization of three-dimensional shape is possible only if the spatial coordinates of the particle surfaces or three-dimensional feature parameters (spatial Feret diameters, spatial moments of inertia, surface areas, and volumes of the particles) are available. Direct acquisition of three-dimensional data is possible by (a) serial sectioning, (b) stereometric measurements using SEM stereo-pair photographs and instrumented stereometers, and (c) additional specialized techniques, including x-ray topography, shadowgraphy, or densitometry (for semi-transparent materials). However, these techniques are time consuming and not yet well established. Stereology of Shape It is generally agreed that three-dimensional shape cannot be quantitatively assessed from planar images without severe limitations. For example, topological parameters, such as the number of separate parts (e.g., the number of pores in a porous particle) or the degree of connectivity (the number of channels between the pores), cannot be obtained from a single cross section. From projections, especially from stereo-pair micrographs, a qualitative shape description can be easily derived. For stereological quantification, however, the data obtained from projections are generally less suitable than those obtained from the cross sections of particles embedded in a resin, a metal, or in a glass prepared by normal metallographic techniques. Thorough analysis of stereological shape parameters obtained from a single cross section through a multitude of particles was proposed in the literature (Ref 38), intended to show that parameters that depend exclusively on the shape of the three-dimensional objects, and not on size or other non-shape-related features, can be traced to one simple combination of field data. This combination is the number of intersecting points between a measuring line and the perimeters of the particle cross section (PL), which for convex particle features, is twice the number of particles per unit length of the measuring traverse--NL (PL = 2NL)--the volume fraction (VV) measured as area fraction (lineal fraction or point fraction) and the number of objects per unit area (NA). Fischmeister (Ref 38) derives the universal shape parameter:

(Eq 1)

The numerical factor aims to normalize the value to 1 for spheres. However, it can be easily demonstrated that this shape parameter is not independent of the size distribution, and deviations from a value of 1 for spheres of varying size are possible. A shape parameter that is actually independent of other geometric properties was proposed by De Hoff (Ref 39) in 1964:

(Eq 2)

where M1, M2, and M3 are the first three moments of the spatial size distribution, which generally are not accessible from planar data without a priori information on shape (the shape of the particles must be simple and known in order to calculate three-dimensional size distributions from planar measurements). FD reduces to FF for uniform size, while logarithmic-normal size distributions with known standard deviation ln yields:

(Eq 3)

The only other parameter that appears to meet the three basic requirements outlined above has been derived by Hilliard (Ref 40) from a geometric theorem relating the total surface and the total volume of three-dimensional objects to the first and the fourth moment of the intercept length distribution--M1(1) and M4(1) (see Ref 34). After normalizing to 1 for spheres again, this yields the shape parameter:

(Eq 4)

NL and VV are defined above. The main limitation of FH is that it fulfills Eq 3 in principle only; in practice, M4(1) is subject to large experimental errors due to the fact that the longest intercepts, which usually are present with a low numerical frequency, contribute most to M4(1). Numerous other shape parameters have been proposed by Underwood (Ref 35, 36), Ministr (Ref 41), and others. However, most equations are in pronounced contradiction to one or more of the basic requirements for true shape factors. Due to the problems outlined in this section, two-dimensional shape characteristics provide practical alternatives and as such are used almost exclusively in practical work. Quantitative Description of Planar Shape Two-dimensional shape parameters provide a means to monitor shape changes occurring in microstructures of P/M products due to variations in processing parameters and a means to compare the particle shape of powders obtained with different materials and production techniques. For this method, Fischmeister (Ref 39) uses the term "fingerprinting," because if applied properly, it yields indirect but accurate information on three-dimensional shape. The basic requirement outlined previously, however, must also be observed for two-dimensional shape factors. Hausner's shape parameters, described earlier in this article in the section on conventional shape parameters, are an example of "fingerprinting," as are other methods of planar shape description, such as Fourier coefficients and fractals of the sectioned or projected particle circumferences. In sectioned or projected particle circumferences, parameters are described that are sensitive to particular shape aspects and are easily determined with semiautomatic or fully automatic image analyzers. In the Hausner method, the fitting of the minimum area rectangle requires repeated fitting, which in some cases leads to ambiguous results. By far the most frequently used planar shape parameter (implemented in most image analyzing devices) combines the area and the perimeter of a planar feature into a dimensionless number normalized to 1 for circles of the same size:

fPL = 4 a/b2

(Eq 5)

where a and b are the area and the perimeter of the planar features (intersects or projections), respectively. Averaging can be completed easily if a and b are measured individually for each feature:

(Eq 6)

where n is the total number of planar features. However, if a and b are averaged, or if total area and perimeter length are used (as provided by field analysis), the average shape parameters obtained depend on size distribution, with values deviating from unity for a system of circles of varying diameter. If Eq 6 is used, deviations from unity are interpreted as deviations from circularity. Elongation (elliptical deformation), as well as concave deformations of the perimeter ("rugged" outlines), yields smaller values approaching zero for highly elongated features or highly rugged perimeters. Thus, F1 combines different aspects of shape. To differentiate between these aspects, simple combinations of moments of inertia, Feret diameters, or curvature have been derived by Schwarz (Ref 42). Exner and Hougardy (Ref 34) propose the following parameters for individual features, which also can be averaged for any number of features in the same way as fPL (see Eq 6):

Elongation, fEL =

(Eq 7) (Eq 8)

Ruggedness (waviness), fRU = bc/b

(Eq 9)

where i1 and i2 are the two principal moments of inertia, bc is the convex perimeter, and b is the actual perimeter. These parameters are sensitive to the respective shape aspects, are easily measured by computer-aided image analysis, and are independent of other geometric properties; their interpretation through visualization is easily accomplished. Consideration should be given to the method of averaging, however. Averaging on a volume basis, rather than on a number basis as in Eq 6, may be appropriate in some practical applications. If three-dimensional measurements are performed, similar shape parameters can be derived for the various aspects of spatial shape. Multidimensional Shape Characterization In order to analyze differences in shape precisely, more than one shape aspect must be monitored. Depending on the number of aspects considered using mutually independent shape parameters, two-, three-, or multidimensional representation of shape can be useful for comparison, and cluster analysis may be appropriate to quantify the significance of differences. Figure 26 shows a two-dimensional shape characterization of idealized particles in a ruggedness/elongation diagram.

Fig. 26 Characterization of the shape of twelve typical particle projections (or sections) in a two-dimensional shape space

These procedures require a computer with a large capacity, but which is modest compared to Fourier analysis, for example. The assessment of the individual parameters involves straightforward arithmetic calculation, and cluster analysis is performed on a few numbers for each feature measured. Simple shape characterization, normally accomplished using Eq 6, prevails in practice, because the multidimensional procedures referred to previously are not yet implemented in commercial image analyzers. At present, the best approach to the difficult problem of quantitatively characterizing the shape of a system of particles in a powder mass is to use one or several shape parameters carefully adjusted to the problem under investigation.

References cited in this section

18. J.E. Hilliard, J. Microsc., Vol 95 (Part 1), Feb 1972, p 45-58 19. T.O. Johari, Res. Develop., Vol 22 (No. 7), 1971, p 12 20. E.E. Underwood, The Stereology of Projected Images, J. Micros., Vol 95 (Part 1), Feb 1972, p 25-44 21. E.E. Underwood, The Mathematical Foundations of Quantitative Stereology, Stereology and Quantitative Metallography, STP 504, ASTM, Philadelphia, 1972, p 3-38 22. E.E. Underwood, Quantitative Stereology, Addison-Wesley, 1970 31. E.E. Underwood, Stereological Analysis of Particle Characteristics, Testing and Characterization of Powders and Fine Particles, J.K. Beddow and T.P. Meloy, Ed., 1980, p 77 32. E.R. Weibel, Stereological Methods, Vol 2, Theoretical Foundations, Academic Press, 1980 33. J. Serra, Image Analysis and Mathematical Morphology, Academic Press, 1982 34. H.E. Exner and H.P. Hougardy, Quantitative Image Analysis of Microstructures, A Practical Guide to Techniques, Instrumentation and Assessment of Materials, Deutsche Gesellschaft für Metallkunde, Oberursel, 1984 35. E.E. Underwood, Quantitative Shape Indices by Stereological Methods, Quantitative Analysis of Microstructures in Medicine, Biology and Materials Development, Riederer-Verlag, 1975, p 223-241 36. E.E. Underwood, "Three-Dimensional Shape Parameters from Planar Sections," Proc. Fourth Int. Cong. Stereology, NBS Special Publication 431, National Bureau of Standards, 1976, p 91-92 37. E.E. Underwood, Stereological Analysis of Particle Characteristics, Testing and Characterization of Powders and Fine Particles, Heyden and Sons, London, 1980, p 77-96 38. H.F. Fischmeister, Shape Factors in Quantitative Microscopy, Z. Metallkd., Vol 65, 1974, p 558-562 39. R.T. DeHoff, The Determination of the Geometric Properties of Aggregates of Constant Size Particles from Counting Measurements Made on Random Plane Sections, Trans. AIME, Vol 230, 1964, p 764-769 40. J.E. Hilliard, private communication, 1975 41. Z. Ministr, The Determination of Parameters for Two and Three Dimensional Microstructures with Aid of Specific Perimeter or Specific Surface Area of the Microstructural Features, Practical Metallogr., Vol 8, 1972, p 333-359, p 407-423 and Vol 12, 1975, p 244-258 42. H. Schwarz, Two Dimensional Feature-Shape Indices, Proc. Fifth Int. Cong. Stereology, Mikroskopie, Vol 37 (Suppl.), 1980, p 64-67 Particle Image Analysis

Morphological Analysis Morphological analysis is primarily the characterization of particle shape. Particle size may also be obtained in the analysis (Ref 30). Typically, a shape analyzer is used to conduct the analysis. This instrument, which consists of a highquality imaging (graphics) system that is augmented with computer equipment, must provide a clear image of the object for effective and accurate morphological analysis.

The image is digitized to obtain the (x, y) data set of the profile. This data set is then converted to the polar form. The polar data set (R, ) constitutes a line, and the Fourier equation that represents this line can be constructed and the set of coefficients extracted from this equation. In turn, the coefficients are converted to invariant forms called "morphic features." Statistical properties of these features are also determined. The rigorous definition of various morphological descriptors are given in Eq 10, 11, 12, 13, and 14, and statistical properties are given in Eq 15, 16, and 17. The size term is defined as the equivalent radius, R0, of the particle profile, such that

is the area of the circle of equivalent area to that of the particle profile.

(Eq 10) The mean radius a0 and R0 are related as follows:

(Eq 11) L1 (n) = 0 for all n

(Eq 12)

The next set of terms, called the L2(n) terms, where n is the order of the Fourier coefficient, is useful in describing morphic features such as aspect ratio, triangularity, squareness, roughness, and other shape aspects.

(Eq 13) (n = 1, 2, 3, . . . N is the order of the coefficient)

(Eq 14)

The mean, standard deviation, and skewness of the radial distribution of the particle are defined below. They are standard statistical properties and each has a distinct physical meaning. For example, the standard deviation is an indicator of the not-roundness of the particle profile. 0

= L0R0 (mean radius)

(Eq 15)

(Eq 16)

(Eq 17)

Morphological Descriptors and Statistical Terms. Morphological descriptors are unique in that they can be used

to regenerate the form of the original particle profile. A set of descriptors therefore, contains all of the information in the original profile. Because morphological analysis has been developed to solve real problems, it is important that the morphological features can be identified with a corresponding physical meaning wherever possible.

The equivalent radius R0 is defined so that a0 is the mean radius of the particle.

is the area of the particle. Therefore, R0 can be interpreted as a size term;

The series of terms L2(n) are defined in terms of the Fourier coefficients an and bn, and they are normalized by dividing by to make them independent of size. Therefore, L2(n) terms are the equivalent of the Fourier coefficients An. In terms of their physical meaning, an indication of the relative strength of the L2(3) term can be obtained by dividing the value of L2(3) by the sum of L2(n). This value provides an indication of the angularity (triangularity) of the particle. There is a relationship between the L3(m, n) terms and the radial skewness as shown in Eq 17. Therefore, if all of the L3(m, n) terms are summed and multiplied by , the skewness of the radial distribution can be obtained. A pear-shaped drop or a teardrop shape, for example, would provide a high value of skewness. However, the physical meaning of the individual L3(m, n) terms is not understood at present. The statistical properties of the morphological features defined in Eq 15 and 17 have been discussed above. The second moment of the radial distribution (the radiance), which is defined in Eq 16, is interpreted as indicating the not-roundness (out-of-roundness) of the particle profile. The equivalent radius squared minus the mean radius squared is equal to the sum of all the L2(n) terms multiplied by . Equation 10, therefore, relates intrinsic and extrinsic properties of the particle profile quantitatively. In addition, the radance (the sum of the L2(n) terms) can be equated to a difference between the square of the size minus the square of the mean radius. This discussion indicates that there is a basis for the physical interpretation of the morphological features of a particle profile. The exception to this is the physical meaning of the L3(m, n) terms. This problem requires further study.

Reference cited in this section

30. J.K. Beddow, Particulate Science and Technology, Chemical Publishers, 1980 Particle Image Analysis

References 1. B.H. Kaye, Particle Shape Characterization, Chap. 2, Handbook of Powder Science and Technology, 2nd ed., M.E. Fayed and L. Otten, Ed., Chapman & Hall, 1997, p 35-52 2. T. Allen, Particle Size Analysis, 4th ed., Chapman & Hall, 1996 3. British Standards 2955 Glossary of Terms Relating to Powders, British Standards Institute, 1965 4. H. Heywood, Size and Shape Distribution of Lunar Fines Sample 12057, 72, Proc. of Second Lunar Science Conf., Vol 13, 1971, p 1989-2001 5. H. Heywood, Numerical Definitions of Particle Size and Shape, Chem. Ind., Vol 15, 1937, p 149-154 6. H. Heywood, Particle Shape Coefficients, J. Imp. Coll. Eng. Soc., Vol 8, 1954, p 25-33 7. H.H. Hausner, Characterization of the Powder Particle Shape, Proc. of the Symp. on Particle Size Analysis (Loughborough, England), The Society for Analytical Chemistry, 1967, p 20-77 8. M.H. Phal, G. Schädel, and H. Rumpf, Zusammenstellung von Teilchenform Beschreilbungsmethoden, Aufbereit.-Tech., Vol 5, 1973, p 7-11 9. R.P. Loveland, ASTM Committee E29.02 on Fineparticle Shape, Report, ASTM, 1975 10. B.H. Kaye, Chap. 10, Direct Characterization of Fineparticles, John Wiley & Sons, 1981 11. B.H. Kaye, G.G. Clark, and Y. Liu, Characterizing the Structure of Abrasive Fineparticles., Part. Part. Syst.

Charact., Vol 9 (No. 1), 1992, p 1-8 12. R. Davies, A Simple Feature Based Representation of Particle Shape, Powder Technol., Vol 12, 1975, p 111-124 13. H.P. Schwartz and K.C. Shane, Measurement of Particle Shape by Fourier Analysis, Sedimentology, Vol 13, 1969, p 213-231 14. R. Ehrlich and B. Weinberg, An Exact Method for Characterization of Grain Shape, J. Sediment. Petrol., Vol 40 (No. 1), March 1970, p 205-212 15. A.G. Flook, A Comparison of Quantitative Methods of Shape Characterization, Acta Stereol., Vol 3, 1984, p 159-164 16. J.K. Beddow, M.D. Nasta, and G.C. Philip, Characteristics of Particle Signatures, Proc. of the Int. Conf. on Powders and Bulk Handling Systems, May 1977 (Chicago, IL), International Powder Institute, 1977 17. B.B. Mandelbrot, Fractals; Form, Chance and Dimension, Freeman, 1977 18. J.E. Hilliard, J. Microsc., Vol 95 (Part 1), Feb 1972, p 45-58 19. T.O. Johari, Res. Develop., Vol 22 (No. 7), 1971, p 12 20. E.E. Underwood, The Stereology of Projected Images, J. Micros., Vol 95 (Part 1), Feb 1972, p 25-44 21. E.E. Underwood, The Mathematical Foundations of Quantitative Stereology, Stereology and Quantitative Metallography, STP 504, ASTM, Philadelphia, 1972, p 3-38 22. E.E. Underwood, Quantitative Stereology, Addison-Wesley, 1970 23. H. Heywood, Symposium on Particle Size Analysis, Trans. Inst. Chem. Eng., Vol 25, 1947, p 14-24 24. H. Heywood, J. Imperial Coll. Eng. Soc., Vol 8, 1954 25. H. Wadel, J. Geol., Vol 40, 1932, p 343-357 26. H. Wadel, J. Geol., Vol 43, 1935, p 250-280 27. W.C. Krumbein, J. Sedimen. Petrol., Vol 11 (No. 2), 1941, p 64-72 28. H.H. Hausner, Planseeber, Vol 14 (No. 2), 1966, p 75-84 29. T. Church, Powder Technol., Vol 2, 1968-1969, p 27-31 30. J.K. Beddow, Particulate Science and Technology, Chemical Publishers, 1980 31. E.E. Underwood, Stereological Analysis of Particle Characteristics, Testing and Characterization of Powders and Fine Particles, J.K. Beddow and T.P. Meloy, Ed., 1980, p 77 32. E.R. Weibel, Stereological Methods, Vol 2, Theoretical Foundations, Academic Press, 1980 33. J. Serra, Image Analysis and Mathematical Morphology, Academic Press, 1982 34. H.E. Exner and H.P. Hougardy, Quantitative Image Analysis of Microstructures, A Practical Guide to Techniques, Instrumentation and Assessment of Materials, Deutsche Gesellschaft für Metallkunde, Oberursel, 1984 35. E.E. Underwood, Quantitative Shape Indices by Stereological Methods, Quantitative Analysis of Microstructures in Medicine, Biology and Materials Development, Riederer-Verlag, 1975, p 223-241 36. E.E. Underwood, "Three-Dimensional Shape Parameters from Planar Sections," Proc. Fourth Int. Cong. Stereology, NBS Special Publication 431, National Bureau of Standards, 1976, p 91-92 37. E.E. Underwood, Stereological Analysis of Particle Characteristics, Testing and Characterization of Powders and Fine Particles, Heyden and Sons, London, 1980, p 77-96 38. H.F. Fischmeister, Shape Factors in Quantitative Microscopy, Z. Metallkd., Vol 65, 1974, p 558-562 39. R.T. DeHoff, The Determination of the Geometric Properties of Aggregates of Constant Size Particles from Counting Measurements Made on Random Plane Sections, Trans. AIME, Vol 230, 1964, p 764-769 40. J.E. Hilliard, private communication, 1975 41. Z. Ministr, The Determination of Parameters for Two and Three Dimensional Microstructures with Aid of Specific Perimeter or Specific Surface Area of the Microstructural Features, Practical Metallogr., Vol 8, 1972, p 333-359, p 407-423 and Vol 12, 1975, p 244-258 42. H. Schwarz, Two Dimensional Feature-Shape Indices, Proc. Fifth Int. Cong. Stereology, Mikroskopie, Vol

37 (Suppl.), 1980, p 64-67

Surface Area, Density, and Porosity of Powders Introduction A LARGE NUMBER of industries deal with powders at some point during their processing procedures. The characterization of surface area, porosity, density, and particle size is of particular importance. Especially in the case of metal powders, those characteristics are important factors for understanding and controlling material properties and processing behavior. Knowledge about the surface area of the powder is helpful in understanding the sintering behavior, because the reduction of surface area is the essential driving force for the sintering process. Porosity and density are closely related, and they may influence, for example, the hardness of the material. Different techniques can be used to determine these powder characteristics. The methods used to determine surface area, density, and porosity are not specific to metal powders: many have been developed for testing other materials in powder form. The major characterization techniques are summarized below. Expanded discussions of several test methods can be found in subsequent sections of this article, including testing parameters, specifications, and instrumentation. When these techniques are used, however, the results of each test method can vary from the results obtained by a different technique. This is only partly due to sample variations. Yet, each test method determines those properties in a very specific way. Pore size and even surface area are ambiguous parameters. For example, the size of an ideal cylindrical pore would be defined as the radius or diameter of this cylinder. However, real pores are hardly ideal cylinders. The odd-shaped geometry of pores makes it difficult to define the size of such a pore, and each measuring technique has its own way of looking at those pores and determining an average pore size of the sample. The pore size also depends on whether the actual void size or the entrance size or neck of the pore is of interest. The latter could be the restricting factor for transport processes. It is also impossible to clearly define the surface area. Depending on the "yardstick" of the observer, a sample may have a larger or smaller surface area. An optical microscope may not show cracks and pores in a sample, whereas the same sample can show a substantial surface roughness under the electron microscope or when analyzed by gas adsorption techniques. Correspondingly, the value of the calculated surface area will be different. •

•

•

•

Gas adsorption method: Determines the surface area of a powder sample by measuring the amount of gas adsorbed on the sample surface at low temperatures. The data are then used to calculate the monolayer capacity from which the surface area is calculated using the "known" size of the adsorbed molecules. Nitrogen at liquid nitrogen temperature is the most common gas used; krypton at liquid nitrogen temperature is used for samples with an area of less than 1 m2. Gas adsorption can also be used to determine pore volume and pore size distribution ranging from one Angstrom to about 100 or 200 nm. Permeametry: Measures the resistance to fluid flow through a compacted powder bed. This information is used to determine related properties of a powder, such as pore size distribution, specific surface area, or the average particle size for a packed powder sample. Pycnometry: Determines density by measuring the difference between the specific and bulk volumes of a sample. This method is based on the displacement principle, using the powder as the solid body and helium, water, or mercury as the displaced medium. Pycnometry can be used to determine total pore volume or density. However, it does not provide quantification of the pore size or the pore size distribution. Mercury porosimetry: Measures the volume of mercury intruded into the pores of a powder sample as a function of the pressure applied to the mercury. This method gives pore size and distribution over a wide range: 0.3 mm to 3 nm depending on the capability (pressure range) of the apparatus used. The method is based on the nonwetting behavior of mercury toward most materials. However, certain metals, for example, gold or copper, can react with mercury, and before those samples are to be

analyzed specific precautions have to be taken.

Although surface area, density, and porosity of powder are interrelated, a given powder sample may require the use of several testing methods to provide a complete analysis of these characteristics, because each technique is more or less sensitive for a specific sample. A comprehensive treatment of each of the methods is given in this article. Surface Area, Density, and Porosity of Powders

Gas Adsorption H. Giesche, School of Ceramic Engineering and Sciences, Alfred University

The Brunauer-Emmett-Teller (BET) (Ref 1) method of measuring specific surface area is based on the determination of the amount of gas that is adsorbed on the surface of the sample. The specific surface area (m2/g) determined by this method includes the external as well as internal (pores) surface area. The surface area of closed pores cannot be determined because the adsorbing gas molecules have no physical path to that surface.

Reference cited in this section

1. S. Brunauer, P.H. Emmett, and E. Teller, J. Am. Chem. Soc., Vol 60, 1938, p 309 Surface Area, Density, and Porosity of Powders

Theory of BET Method The BET model is based on a kinetic model of the adsorption process that was described first by Langmuir in 1916 (Ref 2, 3). Langmuir regarded the solid surface as an array of adsorption sites, and a state of dynamic equilibrium was postulated in which the rate molecules arrive from the gas phase and condense on the bare sites is equal to the rate at which molecules evaporate from occupied sites. In 1938 Brunauer, Emmett, and Teller extended Langmuir's kinetic monolayer adsorption theory to a multilayer adsorption theory. The following relationship was derived to calculate the monolayer capacity, Vm:

In the actual experiment the adsorbed amount of gas, V, is measured as a function of gas pressure, P. P0 is the saturation pressure of the liquefied gas at the corresponding temperature of the adsorption measurement and the ratio P/P0 is often referred to as the relative pressure. The parameters C and n are part of the BET theory. The parameter C is:

The parameters a, v, q, R, and T are the activity coefficient, vibration frequency, heat of adsorption, gas constant, and temperature, respectively (Ref 4). The parameter C can be regarded simply as a measure of the additional or specific heat of adsorption in the first adsorbed layer (q1) compared to the heat of condensation, qc, which is assumed to be valid for the second and following adsorption layers. The parameter n corresponds to the maximum number of adsorbed layers on the surface. Only when pore sizes are very small is there a physical limitation of the number of layers and the factor, n, has to

be considered. Usually n is assumed to be infinite and the "standard" BET equation results, which can then be rearranged to yield a linear relation:

A plot of P/[V(P - P0)] versus P/P0 generally gives a straight line in the relative pressure range from P/P0 = 0.06 to 0.3. From the slope and intercept of this line, the amount of gas adsorbed in a monomolecular layer (Vm) is determined as well as the value of C. By using the specific size of the adsorbing molecule, the total surface area of the sample can then be calculated. It should also be noted that the value of C is quite sensitive to errors in the measurement, and it can vary by 50% or more depending how accurately data points are taken and whether or not certain data points are omitted in the calculation. Frequently, C values for metal surfaces are quite large (>100 or 200); under those conditions, the intercept of the BET line can be taken as 0. The latter approach is used in the so-called single-point BET method, when the monolayer capacity, Vm, is directly calculated from one single data point using the following equation:

or Vm = V (1 - P/P0). Different gases can be used for the surface area determination. The different experimental results can be compared if the size of the adsorbed gas molecules is known. Different values for the so-called cross-sectional area (size) of the adsorbate molecule have been reported in the literature. The following tables lists values as used by Davis, DeWitt, and Emmett (Ref 5):

Adsorbate gas and temperature, K Nitrogen, 78 Krypton, 78 Krypton, 195 Butane, 273 Butane, 195 Freon-21 (CHCl2F), 273

Size according to the density of the liquid, nm2 0.162 0.152 0.297 0.321 0.247 0.264

Revised values recommended in Ref 5, nm2 0.162 0.208 0.434 0.469 0.375 0.401

However, even with the use of such modified cross-sectional areas, anomalies between different samples could be diminished, but not completely eliminated, as shown in the data in Table 1 (Ref 5). The interpretation of surface area results also has to be carefully evaluated when the BET-C parameter is very large. A value of several hundred is a strong indication of either the presence of microporosity, adsorption on active sites at the surface, or even a chemisorption effect. As shown, very clean metal surfaces can produce extremely large C values and the adsorption of the gas molecules occurs in those cases in highly localized positions on the surface. As a result, the calculated surface area may not reflect the true area of the sample.

Table 1 Surface area determinations with different size values of the adsorbed gas Adsorbate solid

Glass spheres (7

Gas (Temperature, K) m)

Tungsten powder

Zinc oxide

Silver foil (geometrical area, 1.56 m2/g)

Monel ribbon (geometrical area 1.56 m2/g)

N2 (78) KR (78) C4H10 (195) CHCl2F (195) N2 (78) Kr (78) C4H10 (273) CHCl2F (273) N2 (78) Kr (78) C4H10 (273) CHCl2F (273) Kr (78) C4H10 (195) CHCl2F (195) Kr (78) C4H10 (195) CHCl2F (195)

Area per gram using liquid density values, m2/g 0.434 0.322 0.333 0.315 2.69 1.96 1.67 1.73 9.40 6.82 6.93 6.63 1.56 1.22 1.13 0.456 0.652 0.577

Area per gram using revised values (see text), m2/g 0.434 0.441 0.489 0.479 2.69 2.68 2.43 2.62 9.40 9.34 10.1 10.1 2.14 1.78 1.72 0.622 0.952 0.878

C-parameter

150 32 7 106 81 290 26 21 155 150 52 215 19 6 11 13 4 7

Source: Ref 5

Values of the BET-C parameter for krypton adsorption at a temperature of 77 K (Ref 4) for various solids are:

Material Organic materials Glass Silica Ferric oxide Nickel oxide Silica gel Ferrites Micas Tungsten powder Carbon black Nickel film, contaminated Nickel film, clean

C-parameter 10-70 20-80 25-75 30-75 70-120 80 60-200 100-130 215, 290 230 400, 1000 1200, 2300

References cited in this section

2. I. Langmuir, J. Am. Chem. Soc., Vol 38, 1916, p 2221 3. I. Langmuir, J. Am. Chem. Soc., Vol 40, 1918, p 1361 4. K.S.W. Sing and D. Swallow, Proc. Br. Ceram. Soc., Vol 39 (No. 5), 1965 5. R.T. Davis, T.W. DeWitt, and P.H. Emmett, J. Phys. Chem., Vol 51, 1947, p 1232 Surface Area, Density, and Porosity of Powders

BET Apparatus

Figure 1 shows the adsorption apparatus originally developed by Emmett. Commercial instruments are now totally automated, and instead of the burette-type mechanism to measure the adsorbed gas volume very precise pressure transducers are used to calculate the adsorbed amount of gas from the known volume of a calibration chamber and the pressure drop that occurs when the valve between sample and calibration chamber is opened. Other instruments use a mass-flow controller to monitor the amount of gas introduced to the sample. In flow through cells, the adsorbing gas is passed over the sample using a carrier (helium, which is assumed to be nonadsorbing at the specific conditions). The composition of the gas stream before and after is analyzed to provide information about the total amount of adsorbed gas. A third technique uses a microbalance to measure the weight change as gases adsorb on the sample. Each technique has advantages and disadvantages. Further details are described in all standard textbooks (Ref 6, 7, 8, 9).

Fig. 1 Brunauer-Emmett-Teller apparatus for determining specific surface area

The isotherm is usually constructed point by point by the admission or withdrawal of known amounts of gas, with adequate time allowed for equilibration at each point. A small amount of heat is produced during the adsorption process, which has to be removed from the sample before final equilibrium is achieved. The amount of adsorbed gas is very sensitive to temperature changes. Prior to the determination of the isotherm all physisorbed material has to be removed. The exact conditions required to obtain a "clean" surface depend on the nature of the sample. For the determination of the surface area and the mesopore size distribution by nitrogen adsorption, outgassing to a residual pressure of about 10-4 torr ( 10-2 Pa) is considered acceptable. The outgassing conditions and the determination of the outgassed mass of the sample is one major source of error (or at least the reason for discrepancies between different experimenters). Inorganic oxides are usually outgassed at temperatures 150 °C, while microporous carbons and zeolites require higher temperatures 300 °C. Depending on the synthesis or other previous processing steps of the metals powders, the sample should be cleaned by washing with an appropriate solvent or treated at sufficiently high outgassing temperatures. Usually nitrogen gas is adsorbed on the sample at liquid nitrogen temperature. The BET surface area determined from those measurements is used as a standard reference point. However, it might be appropriate or advantageous to use a different adsorbing gas in other applications. For example, nitrogen adsorption is limited to samples having a total surface area of more than 1 m2. However, small surface areas can be measured much more precisely by using krypton as the adsorptive gas. A total sample surface area of down to 50 cm2 can be measured with many commercial instruments. The increased sensitivity of the krypton analysis is due to the fact that the saturation pressure or krypton is only about 2.5 torr ( 300 Pa) at liquid nitrogen temperature. Due to this lower saturation pressure, the ratio between the actual amount

of adsorbed gas on the sample surface against the amount of gas that remains in the surrounding gas phase is much higher for krypton compared to the same data point in a nitrogen isotherm and correspondingly the amount of adsorbed krypton can be determined with a much higher precision. The next section demonstrates using a simple calculation why krypton measurements are so much more sensitive for low surface area samples. However, krypton at liquid nitrogen temperature has other problems. The saturation pressure under those conditions cannot be unambiguously defined. According to the phase diagram (Fig. 2), krypton could either exist as a solid or as an undercooled liquid at that temperature. Most publications refer to the supercooled liquid state. This uncertainty severely limits the krypton analysis. A similar condition is also true for argon or xenon at liquid nitrogen temperature. In the case of argon, the experiment could be performed at a different temperature. For example, liquid oxygen would create a temperature of about 90 K, which is above the triple-point temperature of argon (83.8 K). However, liquid oxygen is much more problematic to handle and not as commonly available as liquid nitrogen.

Fig. 2 Krypton phase diagram

References cited in this section

6. S.J. Gregg and K.S.W. Sing, Adsorption, Surface Area and Porosity, Academic Press, 1982 7. S. Lowell and J.E. Shields, Powder Surface Area and Porosity, Chapman & Hall, 1991 8. T. Allen, Particle Size Measurement, Vol 2, Surface Area and Pore Size Determination, Chapman & Hall, 1997 9. A.W. Adamson, Physical Chemistry of Surfaces, John Wiley & Sons, 1982 Surface Area, Density, and Porosity of Powders

Krypton for Low Surface Area Determination The use of krypton or argon rather than nitrogen for adsorbate improves the accuracy of low specific surface area measurements. Assume the sample has a total surface area of 1 m2 and the volume of the sample tube is 20 cm3. The saturation pressure for nitrogen is about P0(N2) = 760 torr and for krypton it is P0(Kr)= 2.5 torr at liquid nitrogen temperature (assuming a super-cooled liquid state for the krypton). The single-point BET equation for those gases yields:

Vm = 4.35 Va (1 - P/P0) for nitrogen and

Vm = 5.64 Va (1 - P/P0) for krypton At a relative pressure of P/P0 = 0.2 the adsorbed volume, Va, is then calculated to:

1.0 = 4.35 Va (1 - 0.2) or

Va = 0.287 cm3 (STP nitrogen) 1.0 = 5.64 Va (1 - 0.2) or

Va = 0.222 cm3 (STP krypton) The remaining volume of gas in the sample cell can be calculated as follows:

As shown in Fig. 3, it becomes obvious that the ratio of adsorbed versus remaining gas allows for a much more accurate measurement in the krypton analysis.

Fig. 3 Ratio of adsorbed versus remaining gas demonstrates more accurate measurements in the krypton analysis compared to N2 analysis for low surface area samples

Surface Area, Density, and Porosity of Powders

Accuracy The high degree of accuracy attainable in some branches of chemistry and physics is out of the question where evaluation of specific surface from adsorption data is concerned. Even in the favorable case of nitrogen or argon, a divergence of at least ±10% from the actual area of the solid should be expected (owing to theoretical factors not yet supported by accurate quantitative assessments). In addition, there are several nonnegligible experimental uncertainties, which can be strikingly illustrated by results of a round-robin test. The project embraced thirteen laboratories, all well experienced in the field (Ref 10). The participants determined detailed isotherms of nitrogen (often 30 to 40 points), calculating both the specific surface, based on am(N2) = 16.2 2 and the value of the BET C-parameter.

Solid Graphitized carbon (Vulcan 3G-2700) Graphitized carbon (Sterling FT-2700) Plasma produced silica (TK 800) Mesoporous silica gel (Gasil (1))

Surface area and standard deviation, m2/g 71.3±2.7 11.00±0.8 165.8±2.1 286.2±3.5

No. of labs 6 5 4 3

Reference cited in this section

10. D.H. Everett, G.D. Parfitt, K.S.W. Sing, and R. Wilson, J. Appl. Chem. Biotechnol., Vol 24, 1974, p 199 Surface Area, Density, and Porosity of Powders

Pore Size Analysis Gas adsorption/desorption experiments can also be used to study the pore size distribution in a sample. Pores of up to 100 or 200 nm can be determined. The measurement is based on the capillary condensation effect, which essentially means that a gas will condense into a liquid phase below its saturation point when it is confined in small pores. The basic relationship between relative pressure and pore size is given by the Kelvin equation:

where lv is the interfacial tension of the liquid-gas interface, V1 is the molar volume of the liquid gas phase, and the other factors are gas constant, R, temperature, T, and the radius of curvature of the liquid-gas interface, r. Numerous theories and models have been developed for the pore size analysis. The most commonly accepted theory was developed by Barrett, Joyner, and Halenda (Ref 11). Frequently, the adsorption-desorption isotherm shows a hysteresis and the shape as well as the relative pressure range where the hysteresis occurs provides information about the pore shape and pore size. Several theories have been developed to explain and analyze the hysteresis effect in more detail including deriving additional information about network and connectivity effects of the pore system in a sample. Details are described in several textbooks (Ref 6, 7, 8, 9).

References cited in this section

6. S.J. Gregg and K.S.W. Sing, Adsorption, Surface Area and Porosity, Academic Press, 1982 7. S. Lowell and J.E. Shields, Powder Surface Area and Porosity, Chapman & Hall, 1991 8. T. Allen, Particle Size Measurement, Vol 2, Surface Area and Pore Size Determination, Chapman & Hall, 1997 9. A.W. Adamson, Physical Chemistry of Surfaces, John Wiley & Sons, 1982 11. E.P. Barrett, L.G. Joyner, and P.H. Halenda, J. Am. Chem. Soc., Vol 73, 1951, p 373 Surface Area, Density, and Porosity of Powders

Permeametry Peter J. Heinzer, Imperial Clevite Technology Center

Permeametry is the measurement of resistance to fluid flow through a compact powder bed. Its main purpose in powder metallurgy is not so much to quantify resistance as it is to measure the related properties of a particle population--namely, specific surface area and average particle size. Fluids may be in liquid or gaseous form. Liquid permeametry was prevalent in the early stages of development and is the simplest method if the minimum diameter of any appreciable size fraction is 5 m (Ref 12). Settling and segregation, aggregation, and the difficulty of removing bubbles make liquid permeametry unsatisfactory for smaller-sized particles. Gas permeametry is now the preferred method because it extends the size measurement capability down to 0.1 to 0.5 m. Commercial permeametry is applicable in the following ranges (Ref 13, 14):

Specific surface area, cm2/g (in.2/g) Particle size, m

70-20,000 (10.8-3100) 0.5-50

In production practice, these ranges are subdivided into smaller segments to improve accuracy, because the behavior of fluid flow through powder beds with diverse characteristics changes considerably. A typical sample occupies a volume of about 5 cm3 (0.30 in.3) and, depending on density, weighs from 5 to 20 g (0.18 to 0.70 oz). Reproducibility in the surface area range cited for commercial instruments is ±1%. Accuracy varies with the type of sample, as discussed later in this section. Inherent problems in the use of permeametry for particle size measurement involve its use of semiempirical relationships with parameters that can only be evaluated indirectly. Additionally, permeametric methods provide only an average particle size--not the more useful size distribution information. Despite these disadvantages, permeametry is popular due to its simplicity of operation, reproducibility, speed of analysis (2 to 15 min compared to 1 h or more with other methods), and low cost (one-fifth to one-tenth of the equipment cost of competing methods).

References cited in this section

12. T. Allen, Permeametry, Particle Size Measurement, Chapman and Hall, 1968 13. C.F. Callis and R.R. Irani, Miscellaneous Techniques, Particle Size: Measurements, Interpretation, and Application, John Wiley & Sons, 1963 14. "Permaran Specific Surface Area Meter," Outokumpu Oy, Instrument Division, Tapiola, Finland, 1973

Surface Area, Density, and Porosity of Powders

History and Theory The original study of fluid flow through compacted particulate matter is attributed to D'Arcy (Ref 15), who examined water flow rates from the public fountains of Dijon, France, through sand beds of varying thicknesses in 1856. He formulated the basic principle behind permeability, showing that the average flow rate is proportional to the pressure gradient and inversely proportional to the thickness of the bed. In 1927, Kozeny (Ref 16) published the first derivation showing correlation among porosity, permeability, and particle surface area. Following the lead of Blake (Ref 17), Kozeny treated the flow of fluid through a particulate bed as being equivalent to the flow of fluid through a comparable volume of parallel pipe channels of circular cross section. This simplification resulted in determining an equivalent diameter of pipe channel to characterize flow rate through the powder bed. In 1938, Carman (Ref 18) and Dallavalle (Ref 19) independently proposed the determination of specific surface area for powders using permeability methods. Carman published related experimental work in 1941 (Ref 20). He developed a liquid flow technique to determine surface area for coarse materials by taking into consideration (1) the dependence of permeability on the number of permeable pores of the particle bed, (2) the pore or void volume fraction contribution to total bed volume, (3) the friction of the gas or liquid flowing through the bed, and (4) the adsorption of immobile liquid layers that effectively reduce the capillary diameter, thus causing less permeability and therefore a greater apparent surface area. The Kozeny-Carman equation, given below, has served as the most widely used basis for all permeability variations:

where S is surface area per unit weight of the powder, P is the pressure drop across the powder bed, fv is void fraction of packed sample, v is velocity of fluid flow, is density of the powder material, is viscosity of the fluid, L is length of the powder bed, and Le is average path length through the powder bed.

References cited in this section

15. H.P.G. De'Arcy, Les Fontaines Publiques de la Ville de Dijon, Victor Dalmont, 1856 16. J.J. Kozeny, Akad. Wiss. Wein. Math. Naturwiss. K.I., Sitzungsbor, Abstr. IIA, Vol 136, 1927, p 271-306 17. F.C. Blake, Trans. Am. Inst. Chem. Eng., Vol 14, 1922, p 415 18. P.C. Carman, J. Soc. Chem. Ind. Lond., Vol 57, 1938, p 225-234; Trans. Inst. Chem. Eng., Vol 15, 1932, p 150-166 19. J.M. Dallavalle, Chem. Met. Eng., Vol 45, 1938, p 688 20. P.C. Carman, Symposium on New Methods for Particle Size Determination in the Sub-Sieve Range, ASTM, 1941, p 24

Surface Area, Density, and Porosity of Powders

Apparatus Lea and Nurse (Ref 21) developed the apparatus shown in Fig. 4 to provide permeability measurements. The powder was compacted in the sample cell to a predetermined porosity. Air was permitted to flow through the bed, and the pressure drop (h1) was measured on the first manometer; the air then passed through a capillary flowmeter, across which another pressure drop was measured as h2 on a second manometer.

Fig. 4 Lea and Nurse permeability apparatus with manometer and flowmeter

The capillary permitted the system to operate under a constant pressure. The volume rate of flow through the flowmeter, the pressure drop across the bed as measured by the manometer, and the constants associated with the apparatus permitted determination of the specific surface area (surface area per unit volume). Gooden and Smith (Ref 22) added a self-calculating chart to a modified Lea and Nurse apparatus to enable direct readout of the specific surface. The commercial version of their modification is known as the Fisher subsieve sizer (Fig. 5).

Fig. 5 Fisher subsieve sizer operation

A simplified version of the air permeameter, known as the Blaine permeameter (Fig. 6) (Ref 23), relied on a variable pressure technique (Ref 24). A vacuum was used to displace the oil in a U-tube connected in series with the powder cell. The resultant pressure caused air to flow through the powder bed, and the time required for the displaced oil to fall back to its equilibrium position was measured. This method resulted in a measured specific surface area, which decreased with porosity. Usui (Ref 25) showed that log t and the void fraction exhibited a linear relationship and that a plot of these parameters gave a value for surface area.

Fig. 6 Blaine air permeability apparatus. Source Ref 23

References cited in this section

21. F.M. Lea and R.W. Nurse, J. Soc. Chem. Ind., Vol 58, 1939, p 277-283; Symposium on Particle Size Analysis, Trans. Inst. Chem. Eng., (suppl.), Vol 25, 1947, p 47-56 22. E.L. Gooden and C.M. Smith, Ind. Eng. Chem. Anal. Ed., Vol 12, 1940, p 479-482

23. K. Niesel, External Surface of Powders From Permeability Measurements--A Review, Silicates Industrials, 1969, p 69-76 24. R.L. Blaine, ASTM Bull., No. 123, 1943, p 51-55; also see ASTM Bull., No. 108, 1941, p 17-20 25. K. Usui, J. Soc. Mater. Sci. Jpn., Vol 13, 1964, p 828 Surface Area, Density, and Porosity of Powders

Limitations For very fine powders, the basic Kozeny-Carman equation is not accurate. This is because the laminar flow assumption on which it is based is no longer valid. Compressed fine particles result in a powder bed with very small channel widths. If these widths are comparable to the mean free path length of the gas molecules, laminar flow conditions are not maintained. Such a situation, involving molecular flow or diffusion conditions, is known as Knudsen flow (Ref 26) and can occur with very fine powders or with coarser particles at low pressures. Figure 7 shows a typical apparatus used to measure find particles under molecular flow conditions (Ref 27, 28). In some powders, both laminar and molecular flow may be significant. This is known as the transitional region.

Fig. 7 Modified Pechukas and Gage apparatus for fine powders. Source: Ref 27

Evaluating surface areas with steady-state flow conditions historically excluded noninterconnected blind pores. A method for including blind pores by utilizing transient-state flow measurements is the principle behind the apparatus shown in

Fig. 8(a) (Ref 29, 30). A typical flow rate curve, showing extrapolation of the steady-state portion to determine the time lag is given in Fig. 8(b). The basic principle of the permeability analysis has not changed much during the last several years; even so, instruments now use more sophisticated pressure transducer and flowmeter or flow controller. Combined with a computerization of the instrument and an improved data analysis software, the instrument offer a wider variety of analysis tasks and computational methods.

Fig. 8(a) Transient flow apparatus

Fig. 8(b) Flow rate curve for the transient flow apparatus

References cited in this section

26. M. Knudsen, Ann. Physik, Vol 28, 1909, p 75-130 27. A. Pechukas and F.W. Gage, Ind. Chem. Eng. Anal. Ed., Vol 18, 1946, p 37 28. P.C. Carman and P.R. Malherbe, J. Soc. Chem. Ind., Vol 69, 1950, p 134 29. R.M. Barrer and D.M. Grove, Trans. Faraday Soc., Vol 47, 1951, p 826, 837 30. G. Kraus, R.W. Ross, and L.A. Girifalco, J. Phys. Chem., Vol 57, 1953, p 330 Surface Area, Density, and Porosity of Powders

Pycnometry Peter J. Heinzer, Imperial Clevite Technology Center

Pycnometry is used to determine the true density of P/M materials. Based on the displacement principle, pycnometry is actually a method of determining the volume occupied by a solid of complex shape, such as a powder sample. For commercial pycnometers, typical sample sizes range from 5 to 135 cm3 (0.30 to 8.24 in.3). A properly prepared specimen can be analyzed in 15 to 20 min. The pycnometric determination of density can be quite useful in P/M applications. In addition to its primary use in measuring the true density of a P/M part or product, it can be used to distinguish among different crystalline phases or grades of material, different alloys, compositions, or prior treatments. Information on the porosity of a material can be obtained from pycnometry if the sample has a uniform geometry, or if the bulk volume is known. Pore volume is the difference between the bulk volume (1/bulk density) and the specific volume (1/true density). Finally, pycnometry can be useful in determining properties that relate to density. Often, P/M materials have no solid counterpart to use for measuring true density, making percentage of theoretical density measurements questionable. Pycnometric measurements of the true density of the powder have provided a good point of reference. Density is one of the most important properties of P/M materials. Critical processing parameters, such as applied force and pressure, and properties of the resulting P/M product, such as strength and hardness, usually depend on the density of the materials being processed. Standard industry practice compares the achieved density of a P/M product with the full, or theoretical, density. Surface Area, Density, and Porosity of Powders

Theory and Apparatus Archimedes devised the first method for determining true density by using the displacement principle. Modern pycnometry represents a refinement of the displacement principle and uses either a liquid or gaseous substance as the displaced medium. Absolute densities of solids can be measured by the displacement principle using either liquid or gas pycnometry. In liquid pycnometry, volume displacement is measured directly, as liquids are incompressible. Inability of the liquid to penetrate pores and crevices, chemical reaction or adsorption onto the sample surface, wetting or interfacial tension problems, and evaporation contribute to errors in density measurement. Therefore, gas pycnometry is usually preferred for P/M applications.

In gas pycnometry, volume displacement is not measured directly, but determined from the pressure/volume relationship of a gas under controlled conditions. Gas pycnometry requires the use of high-purity, dry, inert, nonadsorbing gases such as argon, neon, dry nitrogen, dry air, or helium. Of these, helium is recommended because it: • • •

Does not adsorb on most materials Can penetrate pores as small as 0.1 nm (1 Behaves like an ideal gas

)

In commercial pycnometers, the sample is first conditioned or outgassed to remove contaminants that fill or occlude pores and crevices, thus changing surface characteristics. This is accomplished by evacuating the system and heating to elevated temperatures, following by purging with an inert gas such as helium. The helium-filled sample system (Fig. 9) is "zeroed" by allowing it to reach ambient pressure and temperature. At this point, the sample cell and reference volume are isolated from each other and from the balance of the system by valves.

Fig. 9 Flow chart for typical pycnometer

The state of the system can then be defined by:

PV = nRT

(Eq 1)

for the sample cell and:

PVR = nRRT

(Eq 2)

for the calibrated reference cell. In these equations, P is the ambient pressure, Pa; V is the volume of the sealed empty sample cell, cm3; VR is volume of a carefully calibrated reference cell, cm3; n is moles of gas in the sample cell volume at P; nR is moles of gas in the reference cell volume at P; R is the gas constant; and T is ambient temperature, K. A solid sample of volume (Vs) is then placed in the sample cell:

P(V - Vs) = n1RT

(Eq 3)

where n1 is moles of gas occupying the remaining volume in sample cell at P. The system is then pressurized to P2, about 100 kPa (15 psi) above ambient:

P2(V - Vs) = n2RT

(Eq 4)

where n2 is moles of gas occupying the remaining volume in the sample cell at P2. The valve is then opened the connect the sample cell with the calibrated reference volume, and the pressure drops to a system equilibrium P3:

P3(V - Vs) + P3VR = n2RT + nRRT

(Eq 5)

Substituting PVR from Eq 2 for nRRT in Eq 5 and substituting P2(V - Vs) from Eq 4 for n2RT results in:

P3(V - Vs) + P3VR = P2(V - Vs) + PVR

(Eq 6)

Simplifying:

(P3 - P2) (V - Vs) = (P - P3)VR

(Eq 7) (Eq 8)

Vs = V + VR/[1 - (P2 - P)/(P3 - P)]

(Eq 9)

Because P is "zeroed" at ambient before pressurizing, the working equation becomes:

Vs = V + VR/[1 - (P2/P3)]

(Eq 10)

Over the last few years, the gas pycnometer has been further improved by using a more accurate pressure transducer, a better temperature control of the entire system, and by an automation (computerization) of the actual analysis process. Modern pycnometers can now reach an accuracy of 0.01%. Surface Area, Density, and Porosity of Powders

Mercury Porosimetry H. Giesche, School of Ceramic Engineering and Sciences, Alfred University

Many commercially important processes involve the transport of fluids through porous media and the displacement of one fluid, already in the media, by another. The role played by pore structure is of fundamental importance in understanding of these processes. The quality of powder compacts is also affected by the void size distribution between the constituent particles. For these reasons, mercury porosimetry has long been used as an experimental technique for the characterization of pore and void structure. Gas and mercury porosimetry are complementary techniques with the latter covering a much wider size range from 0.3 mm to 3.5 nm (Fig. 10). Mercury porosimetry consists of the gradual intrusion of mercury into an evacuated porous medium at increasingly higher pressures followed by extrusion as the pressure is lowered. The simplest pore model is based on parallel circular capillaries that empty completely as the pressure is reduced to zero. This model fails to take into account the real nature of more porous media, which consist of a network of interconnecting noncircular pores. The network effects lead to hysteresis and mercury retention during the extrusion cycle.

Fig. 10 Pore radii ranges covered by gas and mercury porosimetry

Surface Area, Density, and Porosity of Powders

General Description Relationship between Pore Radii and Intrusion Pressure. Mercury porosimetry is based on the capillary rise

phenomenon whereby an excess pressure is required to cause a nonwetting liquid to enter a narrow capillary. The pressure difference across the interface is given by the equation of Young (Ref 31) and Laplace (Ref 32), and its sign is such that the pressure is less in the liquid than in the gas (or vacuum) phase if the contact angle is greater than 90° or more if is less than 90°. If the capillary is circular in cross section, and not too large in radius, the meniscus will be approximately hemispherical. The curvature of the meniscus can be related to the radius of the capillary, and the Young-Laplace equation reduces to the Washburn equation (Ref 33):

(Eq 11)

This is the Young-Laplace and Washburn equation where lv is the surface tension of the liquid (e.g., for mercury, 0.485 N/m), r1 and r2 are mutually perpendicular radii of a surface segment. The angle is the angle of contact between the liquid and the capillary walls and is always measured within the liquid (Fig. 11). rP is the capillary radius.

Fig. 11 Contract angle ( ) of a liquid in a capillary

Equipment Fundamentals. The sample is placed into the penetrometer assembly; it is then evacuated to a set vacuum level for a specific time, before the sample cell is filled with mercury. Air is admitted to the low-pressure chamber, and the increasing pressure forces the mercury to penetrate the largest pores of the sample. The amount or volume of mercury penetrating into the sample is recorded at each pressure (or pore size) point; the first reading usually is taken at a pressure of 0.5 psi (0.003 MPa), although readings at a pressure of 0.1 psi (0.7 × 10-4 MPa) are possible. The pressure is then increased to 1 atm, or in some instruments the pressure is actually increased to a slight overpressure (up to 50 psi in some cases). After the low-pressure run is finished, the penetrometer is then inserted into a high-pressure port and surrounded with a special grade of high-pressure oil; it is special with respect to the dielectric constant and viscosity of the oil under high-pressure conditions. The pressure is increased up to a final pressure of 60 ksi (400 MPa). Commercial instruments work either in an incremental or continuous mode. In the former, the pressure is increased in steps and the system allowed to stabilize at each pressure point before the next step. In the continuous mode, the pressure is increased continuously at a predetermined rate. Schematics of low-pressure and high-pressure systems are shown in Fig. 12 and 13, respectively (Ref 34).

Fig. 12 Low-pressure mercury porosimeter. Source: Ref 34

Fig. 13 Micromeritics high-pressure mercury porosimeter

References cited in this section

31. T. Young, Miscellaneous Works, Vol 1, Murray 1855, p 418 32. P.S. Laplace, Mecanique Celeste, Suppl. Book 10, 1806 33. E.W. Washburn, Proc. Nat. Acad. Sci. U.S.A., Vol 7, 1921, p 115 34. AutoPore II 9220 Operator's Manual, Micromeritics, 1993 Surface Area, Density, and Porosity of Powders

Measurement Techniques Measuring Displacement Volumes (Pore Volume). Mercury volume displacements may be measured by direct

visual observation of the mercury level in a glass penetrometer stem (Fig. 14) with graduated markings (Ref 35). However, most (if not all) instruments on the market will measure this volume automatically by one of the following techniques:

• • •

Precision capacitive bridges: measure changes in the capacitance between the column of mercury in a dilatometer stem and a coaxial sheet surrounding the column Mechanical transducer: indicate the change in height of the mercury column by moving a contact wire and measuring the displacement of the mercury interface in the stem Submerged wires: measure changes in resistivity corresponding to the change in length of the mercury column

From a practical point of view, the sample mass (pore volume) and the stem volume of the penetrometer should be adjusted to use the instrument to its highest resolution. In general, larger samples are preferred because they provide a better representation of the overall sample.

Fig. 14 Quantachrome filling mechanism and low-pressure porosimetry system. Source: Ref 35

Pressure. The corresponding pressure at which mercury is filling the pore system is usually measured with electronic pressure transducer or with Heise-Bourdon manometer, used in older manual setups. A series of those pressure transducers ensures that accurate data are determined over the entire range from 0.1 psi (0.7 × 10-4 MPa) to 60 psi (0.4 MPa).

Looking at the Washburn equation (Eq 11) makes it obvious that two additional parameters play a critical role in the calculation of pore size from the applied pressure: contact angle, , and the surface tension, Hg. Contact Angle Determination. Various techniques are available to determine the contact angle:

•

A drop of mercury can be placed on the flat surface of the sample, and the resulting contact angle is visually observed. Problems related to "micro" and "macro" contact angles have been reported (Fig. 15) (Ref 36). Brashforth and Adams (Ref 37) published tables that allow calculation of the contact angle as well as the surface tension of liquids from the shape of a drop of mercury on the substrate surface (Fig.

16). A simplified formula can be used when the maximum height of the drop is reached (Ref 35, 37, 38):

• •

•

with g, the acceleration of gravity, and , the density of the liquid. A powder compact can be pressed in such a way that a well-defined hole is created in a disk. Mercury is now placed on top of this disk, and the contact angle can be calculated from the necessary pressure to force the mercury through this cylindrical pore. The Willhelmy plate method (Fig. 17) can be used to determine the contact angle (Ref 39). Figures 17 and 18 illustrate the critical observation of an advancing and receding mercury interface. Surface roughness (Fig. 18) or the change in surface composition during the contact with mercury can explain the presence of this difference. (Note: there is no thermodynamic reason or explanation for any contact angle hysteresis.) No surface roughness effects are assumed below pore sizes of about 100 nm. This effect also emphasizes the importance of clean samples and clean mercury.

The contact angle between mercury and the sample being tested is frequently assumed to be 130 or 140°. This assumption is probably the largest source of error. Contact angles of different materials may differ significantly, as shown in Table 2 (Ref 40).

Table 2 Contact angle between mercury and select P/M materials Powder Aluminum Copper Glass Iron Zinc Tungsten carbide Tungsten

Source: Ref 40

Angle, degrees 140 116 153 115 133 121 135

Fig. 15 Differences between microscopic and macroscopic measurement of the contact angle ( ) under conditions of (a) wetting and (b) nonwetting

Fig. 16 Change of mercury-drop shape with size

Fig. 17 Wilhelmy plate method showing the effect of contact angle hysteresis for emersion and immersion. Adapted from Ref 39

Fig. 18 Effect of surface roughness on contact angles

The Washburn equation is directly proportional to the cosine of the contact angle; the respective pore size errors for iron ( = 115°) and glass ( = 153°), using the values from Table 2 versus a constant value of 130° for , would be:

or

However, published contact angles differ widely between different research groups, even when presumably the same material was studied. Some materials might react with mercury, for example, zinc, silver, or lead samples. This severely changes the nonwetting behavior of mercury with that sample and may even lead to a contact angle of 800 MPa Ambient Hard High No 1

Die compaction High, 700 MPa Ambient Hard High No 1

Explosive compaction Very high, >1 GPa Very high Soft Very high No 1

Roll compaction Low Ambient Hard Low Yes 1

Warm compaction High, 700 MPa Warm Hard, heated High No 1

Moderate to high None Low Moderate

Moderate None Moderate Low

High Low, 0.5% High Extensive

Low None Low Very low

Low Low, 0.1% High Moderate

Moderate Low, 0.6% High Low

Reference

1. R. German, Powder Metallurgy of Iron and Steel, John Wiley & Sons, 1998 Powder Shaping and Consolidation Technologies B. Lynn Ferguson, Deformation Control Technology, Inc.; Randall M. German, The Pennsylvania State University

Powder Shaping Technologies General categories of powder shaping methods are as follows (Ref 1): •

Binder-assisted extrusion: long structures, small powders, constant cross section, relatively simple shapes • Injection molding: complex, small components, high-performance materials • Slip casting: very large structures, constant wall thickness, low precision • Tape casting: flat sheets, small powders, very simple shapes

In general these methods employ binders that hold the particles together in the desired shape; the particles are then consolidated to higher densities by sintering. The part is shaped at relatively low pressures (compared to die compaction) with green (unsintered) porosity ranging from ~40 to 60% (ignoring the binder that is sacrificial) (Ref 1). Most binders are polymers such as mineral oil or polyethylene. The following binders are used in powder shaping:

Extrusion • •

56% water, 25% methyl cellulose, 13% glycerine, 6% boric acid 72% water, 12% hydroxypropyl methyl cellulose, 8% glycerin, 4% ammonium polyacrylate, 4% ammonium stearate • 65% polyethylene glycol, 30% polyvinyl butyryl, 5% stearic acid Injection molding • • • •

69% paraffin wax, 20% polypropylene, 10% carnauba wax, 1% stearic acid 75% peanut oil, 25% polyethylene 50% carnauba wax, 50% polyethylene 55% paraffin wax, 35% polyethylene, 10% stearic acid

Slip casting • • • •

96% water, 4% sodium lignosulfonate, trace calcium nitrate 93% water, 4% agar, 3% glycerine 99% water, 1% ammonium alginate 97% water, 3% polyvinyl alcohol

Tape casting •

77% water, 9% polyacrylate emulsion, 9% glycerin, 3% ammonium polyacrylate, 2% ammonium hydroxide • 80% toluene, 13% polyethylene glycol, 7% polyvinyl butyral

•

47% mineral spirits, 24% isopropanol, 8% polyvinyl butyral, 8% dibutyl phthalate, 5% polyethylene glycol, 5% stearic acid, 3% menhaden fish oil

In most cases, the shaping process requires four steps: • • • •

Formation of a powder-binder feedstock Shaping of feedstock using customized tooling Extraction of the binder Sintering densification

Feedstock formulation involves small powders and selected binders that are mixed in a ratio dictated by the desired rheology. Shaping may be conducted by several techniques, but powder lubrication by the polymer is key to fabrication of the desired shape. Low-pressure molding concepts, such as slip casting, require an excess of binder. On the other hand, high-pressure shaping is possible with high-viscosity mixtures that contain more powder. If necessary, the polymer viscosity can be controlled by either heat or solvent. After shaping, the binder is removed (debinding), and the remaining powder structure is sintered. These last two steps can be combined into a single thermal cycle. Since significant shrinkage is associated with sintering densification, the final dimensions rely on uniform shaping to hold tight final tolerances. Slip Casting. In this method, a low-viscosity slurry (typically from 10 to 50 Pa · s) is prepared with a liquid carrier and

powder. The slurry mixture is then poured into a porous mold usually fabricated from plaster of paris. The carrier is evaporated and absorbed into the porous mold, and a free standing shape is thus achieved. In the case of cements, a chemical bond can be created during the drying or curing process. For many powders, sintering is required to achieve particle bonding and to produce a part having useful mechanical properties. This process is more applicable to ceramic powders, and the most popular application is the production of ceramic bathroom fixtures. Slip casting is not commonly used for metal powders. For steels, the largest application is rapid fabrication of prototype shapes and pilot tool steels (Ref 1). Slip casting is very cost effective for large shapes due to inexpensive tooling. More information on this method can be found in the article "Slip Casting of Metals" in this Volume. Tape casting uses a similar low-viscosity mixture of powder, water, and polymer, but the forming step is via deposit of

the mixture on a moving Mylar film. As the water evaporates, the particles are glued together into a thin sheet. It is used to generate homogeneous long, thin, and flat structures, such as sheets. Many of the characteristics of the binder and powder are the same as encountered in slip casting. After forming, the solvent or water is evaporated, leaving residual binder to hold the particles in place. Tape casting is widely used to form battery electrodes, brazing layers, microelectronic substrates, and small production levels of steel sheets. In tape casting, the powder-binder slurry is fed onto a moving paper or plastic sheet that passes through a controlled opening. A doctor blade levels the slurry to form a continuous thin sheet (Fig. 1). Subsequently, one constituent of the binder is evaporated, leaving bonds of binder behind to provide strength during separation of the tape from the substrate before final sintering. Binders for tape casting are similar to those discussed earlier. Acrylics, waxes, polyvinyl alcohol, and polyvinyl butyryl are common components.

Fig. 1 In tape casting, a thin layer of powder-binder mixture is formed on a moving Mylar substrate, which carries the tape into a drying oven prior to sintering. The technique is used to form thin sheets. Source: Ref 1

Freeze Casting. A novel shaping technique relies on water and freezing to form a shape. The water contains polyvinyl alcohol or latex. After casting into hard tooling, the powder-binder mixture is frozen in the tool set. Subsequently, the ice is extracted by vacuum evaporation (freeze-drying or sublimation) from the compact, leaving behind a polymer component to hold the particles in place for subsequent handling. Sublimation of the water is slow and requires at least a few hours. Finally, the powder is sintered to a high density. The process is known by several names, including adiabatic forming, freeze-firing, and quick set. This freeze-sublime-sinter approach is mostly used for large structures; stainless steel is a preferred material. Current production targets components >1 kg (2.2 lb) in mass with modest production volumes--typically a few per day. This is dictated by the time to freeze the component in the tooling, which limits the cycle time to 1 h. Injection Molding of Metals and Ceramics. Injection molding is now the most widely employed shaping

technique for steel powders. The molding process competes with many other shaping technologies. It overcomes the property limitations inherent in plastics, the shape limitations of traditional powder compaction, the costs of machining, the productivity limits of isostatic pressing and slip casting, and the defect and tolerance limitations of casting. Part production by injection molding is used with various metals, alloys, and ceramic materials, such as the following (Ref 2):

Elements • • • • • • • • • •

Beryllium Copper Iron Gold Molybdenum Nickel Niobium Silver Titanium Tungsten

• • •

Bronze Cobalt-base Copper-base

Alloys

• • • • • • • • • • • • • • • •

Gold-base Hastelloy Inconel Invar Iron-nickel Iron-silicon Kovar Nickel-base Niobium-base Stainless steel Steel Stellite Sterling silver Superalloys Tool steel Tungsten heavy alloy

Ceramics and compounds • • • • • • • • • • • • • •

Alumina (A12O3) Alumina-chromia (Al2O3-Cr2O3) Aluminum nitride (AlN) Ferrites (MnFe2O4) Hydroxyapatite (Ca10(PO4)6(OH)2) Mullite (3Al2O3-2SiO2) Nickel aluminide (Ni3Al) Silica (SiO2) Silicon carbide (SiC) Silicon nitride (Si3N4) Spinel (MgAl2O4) Titania (TiO2) Yttria (Y2O3) Zirconia (Zr2O3)

Cermets and composites • • • • • • • • • • • • •

Al2O3-SiC Al2O3-ZrO2 Mo-Cu NbC-Ni Ni3Al-Al2O3 Si3N4-SiC SiO2-Si TiC-Ni-Mo W-Cu WC-Ni WC-TaC-Co ZrO2-MgO ZrO2-Y2O3

Noticeably absent are magnesium and aluminum; these reactive metals develop powders with surface oxides that are difficult to remove during sintering. Stainless steel is the most widely used material for injection molded parts, as shown

by the relative ranking in Fig. 2. Stainless steel 316L is the most commonly used alloy. Ferritic and duplex stainless steels are also used.

Fig. 2 Relative material utilization in powder injection molding (PIM) processing on a weight basis, showing the dominant position of steels and stainless steels. Source: Ref 2

Injection molding moved quickly from a laboratory development to a commercial process for two reasons. First, a precision shape could be achieved at a significant cost reduction compared to a conventional manufacturing process. Second, the use of fine powders promoted densification during sintering, and high properties could be achieved. The high potential of the process added a third reason that sustained growth. Availability of fine powders and pellet feedstock from suppliers increased as the process showed early successes. There is still a great deal of development work toward more efficient and economical production of fine powders (92% of full density. This technology opens new applications to pressed and sintered parts because of the achievable density and the improved mechanical and physical properties. In addition, higher green strengths allow machining of green compacts. Usually warm compaction involves the use of a polymer addition that helps bond particles together. The polymer-coated powder is more costly than typical die-compaction grades, unless a simple lubricant is admixed with the powder. Various common stearates or other lubricants work, including, Teflon (E.I., duPont de Nemours & Co., Inc., Wilmington, DE). Depending on which polymer is selected for coating the powder, ejection forces can be highly variable. Close temperature control is necessary, since product uniformity suffers if the polymer is too hot. In tests with various powders, the green density usually increases by 0.15 g/cm3 over room-temperature compaction. After cooling to room temperature, the warm-compacted powder is stronger because chilling the polymer adds strength to the compact. However, there is no evidence of greater strength during ejection, which means that green cracks from ejection stresses are not reduced by warm compaction. Consequently, a hold-down pressure is required during ejection to avoid cracking. Heating of the die and punches requires modifications to the compaction press, and a heater is required in the powder feed mechanism. Both microwave and hot-oil heaters are available for heating the powder. A typical temperature for the powder and tooling is 150 °C (300 °F), and compaction pressures are usually in the range of 700 MPa (50 tsi) for steels. The major role of warm compaction is to lower the pressure required for attaining densities of >7.0 g/cm3 of ferrous P/M compacts. Hot Pressing. The production of large billets can be accomplished by compacting powder in heated dies. The use of

elevated temperatures and long dwell times allows densities of >95% of full density to be achieved at compaction pressures that are one third to one half those needed for cold pressing to lower density levels. Full density is usually not achieved, and 3 to 5% porosity remains in the billet. This porosity somewhat limits the use of hot pressed billets because the properties are reduced from those of fully dense material. For this reason hot pressed billets are often used as stock for upset forging, closed die forging, and other deformation processes that can eliminate this residual porosity. Some metals, such as beryllium, are routinely processed by hot pressing, and acceptable performance levels for many applications are achieved. Isostatic Pressing

Isostatic pressing allows more uniform density compared to uniaxial compaction in rigid dies. These methods rely on flexible molds for application of pressure in all directions, which reduces friction and allows compaction of compact shapes. Cold Isostatic Pressing. CIP uses a flexible membrane to isolate the powder from a liquid medium that is pressurized

to cause densification of the powder. Typical mold materials are latex, neoprene, urethane, polyvinyl chloride, and other elastomeric compounds. Because the mold moves with the powder as it densifies, friction effects are minimized. Also, because the pressure is applied uniformly around the mold, there is no theoretical size limit. Height-to-diameter and overall size are limited by the pressure vessel size. Often a rigid mandrel is part of he tooling; and, because powder must slide along this mandrel, it is coated with a friction reducing material. In comparison to die pressing, cold isostatic pressing can achieve more uniform densities due to minimized friction effects. Pressure vessels are typically limited to pressures of 415 MPa (60 ksi) although units with twice this capacity have been produced. Isostatic pressing equipment can be automated (i.e., dry bag CIP units), but the production rates are lower than those of die pressing. Dimensional control is generally not as tight as with die pressing due to flexible tooling. As stated, however, rigid members can be incorporated into the flexible mold assembly to produce accurate surfaces where desired. Hot isostatic pressing (HIP) is a versatile near-net shape process that has found niches in the production of

aerospace structure and engine markets, high alloy and tool steel mill shapes and individual components, titanium hardware, and monolithic and composite alloy components for the energy industry. The process fundamentals and manufacturing steps are covered in detail in the article "Hot Isostatic Pressing of Metal Powders" in this Volume. The aim of hot isostatic pressing is a near-net shape and full density. Powder is hermetically sealed in a container that is flexible at elevated temperatures; the "canned" powder is heated within a pressurized vessel and held for a specified time. Commercially used containers include low carbon steel sheet formed into a container, stainless steel sheet, and even glass. The pressurized medium is usually an inert gas such as argon, and pressures range between 100 and 300 MPa (15 and 45 ksi). The temperature for HIP is material dependent, of course, but typical production equipment can heat parts 1000 to 1200 °C (2000 to 2200 °F). HIP units for ceramics and carbon-base materials may heat up to 1500 °C (2700 °F). Densification mechanisms active during HIP include bulk deformation (limited amount), sintering, and creep, with the latter accounting for a significant portion of densification. Densities >98% of full density are typical, and full density is routinely achievable with care during powder sealing and strict control of time, pressure, and temperature. The powders used in hot isostatic pressing are usually spherical in shape and very clean. The particle surfaces are free of contaminants, such as oxide films. The sphericity facilitates can loading and handling, and the particle surface cleanliness facilitates particle bonding. Powder handling and avoidance of contamination is critical to the success of the process, and considerable investment in facilities and equipment, followed up by attention to operating procedures and "good housekeeping," is required. In comparison to hot pressing where only billet shapes are produced, hot isostatic pressing is capable of producing complex shapes. As in cold isostatic pressing, the achievable dimensional tolerances are at best near-net due to the flexible mold. Some net surfaces may be achieved if rigid members are incorporated into the mold. Figure 6 compares HIP capabilities with other compaction methods.

Fig. 6 Application areas of HIP based on part size, complexity, and level of densification three variables that dictate the P/M approach are size, density, and performance (as a percentage of wrought). This behavior corresponds to ferrous-base P/M systems, but is representative of many P/M materials. P/S, press and sinter; reP, press, sinter and repress; P/S + F, press and sinter and forge; CIP + S, cold isostatically press and sinter; HIP, hot isostatic press; HIP + F, hot isostatic press plus forge.

Powder Shaping and Consolidation Technologies B. Lynn Ferguson, Deformation Control Technology, Inc.; Randall M. German, The Pennsylvania State University

Bulk Deformation Processes Powder Forging. In powder forging, a preform shape is cold pressed to between 75 and 85% of full density, sintered, heated to a forging temperature, and then forged in trapped dies using one blow to produce a fully dense net or near-net shape. The sintering step is optional, but it is normally included as a particle surface cleaning step (deoxidation) and to improve the workability of the porous preform. Powder forging fundamentals and general applications are presented in the article "Powder Forging and Hot Pressing" in this Volume. More detailed information is also contained in the article "Powder Forged Steel."

Normally, powder forging is performed hot (1000 to 1200 °C, or 1800 to 2200 °F, for steel powders), but it can also be conducted at warm or cold forging temperatures. Typical steel powder forging pressures are 550 to 950 MPa (40 to 70 tsi). Because the workability of a porous preform is poor, the design of the preform is critical to the process in terms of avoiding defects. Local surface tensile stresses and internal hydrostatic tension must be avoided. Nonetheless, with a correct preform shape and a well controlled process, powder forged parts have sound microstructures, good hardenability, and performance that meets or exceeds cast-wrought part performance. There are two classes of forging practice. In repressing, the preform shape is nearly identical to the forged shape. In true forging, considerable shear deformation is involved since the preform is different in shape from the forging. The dynamic properties of toughness and fatigue resistance are higher for forged parts than for repressed parts. Powder Extrusion. Loose powder can be containerized and extruded to full density, either with or without heat.

Extrusion ratios of at least 9:1 have been shown to produce full density, and many materials are extruded to full density using much higher extrusion ratios. Extrusion at such high reductions subjects the powder particles to high levels of shear deformation and compression as they pass through the die. The result is sound particle bonding. As in hot isostatic pressing, the powder is usually hermetically sealed in a container prior to extrusion. In some alloy systems, residual air provides oxygen for selection oxidation of alloy ingredients prior to extrusion, and an oxide dispersion strengthened material is produced. In some other cases, particulates or chopped fibers may be added to the matrix powder, and extrusion produces a composite material. In still other cases, combinations of powders or powder plus wrought pieces are

extruded to produce multimetal parts, e.g., bimetallic tubing. These and other uses of extrusion are discussed in the article "Extrusion of Metal Powders" in this Volume. Powder rolling or roll compaction of powders consolidates loose powder(s) into a porous strip as it passes through a roll gap. It is possible to produce a monolithic strip or a multilayer strip. Further processing may include sintering of the strip and additional rolling to densify the strip. Applications of this type of processing include such diverse products as clad metal for coin stock, automotive sleeve bearings, and electrode stock. Powder Shaping and Consolidation Technologies B. Lynn Ferguson, Deformation Control Technology, Inc.; Randall M. German, The Pennsylvania State University

Sintering For sintering to be effective, the powder particles must be in intimate contact. For this reason, sintering is typically performed on compacted or molded powder and not on loose powder. While most powder parts are sintered during their manufacture, the densification stage of sintering may not be utilized, and in these cases the primary use of sintering is to achieve metallurgical bonding of particles. Die pressed parts typically fall into this category because dimensional control is of primary importance. Process routings that rely on sintering specifically to achieve high densities usually fall into three categories. A very fine particle size is used so that bulk diffusion paths are minimized at the last stage of sintering, and there is a high initial surface area to drive sintering in its early stages. This is one reason for the use of ultrafine powders in injection molding where sintering is the primary mechanism for development of high property levels. A liquid phase, either permanent or transient, is present during sintering so that particle rearrangement is promoted. A high temperature is used so that local melting may accompany sintering. Liquid Phase Sintering. The use of blended powders allows the combination of a low melting powder with a higher melting powder so that a liquid is present during sintering. This liquid can aid powder rearrangement and densification of a green compact. Cermets are typically produced by this technique. Nearly pore-free microstructures may be achieved by this technique. Transient Liquid Phase Sintering. A subclass of liquid phase sintering is transient liquid phase sintering. Here, a

blend of powders contains a lower melting point phase, as above. Only in this case, the liquid alloys with the solid phase which remains solid, and the liquid is present only for a brief period. This process is not used as the primary densification method, but it is used to achieve another increment in density. For example, it may be used in high alloy steel or tool steel production to increase the density of a green compact from 80% to >95% of full density. High-Temperature Sintering. If the temperature of the green compact is raised to levels very close to the solidus

temperature of the equilibrium diagram, small pools of liquid can form due to microsegregation of alloying elements. This small amount of liquid accelerates densification during sintering. Although this represents a somewhat extreme use of high temperature sintering, this technique is useful for a wide range of alloys, and it is used especially for tool steels and other high alloy steel components. The article "High-Temperature Sintering of Ferrous Powder Metallurgy Components" in this Volume contains more information on high-temperature sintering. Reactive Sintering. By taking advantage of the flexibility of powder metallurgy, a blend of powders may be

compacted and then sintered so that during sintering a self propagating reaction occurs (see "Combustion Synthesis of Advanced Materials" in this Volume). This reaction, if controlled properly, can densify the compact rapidly while forming the new alloy. If not controlled, the component may revert to powder!

Powder Shaping and Consolidation Technologies B. Lynn Ferguson, Deformation Control Technology, Inc.; Randall M. German, The Pennsylvania State University

Compaction to Higher Density Many methods have been investigated and developed for pressing metal powders to higher densities. Most of the commercial methods involve high-temperature compaction (such as hot isostatic pressing and powder forging, as described in detail in other articles in this Section). In general, these methods are used to develop fully dense or nearly fully dense P/M products. For example, conditions for full-density iron and steel by HIP are summarized in Table 2.

Table 2 Hot isostatic pressing conditions for full-density iron and steel Material 4 m carbonyl iron 75 m sponge iron 70 m low-alloy steel 190 m maraging steel 100 m austenitic stainless steel 120 m martensitic stainless steel 65 m Fe-10Al-5Si 80 m tool steel 165 m iron superalloy

Pressure, MPa 200 98 150 210 160 150 200 100 69

Temperature, °C 805 1000 800 1200 1150 1150 1000 1100 1200

Time, h 1 1 1 3 3 3 1 1 3

Source: Ref 1

However, other methods besides high-temperature compaction are used to achieve higher densities in green or consolidated form. Some of these miscellaneous methods are described here. These methods are not necessarily fulldensity methods, and none have reached any significant commercial significance. They are described for general reference. Many experiments also have been reported that attempt to press compacts to higher densities or to produce a more uniform density and stress distribution in pressed compacts. Although none of the techniques developed has led to large-scale industrial use, they are discussed here. Die Barrel Rotation. Rotating the die barrel while the powder is being pressed using a fine atomized aluminum

powder was reported by Hammond and Schwartz (Ref 3). They lubricated the die barrel with a suspension of lithium stearate in didecyl alcohol. Annular compacts, 12.7 mm ( in.) high, were pressed with outer diameters of 38 mm (1 in.) and inner diameters of 25 mm (1 in.). The core rod was stationary, but the die barrel could be rotated during compression. Compaction with a stationary and a rotating die barrel was compared. While in static compression, 20% of the applied stress was consumed in die wall friction, while only 2% of the stress was consumed when the die barrel was-rotated. In addition, the pressure necessary to eject the compact from the die was reduced to approximately one half of the pressure for compacts pressed with a stationary die barrel. Similar experiments on the effect of die barrel rotation on the density of iron powder compacts were reported by Rutkowski et al. (Ref 4). Triaxial compression by simultaneous isostatic and uniaxial compression is obtained by applying pressure

to the circumference of a cylindrical specimen confined in a flexible envelope while an axial load is superimposed by a vertical piston. With this method, the level of pressure necessary to obtain a given density is less than with isostatic or uniaxial compression alone. For example, to compact an atomized iron powder (Ancor 1000) to a relative density of 85%, uniaxial compression at 540 MPa (78 ksi) or isostatic compression at 415 MPa (60 ksi) is necessary. The same density can be obtained by combining a confining (isostatic) pressure of 83 MPa (12 ksi) with a uniaxial pressure of 470 MPa (68 ksi). The principles involved in this method of compaction have been reviewed by Broese van Groenou (Ref 5).

High-Energy-Rate Compacting. The rate at which pressure is applied in compacting in a hydraulic press is slow.

Compacting in certain mechanical presses is somewhat faster. The effects of the rate of pressure application in compacting were studied by Davies and Elwakil (Ref 6). They found that somewhat higher densities can be obtained for a given pressure and for a given energy input when compacts from iron powder are pressed in high-speed presses (petroforge-presses). They also determined the effects of multiple blows during pressing. In the fabrication of sheet metal products, techniques were developed that formed sheet metal at rates considerably higher than those obtained in fast-acting mechanical presses. This is the high-energy-rate forming technique, which generally uses explosives (see also Ref 1). The success of high-energy-rate forming in fabricating sheet metal led to extensive experimental work on high-energy compacting of metal-powders. The most common means to achieve high velocity is by the use of explosives. In one experiment, for example, compaction was done in a rigid die with pressure applied by a projectile propelled by an explosive charge that moves through a barrel (Ref 7). In a similar experiment compressed gas actuated the projectile (Ref 8). This experiment showed that the density of copper powder is not so much a function of projectile velocity, but depends primarily on the kinetic energy of the projectile, which can be varied by proper selection of gas pressure and projectile mass (Ref 8). For copper compacts with a volume of 0.86 cm3 (0.052 in.3), relative densities of 95% were obtained with energies of 150 J (110 ft · lb). The most widely used method of explosive compacting is shown schematically in Fig. 7. The powder to be compacted is placed in a steel tube, which is closed at each end by steel plugs. The steel tube is surrounded by an explosive that is set off by a detonator located so that on detonation the tube collapses uniformly inward. Experiments by Lennon et al. (Ref 9) showed that density is a function of energy. They developed the equation:

Dc = DT -

D

where Dc is the compacted density, DT is the full density of the material, D is the difference between full and compacted density, and are constants, and E is the net energy absorbed by the powder. The highest relative densities obtained for iron, nickel, copper, and aluminum powders and the corresponding net energies absorbed per unit volume of compact in their experiments were:

Powder Iron Nickel Copper Aluminum

(a)

Density(a), % 98.1 98.1 98.5 99.0

Net energy, J/cm3 261 556 285 182

Percent of theoretical density

This method of explosive compacting is not necessarily confined to cylindrical compacts. Cones and hollow cylinders have also been explosively compacted by this technique (Ref 10).

Fig. 7 Explosive compacting with powder contained in a steel tube

Vibratory Compacting of Powders. Vibration can be very effective in obtaining higher packed densities in powders. The relative densities of powders vibrated under carefully controlled conditions are much higher than those obtained by simply pouring the powder into the container. Therefore, much lower compaction pressures are required to reach a given density for a vibrated powder than for a poured powder. This is illustrated in Fig. 8 for a carbonyl iron powder. The density of 5.53 g/cm3 (71% relative density), reached by compacting under a pressure of 245 MPa (35 ksi), is due to the plastic deformation of the iron powder particles, while the 5.37 g/cm3 (69% relative density) obtained by vibrating at 167 oscillations per second is due mainly to vibratory packing. Plastic deformation during the simultaneous compacting at 2.4 MPa (0.36 ksi) is minimal. The method of consolidating powders by vibrating and simultaneous compacting is, therefore, primarily applicable to hard powders, such as refractory metal and cemented carbide powders, which can be densified relatively little by pressure application alone.

Fig. 8 Effect of powder vibration on densities of carbonyl iron compacts. Obtained in static pressing. • Vibratory compacting at a frequency of 233 oscillations per second. Vibratory compacting at a frequency of 167 oscillations per second

Melt-spray deposition of powders encompasses a wide variety of materials and product forms. It can be used to produce monolithic shapes by build up of the spray deposition, or it can be used to form coatings by deposition of a thin layer. One of the largest applications for melt-spray deposition is welding, including hardfacing and plasma spraying techniques for coatings. Spray forming is also a method for producing preforms by a buildup of sprayed metal powder. These preforms and billets subsequently can be consolidated into various mill shapes. Of these processes, the Osprey process, developed in Wales by Osprey Metals Ltd., and the controlled spray deposition process are in commercial use. Several other methods are being developed; plasma spray buildup has high commercial potential. Laser techniques, such as laser glazing, also have commercial potential, especially when combined with rapid solidification technologies. Osprey Process. Facilities for production of preforms made by the Osprey process consist of induction melting

equipment and a preform production unit. In the Osprey process, an alloy is melted and subjected to gas atomization under inert conditions (usually nitrogen or argon is used). The atomized droplets are collected in a mold or group of molds, in which final solidification occurs. Molds are normally copper cooled by water. High-temperature ceramics offer

other material options for molds. During solidification, welding of particles causes buildup of alloy in the mold. A preform having a density >96%, and normally >99%, of theoretical is generated by this buildup of alloy. The preform then can be consolidated to full density and formed into a mill or near-net shape part. Alloys that have been processed by the Osprey process include stainless steels, high-speed steels, and nickel-base superalloys, although many materials appear to be compatible with the process. Alloy development has centered on highalloy ferrous metals, Stellite alloys, superalloys, and composite materials. Because an inert atmosphere is maintained during spraying, oxygen levels similar to conventional ingot metallurgy products are attained, typically 20 to 40 ppm for superalloys. The high preform density ensures that no interconnected porosity is present in the preform, preventing internal oxidation during transfer of the material to subsequent consolidation and forming operations. The Osprey process is used to produce a wide variety of preform shapes and sizes. Typical preform shapes are tubes, rings, cylinders, disks, or simple billets. Size is dictated by economics, with the melt facility, atomizer, and inert chamber sized for a particular product line. The largest preform size produced weighed 540 kg (1200 lb). Typical deposition rates range from 10 to 90 kg/min (20 to 200 lb/min). More information on the process is described in the article "Spray Forming" in this Volume. The controlled spray deposition process is similar to the Osprey process in principle, but uses different machinery. Controlled spray deposition uses centrifugal atomization, while the Osprey process involves gas atomization. This process is used for production of mill shapes from high-alloy steel, which utilizes the enhanced workability of P/M workpieces. Highly alloyed metals may suffer from macrosegregation, which reduces the material workability. Elimination of segregation on a macroscale, coupled with a uniform distribution of fine carbides (2 to 3 m range for M2 high-speed steel), results in improved workability for P/M workpieces produced by spray deposition. These billets are processed subsequently into mill shapes and sheet products.

By atomizing liquid metal into droplets 0.5 to 1.5 mm (0.02 to 0.06 in.) in diameter, solidification rates three or more orders of magnitude higher than those of conventional ingot solidification are achieved. Impacting liquid droplets of metal onto a cooled substrate increases the solidification rate, resulting in solidification rates of 10,000 to 1,000,000 °C/s (18,000 to 1,800,000 °F/s). Controlled spray deposition relies on this type of splat solidification to build up a solidified deposit that becomes a workpiece for subsequent deformation processing. As the thickness and temperature of the builtup material increases, the solidification rate decreases, but it remains much higher than that of conventional ingot solidification. Heating prior to hot working can remove any microstructural variations that may exist throughout the thickness of the built-up deposit. Along with the metallurgical benefits of controlled spray deposition, economic advantages of direct spraying of powder into preform shapes are attained by eliminating sieving, blending, and other powder preparation steps. Also, primary compaction of powder into a green shape is eliminated. Controlled spray deposition proponents claim that these reductions in equipment needs and processing steps allow more efficient utilization of energy, compared to conventional pressing and sintering P/M technology.

References cited in this section

1. R. German, Powder Metallurgy of Iron and Steel, John Wiley & Sons, 1998 3. L.F. Hammond and E.G. Schwartz, The Effect of Die Rotation on the Compaction of Metal Powders, Int. J. Powder Metall., Vol 6 (No. 1), 1970, p 25-36 4. W. Rutkowski, D. Bialo, and J. Duszczyk, Problems of Increase of Homogeneity and Density of Powder Metal Compacts Produced under Conditions of Pressing Process Assisted by Additional Movements of the Die, Planseeberichte für Pulvermetallurgie, Vol 28, 1980, p 194-203 5. A. Broese van Groenou, Pressing of Ceramic Powders, A Review of Recent Work, Powder Metall. Int., Vol 10, 1978, p 206-211 6. R.I. Davies and S. Elwakil, "Comparison of Slow Speed, High Speed and Multiple Compaction in Ferrous Powders," Proc. of 17th Machine Tool Design and Research Conf. sponsored by University of Birmingham, Vol 3, 1976, p 483-488 7. J.W. Hagemeyer and J.A. Regalbuto, Dynamic Compaction of Metal Powders with a High Velocity Impact Device, Int. J. Powder Metall., Vol 4 (No. 3), 1968, p 19-25

8. R.M. Rusnak, Energy Relationship in High Velocity Compacting of Copper Powder, Int. J. Powder Metall., Vol 12 (No. 2), 1976, p 91-99 9. C.R.A. Lennon, A.K. Bhala, and J.D. Williams, Explosive Compacting of Metal Powders, Powder Metall., Vol 21, 1978, p 29-34 10. S.W. Porembka, Explosive Compacting, Ceramic Age, Dec 1963 Powder Shaping and Consolidation Technologies B. Lynn Ferguson, Deformation Control Technology, Inc.; Randall M. German, The Pennsylvania State University

Selecting a Process Selecting the proper process for consolidating powder to produce a part requires making many decisions. Using the performance requirements as the accept-reject criteria, the wide range of possible processes can be narrowed down, but usually the designer/manufacturer still finds that several methods are possible for producing a particular part. The final selection is then usually based on the availability of equipment, the experience with particular processes, and the marketplace requirements of cost, delivery, and quantity. Some guidelines and corollaries to these guidelines are presented here as an example of process selection decision making. These are not hard-and-fast rules for several reasons. First, the craft or skill aspect of parts making will alter these guidelines for each and every company. Second, new methods, new powders, and new equipment constantly require that any such set of guidelines be adjusted. Third, the marketplace itself dictates changes and additions to such guidelines. The reader is encouraged to take these guidelines and corollaries as a starting point, then modify and add new statements with quantified details wherever possible so that the decision making process within each organization is captured. Guideline 1: Control of porosity is the most important aspect of a consolidation process. The size,

shape, distribution, and volume fraction of pores is the singlemost important property of a powder metallurgy part. The performance of the part is directly dependent on these aspects of porosity. Therefore, the partsmaker must control porosity in order to achieve a usable part, and the consolidation process is critical to controlling part porosity. Corollary: Dynamic properties such as toughness and fatigue resistance improve dramatically as porosity is eliminated. Higher part stresses during application require higher densities. If a powder metallurgy part

must compete with a wrought part, full density must be achieved in all critically stressed regions. This is especially true for applications that require high toughness or where cyclic loading is a life limiting condition. Guideline 2a: Control of dimensions is the second most important aspect of the consolidation process. The successful implementation of parts production using powder metallurgy techniques is due to the

minimization or elimination of machining operations. Therefore, the consolidation process must offer substantial control of the final part dimensions so that secondary operations are minimal. Corollary: The use of rigid tooling provides superior control of dimensions in comparison to flexible tooling. Therefore, die pressing, injection molding, and powder forging are preferred over cold and hot isostatic pressing

in terms of dimensional control. The shape of specific sections of isostatically pressed components can be better controlled by the selective use of rigid fixtures such as mandrels or pressing plates. Corollary: The effect of friction on porosity distribution must be taken into account in order to maintain control of dimensions. Porosity gradients due to friction can cause nonuniform dimensional change during

sintering. This is especially critical in die compaction where friction along vertical die walls and punch faces must be minimized through the use of admixed lubricants or die wall lubricants. Corollary: Uniform powder loading is required to achieve a uniform dimensional change during sintering. Because the sintering rate is density dependent, a nonuniform distribution of powder will produce a

nonuniform dimensional change during sintering. For injection molding, this means that the volume fraction of binder should be uniform throughout the green part. For multilevel die pressed compacts, this means that multiple pressing

motions should be used so that the proper powder distribution is achieved before consolidation starts. In die pressing there is negligible transfer of powder between cavity sections while pressure is applied. Corollary: Prior to sintering but after debinding, injection molded parts must be supported because particle bonding at this point is simple cohesion. After ejection from the mold, an injection molded shape is

held together by the binder. Prior to sintering, the binder is removed, leaving a time period when the only force holding the powder mass together is particle cohesion plus residual binder. A supporting bed of an unreactive matter such as alumina spheres may be necessary. Guideline 2b: Ignore Guideline 2a if powder metallurgy is the only method that can be used to produce a particular material. In some cases powder metallurgy offers a way to process a unique material. In these

cases, the uniqueness is the material often outweighs the economics of dimensional control so that substantial machining or secondary processing is acceptable. Guideline 3: Metallurgical bonding of particles must be achieved during the consolidation process.

After porosity, the next most important characteristic for achieving a usable part is metallurgical bonding or welding of the particles. Sintering is the most common process for accomplishing this bonding. The alternative to sintering is shear deformation of particles while under a general state of compression. In processes such as forging and extrusion, the particle deformation and local pressures are sufficiently high that strong welds develop between particles. Corollary: Spherical particles should not be cold pressed if they are to serve a structural function.

Cold pressing of spherical powders does not result in significant mechanical bonding of particles and green strength is poor. With careful handling, cold pressing of spherical powder can be accomplished, followed by sintering to achieve particle bonding; P/M filters are an example. Corollary: The particle surface quality directly effects the strength of the particle bond. Sintering is not

effective in improving mechanical properties if particle surface contaminants such as carbides or oxides block diffusion of alloy species. Therefore, sintering requires clean surfaces or it must include a surface cleaning step such as deoxidation. Corollary: The mode of densification affects the quality of the interparticle bonding. The strength of the

bond between powder particles is improved by shear deformation in comparison to repressing or isostatic compaction. Shear plus pressure breaks up surface contaminants and promotes particle bonding. As surface cleanliness is improved toward being contaminant free, there is less difference between consolidation processes involving high or low amounts of shear. Powder Shaping and Consolidation Technologies B. Lynn Ferguson, Deformation Control Technology, Inc.; Randall M. German, The Pennsylvania State University

References 1. R. German, Powder Metallurgy of Iron and Steel, John Wiley & Sons, 1998 2. R. German and A. Bose, Injection Molding of Metals and Ceramics, Metal Powder Industries Federation, 1997 3. L.F. Hammond and E.G. Schwartz, The Effect of Die Rotation on the Compaction of Metal Powders, Int. J. Powder Metall., Vol 6 (No. 1), 1970, p 25-36 4. W. Rutkowski, D. Bialo, and J. Duszczyk, Problems of Increase of Homogeneity and Density of Powder Metal Compacts Produced under Conditions of Pressing Process Assisted by Additional Movements of the Die, Planseeberichte für Pulvermetallurgie, Vol 28, 1980, p 194-203 5. A. Broese van Groenou, Pressing of Ceramic Powders, A Review of Recent Work, Powder Metall. Int., Vol 10, 1978, p 206-211 6. R.I. Davies and S. Elwakil, "Comparison of Slow Speed, High Speed and Multiple Compaction in Ferrous Powders," Proc. of 17th Machine Tool Design and Research Conf. sponsored by University of Birmingham, Vol 3, 1976, p 483-488

7. J.W. Hagemeyer and J.A. Regalbuto, Dynamic Compaction of Metal Powders with a High Velocity Impact Device, Int. J. Powder Metall., Vol 4 (No. 3), 1968, p 19-25 8. R.M. Rusnak, Energy Relationship in High Velocity Compacting of Copper Powder, Int. J. Powder Metall., Vol 12 (No. 2), 1976, p 91-99 9. C.R.A. Lennon, A.K. Bhala, and J.D. Williams, Explosive Compacting of Metal Powders, Powder Metall., Vol 21, 1978, p 29-34 10. S.W. Porembka, Explosive Compacting, Ceramic Age, Dec 1963

Powder Treatments and Lubrication Erhard Klar, Consultant, and C.B. Thompson, OMG Americas

Introduction MOST POWDERS receive at least one treatment prior to compaction. These treatments are tailored to the use of the powder and may include (a) particle size distribution adjustment through screening and/or air classifying, (b) annealing for the purpose of improving compacting properties, (c) lubricant addition for compacting grade powders, (d) mixing of different powders for premixes, and (e) blending of powders and powder mixes to homogenize their various components. These treatments are usually performed by the powder producer. The quality of these treatments can greatly affect the uniformity and consistency of sintered part properties. With increasing emphasis on zero defect manufacture, the treatments have received more attention in recent years.

Acknowledgement Portions of this article are based on C.B. Thompson's article "Lubrications of Metal Powders" from Powder Metallurgy, Vol 7, ASM Handbook, 1984. Powder Treatments and Lubrication Erhard Klar, Consultant, and C.B. Thompson, OMG Americas

Classifying/Screening Many powder production processes yield relatively broad particle size distribution, and because of the many manufacturing variables, the distributions and average particle sizes may exhibit marked lot-to-lot variations, which contribute to the variation of important powder and sintered part properties. Classifying and screening are used to render the particle size distributions of powders more uniform with well-defined upper particle size limits. Powder producers often manufacture series of powders that differ only in particle size distribution. Such powders exhibit "graded" differences in dimensional change during sintering. For filter manufacture, classifying and screening are used to generate narrowly sized powder fractions for controlled pore size in filters and flow restrictors.

Powder Treatments and Lubrication Erhard Klar, Consultant, and C.B. Thompson, OMG Americas

Mixtures and Segregation Most P/M powders are multicomponent systems and, therefore, are subject to segregation. Segregation is even possible in a one-component metal powder if, for instance, coarse and fine powder particles "demix" as a result of vibration. The opportunity for powder segregation exists in processes such as shipping and the filling of hoppers and compaction dies, where individual components exhibit different flow rates due to differences in particle size, shape, density, surface roughness, and other properties. The most widely used P/M compacting grade powders are mixtures of iron with graphite, copper, nickel, and/or molybdenum. The preferred use of powder mixtures rather than prealloyed powders is related to several factors. The most important one is that elemental powder mixtures generally possess significantly higher compressibilities than their prealloyed counterparts. Secondly, powder mixtures are usually less expensive than prealloyed powders. Finally, powder mixtures more often can be formulated to provide transient liquid phases during sintering, which can reduce sintering times and improve mechanical properties. In most powder mixtures, the base powder typically comprises 90% or more of the powder mixture. Since the other components are present in only small amounts, and their alloying effects are powerful, it is very important that they are uniformly distributed; homogeneity of the powder must be preserved during shipping and handling at the parts producer's site until compaction is completed. Without such precautions, the properties of the sintered parts are not optimal and the standard deviation, that is, the scatter of the properties, can be quite large. Powder Treatments and Lubrication Erhard Klar, Consultant, and C.B. Thompson, OMG Americas

Stabilization of Powder Mixtures Because of the wide use of powder mixtures and their sensitivity to demixing, much work has been done towards the stabilization of powder mixtures. Stabilizers. A powder mixture is optimally mixed if its components approach a random (statistical) distribution free of

agglomeration (Fig. 1). The quality of mixing can be measured by the number of particle contacts (in green compacts) between identical or different powder components as illustrated in Fig. 2 for iron-copper mixtures, or by the chemical analysis of appropriate samples taken from the powder mixture.

Fig. 1 Schematic representation of particle patterns in a powder mixture. (a) Ordered. (b) Agglomerated. (c) Statistical (random) distribution. (d) Demixed or segregated

Fig. 2 Effect of stabilizer on iron-iron contact formation in binary iron-copper system. Lower curve represents theoretical random mixture.

Blenders and mixers that rely mainly on gravity (tumblers) are suitable for powders that mix readily. More intense

mixing is accomplished with low-shear agitated-type blenders that use ribbons, slow-speed paddles, screw-type augers, or other means of motion. Figure 3 illustrates how spherical powders mix quite readily, but also are subject to demixing or overblending. Consequently, mixing should be stopped once a near-random distribution has been achieved. The variability coefficient in Fig. 3 represents the standard deviation of the measured degree of mixing divided by the average value of the measured property. The quality of the mixture improves with decreasing variability coefficient.

Fig. 3 Effect of particle size and shape of components of 90%Fe-10%Cu mixtures on degree of blending. Quality of blending improves as variability coefficient decreases. Particle size and shape for components: (a) Cu, 200 to 300 m; Fe, 0. Then, differentiating (Eq 7) gives:

(Eq 19)

Combining Eq 18 and 19 gives

(Eq 20)

The individual strain increments can be obtained by combining Eq 20 with Eq 12 and 14. In the above model, it is assumed that the material remains isotropic as it deforms plastically, so that the yield surface maintains the same shape and simply changes in size as plastic strain is accumulated. In practice, the material response can become anisotropic. One means of representing anisotropy is through the use of a kinematic hardening model, wherein the center of the yield surface in stress space is allowed to translate according to some rule. See Ref 12 for further information. As a different approach towards anisotropy, Hill (Ref 13) presents a generalized theory of plasticity that allows for a change in the shape of the yield surface, while the basic requirements of convexity and incompressibility are maintained. There are many other types of constitutive laws that have been developed to model the response of plastically deforming materials. As in the models described here, these alternatives mainly consist of identifying a yield surface and developing appropriate laws for the expansion, translation, and change in shape of the surface as the material deforms plastically. These laws can contain many state variables, together with complex evolution laws. The choice of constitutive law depends on the type of loading histories being evaluated. If the stress history experienced by the material point is simple, then relatively simple constitutive laws adequately describe the material response. For example, for monotonic proportional loading, the kinematic and isotropic hardening models predict the same response provided they are fit to the same base data. A detailed knowledge of the shape and orientation of the entire yield surface and how these evolve with time is only required if complex loading histories are being considered. It should also be remembered that the more complex the constitutive law, the wider the range of tests that need to be devised to determine all the material parameters and the evolution equations for each of the state variables. When selecting the structure of a constitutive law, it is therefore important to first determine the type of stress history that each material point is likely to experience and to select the equations capable of capturing the major feature of the response. In the following section, assume that the individual particles of a compact deform according to the constitutive models described in this section. This understanding provides a structure for the development of the constitutive laws for the compact. The Structure of Constitutive Laws for Powder Material Consider an element of a powder compact (Fig. 7), subjected to a macroscopic stress state Eij. The strain can be decomposed into macroscopic elastic and plastic components, as before:

Eij =

+

ij,

and experiencing strains

(Eq 21)

Assume that the elastic response is isotropic. The elastic constitutive law can then be expressed in the form of Eq 5:

(Eq 22)

where G and K are functions of the state of the material. Several techniques for determining the elastic moduli G and K are summarized in Table 1. It is sometimes more practical to measure Young's modulus E instead of shear modulus G. Shear modulus can be determined using G = 3KE/9(K - E).

Table 1 Experimental techniques for the determination of elastic properties of powder compacts Experimental technique Triaxial test unloading Resonant frequency

Elastic property E(D)

Ultrasound

E(D)

Hydrostatic test unloading

K(D)

E(D)

Comments

Reference

Can determine E at low density (D is 0.65 for atomized steel powders). Experimental apparatus is involved. Very accurate and repeatable. Simple experimental setup. Simple sample preparation. Commercial laboratories available that use this technique. Minimum density D is 0.75 for atomized steel powders. Higher scatter than resonant frequency. Able to determine anisotropic properties in a transverse rupture bar. Can determine K at low density (D is 0.65 for atomized steel powders). Experimentally difficult at very high density.

Ref 14

Fig. 7 Macroscopic element of powder compact subjected to stresses

Ref 15

Ref 15 Ref 14

ij

The variables E and K can be measured during the unloading of a triaxial compaction test and a hydrostatic compaction test, respectively. Another technique used to determine E is resonant frequency. The powder is compacted into a beam, typically a transverse rupture bar. The beam is excited through an impact load, and the natural vibration frequency is measured. This frequency can be used to compute E. Ultrasound has also been applied to the determination of E in compacted specimens. Watson and Wert (Ref 9) measured the variation of G and K as a function of D for samples which had been compacted in closed die. They found that:

G = Gm exp(8.15(D - 1))

(Eq 23)

=

m

(Eq 24)

exp(5(D - 1))

with

(Eq 25) and Gm and

m

representing the shear modulus and Lamé constant for the fully dense material, respectively.

If a microscopic element deforms plastically according to any of the models described, and there are no other dissipative mechanisms operating, then it is possible to identify a yield surface in stress space:

F = F(

ij,S

(Eq 26)

)=0

which is a function of stress and the state of the material, described in terms of a number of state variables, S . Plastic flow can only occur if the stress state is on the yield surface. It can further be shown that associated flow at the microlevel guarantees associated flow at the macroscopic level, thus:

(Eq 27)

where

is the macroscopic plastic multiplier.

If frictional sliding contributed to the internal dissipative processes, then it is still possible to identify a yield surface, but associated plastic flow cannot be guaranteed. Fleck (Ref 16) has examined the two extremes of frictional behavior: free sliding; and sticking, whereby sliding can only occur by shear yielding in the vicinity of a contact. In each case, there is no frictional dissipation and associated flow can be guaranteed. The yield surface for intermediate frictional conditions should lie between the yield surfaces for these two situations. Fleck (Ref 16) found that there is only a small difference between the surfaces for a given assumed state, indicating that only a small amount of energy is dissipated by shearing between the particles. Thus, even if frictional sliding occurs it will only have a small influence on the macroscopic response, and it is appropriate to assume that an associated flow rule is valid. During compaction the stress point remains on the yield surface, therefore, following Eq 19:

(Eq 28) In order to complete the model, evolution laws for the state variables in terms of the stress, strain, and the state of the material are needed. In the models, the state could be described in terms of suitable measures of strain, or alternatively, plastic strains can be considered the state variables. Then Eq 28 becomes:

(Eq 29)

Combining Eq 29 with Eq 27 gives

(Eq 30)

with the individual strain components given by (Eq 27).

To simplify constitutive models, isotropic behavior is often assumed. This consideration dictates the use of mechanical and material descriptions independent of a particular choice of coordinate system. This freedom allows one to use rotation invariant or coordinate system independent quantities in the constitutive model. For the mechanical stress or strain, which are mathematically represented as tensors of the second order, these quantities are the first and second tensor invariants. The yield function can then be expressed as a function of

e

=

, the macroscopic von Mises effective

stress in the powder aggregate, a form of the second invariant of the macroscopic stress and m = kk, the mean stress in the powder aggregate, a form of the first invariant of the macroscopic stress. Equation 26 then assumes the form: F = F( e, m, S ) = 0. If appropriate forms for the yield function in terms of the stress and accumulated plastic strain can be found, the above equations can be employed to determine the full constitutive response. In the following sections, different possible forms of the yield function are examined. Yield Functions from Micromechanical Models. The section "Deformation of Powder Compacts: Experimental Observations" states the nature of the porosity changes as a material densifies. Initially, the porosity is open and distinct necks exist between the contacting particles (stage 1). At high relative densities (D > 0.9), the porosity is closed, which is referred to as stage 2. Different forms of micromechanical models have been developed for these two stages. In the stage 1 models, it is assumed that there is no interaction between the deformation zones that form in the vicinity of the different contacts. Analyses of the contact of two isolated bodies can then be used to obtain the appropriate contact law for a given microplasticity model. The macroscopic response is then determined by combining the contributions from each contact. Fleck et al. (Ref 16) employed a perfectly plastic microplasticity model and assumed that the material was isotropic and could be described using a single state variable, which they took to be the relative density, D. This assumption is equivalent to assuming that the material had previously been compacted isostatically to a given density. From their analysis, Fleck et al. (Ref 16) propose an approximate yield function:

(Eq 31)

where Py is the yield strength in hydrostatic compression;

(Eq 32)

where D0 is the initial dense random packing density, which is generally taken to be 0.64. The full response can be obtained from Eq 26 to Eq 30 by noting that the densification rate is given by (neglecting elastic volume changes):

(Eq 33)

=-

In practice, the matrix material can strain harden as the contact zones deform plastically. Fleck et al. (Ref 16) demonstrate how the effects of strain hardening can be taken into account. A more compact form of model can be obtained by rewriting Eq 31 in the form of Eq 7, that is,

F=

-

y

=0

(Eq 34)

where is an effective stress for the porous material, which is a function of the macroscopic stress and internal geometry described by D. Rearranging Eq 31 then gives:

(Eq 35)

For loading conditions on the yield surface, Eq 29 becomes:

(Eq 36)

Fleck et al. (Ref 16) take

y

to be the average yield strength in the plastically deforming material, and they define an

average internal effective strain increment, terminology,

=

, by equating the internal and external work. Using the current

=

y

= G(D)

y

(Eq 37)

where G(D) is the effective volume fraction of plastically deforming material, given by Fleck (Ref 17) as:

(Eq 38)

Combining Eq 36, 37, 33, and 18 gives:

(Eq 39)

where d is a macroscopic effective plastic strain increment not to be confused with the von Mises effective strain increment, dEe. The advantage of this form is that it separates out the effects of geometric and material hardening and it leads to a more compact expression than the relationships proposed by Fleck et al. (Ref 16). In the development of the above model, two major assumptions were made: the strengths of the contact zones were assumed to be the same in tension and compression, and the material was assumed to be isotropic. It was demonstrated in the section "Deformation of Powder Compacts: Experimental Observations" that the strengths of the contacts are likely to be different in compression and tension, particularly for smooth particles. Also, when a compact is loaded along stress paths other than hydrostatic, the microstructure (the distribution of necks) becomes anisotropic. Fleck (Ref 17) has proposed a model in which the tensile strength of a contact patch is less than the compressive strength by a factor . If Py (Eq 33) is the magnitude of the pressure required to plastically deform a compact, then the magnitude of the hydrostatic tensile stress required to initiate plastic flow is Py. Fitting a quadratic to Fleck's results and adopting the form of expression presented in Eq 34, the following equation is obtained:

(Eq 40)

This expression reduces to Eq 35 when = 1. When = 0, it becomes:

(Eq 41)

The yield surfaces are plotted in Fig. 8 for = 0 and 1, where they are compared with the surfaces obtained by Fleck (Ref 17). It should be noted that these surfaces have vertices where they meet the m axis. The direction of the strain increment vector is nonunique for pure hydrostatic stress states, which can lead to computational difficulties when implementing this model. In practice, this nonuniqueness has been bypassed by inserting a circular arc at the vertex to ensure a smooth continuous yield surface (Ref 19).

Fig. 8 Yield surfaces predicted by Fleck's (Ref 17) isotropic model for (a) the prediction of Eq 41 and 35

= 0 and (b)

= 1, compared with

Fleck (Ref 17) has also examined the influence of material anisotropy by relating the size and number of contacts at a given orientation to the plastic strain . The surface for frictionless closed-die compaction, normalized by the hydrostatic stress, P, required to compact the material to the same density is shown in Fig. 9 for = 0, where it is compared with the comparable surface for hydrostatic compaction. There are a number of important points to note:

•

The surface for closed-die compaction is extended in the direction of loading and contracted in the transverse direction compared to the surface for hydrostatic compaction. Thus, the yield behavior cannot be described in terms of the relative density alone. • The vertex on the yield surface is at the loading point. Similar results are obtained for crystal plasticity theory applied to fully dense materials. • The yield surfaces expand in a self-similar manner. Thus, the different components of stress increase in proportion to each other when the straining is proportional. • It is not possible to obtain a closed form solution for arbitrary loading paths. Thus, the full micromechanical model is required in order to compute the constitutive response.

The predictions of this model are compared with the experimental results of Akisanya et al. (Ref 10), shown in Fig. 10 for hydrostatic and closed die compaction. It is evident that the general form of these surfaces is well represented by this model.

Fig. 9 Yield surfaces for isostatic and closed-die compaction predicted by the anisotropic model of Fleck (Ref 17)

Fig. 10 Comparison of Fleck's (Ref 17) model with the experimental results of Akisanya et al. (Ref 10)

So far only the early stages of compaction, when distinct necks exist between the particles, have been discussed. In the latter stages of compaction (D > 0.9), the pores are isolated. Constitutive models employed in this regime have been taken from studies originally concerned with ductile failure. In these models, the pores are assumed to remain spherical. The most widely used model is developed by Gurson (Ref 20), which predicts a yield surface given by:

(Eq 42)

This expression suffers from the same problem as Eq 31, in that the effects of material and geometric hardening are coupled. Cocks (Ref 21) proposed that a more appropriate form for the constitutive model, which does not differ significantly from that proposed by Gurson (Ref 20), is given by Eq 34, with

(Eq 43)

where

(Eq 44)

(Eq 45)

Equations 36, 37, 38, and 39 for the constitutive response again apply, where, now, G(D) = D. In practice, the pores do not achieve a spherical shape during the latter stages of compaction. Instead, they maintain a cusp shaped profile. Liu et al. (Ref 22) and Akisanya et al. (Ref 23) have demonstrated that bodies with cusp-shaped pores have greater compressibility than bodies with spherical pores at the same density. Qian et al. (Ref 24) have suggested how the influence of the cusp-shaped pores can be incorporated into the constitutive models for creeping materials. They still assume that the body is isotropic macroscopically. Extending their analysis to a perfectly plastic material would simply result in different definitions of G1(D) and G2(D) in Eq 43. Under general loading situations, the expectation might be that the pores would become squashed in the main direction of compaction. Ponte Castenada (Ref 25) has examined the effect of oriented ellipsoidal-shaped pores on the constitutive response, but there have not been any studies of the influence of distorted cusp-shaped pores on the material behavior. In all the stage 2 models, it has always been assumed that the contact patches between the distorted particles are able to support tensile tractions and the contacting particles are unable to slide over each other. There has been no attempt to analyze the effect of the strength of these contacts on the constitutive response. The yield surfaces for stage 1 and stage 2 compaction can have different shapes. In order to ensure a smooth transition, a gradual change from one shape to another is assumed to occur over a given range of relative density. Assuming a linear variation from the yield expression for stage 1 to that for stage 2 over the range D1 to D2 gives:

(Eq 46)

where 1 is the effective stress for stage 1 compaction and 2 is the effective stress during stage 2. In practice, the transition is generally assumed to occur over the range D1 = 0.77 to D2 = 0.9. Empirical Yield Functions. In the empirical models of powder compaction, the different form of porosity in stage 1

and stage 2 is not considered. Simple continuous expressions for the yield function are developed that satisfy the condition that the material response is incompressible when D = 1. The most widely used expressions are those due to Kuhn and Downey (Ref 6) and Shima and Oyane (Ref 7). These models have the same general structure as those described in the section "Yield Functions from Micromechanical Models." The yield function can again be represented in the form of Eq 34. For simplicity, a quadratic function of the von Mises effective stress, e, and mean stress, m, is assumed for . It is further assumed that the material response can be described in terms of a single state variable, the relative density, D. can then be expressed in the form of Eq 43, where G1(D) and G2(D) are determined experimentally by assuming that the material yield strength y remains constant. Shima and Oyane (Ref 7) determined these quantities by compacting samples in a closed die to a given density, sintering the samples and then performing uniaxial tensile and compressive tests to determine the yield properties. By measuring the uniaxial yield strength and Poisson's ratio, G1(D) and G2(D) could be determined at the chosen density; and by repeating these tests at different densities, the complete functions G1(D) and G2(D) could be determined. Their results for copper and iron powders are well approximated by:

G1(D) = D5

(Eq 47) (Eq 48)

The above form has been employed by a number of authors in experimental and computational studies of powder compaction. Kim and Suh (Ref 26) recognized that a material could harden as a result of shear deformation as well as a result of densification. They proposed a modified form of expression for the yield function:

F=

-

(Eq 49)

=0

with

(Eq 50) Here

is the effective yield strength of the compact. The evolution law for

d

=

1

ed e

+

2

is assumed to be of the form:

md kk

(Eq 51)

which takes into account geometric and material hardening. Now, from Eq 28 and 51, find:

(Eq 52)

and the individual plastic strain increments are given by Eq 27. The yield function of Eq 49 was found to give a slightly better fit of the tension/torsion experiments of Kim et al. (Ref 5) on sintered iron compacts discussed earlier than the model of Shima and Oyane (Ref 7). These models were validated by performing experiments on sintered compacts. The effect of the sintering process is to anneal the particles and to bond them together. As a result, the yield strength is expected to be the same in tension and compression, as observed by Shima and Oyane (Ref 7), and it is not unreasonable for a quadratic function, such as those presented in Eq 43 and 50, to adequately describe the shape of the yield surface. A consequence of this choice of yield function is that it will always predict a negative volumetric strain-rate when the mean stress is negative (i.e., compaction will always occur under net compressive stress states). In experiments on compacted, but not sintered, powders,Brown and Abou-Chedid (Ref 27) and Watson and Wert (Ref 9) observed that the volumetric strains were positive in uniaxial compression. Based on their experimental observations, Watson and Wert (Ref 9) proposed the adoption of Drucker and Prager's (Ref 28) two part yield surface, which has been employed to model the response of soils. This surface is shown schematically in Fig. 11 and consists of a shear surface and a spherical cap. Watson and Wert (Ref 9) assumed associated flow for both surfaces (i.e., the strain-increment vector is normal to the surface). Thus, dilation occurs for stress states on the shear surface and compaction occurs if the loading point lies on the spherical cap.

Fig. 11 Schematic of Watson and Wert's (Ref 9) two surface Drucker-Prager model

Following Gurson and McCabe (Ref 29), Brown and Abou-Chedid (Ref 27) proposed a modified form of the Cam-Clay model (Ref 30, 31) to model their experiments. In this model the yield surface is represented by Eq 34, with

(Eq 53) Brown and Abou-Chedid (Ref 27) did not propose forms for the functions A(D) and B(D). More detailed experiments are required to determine the exact form of these expressions. The macroscopic effective stress of Eq 53 is similar to the quadratic forms of Eq 43 and 50, but now the elliptic surface is centered on the point (0, C) in e - m space. The quantity C represents the degree of cohesion between the particles. If they are perfectly bonded, C = 0, and the surface reduces to the symmetric expressions of Eq 43 and 50, which is appropriate for sintered materials. If the particles are smooth and the contacts are unable to support a normal tensile stress, then the value of C can be chosen so that the yield surface passes through the origin, as in the micromechanical model of Eq 41 for = 0. Then,

(Eq 54)

and Eq 53 becomes

(Eq 55)

Equation 55 is equivalent to the most common form of Cam-Clay model used to evaluate the response of soils (Ref 31). These last two models were derived from constitutive laws that have been developed for soils. A wide range of models have been proposed in the literature for granular materials, in which the yield surfaces are allowed to adopt different shapes, as well as being able to translate in stress space, in a similar manner to the kinematic models (Ref 31).

References cited in this section

5. K.T. Kim, J. Suh, and Y.S. Kwon, Plastic Yield of Cold Isostatically Pressed and Sintered Porous Iron under Tension and Torsion, Powder Metall., Vol 33, 1990, p 321-326 6. H.A. Kuhn and C.L. Downey, Material Behavior in Powder Preform Forging, J. Eng. Mater. Technol., 1990, p 41-46 7. S. Shima and M. Oyane, Plasticity Theory for Porous Metals, Int. J. Mech. Sci., Vol 18, 1976, p 285-291 8. S.B. Brown and G.A. Weber, A Constitutive Model for the Compaction of Metal Powders, Modern Developments in Powder Metallurgy, Vol 18-21, 1988, MPIF, p 465-476 9. T.J. Watson and J.A. Wert, On the Development of Constitutive Relations for Metallic Powders, Metall. Trans. A, Vol 24, 1993, p 2071-2081 10. A.R. Akisanya, A.C.F. Cocks, and N.A. Fleck, The Yield Behaviour of Metal Powders (1996), Int. J. Mech. Sci., Vol 39 (No. 12), 1997, p 1315-1324 11. S. Brown and G. Abou-Chedid, Yield Behaviour of Metal Powder Assemblages, J. Mech. Phys. Solids, Vol 42 (No. 3), 1994, p 383-399 12. W. Prager, Proc. Inst. Mech. Eng., Vol 169, 1955, p 41 13. R. Hill, The Mathematical Theory of Plasticity, Oxford University Press, 1950 14. E. Pavier and P. Doremus, Mechanical Behavior of a Lubricated Powder, Advances in Powder Metallurgy & Particulate Materials-1996, Vol 2 (Part 6), Metal Powder Industries Federation, 1996, p 27-40 15. C.J. Yu, R.J. Henry, T. Prucher, S. Parthasarathi, and J. Jo, Advances in Powder Metallurgy & Particulate Materials, Vol 6, Metal Powder Industries Federation, 1992, p 319-332 16. N.A. Fleck, L.T. Kuhn, and R.M. McMeeking, Yielding of Metal Powder Bonded by Isolated Contacts, J. Mech. Phys. Solids, Vol 40, 1992, p 1139-1162 17. N.A. Fleck, On the Cold Compaction of Powders, J. Mech Phys. Solids, Vol 43 (No. 9), 1995, p 1409-1431 19. R.M. Govindarajan and N. Aravas, Deformation Processing of Metal Powders, Part 1: Cold Isostatic Pressing, Int. J. Mech. Sci., Vol 36, 1994, p 343-357 20. A.L. Gurson, Continuum Theory of Ductile Rupture by Void Nucleation and Growth, Part 1: Yield Criteria and Flow Rules for Porous Ductile Media, J. Eng. Mater. Technol., Vol 99, 1977, p 2-15 21. A.C.F. Cocks, The Inelastic Deformation of Porous Materials, J. Mech. Phys. Solids, Vol 37 (No. 6), 1989, p 693-715 22. Y-M. Liu, H.N.G. Wadley, and J. Duva, Densification of Porous Materials by Power-Law Creep, Acta Metall. Mater., Vol 42, 1994, p 2247-2260 23. A.R. Akisanya, A.C.F. Cocks, and N.A. Fleck, Hydrostatic Compaction of Cylindrical Particles, J. Mech. Phys. Solids, Vol 42 (No. 7), 1994, p 1067-1085 24. Z. Qian, J.M. Duva, and H.N.G. Wadley, Pore Shape Effects during Consolidation Processing, Acta Metall. Mater., Vol 44, 1996, p 4815 25. P. Ponté Castañeda and M. Zaidman, Constitutive Models for Porous Materials with Evolving Microstructure, J. Mech. Phys. Solids, Vol 42, 1994, p 1459-1497 26. K.T. Kim and J. Suh, Elastic-Plastic Strain Hardening Response of Porous Metals, Int. J. Eng. Sci., Vol 27, 1989, p 767-778 27. S. Brown and G. Abou-Chedid, Appropriate Yield Functions for Powder Compacts (1992), Scr. Metall. Mater., Vol 28, 1993, p 11-16 28. D.C. Drucker and W. Prager, Q. Appl. Math., Vol 10, 1952, p 157-165 29. A.L. Gurson and T.J. McCabe, Experimental Determination of Yield Functions for Compaction of Blended Powders, Proc. MPIF/APMI World Cong., on Powder Metallurgy and Particulate Materials (San Francisco), Metal Powder Industries Federation, 1992 30. A. Schofield and C.P. Wroth, Critical State Soil Mechanics, McGraw-Hill, 1968 31. D.M. Wood, Soil Behavior and Critical State Soil Mechanics, Cambridge University Press, 1990

Mechanical Behavior of Metal Powders and Powder Compaction Modeling J.R.L. Trasorras and R. Parameswaran, Federal-Mogul, Dayton, Ohio; A.C.F. Cocks, Leicester University, Leicester, England

A Constitutive Model for Metallic Powders with Ductile Particles Several constitutive models have been applied to the numerical simulation of powder compaction in dies. The most widely used are variations of two types of models: models with an empirical quadratic yield function (Ref 32, 33, 34, 40, and 41) and Cap models (Ref 35, 36, 38, and 39). The main practical advantages of a model of the first type as developed in Ref 34, 36, 40, and 41 include: •

A reasonable representation of the behavior of metal powders with ductile particles under monotonic loading • A continuous yield function that facilitates its numerical implementation • Experimental calibration with a relatively small number of tests

Furthermore, this model has been implemented in the compaction modeling finite element code PCS (Ref 42) that is in use by several parts manufacturers in the United States. The overall strategy presented here can be employed for any of the models described in the section "Constitutive Models for Metal Powder Compaction." The only difference lies in the range of experiments that need to be performed in order to determine any unknown functions in the models. Model Formulation The model considered in detail here uses an empirical yield function and fits the general framework presented earlier with assumptions: •

As the powder aggregate is compacted, the particles deform plastically according to the behavior described by classical plasticity with isotropic hardening • The powder compact displays macroscopic elastic-plastic behavior that is isotropic and independent of strain rate • The total strain can be decomposed into elastic and plastic components • The contributions of particle sliding to the overall deformation are negligible • The state of the powder aggregate is represented by two state variables, the relative density, D, and the yield strength of the powder particle y

Rigorously, for elastic-plastic models, a slightly different definition of the relative density D is necessary when D is used as a state variable. Earlier D was defined as the ratio of the density of powder to the density of the fully dense material. However, the plastic state of the material should not be affected by the changes in D that are due to elastic deformation. A more appropriate definition of D is the ratio of the density of unloaded powder to the density of the fully dense material, the relaxed relative density of the powder aggregate. Following the general approach presented earlier, consider an element of a powder compact, which is subjected to a macroscopic stress state ij and experiences strains Eij. The strain can be decomposed into macroscopic elastic and plastic components as per Eq 21:

Eij =

+

(Eq 56)

Here assume that the elastic response is isotropic. The elastic constitutive law can then be expressed by Eq 22:

(Eq 57) where G and K are functions of D. An empirical yield function F of the general form of Eq 26, which is a function of the stress state and two state variables, is given by:

F = F( where P = -

kk

=-

m,

e,

P,

y,

D) = 0

(Eq 58)

is the pressure, a form of the first invariant of the macroscopic stress.

The specific form of the yield function is as defined in Eq 34 and 43:

(Eq 59)

As the powder aggregate densifies and reaches high relative density, its response will approach the incompressible plastic behavior of fully dense metals. Therefore, the function G1(D) is expected to increase monotonically with an increase in relative density and the function G2(D) is expected to decrease monotonically with an increase in relative density. The yield functions developed by Trasorras et al. (Ref 34) involved functions b(D) and c(D) that are expressible in terms of the functions G1(D) and G2(D) in the following manner: G1(D) = c(D) and G2(D) = 3c(D)/2b(D). In the following development, the yield function is cast in the form used in Ref 34.

(Eq 60) The functions b(D) and c(D) are to be determined through experiments. The associated flow at the macroscopic level (from Eq 27) yields:

(Eq 61)

From Eq 61, the volumetric and deviatoric components of the plastic strain are given by:

(Eq 62) (Eq 63) Eliminating d

allows

(Eq 64)

The conservation of mass is expressed in the form of the continuity equation (from Eq 33):

(Eq 65)

=

On integration, the continuity equation yields the following result for the evolution of relaxed relative density,

(Eq 66)

D= The isotropic hardening law for the powder particle (from Eq 18) is:

d

y

(Eq 67)

=

The particle hardening h may be constant (isotropic linear hardening), or may be a function of equivalent plastic strain in the particle. If the energy dissipated by particle sliding is negligible, as assumed earlier, the external and internal plastic work of deformation can be equated (from Eq 37 with G(D) = D):

=

(Eq 68)

Application to Powder Blends In the previous sections, the constitutive models have assumed that the powder consists of an aggregate of metal particles of a single kind. In powder compaction in dies, the powder blend will contain lubricant, and in the case of ferrous alloys, often graphite and other alloying elements as well. The theories presented here are expected to be valid for powder blends, provided that the additions to the base metal powder occupy a small volume of the total aggregate (typically a few volume percent). That being the case, the models can be used with the constraints that: •

The relative density D be defined as the ratio of the total density of the powder to the pore-free density of the powder to enforce a plastically incompressible behavior when the material reaches the pore-free density, as opposed to the density of the fully dense metal particle. • The elastic properties, which are strongly dependent on the presence of nonmetallic constituents, be determined experimentally for the specific blend. • The constitutive functions b(D) and c(D) be experimentally determined for the specific blend.

References cited in this section

32. S. Shima, "A Study of Forming of Metal Powders and Porous Metals," Ph.D. thesis, Kyoto University, 1975 33. Y. Morimoto, T. Hayashi, and T. Takei, Mechanical Behavior of Powders in a Mold with Variable Cross Sections, Int. J. Powder Metall. Powder Technol., Vol 18 (No. 1), 1982, p 129-145 34. J.R.L. Trasorras, S. Armstrong, and T.J. McCabe, Modeling the Compaction of Steel Powder Parts, Advances in Powder Metallurgy & Particulate Materials-1994, Vol 7, American Powder Metallurgy Institute, 1994, p 33-50 35. J. Crawford and P. Lindskog, Constitutive Equations and Their Role in the Modeling of the Cold Pressing Process, Scand. J. Metall., Vol 12, 1983, p 271-281 36. J.R.L. Trasorras, T.M. Krauss, and B.L. Ferguson, Modeling of Powder Compaction Using the Finite

Element Method, Advances in Powder Metallurgy, Vol 1, T. Gasbarre and W.F. Jandeska, Ed., American Powder Metallurgy Institute, 1989, p 85-104 38. H. Chtourou, A. Gakwaya, and M. Guillot, Assessment of the Predictive Capabilities of the Cap Material Model for Simulating Powder Compaction Problems, Advances in Powder Metallurgy & Particulate Materials-1996, Vol 2 (Part 7), Metal Powder Industries Federation, 1996, p 245-255 39. D.T. Gethin, R.W. Lewis, and A.K. Ariffin, Modeling Compaction and Ejection Processes in the Generation of Green Powder Compacts, Net Shape Processing of Powder Materials, 1995 ASME Int. Mechanical Engineering Congress and Exposition, AMD-Vol 216, S. Krishnaswami, R.M. McMeeking, and J.R.L. Trasorras, Ed., The American Society of Mechanical Engineers, 1995, p 27-45 40. J.R.L. Trasorras, S. Krishnaswami, L.V. Godby, and S. Armstrong, Finite Element Modeling for the Design of Steel Powder Compaction, Advances in Powder Metallurgy & Particulate Materials-1995, Vol 1 (Part 3), Metal Powder Industries Federation, 1995, p 31-44 41. S. Krishnaswami and J.R.L. Trasorras, Modeling the Compaction of Metallic Powders with Ductile Particles,Simulation of Materials Processing: Theory, Methods and Application, Shen and Dawson, Ed., Balkema, Rotterdam, 1995, p 863-858 42. Powder Compaction Simulation Software (PCS Elite) User's Manual, Concurrent Technologies Corp., Johnstown, PA Mechanical Behavior of Metal Powders and Powder Compaction Modeling J.R.L. Trasorras and R. Parameswaran, Federal-Mogul, Dayton, Ohio; A.C.F. Cocks, Leicester University, Leicester, England

Experimental Determination of Powder Material Constitutive Properties and Functions In this section, the experimental procedures that can be applied to determine the constitutive parameters that appear in the model of the preceding section are presented. Techniques used to determine the elastic parameters and the form of the yield function have been described in the section "Deformation of Powder Compacts: Experimental Observations." Triaxial Tests A brief description of the triaxial test and apparatus is presented in this section. For a more detailed treatment of the test, stemming from a soil mechanics perspective, see Ref 31. The triaxial test is important as it provides a convenient means of examining a variety of situations involving different ratios of the deviatoric and hydrostatic stress measures. A schematic of the triaxial test cell is shown in Fig. 12. The powder is placed in a container with a flexible wall, typically made of an elastomer, and immersed in a fluid under a pressure, p. Take care not to confuse this notation for the magnitude of fluid pressure with the one used elsewhere for the hydrostatic component of stress (P). A ram at the top of the apparatus provides a means of increasing the axial load on the powder while maintaining fluid pressure. Knowledge of the cell pressure as well as the applied ram load allows for the calculation of the resulting axial stress in the test sample. In addition, appropriate instrumentation on the apparatus ensures measurement of axial and radial dimensional changes. These measurements can then be translated into axial and radial strains. With the added assumption of a uniform stress state in the specimen and isotropic behavior of the material, it is reasonable to conclude that the radial and tangential strains, as well as radial and tangential stresses, are equal. Macroscopically, the stress state of the powder can be stated as:

(Eq 69) r

= -p

(Eq 70)

Fig. 12 Triaxial compaction cell. Source: Ref 31

The quantities A and a are the areas of cross section of the specimen and ram, respectively. The strains are calculated as:

(Eq 71)

(Eq 72)

where r0 and l0 are, respectively, the initial radius and length of the compact, and r and l, the current values of these quantities. The deviatoric and hydrostatic components of stress reduce to:

(Eq 73) (Eq 74) In soil mechanics, triaxial tests are often performed in one of three ways, as depicted in Fig. 13. The first, the isostatic compaction test, is used to characterize soils under hydrostatic loading and to determine the elastic bulk modulus. In this test, the ram load is maintained insofar as to ensure a hydrostatic state of stress in the test specimen. The second test, the consolidated triaxial test, involves initial pressurization of the test cell followed by an increase in ram load, the cell pressure being maintained constant. The path of stress on the P- e plane has a slope of 3 as the axial load increases at constant containing pressure. This path can be verified by taking the derivatives of the expressions for P and e while keeping p constant, and then calculating the slope using the ratio of the derivatives. Finally, the overconsolidated tests are employed to help determine the shape of the yield surface on the P- e plane. The test consists of subjecting the specimen to a history of deformation under pure pressure, following this by unloading to an intermediate pressure, and subsequently increasing the ram load while maintaining cell pressure to determine a different point on the yield surface.

Fig. 13 Stress paths in isostatic compaction test, consolidated triaxial compaction test, and overconsolidated compaction test. Source: Ref 43

As presented in the section "Deformation of Powder Compacts: Experimental Observations," the size and shape of the yield surface are functions of the history of loading. Thus, if the strain history only involved hydrostatic loading, subsequent probing of the yield surface in the manner of an overconsolidated triaxial test will only identify a yield surface defined by this initial loading. A different picture will emerge if the yield surface is identified following a stress path, such as the one used in a consolidation test, followed by unloading and subsequent probing of the yield surface (Fig. 3). This is of particular importance when considering tests to be used for material parameter extraction; the best tests shall be those involving stress paths as close to the real situation as possible. In the case of compaction in dies, the compact is constrained radially. The triaxial test setup should then include control of cell pressure to ensure zero radial strain in the sample to simulate actual compaction conditions. Calibration of Material Parameters for an Iron Powder Blend The procedures of the previous section are applied below to the calibration of the two-state variable model introduced earlier. The data is for a powder blend comprising 99.5% by weight of Distalloy AE, 0.5% by weight of graphite, and 1% wax Hoechst micropulver. This last component is admixed as internal lubricant. Distalloy AE is a diffusion alloyed iron powder with composition 4 wt% Ni, 1.5 wt% Cu, and 0.5 wt% Mo. Particle sizes for this powder range from 20 to 180

m. The apparent density of the powder is 3.04 g/cm3; the pore free density of the material is 7.33 g/cm3. The data used are from experiments reported by Pavier and Doremus (Ref 14). Consider the yield function of the model introduced in the section "A Constitutive Model for Metallic Powders with Ductile Particles" :

To fully determine the yield function F the functions b(D) and c(D) must be identified. Figure 14 illustrates the yield surface for a lubricated atomized iron powder (Hoeganaes Ancorsteel 1000). The surface is plotted for different values of relative density and particle yield stress pairs. For this powder, the functions b(D) and c(D) were determined by Trasorras et al. (Ref 40) to be:

(Eq 75) c(D) = D6

(Eq 76)

The constitutive model has two-state variables and requires two different kinds of tests for its calibration.

Fig. 14 Yield surface for lubricated atomized iron powder (Hoeganaes Ancorsteel 1000). Surface plotted for different values of relative density and particle hardening

Triaxial Consolidation Test for b(D). The state of stress in the test specimen is assumed to be uniform throughout (homogeneous). This assumption leads to the result that the radial r and tangential z stresses are equal and given by the cell pressure. If the specimen is cylindrical in shape, the components of stress in the powder aggregate can then be appropriately defined using a cylindrical coordinate system as follows:

(Eq 77)

The negative signs are assigned to indicate the compressive nature of the stresses in the aggregate. The deviatoric and pressure components are then:

(Eq 78)

(Eq 79)

Calculating

e

=

find

=

( (

e

z

-

r)

2

+

(

r

-

z)

(Eq 80)

2

) (Eq 81)

=

The increment of plastic strain given by Eq 61 is:

(Eq 82)

The deviatoric and volumetric components of plastic strain increment become:

Eliminating

=3

[- (2

-

= -3

r

+ (

z)b(D)] z

-

r)

(Eq 83) (Eq 84)

obtain,

(Eq 85) Solving for b(D),

(Eq 86)

Thus, the function b(D) may be determined using data from the triaxial tests for values of axial stress strain increment

, and radial strain increment

z,

corresponding to a range of values of cell pressure

axial plastic r.

The results of applying Eq 86 to the experimental data of Pavier and Doremus (Ref 14) for the Distalloy AE blend are shown in Fig. 15. Because noise is inherently present in such information, a cursory smoothing using a moving average scheme was performed prior to computing the increments of strain. In addition, the elastic components of strain were

assumed to be negligible. For low relative densities, this assumption is reasonable. However, as plastic incompressibility is approached, the validity of the assumption must be checked. A curve fit for b(D) results in the following expression:

(Eq 87)

Fig. 15 Function b(D) for an iron powder blend comprising 99.5% by weight Distalloy AE, 0.5% by weight graphite, and 1% wax Hoechst micropulver

Isostatic Compaction Test for c(D). The state of stress in the isostatic compaction test is one of pure pressure, with no shear or deviatoric stress present. The isotropic test provides a convenient means of evaluating the evolution of relative density as a function of increasing cell pressure. Data are obtained by increasing cell pressure to a point before unloading the cell to determine the resulting powder relative density. Following Pavier and Doremus (Ref 14), the relation between the relative density for the powder under hydrostatic loading and the applied pressure or hydrostatic stress was determined to be:

(Eq 88) where Py is expressed in MPa and represents the pressure at yield given a value of relative density, D in the isostatic test. It is reasonable to assume that the material is continuously yielding with load. Further, since shear stress is completely absent, Eq 60 reduces to:

(Eq 89) If particle hardening is considered, the calculation of c(D) involves the simultaneous integration of partial differential equations for the evolution of particle yield stress and the functions b(D) and c(D). However, for simplicity, assume the powder particles are perfectly plastic and obtain c(D) from (Eq 89):

(Eq 90)

Using the forms for pressure and b(D) from Eq 88 and 87, and assuming a mean value of 375 MPa for particle yield stress, the function c(D) is calculated and plotted in Fig. 16.

Fig. 16 Functions b(D) and c(D) for a iron powder blend comprising 99.5% by weight Distalloy AE, 0.5% by weight graphite, and 1% wax Hoechst micropulver

Verification of Calibrated Model. This section evaluates the accuracy of the calibration performed for functions

b(D) and c(D) for the Distalloy AE blend considered using triaxial test data. Analysis data were generated through a finite element simulation of the triaxial tests conducted. Because hardening of the powder particles is a likely event, it is considered here. Specifically, the material of the powder particle was modeled as having a yield stress of 250 MPa with a hardening of 333 MPa. These values were obtained from data reported by Trasorras et al. (Ref 36). Figure 17 shows a comparison of the predicted powder density as it evolves during the triaxial tests against data measured by Pavier et al. (Ref 14). There is very good agreement between simulation and experiment. Figure 18 shows a plot of axial strain versus radial strain for a triaxial cell pressure of 250 MPa. The strains are measured relative to the onset of shear in the specimen, marked by an increase in axial stress over the cell pressure. Figure 19 shows a plot of axial stress versus axial strain for the same triaxial cell pressure of 250 MPa. Again, the simulations compare well with the experiments.

Fig. 17 Density evolution in a triaxial test: comparison of FEA results with experiment

Fig. 18 Axial versus radial strain in a triaxial test with 250 MPa cell pressure: comparison of FEA results with experiment

Fig. 19 Axial stress versus strain in a triaxial test with 250 MPa cell pressure: comparison of FEA results with experiment

References cited in this section

14. E. Pavier and P. Doremus, Mechanical Behavior of a Lubricated Powder, Advances in Powder Metallurgy & Particulate Materials-1996, Vol 2 (Part 6), Metal Powder Industries Federation, 1996, p 27-40 31. D.M. Wood, Soil Behavior and Critical State Soil Mechanics, Cambridge University Press, 1990 36. J.R.L. Trasorras, T.M. Krauss, and B.L. Ferguson, Modeling of Powder Compaction Using the Finite Element Method, Advances in Powder Metallurgy, Vol 1, T. Gasbarre and W.F. Jandeska, Ed., American Powder Metallurgy Institute, 1989, p 85-104 40. J.R.L. Trasorras, S. Krishnaswami, L.V. Godby, and S. Armstrong, Finite Element Modeling for the Design of Steel Powder Compaction, Advances in Powder Metallurgy & Particulate Materials-1995, Vol 1 (Part 3), Metal Powder Industries Federation, 1995, p 31-44 43. H-A. Haggblad, P. Doremus, and D. Bouvard, An International Research Program on the Mechanics of Metal Powder Forming, Advances in Powder Metallurgy & Particulate Materials-1996, Vol 2 (Part 7), Metal Powder Industries Federation, 1996, p 179-192

Mechanical Behavior of Metal Powders and Powder Compaction Modeling J.R.L. Trasorras and R. Parameswaran, Federal-Mogul, Dayton, Ohio; A.C.F. Cocks, Leicester University, Leicester, England

Numerical Modeling of Powder Compaction in Dies In this section, practical illustrations of finite element analysis of powder compaction are presented. The first one is the compaction of a two-level component with emphasis on the effect of compaction kinematics on punch loads. The second example, the compaction of a long bushing, illustrates density predictions. Compaction of a Two-Level Part The model presented in the section "A Constitutive Model for Metallic Powders with Ductile Particles" is used to examine the compaction of an axisymmetric two-level component. This part was studied experimentally by Kergadallan et al. (Ref 4). The part was pressed with five different tooling motions; two of those cases, labeled as part 30 and part 34, are presented here. The compaction experiments and the finite element model are described, and numerical predictions are compared to experimental results. Compaction Experiments. Following are the conditions of the compaction tests used to study an axisymmetric two-

level component. Part Geometry. The part is axisymmetric with a thin outer rim, a hub, and a bore. This geometry is representative of

many common powder metal parts (e.g., engine camshaft timing pulleys, one-way mechanical diode clutch plates). Geometry and dimensions are shown in Fig. 20. The outer diameter is 78 mm and the overall height is 26 mm.

Fig. 20 Nominal dimensions of axisymmetric part used in compaction experiments

The powder blend is based on the diffusion alloyed iron powder, Hoeganaes Distalloy AE, as described in section

"Calibration of Material Parameters for an Iron Powder Blend." Press and Tooling. Figure 21 shows a schematic of the press and tooling. The part was pressed in a hydraulic COSMO

press. The tooling consists of four moving components: an outer die and an inner core that move together, a top punch, and a lower inner punch. The tooling also includes a stationary lower outer punch. The press was instrumented with strain gauges to measure loads on tooling members (see Fig. 21) and potentiometer displacement transducers to measure tooling displacements.

Fig. 21 Schematic of hydraulic COSMO press, tooling and strain gauges for load measurement. Source: Ref 4

Initial Conditions. The fill positions of the punches are given by R 1 and R2 as shown in Fig. 22 and Table 2. The initial

density in the die cavity was estimated by assuming uniform density within the rim section and the hub section of the part, as shown schematically in Fig. 23. The initial density values (Table 2) were estimated by weighing the different sections of the part and assuming no mass transfer between the top and bottom sections.

Table 2 Initial conditions for two-level compaction experiments Conditions Rim fill (R1), mm Hub fill (R2), mm pHub, g/cm3 pRim, g/cm3 Mass, g

Part 30 54.71 27.23 3.51 3.21 498.09

Part 34 55.82 25.18 3.63 2.97 482.42

Fig. 22 Actual tooling motions for axisymmetric component with piecewise linear approximations. Source: Ref 4

Fig. 23 Estimated initial fill density distribution for parts used in compaction experiments. Source: Ref 4

Compaction Kinematics. The displacement histories for each of the moving tooling members are shown in Fig. 22.

Tabulated information on punch displacements for part 30 and part 34 can be found in Ref 4. As explained later in this section, compaction of part 30 results in low density in the rim section. Also, cracks appear on the inner and outer surface of the rim. The aim of the compaction kinematics used with part 34 was to reduce the density imbalance between the hub and rim sections and to eliminate the cracks. The Finite Element Model. The powder material model described in the section "A Constitutive Model for Metallic

Powders with Ductile Particles" was implemented in the finite element code Abaqus/Standard (Ref 2) through a userdefined material subroutine. Details of the numerical implementation are beyond the scope of this article. Aravas (Ref 3) has described the numerical integration of constitutive models of this class. Finite Element Model. An axisymmetric model of the compact was set up using four noded axisymmetric elements

(Ref 2). Figure 24 shows a three-dimensional representation of the finite element mesh obtained by revolving the part about the axis of symmetry.

Fig. 24 Three-dimensional representation of the axisymmetric finite element discretization of the powder compact

Tooling. For simplicity, the tooling components were modeled as rigid surfaces. In a more detailed model, the tooling can also be modeled assuming elastic material behavior. Tooling stresses can be computed this way. However, modeling the interfaces between the powder and the elastic tooling is numerically very intensive. Material Model. Constant values were used for the elastic properties. A fixed value of Young's modulus (141 GPa),

corresponding to the final average density, was calculated using the expression for elastic modulus determined by Pavier et al. (Ref 14). It was mentioned earlier that the elastic properties are strong functions of density. However, in an analysis such as this, severe numerical instabilities can result as a consequence of the soft elastic properties at low density. Further, the effect of elastic behavior is dominant only when the relative density approaches its pore-free value. The friction coefficient between iron powder compacts and steel/carbide tooling varies with the normal stress (Ref 44,

45). For lubricated iron powders, the coefficient of friction varies in the range 0.1 to 0.2 for normal stresses in the range 100 to 700 MPa. For the present model, a Coulomb friction model was used and a coefficient of friction of 0.15 was assumed.

Compaction Kinematics. A piecewise linear approximation to the actual tooling motions was used as depicted in Fig.

22. Part Fill Density. The initial fill density was set as represented schematically in Fig. 23. Experimental and Modeling Results. Figure 25 shows the compaction forces measured for part 30. The forces

exerted on the bottom and upper punches increase from the initial instant corresponding to the moment of contact between the powder and the upper punch. At instant tL1 (Fig. 22), the speed of the lower inner punch (LIP) is modified such that it descends at practically constant force over the rest of its travel through time tL2. While the LIP is in motion, the upper punch (UP) and the lower outer punch (LOP) withstand forces that increase progressively up to their peak values. The rest of the cycle, for time >tL2, continues while the LIP remains at a fixed position. The forces on the LOP and the UP increase rapidly to their maximum values. Compression finishes at time tF, corresponding to the maximum displacement of the UP. The maximum loads recorded are shown in Table 3. Densities were measured in five sections of the part as shown schematically in Fig. 26. Measurements are shown in Table 4. For part 30, the hub and rim densities are approximately 7.06 g/cm3 and 6.90 g/cm3.

Table 3 Measured and computed peak compaction loads Punch LIP LOP

Part 30 loads, MN Measured Predicted 2.10 2.12 0.59 0.36

Part 34 loads, MN Measured Predicted 1.61 1.75 0.61 0.5

Table 4 Measured sectional densities Section of Fig. 26 1 2 3 4

Density, g/cm3 Part 30 Part 34 7.01 6.9 7.07 6.93 7.09 7.03 6.90 6.91

Fig. 25 Comparison of compaction tooling loads for part 30: predicted versus experimental

Fig. 26 Comparison of compaction tooling loads for part 34: predicted versus experimental

After ejection, the parts were visually inspected for cracks. Part 30 presented a distinct crack around the rim ID, 2 mm below the hub. Very fine cracks were also present on the outer surface of the rim, close to the bottom end. Although it is not possible to state exactly when the cracks were formed, it is very likely that they appeared during load removal/ejection. For part 30, the maximum loads on the inner and outer lower punches are 2.1 MN and 0.59 MN, respectively. The high load on the lower inner punch results in high punch deflection. During ejection, the elastic recovery of the lower inner punch results in tension on the rim causing cracking. To eliminate the cracks that appeared in part 30, part 34 was pressed with higher fill in the rim section and reduced fill in the hub section (Table 2). The tool motions applied were modified to accommodate the different fill positions, otherwise they were very similar to the motions used with part 30. The forces measured are shown in Fig. 27. The peak load values are given in Table 3. The load patterns for parts 30 and 34 are similar; however, for part 34, the lower inner punch load is reduced to 1.61 MN, while the outer lower punch force increases slightly to 0.61 MN. The hub and rim densities are approximately 6.97 g/cm3 and 6.94 g/cm3 (Table 3). The reduced load in the lower inner punch results in lower deflection and eliminates tension during load removal/part ejection. Part 34 was defect free.

Fig. 27 Schematic of sections for density measurements for test parts

Figure 28 illustrates the compaction process as predicted by the numerical model. The deformed mesh is plotted for part 34. The model shows that there is very limited transfer of powder between the top and bottom (rim) sections of the part. This behavior is expected because the tooling motions were designed to minimize powder transfer to avoid the formation of a defect at the corner between the two sections. Figure 29 plots the evolution of density in the compact during pressing of part 34.

Fig. 28 Deformed mesh during stages of compaction of part 34

Fig. 29 Computed density fields at different stages of compaction of part 34

The loads computed for parts 30 and 34 are plotted in Fig. 26 and 27 and compared to the measured loads. There is good overall agreement between the experiments (Ref 4) and the calculations. The computed maximum load values for each punch are shown in Table 3. Again the agreement with the experiments is very good, showing that the model is capable of predicting the effects of punch kinematics and fill position on punch loads. One observation on the nature of the load computations is pertinent. As the compact approaches its pore-free density, it becomes incompressible. Consequently, towards the end of the compaction stroke, minor variations in tool motions result in sharp changes in load. Thus, it is very important to ascertain the exact displacements of tooling members from their reference positions if accurate load predictions are sought. Compaction of a Long Bushing This example illustrates how the numerical model can be applied to predict density distributions. This part was studied experimentally by Trasorras et al. (Ref 40). A steel bushing was pressed in a 150 ton mechanical press. The green bushing dimensions are outside diameter (OD) = 19.05 mm, inner diameter (ID) = 12.7 mm, and height = 25.4 mm. The powder used was a blend of atomized steel powder (Ancorsteel 1000) with 0.75 wt% zinc stearate as lubricant. The punch motions comprise the following sequence. After powder filling, the top punch moves down to compact the powder, then rises and exits the die cavity. At the end of top punch motion, the die is stripped to eject the compact. Finally, the core rod is stripped. The lower punch and the die remain stationary throughout the compaction. As compaction starts, a density gradient develops in the bushing due to the friction between the compact and the tooling members. With continued top punch motion, the bushing densifies with the top always being at a higher density than the bottom. During ejection, there is some additional densification of the bottom end of the compact. Finally, the compact expands as it exits the die cavity thereby reducing the overall density. The axial density distribution in the bushing was determined by successively sectioning and weighing the compact.

Compaction of the bushing was modelled in PCS (Ref 42), a powder compaction modeling system based on the finite element code NIKE2D (Ref 46). Figure 30 shows the finite element discretization of the tooling and powder, with the punches shown at their fill position. The material model described earlier was used with the constitutive functions b(D), c(D), and elastic properties calibrated for the atomized iron powder. The initial apparent density of the powder was 3.2 g/cm3. A complete model of the tooling was used (Fig. 30) and elastic isotropic behavior was assumed. The friction between the compact and the tooling members was assumed to follow Coulomb's model with a friction coefficient of 0.2. Figure 31 compares the axial density distribution predicted by the numerical model with the experimental results. The model properly represents the densification that takes place during both compaction and ejection and the predicted final density distribution is in good agreement with the experiments. The experiments show a sharp increase of density at the powder layer in contact with the top and bottom punches. The numerical model, with the discretization level used, was not able to predict this effect.

Fig. 30 Finite element discretization of powder and tooling used in the compaction of cylindrical bushing

Fig. 31 Density distribution along axis of cylindrical bushing: FEA predictions versus experimental results

This article has examined the general structure of constitutive laws for the compaction of powder compacts and demonstrated how these material models can be used to model the response of real world components to a series of complex die operations. It identified the general structure of the constitutive law and described a number of models that have been proposed in the literature. This field is still evolving, and it is evident that there will be significant developments in this area over the next few years as a wider range of experimental studies are conducted, providing greater insights into the compaction process. At the current time, there is no universally accepted model. Therefore, a pragmatic approach and a relatively simple form of empirical model were adopted requiring, for the determination of the unknown functions, a limited range of experiments. This selection allowed an examination of the compaction of axisymmetric components in detail and a comparison of general features of the component response with practical measurements. Similar procedures could have been adopted for any of the methods described, although in general, more sophisticated experiments are required in order to determine any unknown function or coefficients, particularly if the shape of the yield function is not known, or assumed, a priori.

References cited in this section

2. ABAQUS/Standard User's Manual, Version 5.7, Vol 1-3, Hibbitt, Karlsson, & Sorensen, Inc., Providence, RI, 1997 3. N. Aravas, On the Numerical Integration of a Class of Pressure-Dependent Plasticity Models, Int. J. Numer. Meth. Eng., Vol 24, 1987, p 1395-1416 4. Y. Kergadallan, G. Puente, P. Doremus, and E. Pavier, Compression of an Axisymmetric Part, Proc. of the Int. Workshop on Modelling of Metal Powder Forming Processes (Grenoble, France), 1997, p 277-285 14. E. Pavier and P. Doremus, Mechanical Behavior of a Lubricated Powder, Advances in Powder Metallurgy & Particulate Materials-1996, Vol 2 (Part 6), Metal Powder Industries Federation, 1996, p 27-40 40. J.R.L. Trasorras, S. Krishnaswami, L.V. Godby, and S. Armstrong, Finite Element Modeling for the Design of Steel Powder Compaction, Advances in Powder Metallurgy & Particulate Materials-1995, Vol 1 (Part 3), Metal Powder Industries Federation, 1995, p 31-44 42. Powder Compaction Simulation Software (PCS Elite) User's Manual, Concurrent Technologies Corp., Johnstown, PA 44. B. Wikman, H.A. Häggblad, and M. Oldenburg, Modelling of Powder-Wall Friction for Simulation of Iron Powder Pressing, Proc. of the Int. Workshop on Modelling of Metal Powder Forming Processes (Grenoble, France), July 1997, p 149-158

45. E. Pavier and P. Dorémus, Friction Behavior of an Iron Powder Investigated with Two Different Apparatus, Proc. of the Int. Workshop on Modelling of Metal Powder Forming Processes (Grenoble, France), July 1997, p 335-344 46. J. Hallquist, "NIKE2D-A Vectorized, Implicit, Finite Deformation, Finite-Element Code for Analyzing the Static and Dynamic Response of 2-D Solids," Technical report UCRL-19677, Lawrence Livermore National Laboratory, Livermore, California, 1993 Mechanical Behavior of Metal Powders and Powder Compaction Modeling J.R.L. Trasorras and R. Parameswaran, Federal-Mogul, Dayton, Ohio; A.C.F. Cocks, Leicester University, Leicester, England

References 1. R. German, Particle Packing Characteristics, Metal Powder Industries Federation, 1989 2. ABAQUS/Standard User's Manual, Version 5.7, Vol 1-3, Hibbitt, Karlsson, & Sorensen, Inc., Providence, RI, 1997 3. N. Aravas, On the Numerical Integration of a Class of Pressure-Dependent Plasticity Models, Int. J. Numer. Meth. Eng., Vol 24, 1987, p 1395-1416 4. Y. Kergadallan, G. Puente, P. Doremus, and E. Pavier, Compression of an Axisymmetric Part, Proc. of the Int. Workshop on Modelling of Metal Powder Forming Processes (Grenoble, France), 1997, p 277-285 5. K.T. Kim, J. Suh, and Y.S. Kwon, Plastic Yield of Cold Isostatically Pressed and Sintered Porous Iron under Tension and Torsion, Powder Metall., Vol 33, 1990, p 321-326 6. H.A. Kuhn and C.L. Downey, Material Behavior in Powder Preform Forging, J. Eng. Mater. Technol., 1990, p 41-46 7. S. Shima and M. Oyane, Plasticity Theory for Porous Metals, Int. J. Mech. Sci., Vol 18, 1976, p 285-291 8. S.B. Brown and G.A. Weber, A Constitutive Model for the Compaction of Metal Powders, Modern Developments in Powder Metallurgy, Vol 18-21, 1988, MPIF, p 465-476 9. T.J. Watson and J.A. Wert, On the Development of Constitutive Relations for Metallic Powders, Metall. Trans. A, Vol 24, 1993, p 2071-2081 10. A.R. Akisanya, A.C.F. Cocks, and N.A. Fleck, The Yield Behaviour of Metal Powders (1996), Int. J. Mech. Sci., Vol 39 (No. 12), 1997, p 1315-1324 11. S. Brown and G. Abou-Chedid, Yield Behaviour of Metal Powder Assemblages, J. Mech. Phys. Solids, Vol 42 (No. 3), 1994, p 383-399 12. W. Prager, Proc. Inst. Mech. Eng., Vol 169, 1955, p 41 13. R. Hill, The Mathematical Theory of Plasticity, Oxford University Press, 1950 14. E. Pavier and P. Doremus, Mechanical Behavior of a Lubricated Powder, Advances in Powder Metallurgy & Particulate Materials-1996, Vol 2 (Part 6), Metal Powder Industries Federation, 1996, p 27-40 15. C.J. Yu, R.J. Henry, T. Prucher, S. Parthasarathi, and J. Jo, Advances in Powder Metallurgy & Particulate Materials, Vol 6, Metal Powder Industries Federation, 1992, p 319-332 16. N.A. Fleck, L.T. Kuhn, and R.M. McMeeking, Yielding of Metal Powder Bonded by Isolated Contacts, J. Mech. Phys. Solids, Vol 40, 1992, p 1139-1162 17. N.A. Fleck, On the Cold Compaction of Powders, J. Mech Phys. Solids, Vol 43 (No. 9), 1995, p 1409-1431 18. J. Gollion, D. Bouvard, P. Stutz, H. Grazzini, C. Levaillant, P. Baudin, and J.P. Cescutti, On the Rheology of Metal Powder during Cold Compaction, Proc. Int. Conf. on Powders and Grains, Biarez and Gourves, Ed., Clermont-Ferrand, France, 4-8 September 1989, p 433-438 19. R.M. Govindarajan and N. Aravas, Deformation Processing of Metal Powders, Part 1: Cold Isostatic Pressing, Int. J. Mech. Sci., Vol 36, 1994, p 343-357 20. A.L. Gurson, Continuum Theory of Ductile Rupture by Void Nucleation and Growth, Part 1: Yield Criteria

and Flow Rules for Porous Ductile Media, J. Eng. Mater. Technol., Vol 99, 1977, p 2-15 21. A.C.F. Cocks, The Inelastic Deformation of Porous Materials, J. Mech. Phys. Solids, Vol 37 (No. 6), 1989, p 693-715 22. Y-M. Liu, H.N.G. Wadley, and J. Duva, Densification of Porous Materials by Power-Law Creep, Acta Metall. Mater., Vol 42, 1994, p 2247-2260 23. A.R. Akisanya, A.C.F. Cocks, and N.A. Fleck, Hydrostatic Compaction of Cylindrical Particles, J. Mech. Phys. Solids, Vol 42 (No. 7), 1994, p 1067-1085 24. Z. Qian, J.M. Duva, and H.N.G. Wadley, Pore Shape Effects during Consolidation Processing, Acta Metall. Mater., Vol 44, 1996, p 4815 25. P. Ponté Castañeda and M. Zaidman, Constitutive Models for Porous Materials with Evolving Microstructure, J. Mech. Phys. Solids, Vol 42, 1994, p 1459-1497 26. K.T. Kim and J. Suh, Elastic-Plastic Strain Hardening Response of Porous Metals, Int. J. Eng. Sci., Vol 27, 1989, p 767-778 27. S. Brown and G. Abou-Chedid, Appropriate Yield Functions for Powder Compacts (1992), Scr. Metall. Mater., Vol 28, 1993, p 11-16 28. D.C. Drucker and W. Prager, Q. Appl. Math., Vol 10, 1952, p 157-165 29. A.L. Gurson and T.J. McCabe, Experimental Determination of Yield Functions for Compaction of Blended Powders, Proc. MPIF/APMI World Cong., on Powder Metallurgy and Particulate Materials (San Francisco), Metal Powder Industries Federation, 1992 30. A. Schofield and C.P. Wroth, Critical State Soil Mechanics, McGraw-Hill, 1968 31. D.M. Wood, Soil Behavior and Critical State Soil Mechanics, Cambridge University Press, 1990 32. S. Shima, "A Study of Forming of Metal Powders and Porous Metals," Ph.D. thesis, Kyoto University, 1975 33. Y. Morimoto, T. Hayashi, and T. Takei, Mechanical Behavior of Powders in a Mold with Variable Cross Sections, Int. J. Powder Metall. Powder Technol., Vol 18 (No. 1), 1982, p 129-145 34. J.R.L. Trasorras, S. Armstrong, and T.J. McCabe, Modeling the Compaction of Steel Powder Parts, Advances in Powder Metallurgy & Particulate Materials-1994, Vol 7, American Powder Metallurgy Institute, 1994, p 33-50 35. J. Crawford and P. Lindskog, Constitutive Equations and Their Role in the Modeling of the Cold Pressing Process, Scand. J. Metall., Vol 12, 1983, p 271-281 36. J.R.L. Trasorras, T.M. Krauss, and B.L. Ferguson, Modeling of Powder Compaction Using the Finite Element Method, Advances in Powder Metallurgy, Vol 1, T. Gasbarre and W.F. Jandeska, Ed., American Powder Metallurgy Institute, 1989, p 85-104 37. B.L. Ferguson, et al., Deflections in Compaction Tooling, Advanced in PM & Particulate Materials, Vol 2, Metal Powder Industries Federation, 1992, p 251-265 38. H. Chtourou, A. Gakwaya, and M. Guillot, Assessment of the Predictive Capabilities of the Cap Material Model for Simulating Powder Compaction Problems, Advances in Powder Metallurgy & Particulate Materials-1996, Vol 2 (Part 7), Metal Powder Industries Federation, 1996, p 245-255 39. D.T. Gethin, R.W. Lewis, and A.K. Ariffin, Modeling Compaction and Ejection Processes in the Generation of Green Powder Compacts, Net Shape Processing of Powder Materials, 1995 ASME Int. Mechanical Engineering Congress and Exposition, AMD-Vol 216, S. Krishnaswami, R.M. McMeeking, and J.R.L. Trasorras, Ed., The American Society of Mechanical Engineers, 1995, p 27-45 40. J.R.L. Trasorras, S. Krishnaswami, L.V. Godby, and S. Armstrong, Finite Element Modeling for the Design of Steel Powder Compaction, Advances in Powder Metallurgy & Particulate Materials-1995, Vol 1 (Part 3), Metal Powder Industries Federation, 1995, p 31-44 41. S. Krishnaswami and J.R.L. Trasorras, Modeling the Compaction of Metallic Powders with Ductile Particles,Simulation of Materials Processing: Theory, Methods and Application, Shen and Dawson, Ed., Balkema, Rotterdam, 1995, p 863-858 42. Powder Compaction Simulation Software (PCS Elite) User's Manual, Concurrent Technologies Corp., Johnstown, PA

43. H-A. Haggblad, P. Doremus, and D. Bouvard, An International Research Program on the Mechanics of Metal Powder Forming, Advances in Powder Metallurgy & Particulate Materials-1996, Vol 2 (Part 7), Metal Powder Industries Federation, 1996, p 179-192 44. B. Wikman, H.A. Häggblad, and M. Oldenburg, Modelling of Powder-Wall Friction for Simulation of Iron Powder Pressing, Proc. of the Int. Workshop on Modelling of Metal Powder Forming Processes (Grenoble, France), July 1997, p 149-158 45. E. Pavier and P. Dorémus, Friction Behavior of an Iron Powder Investigated with Two Different Apparatus, Proc. of the Int. Workshop on Modelling of Metal Powder Forming Processes (Grenoble, France), July 1997, p 335-344 46. J. Hallquist, "NIKE2D-A Vectorized, Implicit, Finite Deformation, Finite-Element Code for Analyzing the Static and Dynamic Response of 2-D Solids," Technical report UCRL-19677, Lawrence Livermore National Laboratory, Livermore, California, 1993

Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Introduction POWDER METAL COMPACTING PRESSES, equipped with appropriate tooling, frequently are used for producing P/M components. Although commonly called P/M presses, use is not limited to the pressing of metal powders. Almost any alloy or mixture of materials produced in powder form can be compacted into suitable end products. The majority of components fabricated by P/M presses, in number of pieces and pounds of product produced, consists of compacted metals. Ferrous-base metals constitute the largest usage. Powder metallurgy compacting presses usually are mechanically or hydraulically driven, but they can incorporate a combination of mechanically, hydraulically, and pneumatically driven systems. Table 1 summarizes some of the developments for P/M presses in the last 40 years. Other recent improvements in compaction technology include: •

Split-die techniques to make multilevel parts having different peripheral contours at different

levels • •

Punch rotation capability to facilitate production of helical gears and other helical shapes Higher compaction pressures by using stronger tool materials, advanced pressure control methods, and die wall lubricants • Better process control with computerized tool motion monitoring • Warm compaction and improved "segregation-free" powders with enhanced flow characteristics

Table 1 History of development in P/M presses Years 1955-1959 1960-1964 1965-1969 1970-1974 1975-1979 1980-1989 1990-1994 1995-present

Compacting press Cam press, HP Toggle press, MP Large size HP (500 +), large size MP (500 +) Multistepped MP, double die compacting press Large size MP (750 +), tool holder quick change NC press, multistepped HP (800 +), large size rotary press Large size MP, automatic P/M manufacturing line Hybrid (mechanical/hydraulic) presses (800 tons)

HP, hydraulic press; MP, mechanical press; NC, numeric controlled

Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Compacting Press Requirements Although P/M presses resemble stamping and forming presses, several significant differences exist. Press frames generally have straight sides. Gap-type or "C" frame presses are not suitable because the frame deflects in an arc under load, resulting in a slight out-of-alignment condition between the bed and side of the press. This arrangement produces a compacted part that is slightly out of parallel, top to bottom. Because P/M tooling clearances are generally 0.025 mm/25 mm (0.001 in./1 in.) total, bending deflection can cause broken tooling or excessive tool wear. Powder metallurgy presses apply sufficient pressure from one or both pressing directions (top and bottom) to achieve uniform density throughout the compact. Design should include provision for ejecting the part from the tooling. Pressing and ejection occur during each cycle of the press and must be accurately synchronized. Presses need sufficient connected horsepower to compact and eject the part. In most press-working applications, the working stroke is a small portion of the total stroke of the press. In P/M presses, the working stroke during the compaction portion of the cycle is usually greater than the length of the part being produced, and the ejection portion of the cycle has a working stroke equal to or greater than the length of the part by a factor of approximately two or three. In some cases, the power required during the ejection cycle is greater than that required during compaction. Presses should provide for adjustable die filling (the amount of loose powder in the tooling cavity). Automatic powder feeding systems that are synchronized with the compaction and ejection portion of the press cycle are desirable. Finally, P/M presses must meet federal, state, and local design and construction safety laws. Metal Powder Industries Federation (MPIF) standard 47 details safety standards for P/M presses. Mechanical presses are available in top-drive and bottom-drive arrangements. In top-drive presses, the motor, flywheel, and gearing system are located in the crown or upper structure of the press. Presses with pressing capacities of 1780 kN (200 tons) are floor mounted, requiring little or no pit. Top-drive presses with pressing capacities >1780 kN (200 tons) usually require a pit to maintain a convenient working height for the operator. In bottom-drive presses, the drive mechanism, motor, and flywheel are located in the bed of the press. These presses usually are "pulled down"; that is, the top ram of the press is pulled downward by draw bars or tie rods. Bottom-drive presses with pressing capacities of >445 kN (50 tons) usually require pits. Top-drive and bottom-drive presses are comparable in terms of partmaking capability, reliability, and equipment cost. Press Tonnage and Stroke Capacity. Required press capacity to produce compacts in rigid dies at a given pressure

depends on the size of the part to be pressed and the desired green density of the part, which in turn is determined by requirements for mechanical and physical properties of the sintered part. Compacting pressures can be as low as 70 to 140 MPa (5 to 10 tsi) for tungsten powder compacts or as high as 550 to 830 MPa (40 to 60 tsi) for high-density steel parts. When a part is pressed from the top and bottom simultaneously, the press should apply the required load to the upper and lower ram of the press. To eject the pressed compact, an ejection capacity must be available that is sometimes divided into the load for the breakaway stroke (which is the first 1 to 12 mm ( to in.) of the ejection stroke and the load for a sustained stroke). The load for a sustained stroke is generally one-fourth to one-half of the breakaway load. The stroke capacity of a press, or the maximum ram travel, determines the length of a part that can be pressed and ejected. In presses used for automatic compacting, the stroke capacity is related to the length available for die fill and ejection stroke. Load Requirements. The total load required for a part is determined by the product of the pressure needed to compact

the part to the required density and the projected area of the part. Compaction curves relate pressure, P, to the required density, q, and are usually obtained from compacting tests on cylindrical shapes with the height, L, equal to the diameter,

D. For thicker parts the load must be increased, by as much as 25% for a length to diameter ratio of 4 to 1, to give the required density. Required compacting pressures can be estimated with a correction factor, k, such that (Ref 1):

P = P1 (1 + k) where P is the compaction pressure for a larger part and P1 is the compaction pressure for a "standard" part (i.e., L = D). The correction factor is:

k = (0.25/3)(L/D - 1) for L/D >1 k = 0 for L/D < 1 For parts that are not cylindrical, an equivalent L/D ratio can be used:

Le/De = (V · p)/(2 · A2) where V is the part volume and A is the projected area. The press load required is then obtained by multiplying the required compaction pressure by the projected area of the part.

Reference cited in this section

1. W.A. Knight, Design for Manufacture Analysis: Early Estimates of Tool Costs for Sintered Parts, Annals of the CIRP, Vol 40 (No. 1), 1991, p 131-134 Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Mechanical Presses In most mechanical P/M compacting presses, electric motor-driven flywheels supply the main source of energy used for compacting and ejecting the part. The flywheel normally is mounted on a high-speed shaft and rotates continuously. A clutch and a brake mounted on the flywheel shaft initiate and stop the press stroke. To initiate a press stroke, the brake is disengaged and the clutch is engaged, causing the energy stored in the rotating flywheel to transmit torque through the press gearing to the final drive or press ram. Clutch and brake systems should be of the partial revolution type that can be engaged and disengaged at any point in the pressing cycle. The clutch usually is pneumatically engaged with a spring release, and the brake is pneumatically released with a spring set, thereby providing full stopping ability in the event of loss of air pressure. An adjustable speed device normally is supplied with electric drive motor, providing production rate adjustment as indicated by pressing and ejection conditions. On presses that have main motor capacities up to 19 kW (25 hp), the adjustable speed drive is usually of the variablepitch pulley or traction-drive type. Above 19 kW (25 hp), direct-current or eddy-current control devices are preferred. The motor and drive must be totally enclosed to prevent contamination by metal powder dust. Gearing systems usually are either single-reduction (Fig. 1) or double-reduction (Fig. 2) arrangements. Single-reduction gearing frequently is used in lower tonnage presses ( 445 kN, or 50 tons) that have stroking rates of 50 strokes/min. Higher tonnage presses use double-reduction gearing and commonly have maximum stroking rates of 30 strokes/min.

Fig. 1 Single-reduction gearing systems for P/M compacting press

Fig. 2 Double-reduction gearing systems for P/M compacting press

The low-speed shaft of the press, normally called the main shaft, is linked to the press ram, causing motion of the tooling for the compacting and ejection cycles. Ram driving mechanisms can be either cam- or eccentric-driven arrangements. Cam-driven presses generally are limited to pressing capacities 890 kN (100 tons). The main shaft of the press has two cams--one cam operates the upper ram, and the other cam operates the lower ram for compacting the part. The cam that operates the lower ram also controls the powder feed into the die and ejects the part from the die after compacting. Cams normally operate linkages that convert the main shaft rotary motion into the linear motion of the tooling.

Figure 3 shows a schematic of a cam-driven press. The cams in this type of press can be adjusted or arranged with removable sections, thus allowing cam motion to be varied to produce special motions to compact the part. Pressure can be applied either simultaneously or sequentially to the top and bottom of the compact. Anvil and rotary presses are types of cam-driven machines. These presses are described in more detail later in this article.

Fig. 3 Schematic of cam-driven compacting press

Eccentric-Driven Presses. Presses that have a final drive mechanism consisting of an eccentric or crank on the main shaft are the most widely used type of mechanical press. A connecting rod is used to convert the rotary motion of the main shaft into the reciprocating motion of the press ram. Generally, an adjustment mechanism is built into the connecting rod or press ram assembly, thus permitting the height position of the press ram to be changed with respect to the main shaft or press frame, thereby controlling the final pressing position of the ram. This adjustment mechanism can be used to control the length of the compacted part. Standard eccentric-driven presses have pressing capacities ranging from 6.7 to 7830 kN (0.75 to 880 tons). Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Hydraulic Presses Hydraulically driven compacting presses are available with pressing capacities ranging from 445 to 11,100 kN (50 to 1250 tons) as standard production machines, although special machines with capacities 44,500 kN (5000 tons) have been used in production. Hydraulic presses normally can produce longer parts in the direction of pressing than mechanical presses, and longer stroke hydraulic machines are less expensive compared to an equivalent stroke produced in a mechanical press. The maximum depth of powder fill in mechanical presses is 180 mm (7 in.), while 380 mm (15 in.) of powder fill is common in hydraulic presses. The maximum production rate for hydraulic presses producing a single part per stroke is 650 pieces per hour. The slower speed of a hydraulic press when pressing long parts is preferable, because the longer time during pressing permits trapped air within the powder to escape through the tooling clearances. Most hydraulic presses are considered top-drive machines because the main operating cylinder is centrally located in the top of the press. This main cylinder provides the force for compacting the part. Hydraulic presses have three distinct downward speeds: • •

Rapid advance: Produces minimal pressing force, used for rapid closing of the die cavity Medium speed: Pressing capacities 50% of full-rated capacity, used during initial compaction when lower pressing force is required

•

Slow speed: Maximum capacity available for final compaction

Two types of hydraulic pumping systems are commonly found in P/M presses: the high-low system and the filling circuit system. The high-low system has a double-acting main cylinder. A regenerative circuit is used for rapid approach. Initially, the piston of the cylinder is activated by a high-volume, low-pressure pump; the fluid from the bottom of the cylinder is directed into the top of the cylinder in addition to the low-pressure pump volume. At medium speed, the regenerative circuit is deactivated, while the piston remains activated by the low-pressure pump. In full-tonnage press, the low-pressure pump is deactivated, and a high-pressure pump activates the piston. The filling circuit hydraulic pumping system has a single-acting main cylinder, and ram motion is controlled by small double-acting cylinders. The ram control cylinders are smaller than the main cylinder, so only a low flow rate of fluid is needed to cause rapid movement of the ram. During approach and return cycles, however, the fluid flow rate into and out of the main cylinder is high. The main cylinder is fitted with a large two-way valve that allows fluid to flow at low pressures (usually gravity feed). During pressing, the two-way valve is closed, and high pressure from the pump is applied to the main cylinder piston. Ejection of the part usually is accomplished by a cylinder that is centrally located in the bed of the press. The cylinder either upwardly ejects the part or pulls the die downward from the part, depending on the type of tooling used. When pressing parts to a given thickness, positive mechanical stops are used on hydraulic presses to control downward ram movement. When pressing parts to a desired density, downward ram movement is controlled by adjustment of the pressure to the cylinder. When the part is pressed to the desired unit pressure, the press ram stops and returns to the retracted position. Some types of P/M materials, such as P/M friction materials, are always pressed to density rather than size, because uniform density provides uniform friction and wear properties. The drive-motor horsepower on a hydraulic press is considerably larger than on an equivalent mechanical press. A mechanical press has a flywheel from which energy is taken during the pressing and ejection of the part. Energy is restored to the flywheel during the die feeding portion of the cycle. The motor on a hydraulic press must supply energy directly during the pressing and ejection portion of the cycle. Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Comparison of Mechanical and Hydraulic Presses In terms of partmaking capability, no distinct advantage is gained by using either a mechanical press or a hydraulic press. Any part can be produced to the same quality on either type of machine. However, the following parameters influence press drive selection. Production Rate. A mechanical press produces parts at a rate one and one-half to five times that of a hydraulic press as

a result of inherent design of the energy transfer systems and stroke length. Operating cost of a hydraulic press is higher, because the total connected horsepower of a hydraulic press is one and

one-half to two times that of an equivalent mechanical machine. Theoretically, the required energy to compact and eject a part is the same for a hydraulic or a mechanical press, except that the overall efficiency of a mechanical press is slightly higher than that of a hydraulic press. Also, the kilowatt usage of the larger motor on a hydraulic press is greater than that of a mechanical press during the idle portion of the machine cycle. Machine overload protection is an inherent feature of a hydraulic press. If the hydraulic system is operating properly, the machine cannot create a force greater than the rated capacity. Consequently, overload of the machine frame is not possible, even if a double hit or operator error occurs in adjusting the machine. Misadjustment or double hits can cause a mechanical press to overload, can damage the machine, or may cause tooling overload and failure if the tooling

cannot withstand full machine capacity. Some new mechanical presses are equipped with hydraulic overload protection systems. Equipment cost of a hydraulic press generally is one-half to three-quarters that of an equivalent mechanical press.

Facility, foundation, installation, and floor space costs generally are comparable. Die Sets. The mounting into which the tooling is installed is known as the die set. Generally, the die set must be well guided because of the close tooling clearances used. Guide bearings must be protected with boots or wipers to prevent powder particles from entering guiding surfaces. Tooling support team members should have high stiffness to minimize deflection.

The die set must be free of residual magnetism. The maximum acceptable level is 2 G. To ensure press operator safety, die sets should be adequately guarded. In a complex tooling arrangement, as many as seven independent tooling members and supports are moving relative to one another during the pressing and ejection cycles. Die sets can be classified as removable or nonremovable. Both types are used in mechanical and hydraulic presses. Nonremovable die sets are used throughout the entire tonnage requirements of available presses. Manually removable die sets are used primarily in presses with pressing capacities up to 2670 kN (300 tons). Above this press size, the die set assembly is moved by a powered system, and removable die set presses with capacities of 17,800 kN (2000 tons) are available. The major advantage offered by nonremovable die sets is flexibility in setup and operation. Presses equipped with nonremovable die sets usually have all adjustments required for setup and operation built into the press and die set, including: • Part length adjustment: Any dimensions of the part in the direction of pressing can be quickly changed during production. • Part weight: Material weight in any level of the part can be changed easily during production. • Tooling length adjustment: Adjustments are provided to accommodate shortening of punch length due to sharpening or refacing.

Another advantage of nonremovable die sets is the greater space available for tooling, compared to the removable type. This space provides more freedom in tooling design. However, presses incorporating nonremovable die sets must be shut down during tooling changes or maintenance. Tooling change and setup time generally is from 1 to 4 hours--but sometimes substantially longer, depending on the complexity of tooling. Nonremovable die sets are well suited for developing new P/M parts, because press and tooling adjustments can be made quickly to achieve the desired weight, density, and part dimension. Adjustment features of nonremovable die sets make them desirable on long production runs, where changes in powder quality among lots require frequent tooling adjustment to maintain part quality. Users of removable die sets normally have two or more die sets per press. Tooling can be set up in a spare die outside the press. Removable die sets normally can be changed in less than 30 min, so loss of production time is minimal. On small presses where the die set is also small, the die set is restricted to a given set of tools and is considered semidurable tooling. One disadvantage of many removable die sets is that pressing is controlled by pairs of pressing blocks made of hardened tool steel, such as D-2. The height of the pressing block controls the height of the part. If the part length dimension is changed due to design, or if the tooling length is changed due to repair, the pressing blocks must be changed accordingly. Removable die sets are ideally suited for shorter production runs. On newer presses with removable die sets, complete powder adjustment is available, even when the die set is outside the machine.

Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Part Classification The Metal Powder Industries Federation has classified P/M parts according to complexity. Class I parts are the least complex, and class IV parts are the most complex. To better understand the types of commercially available P/M compacting presses, and their advantages and limitations, an understanding of P/M part classification and tooling systems used to produce parts is necessary. Part thickness and number of distinct levels perpendicular to the direction of powder pressing determine classification--not the contour of the part. Class I parts are single-level parts that are pressed from one direction, top or bottom, and that have a slight density variation within the part in the direction of pressing (Fig. 4a). The highest part density is at the surface in contact with the moving punch, and the lowest density is at the opposite surface. Parts with a finished thickness of 7.5 mm (0.3 in.) can be produced by this method without significant density variation.

Fig. 4 Basic geometries of (a) MPIF class I (simple) and (b) MPIF class IV (complex) parts

Class II parts are single-level parts of any thickness pressed from both top and bottom. The lowest density region of

these parts is near the center, with higher density at the top and bottom surfaces.

Class III parts have two levels, are of any thickness, and are pressed from both top and bottom. Individual punches are

required for each of the levels to control powder fill and density. Class IV parts are multilevel parts of any thickness, pressed from both top and bottom (Fig. 4b). Individual punches are

required for each level to control powder fill and density. Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Shape of Rigid Tooling Rigid tool compaction differs from roll compaction, isostatic compaction, hot isostatic pressing, and injection molding in that a quantity of powder (fill) is confined in a rigid die cavity at ambient temperature. The die cavity is entered by one or more punches, which apply compaction pressure to the fill powder. As a result of the compaction pressure, the fill powder densifies, develops green strength, and assumes the exact shape of the die cavity and punch faces. Following the pressure cycle, the shaped powder fill, now a piece part, is ejected (stripped) from the die cavity. The physical size of parts made in rigid tool compaction systems is a function of press tonnage capacity, fill depth, and also the length of a green powder fill that can be effectively compacted in terms of a maximum density variance. Parts vary in size from those weighing 1 g (0.035 oz) that are made in presses with capacities as small as 35 kN (4 tons) to those weighing 10 kg (22 lb) that are made in presses with capacities of 8900 kN (100 tons). Rigid tools must also be constructed oversize, with exact linear dimensions, to compensate for the final volume change. Although theoretical computations are useful, most successful rigid tool sets are based on shrinkage allowances developed from existing tooling and the dimensional histograms developed for particular powders. However, shrinkage allowances can be complex depending on subsequent sintering and binder additives. For example, some metallic powders, such as the carbide and tool steel types, and some gas and centrifugally atomized specialty powders, such as spray-dried tungsten carbide, do not develop significant green strength, because their individual particles are predominantly spherical or they lack plasticity. To compact such powders in rigid tool systems, wax or wax-stearate binders are added, which can occupy up to 20 vol% of the green compacted shape. The development of full metallic properties during sintering also requires a volume shrinkage. Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Powder Fill The important consideration in P/M part production is the fill ratio required to produce parts to a density that is compatible with end use requirements. The fill ratios must remain constant for a given part to maintain dimensional reproducibility. Parts can be of single-level or multilevel design. Single-level parts, designated as class I by the Metal Powder Industries Federation (MPIF), present the least difficulty to the tool designer, regardless of the size or part configuration. The main consideration is designing a die that is long enough guidance for the lower punch (usually 25 mm, or 1 in.) and providing adequate fill depth for compacting the powder to the required density. This challenge, coupled with the primary mechanical consideration of locating the center of mass in the press center, provides the best potential for producing a uniform quality part. Figure 4(a) shows basic geometries of MPIF class I parts. Multilevel parts, with industry classifications II through IV, present two additional complications to the tool designer:

powder fill and part ejection. Because metal powders tend to compact in vertical columns and generate little hydraulic flow, the tool designer must create fill levels in the tools that compensate for the thickness variations present in the final

part configuration. Uniform density, neutral axis of compaction, and part ejection should be considered to determine the need to vary fill levels and the manner in which these variations are achieved. Excessive density variations contribute to green cracks and sintered distortion. A common method of varying fill levels is by using multiple lower punches, which are timed to react to one another either through the use of springs or air, or by mounting on separate press platens. Other methods are less effective, because punches are not adjustable and are fixed on one of the tool members, such as the die or core rod. Fixed levels are commonly referred to as die chokes, core rod steps, or splash pockets (Fig. 5). Fixed fills are sensitive to the apparent density of the material being compacted. In operations that control compacting pressure, such as in hydraulic pressing, fixed fills cause dimensional variations in part thickness. Because mechanical presses are set to operate to a fixed position relative to the die, the variation created by the apparent density of the powder causes overdensification or underdensification, resulting in a corresponding oversize or undersize peripheral area on the part. Green expansion occurs as a part is stripped from the die. Ideally, the part returns to die size through shrinkage during sintering.

Fig. 5 Methods of achieving fixed fill levels. (a) Fixed fill on an upper level using a step die. (b) Fixed fill using a splash pocket to permit a projection feature on an upper punch. (c) Stepped core rod forming an internal shoulder

When a part has more than one level in the compacting direction, the step height should be limited to one-quarter of the overall height for a single punch (Fig. 6a). If a larger step is required, multiple punches should be considered (Fig. 6b).

Fig. 6 Two-level compaction. (a) Single lower punch when h

H/4. (b) Double lower punches when h > H/4.

Fill Height. The fill height is the depth of the loose powder required to give the required part thickness after compaction.

The value is determined by the compressibility of the loose powder at the required density. The fill height, hf, is obtained by multiplying the finished part height by the compression ratio of the powder:

hf = tkr In this equation, t is the part thickness, kr is the compression ratio, and kr = q/qa, where q is the part required compaction density and qa is the apparent density of the loose powder. If the fill height is greater than the maximum fill height that can be accommodated in the press selected on the basis of the compacting load required, a larger capacity machine should be selected, which has the required fill height capacity. Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Tooling Systems High-production P/M compacting presses are available as standard production machines in a wide range of pressing capacities and production rate capabilities. Presses are designed to produce parts of a specific classification, as discussed previously. Single-action tooling systems generally are limited to production of class I parts. During the compacting cycle, the die, core rod, and one of the punches (usually the lower punch) remain stationary. Compacting is performed by the moving punch, which is driven by the action of the press. One or more core rods may form any through holes in the part.

During ejection, the upper punch moves away from the formed part, and the part is ejected from the die by the lower punch. The core rod (Fig. 7) is stationary, and the part is ejected from the die and core rod simultaneously. On some presses, the core rod is arranged so that it is free to move upward (float) with the part as it is ejected. The compacted part experiences slight elastic expansion on ejection from the die, which causes the part to free itself from the core rod. The core rod is then free to move downward to the fill position. This floating core rod arrangement reduces ejection forces and core rod wear.

Fig. 7 Compacting sequence utilizing single-action tooling. Dashed line indicates motion of lower punch.

Double-action tooling systems primarily are used to produce class I and II parts. Force is applied to the top and

bottom of the part simultaneously, because the punches have the same travel rate. The die and core rod are stationary. Densification takes place from the top and bottom, with the lowest density region near the center of the part. Although the

core rod is fixed in this system, it can be arranged in a floating position. Figure 8 shows the compacting sequence of a double-action tooling system.

Fig. 8 Compacting sequence utilizing double-action tooling. Dashed line indicates motion of component parts.

Floating die tooling systems are similar to double-action arrangements. As shown in Fig. 9, the die is mounted on a yielding mechanism (springs). However, pneumatic or hydraulic cylinders usually are used, because they offer an easily adjustable resisting force. As the upper punch enters the die and starts to compact the powder, friction between the powder and die wall causes the die to move down. This has the same effect as an upward-moving lower punch. After pressing, the die moves upward to the fill position, and the upward-moving lower punch ejects the part. The core rod can be fixed or floating.

Fig. 9 Compacting sequence utilizing floating die tooling. Dashed lines indicate motion of component parts.

Withdrawal tooling systems use the floating die principle, except that the punch forming the bottommost level of the part remains stationary and that the die motion is press activated rather than friction activated. The die and other lower tooling members, including auxiliary lower punches and core rods, move downward from the time pressing begins until ejection is complete.

Figure 10 shows the compacting sequence in a multiple-motion withdrawal tooling system. During compaction, all elements of the tooling system except the stationary punch move downward. The die is mounted on the top press member of the platen and is supported by pneumatic or hydraulic cylinders. Auxiliary punches are mounted on additional platens, which are similarly supported and have positive pressing stops. The stops control the finished length of each of the levels within the compacted part. Before ejection, these stops are released or disengaged so that the platens can be moved further downward. During ejection, the upper punch moves upward, away from the compact, while the die and lower punches move sequentially downward until all tool members are level with the top of the stationary punch. The compact is fully supported by the tooling members during ejection, resting on the stationary punch as the die and lower punches are lowered to release it.

Fig. 10 Compacting sequence utilizing floating die withdrawal double-action tooling. Dashed lines indicate motion of component parts.

The core rod can be provided with pressing position stops to allow a part to be produced with blind or counterbored holes. The core rod is held stationary until the part is free of all other tooling members before moving downward to the ejection position. At this point in the machine cycle, the feeder moves across the die, pushing the compacted part from the die area and covering the die cavity. The die and auxiliary lower punches move upward to their respective fill positions. The core rod then moves upward, displacing the excess powder into the partially empty feed shoe. The feeder retracts, wipes the top fill level, and readies the press for the next cycle.

Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Types of Presses Anvil presses generally are limited to compaction of class I parts in a single direction. Anvil presses do not have an upper punch; a moveable, solid, flat block seals the top of the die. Compacting is done by the lower punch, which, after the anvil is released and moved, moves farther to eject the compact from the die.

Anvil presses are available with pressing capacities ranging from 6.7 to 310 kN (0.75 to 35 tons), with maximum depth of fill ranging from 1 to 75 mm (0.040 to 3 in.). Multiple-cavity pressing frequently is used in anvil presses, with possible production rates of >100,000 pieces per hour. Some anvil presses can be converted to double action, using an upper punch entry system. Anvil presses usually are mechanically driven. Figure 11 shows a schematic of an anvil press operation.

Fig. 11 Compacting sequence utilizing sliding anvil single-action tooling. Dashed line indicates motion of component parts.

Rotary presses generally are limited to compaction of single-level class II parts, although some class III parts, such as

flanged bushings, are produced. Rotary machines are available with pressing capacities ranging from 36 to 590 kN (4 to 66 tons), with a depth of fill up to 75 mm (3 in.). Production rates of >60,000 pieces per hour are possible, depending on machine size and the number of tooling stations. Rotary presses are mechanically driven. Single-Punch Opposing Ram Presses. Like rotary presses, these machines are limited to production of class II and some class III P/M parts. These presses are available in top- and bottom-drive models, with pressing capacities ranging from 36 to 980 kN (4 to 100 tons) and with a maximum depth of fill up to 100 mm (4 in.).

Production rates of up to 3000 parts per hour are possible using mechanical presses with single-cavity tooling, although production rates of 900 to 1800 pieces per hour are more common. Hydraulic presses produce 900 pieces per hour. Ejection of the part is accomplished by the lower punch moving upward. Mechanical and hydraulic presses are available. Single-punch withdrawal presses have essentially the same partmaking capabilities as the single-punch opposing

ram system in terms of pressing capacity, depth of fill, and production rate. The major difference is that floating dies are used to achieve top and bottom pressing. The die is moved downward to eject the part.

Multiple-motion die set presses can be designed to produce the most complex P/M parts. These presses use floating

die and withdrawal tooling methods. Machines are available with either bottom- or top-drive arrangements. Pressing capacities range from 27 to 7830 kN (3 to 880 tons), with a maximum depth fill of 180 mm (7 in.). Production rates vary from more than 6000 pieces per hour on smaller machines to 1800 pieces per hour for 1960 kN (220 ton) presses. In addition to producing complex parts, the removable die set (tool holder) minimizes press downtime for part changeover if the die set for the next part to be produced is set up outside the press and is ready for installation. Pressing position for each level being produced by a separate tooling member is controlled by fixed-height tooling blocks (stop blocks), which usually are ground to the proper height to produce a given dimension on the part. A small adjustment in the block mounting member allows for minor changes to part dimension. Full range adjustments are available on more recent presses. Multiple-motion adjustable stop presses have the same partmaking capability as multiple-motion die set presses

and use the same tooling methods. Pressing capacities range from 980 to 7340 kN (110 to 825 tons), with a maximum depth of fill of 150 mm (6 in.). These presses do not incorporate removable die sets; however, press stop positions are adjustable, and a change in any dimension of the part in the direction of pressing is easily accomplished. Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Advanced Tool Motions A common limitation of some rigid tooling systems is that part features not perpendicular to the direction of pressing cannot be compacted and stripped. Frequently, it is cost effective to form features such as cross holes and threads by machining. Other nonperpendicular features, notably helix shapes and hidden flanges, can be formed using complex tool motions. Another type of advanced tooling system permits production of complex shapes with magnetic orientation of the microstructure. Helical shapes, typically helical spur gears, are produced in rigid compaction tool sets with punch rotation capability.

In a simple system, a helical form lower punch is engaged in a die with a matching gear form. In such a system, the lower punch remains engaged in the die at all times, as is common practice for all rigid tool systems, so that indexing rotation of the punch to the die is avoided. The die acts as a guide. Rotation is carried out on a thrust bearing, which rests on the punch platen that supports the lower punch. An upper punch is not required, because the top of the die cavity is closed by an upper anvil, which does not enter the die cavity. Central core rods, with or without additional features such as splines and key forms, are commonly operated in this helical tool system. Helical gears made in this manner are limited to helix angles of 25° and a thickness of 32 mm (1 in.) due to fill limitations along the helix tooth form. More complex helical gear tooling systems have been developed for routine production using helical upper punches, driven by follower cams for indexed die entry, with inner and outer lower helical punches for stepped helical gears. Split Die Systems. Another rigid tooling system that avoids some through-cavity limitations is known as the split die,

or "double die," system. It enables the compaction of parts with completely asymmetric upper and lower sections in the pressing direction. Figure 12 shows typical tool motions in split die compaction. This system requires two die-holding platens to carry the upper and lower die. Each platen is controlled and moved independently.

Fig. 12 Split die compaction sequence

Wet magnetic compaction (Fig. 13) has enjoyed wide usage in the production of magnetically oriented ferrite

shapes. In this production process, a feed shoe is not required. Instead, the die cavity is injected with an aqueous slip (slurry) that has a high concentration of ferrite powder, with the addition of green binders as required. Typically, the die filling pressure is 35 MPa (5000 psi). By using an aqueous slip, many of the gravity die fill problems, such as attainment of uniform powder density and filling the areas that are difficult for the powder to reach, are avoided.

Fig. 13 Wet magnetic compaction. (a) Force-time diagram for magnet presses. (b) Schematic of press tool for chamber-filling method designed for withdrawal operation

Following die fill injection, an orienting magnetic field is applied to the slip, resulting in magnetic polarization of the individual ferrite particles, which remain mobile at this point. The optimal orientation of the ferrite particles directly

determines the quality of the finished permanent magnet. After magnetic orientation, the main pressing load is applied, densifying the ferrite mass and causing the suspending aqueous carrier to be expelled through drainage ports. The compact is imparted with the precision shape and dimensions of both the upper and lower dies, plus any core rods that may be inserted. The cycle is completed by separation of the press platens and ejection of the compacted ferrite shape. Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Tooling Design Traditionally, P/M tooling was designed on the basis of production experience. In simple parts, such as single-level class I and II parts, these determinations proved successful. As state-of-the-art materials and presses advanced to the production of complex, multilevel parts, the "cut-and-try" method of tool design became obsolete. The high cost of complex tooling and adapters, plus downtime to redesign and rebuild tooling, requires the partmaking system, including the press, to be carefully analyzed in terms of load, stress, and deflection. Tooling layout is required to design a suitable set of tools and to determine the physical dimensions (length and

thickness) of tooling members. A preliminary layout helps to determine fill, pressing, and ejection positions and to eliminate interference at these positions. The die space drawing supplied with every compacting press, which usually starts with the ejection position, is the basis of the tooling assembly layout. Generally, tooling members are never closer than in the ejection position, which constitutes the minimum space available to contain all components and their adapters. Die Design. Dies are commonly constructed by using inserts that are held in the die case by shrink fitting. The amount

of interference between the insert and the die case depends on the inside and outside diameter of each member and on the compacting pressure used. The powder can be considered a fluid in a closed container that transmits the compacting pressure in all directions; therefore, the die must be designed as though it were a pressure vessel with internal pressure. In actual practice, radial pressure on the die walls due to compacting rarely exceeds 50% of the compacting pressure. The interference fit of the die case and die insert should be such that the stress on the insert always remains in compression for round dies. However, for shaped dies such as gears, cams, and levers, the use of finite element analysis is the best method for accurately determining stress and deflection. In P/M tooling, the die normally controls the outer peripheral shape and size of the piece part. Typically, it is constructed from materials such as tungsten carbide or high alloy tool steels, such as T15, D2, CPM-10V, or CPM-15V with high hardness and good wear resistance. Dies are usually constructed in one or more sections and compressed into a retaining ring made of a low-alloy steel, such as AISI 4340 or 6150. Considerations in die design and material selection include initial tool cost, shear strength of the die material, and die shape. A large die may require tungsten carbide, which costs ten times as much as tool steel materials. Tungsten carbide may be the best material for a set of gear tools with a relatively steep helical angle. Sectional die construction may be required for specific shapes such as sharp corners or projections into the die cavity. Die Wall Thickness. An exact calculation of the stress on die walls is almost impossible from a practical point of view because stress distributions in the compact are extremely complicated and include variables such as part shape, particle size distribution, and other factors that affect transmission of compressive stress in the lateral direction (Ref 2). The vertical axial load can exert a horizontal force after a certain degree of consolidation has been attained. For example, when a simple shape is compacted at 400 MPa, as much as 120 MPa pressure can be exerted radially against the die walls.

If for purposes of simplification, the internal pressure is considered strictly hydrostatic in nature and the confined material is an incompressible liquid, then the die wall thickness for a cylindrical die could be determined by using Lame's formula:

where S is the maximum allowable fiber stress for the material of the die, D is the outer diameter of the die, d is the compact diameter, and p is the radial stress acting on the die wall. This is a simplification because during metal powder compaction the pressure is not hydrostatic and the material is not incompressible. Initially, the powder is compressed with a consequent reduction in the vertical height of the space filled by the powder. The compressed material begins to resemble a solid after a certain degree of compaction has been reached. Poisson's ratio is 0.3 for fully dense and isotropic steel. While this wrought form value cannot apply to powder metal, it is assumed to be applicable in the fully compacted condition. Thus, the Poisson's ratio is introduced into the previous equation, and the following modified Lame's formula is used for estimating the die wall thickness for metal powder compaction.

where = Poisson's ratio = 0.3. This formula, however, does not take into consideration that the internal pressure acting over the length of the compact is balanced by the strength of the die having a larger length. The formula does address the friction at the tooling/powder interfaces resulting in nonuniform pressure distribution in the compact. Generally speaking, the formula produces more conservative results than are necessary. The interference fit between the shrink ring and the die insert should be such that the stress on the insert always remains compressive for round dies. For shaped dies such as those used for production of gears and cams, the use of finite element analysis is the best method for accurately determining the stress and deflection. Core Rods. Basically, the core rod is an extension of the die that controls the inner peripheral shape and size of the piece

part. Tungsten carbide and M2 or M4 high-speed steels are the most common materials used for core rods. Primary factors in materials selection include wear resistance and hardness, which enable the core rod to resist the high compressive force exerted during compaction and the abrasive action sustained during part ejection. Core rods >25 mm (1 in.) in diameter or area are held to a base by mechanical means, such as a screw, while smaller core rods are held by means of silver solder or braze. Punches can perform the function of a die or a core rod and carry the full load of the compressive force required to compact the P/M part. Wear resistance and toughness are the most important factors in materials selection. The most commonly used materials are A2, D2, S7, and H13 tool steels. Dimensional control, especially in areas such as concentricity and hole-to-hole location, depends on the amount of clearance that can be maintained between the punches, die, and core rods. Clearance should be calculated for each specific range and size of part. It is important to note that thermal size changes occur during operation, primarily because of the friction created by stripping the compacted part and the speed of the pressing cycle. Punch Component Stress. Compacting powder causes compressive stress in the punch. This stress must be below the

yield strength of the punch material. Calculation of buckling stability should be made for long, thin-walled punches. Figure 14 shows the effect of axial compressive force on a tubular punch. A tubular punch is subjected to internal pressure during compacting of multilevel parts. In this case, the resulting circumferential tensile stress in the punch wall should be calculated. If the stress and accompanying deflection is excessive, tooling clearances should be designed so that when the outer punch wall expands, it is supported by the die wall before the stress reaches the yield limit (Fig. 15).

Fig. 14 Effect of compressive stress on tubular punch

Fig. 15 Tensile stresses in a tubular punch during compacting. Large arrows indicate action of powder on walls of punch.

During ejection, the punch is subjected to compressive stresses by resisting the stripping action of the die and to tensile stresses from the stripping action of punch. These stresses normally are lower than compacting stresses. Components of the punch subjected to stress include the punch clamp ring and bolts, which should resist the ejection of the punch without permanent deformation. Punch adapters are subjected to bending loads that create a tensile stress around the center hole during compacting. This stress should not exceed the fatigue limit of the adapter material.

Tubular adapters must have sufficient cross-sectional area to withstand the pressing load without permanent deformation. A stepped core rod, or a core rod forming a blind hole, must not buckle during compacting. The base of the core rod must resist, without permanent deformation, whatever ejection loads are imposed on the core rod. The core rod clamp ring and retaining bolts should be sized to withstand the ejection force on the core rod without permanent deformation. The core rod adapter generally is strong enough to resist both pressing and ejection loads, due to the size of the adapter when space is provided for clamp ring fasteners. Deflection Analysis. Pressing of P/M parts at pressures >690 MPa (50 tsi) presents unique considerations for size and

tolerance in multilevel parts. A variety of tool members should be utilized to establish proper fill ratios, and deflection and springback can occur. Deflection occurs because of the column loading effect on the compacting tools during the briquetting cycle. For column load consideration, the bottom section of the lower punch is considered fixed, while the top section or working end of the lower punch can be considered free to rotate. The amount of deflection on the tool member will be determined by the column slenderness ratio of the punch and the adapter. When the column load is released after the press goes through the bottom dead center compaction point, the deflected punches will return to their original lengths, if their elastic material property limits have not been exceeded. This return movement is generally called springback and can be deleterious to the green part, depending on the fragileness of the green part section geometry involved. Deflection can be minimized by strengthening the various tool members through changes in physical size or shape and/or by changes in material selection. The most common method of minimizing deflection effects is to equalize deflection using tool members and adapters that are designed to match the deflection characteristics of the most critical member. The ability of the tool designer to find the proper balance is paramount for production of crack-free parts. When designing tools for production of parts other than single-level class I or II parts, deflection analysis of the tooling, tooling adapters, and press is desirable. These members are essentially stiff springs, each with a different spring rate or modulus. When the compacting load is applied, the parts deflect. When the load is released, they return to their original length. If the press contains two or more separate lower punches, the total deflection of each punch and the supporting members must be the same. Otherwise, the compacted part will move with the punch that has the greatest total deflection, leaving a portion of the part unsupported. This condition is likely to cause cracking during part ejection. A punch under load normally is in pure compression and therefore will follow Hooke's law. If the punch has varying cross-sectional areas, each length having the same cross-sectional area is calculated individually. The total punch compression is the sum of these calculations. For a long, thin-walled punch, local buckling of the punch wall under load should be investigated. Compression of punches and their supporting members may be calculated using the equation given in Fig. 16.

Fig. 16 Punch compression. P is total punch load, L is length, Y is deflection, A is area of punch, and E is Young's modulus.

Adapter Bending. The adapter, on which the punch is mounted, usually is a flat plate with the punch and the load

positioned at the center, around a hole through which either another punch or a core rod passes. This plate, if supported at the outer edge, is subjected to the pressing load around the center hole. Two forms of deflection--bending and shearing-occur in this area. Adapter deflection is linearly proportional to force. Calculated adapter stress should be compared with the allowable adapter material stress to evaluate design suitability. Press Deflection. Like the tooling and support, the press is subject to deflection. This tendency is considerably less

than that of punch compression or adapter bending, but it must be considered in total tool design. Data regarding press deflection should be obtained from the press manufacturer. Deflection is linearly proportional to the amount of force exerted:

Y=C×W where C is the equipment constant, W is pressing force, and Y is deflection.

Reference cited in this section

2. S.D.K. Saheb and K. Gopinath, Tooling for Powder Metallurgy Gears, Powder Metall. Sci. Technol., Vol 2 (No. 3), 1991, p 25-42 Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Tool Materials Dies. In the most common type of die construction, wear-resistant inserts or liners are held in place by clamping or

shrink fitting. The amount of interference between the insert and the die case depends on the inside and outside diameter of each member and on the compacting pressure used. The powder can be considered a fluid in a closed container that transmits the compacting pressure in all directions; therefore, the die must be designed as though it were a pressure vessel with internal pressure. In actual practice, radial pressure on the die walls due to compacting rarely exceeds 50% of compacting pressure. The interference fit of the die case and die insert should be such that the stress on the insert always remains in compression for round dies. However, for shaped dies, such as gears, cams, and levers, finite element analysis is the best method for accurately determining stress and deflection. Die inserts for compaction of carbide, ceramic, or ferrite powder most frequently are the medium- or coarse-grain 94WC6Co grades of cemented carbide. Cemented tungsten carbide containing 12 to 16% Co can be used to make inserts for compacting metal powders in medium-to-long production runs. The elastic moduli of carbides are considerably higher than those of steels, a fact that should be considered when designing composite steel and carbide die assemblies. Because carbide will deflect only 33 to 40% as much as steel, the steel portion generally should be designed with enough stiffness to support three times the expected loading in order to match the deflection of the carbide. The shrink-fit allowance should be 1.0 mm/m (0.0010 in./in.). Shrink rings and similar supporting parts of the tooling can be made from medium-carbon alloy steel, such as AISI/SAE 4340 or 6150, quenched and tempered to 42 to 46 HRC. It is especially important that supporting parts for carbide tools provide sufficient support; otherwise, the carbide tools are likely to break in service. Cemented carbides are relatively expensive, and shaping of parts to the required form must be done either by electrical discharge machining or by specialized methods of grinding.

Wear-resistant tool steel inserts are sometimes used instead of carbide inserts. Tool steel inserts are tougher and easier to fabricate than carbide inserts. Powder metallurgy tool steels such as CPM 10V, Vanadis 4, and Vanadis 10 are frequently chosen for medium-to-long production runs because they have wear resistance approaching that of carbides. Other wearresistant tool steels, usually D2 or D3 or a high-speed steel such as M2 or M4, have been used for short-run applications. Tool steel inserts generally are heat treated to a working hardness of 62 to 64 HRC. For increased wear resistance, a nitrided case may be specified for dies made of CPM 10V, Vanadis V, Vanadis 10, or D2. For certain part designs, a solid die rather than an insert die is a more practical choice; an air-hardening 5% Cr tool steel such as A2 is generally used for such applications. Punches. The stresses imposed on punches during service are such that toughness is a much more important material requirement than wear resistance, although wear resistance cannot be ignored. Type A2, and sometimes the shockresisting type S7, are preferred for punches. Wear-resisting grades such as D2 and CPM 10V often lack the required toughness, particularly for solid punches. In type A2, which is deep hardening upon air quenching, an as-quenched hardness of 60 HRC can be developed in the center of a section 125 mm (5 in.) square, even though a solid punch this large would seldom be used. In contrast, for type S7 the maximum section size in which such hardness can be obtained is

65 mm (2 in.) square. If sections >125 mm (5 in.) square are required, type A2 should be oil quenched from the austenitizing temperature to 540 °C (1000 °F), then air quenched to 65 °C (150 °F) before tempering. S7 can be carburized or nitrided for added wear resistance. For applications in which A2 or S7 punch faces become severely abraded, a more wear-resistant grade, such as CPM 10V, D2, D3, or M2, should be considered. Cemented carbide punches and core rods employ a higher cobalt grade ( 11% Co) with a hardness of 90 HRA. They can be made of solid carbide or a composite that uses tungsten carbide in the wear areas. More recently, fine-grain carbides with 10 wt% Co have also been employed in these applications. Core Rods. Both toughness and wear resistance are important criteria in the selection of core-rod materials, but

generally the primary consideration is wear resistance. Tungsten carbide and high-speed steels (M-grades) are the most common materials for core rods. For particularly abrasive conditions, CPM 10V has been used successfully, as have D2, M2, and A2 tool steels that have been nitrided or coated with tungsten carbide. Crucible P/M (CPM) tool steel CPM 15V has recently become popular for these applications. Tooling support adapters normally are made from medium-carbon alloy steel, such as AISI/SAE 4140 or 6150, heat

treated to a hardness sufficient to resist brinelling of the punches into the adapter surface without failing due to fatigue. Adapters should be heat treated to a minimum hardness of 28 to 32 HRC to reduce damage to critical mounting surfaces during handling. Punch clamp rings normally are not highly stressed members, but they should be made from a heat-treatable steel to

prevent damage during handling. Heat treating of the clamp ring is optional. Operational Factors. Die working surfaces and core rods should be polished or lapped to a mirror-like surface finish,

and final polishing should be done in a direction parallel to the axis of the tool. The faces and lands of the punches should also be given a fine finish. An exceptionally smooth surface finish reduces friction, thereby reducing some of the load on the tooling. It also makes it easier to eject the compacts, and it eliminates minute scratches and other stress raisers that could lead to premature fatigue failure. Hard chromium plating is sometimes recommended to improve the life of tool steel punches and core rods, particularly when abrasive powders are involved. Some users claim that nitrided or chromium-plated die parts have up to ten times the wear resistance of untreated tool steel die parts; others claim that chromium plating is not very effective. Both nitrided and chromium-plated die parts are subject to chipping or flaking, especially at sharp edges. When this is a problem, a diffused surface layer such as that produced by chromizing may prove to be an effective alternative.

Powder Metallurgy Presses and Tooling Revised by John Porter, Cincinnati Incorporated

Tooling Clearances and Design As in many other manufacturing operations, process variables (e.g., the type of materials being processed, the density of the part being produced, the amount and type of powder or die lubrication, and production rate) dictate operating conditions. Density and production rate greatly affect tool clearances during a continuous production run in which tooling temperature increases as compacted density and/or production rate increases. Temperature variations and corresponding dimensional changes within the various tooling members must be considered. Standard tooling clearance is 0.016 mm/25 mm (0.0006 in./1 in.) on the diameter total. Minimum clearance should be used initially, because materials can always be removed from the punch or die to provide additional clearance as needed. In addition, the clearances must be smaller than the size of the powder particles to prevent their entrapment. Smaller clearances will also help reduce possible variations in the dimensions of the parts. The density of the compact produced and the production rate have a great influence on the determination of clearances. Compacting loads will be higher for increased densities. These, as well as higher production rates, increase the tooling temperature. As the temperature increases, dimensional changes occur in the tooling members. For this reason, the clearances must be sufficiently large to prevent seizure of the tools. A representative value for the clearances between die walls and punches is 0.005 to 0.008 mm for precision parts and an upper limit of 0.013 mm for other parts. Minimum possible clearance should be used initially, because it can be increased as needed by removing material from the punch or the die. The expansion ("pop out" or "springback") of the compact upon ejection makes it essential that the top edge of the die cavity be properly rounded or flared to allow the compact to make a smooth transition during ejection. Provision of a shallow chamfer is a more practical solution in the case of gears. Die cavities and punch faces should be lapped and polished to a very high degree of surface finish, preferably 775 >600 >600 >600

Mean torque, in. · lbf 16.56 13.57 7.82 14.3 15.58

Mean thrust, lbf 276 244 142.3 186.8 177.7

Wear at 500 holes, 0.001 in. 29.1 26.5 14.2 26.5 19.1

The machinability of the FC-0205 plus MnS is significantly better than that of the FC-0205 and appears comparable to that of the wrought 1215 and 12L14. Drilling the FC-0205 plus MnS requires similar torque but higher thrust than the wrought steels. It appears to produce similar tool wear to that of the 12L14 but somewhat higher wear than the wrought 1215. The test shows under the drilling conditions chosen that the 1215 can possess slightly better machinability than the 12L14. The test also shows that by evaluating free-machining agents and machining conditions, the machinability of a P/M steel can be improved to that of a wrought steel.

References cited in this section

15. P.J. James, Factors Affecting Quality of Drilled Holes in Sintered Steels, Powder Metall., Vol 37 (No. 2), p 133

16. Machinery Handbook, Industrial Press, 1984, p 1803 17. S. Berg, Machinability of Sintered Steels: Guidelines for Turning, Drilling, and Tapping, Advances in Powder Metallurgy and Particulate Technology, Vol 2, Metal Powder Industries Federation, 1997, p 15-145 18. G.T. Smith, Surface Integrity Aspects of Machinability Testing of Fe-C-Cu Powder Metallurgy Components, Powder Metall., Vol 3 (No. 2), 1990, p 157 19. I. Sharif and K. Boswell, Prediction and Modelling of Surface Finish in Drilling of P/M Parts, Advances in Powder Metallurgy and Particulate Technology, Vol 2, Metal Powder Industries Federation, 1997, p 15-155 20. Y.T. Chen et al., Free-Machining P/M Alloy Optimization Using Statistical Analysis Techniques--The Effect of MnS Content and Particle Size, Advances in Powder Metallurgy and Particulate Materials, Vol 4, Metal Powder Industries Federation, 1992, p 269 Machinability of P/M Steels R.J. Causton and T. Cimino, Hoeganaes Corporation

Machinability Improvement Machining P/M steels does present problems. Several different approaches to improve machinability are: • • • • • • •

Closure of porosity Green machining Presintering Microcleanliness improvement Free-machining additives Microstructure modification Tool materials

The effects of free-machining additives, microstructure modification, and tool materials are illustrated by controlled drilling tests conducted under laboratory conditions. Closure of Porosity. Closing or sealing porosity improves the machinability of P/M steels significantly by changing the

cutting process from intermittent to continuous. The reduction in vibration and chatter improves tool life and surface finish. Copper infiltration (Ref 8) and polymer impregnation (Ref 19) are efficient means to close porosity and can require an additional process step. Thus, they are most efficient when dictated by the end use, such as fluid power applications, that require a pore-free structure. However, the improvement in machinability can justify their use in severe machining operations or when a machining operation is the rate-limiting step in a process sequence. Microcleanliness Improvement. The increase in the production and use of atomized rather than reduced iron powders has improved the microcleanliness of iron and low-alloy steel powders. Driven largely by the requirements of powder forging, the content of coarse nonmetallic inclusions in atomized powders has been reduced significantly (Ref 21). For an atomized FL-4600, the median frequency of inclusions greater than 100 m in size (F4) has been reduced from approximately 2.5 to 0.25 per 100 mm2. The maximum frequency of inclusions greater than 100 m was reduced from 9 to 1.3 inclusions per 100 mm2. These improvements suggest that the incidence of edge damage due to the presence of coarse inclusions should be reduced significantly. Because powder forging practices are now employed to produce all atomized steel powders, P/M users of these powders have benefited. Green Machining. One way to reduce the machining problems of P/M parts is to machine them in the green (i.e., aspressed) condition prior to sintering. The lack of bonding between particles in green compacts results in low cutting forces.

Such techniques are used in the processing of ceramic and hard metal powders. However, the green strength of metal powder compacts has been too low to withstand the cutting and clamping forces employed in machining operations. The introduction of warm compaction technology (Ref 22) can change this perspective. The green strength of warm compacted parts is two to four times higher than that of conventional ferrous P/M parts (Table 5). This is sufficient to withstand both the cutting and clamping forces of modern machine tools. Research confirms (Ref 23) that green compacts produced with warm compaction can be machined with conventional cutting tools with low cutting forces. Drill testing (Table 6) shows that the cutting forces are relatively low. Both cutting forces and surface finish can be improved by changes to drill type and profile. These changes also alter the accuracy and surface finish of the drilled hole. Thus, the choice of tool will be a compromise between low cutting forces, surface finish, and tolerances.

Table 5 Green strength of ANCORDENSE premixes Material M-1 M-2 M-3 M-4 M-5

Green density, g/cm3 7.29 7.33 7.37 7.31 7.15

Green strength, psi 4221 4800 7703 9454 6286

Table 6 Mean drilling forces for warm compacted test pieces Drill type 118° parabolic geometry 135° split point 135° split point-wide land parabolic flute 135° split point-wide land parabolic flute, coated

Mean force, lbf 100.4 45.3 49.3 55.5

Material: Ancorsteel 85HP, 2% Ni, 0.4% graphite; green density, 7.33 g/cm3. Machining: 0.375 in. HSS drill; speed, 3285 rpm; feed, 0.012in./revolution Presintering the green compact at a lower temperature than the final sintering operation produces a compact of relatively

low hardness and strength but with sufficient edge retention to be handled and machined. Thus, the machinability of a presintered compact can be substantially better than that of the sintered part. However, presintering introduces an additional step to the manufacturing process and increases cost. It can be justified where the properties of the as-sintered part make its machining difficult or impossible. For example, high-carbon sinter hardened steels can require grinding rather than machining. In this case, providing that the part application and tolerances permit, it can be desirable to machine the part in the presintered condition rather than perform a grinding operation. Similarly, if part design calls for a through hole normal to the compaction axis, drilling the presintered preform followed by final sintering can be the only way to produce the hole economically in a high performance P/M steel. Free-machining agents are added to P/M steel to improve machinability (Ref 8, 24). These agents are thought to perform

several functions during the cutting process (Ref 9), including initiation of microcracks at the chip/workpiece interface, chip formation, lubrication of the tool/chip interface, and prevention of adhesion between the tool and chips (Fig. 5).

Fig. 5 Potential benefits of a machining agent

Several materials including sulfur, molybdenum disulfide (Ref 24), manganese sulfide (Ref 11), and boron nitride (Ref 25) are used as free-machining agents for P/M steels. They are most frequently introduced as fine powder to powder premixes, but sulfur and manganese sulfide are also available as prealloyed powders (Ref 26, 27, 28). Sulfur and molybdenum disulfide can have strong effects upon the dimensional change and strength of P/M steels (Fig. 6, 7). Their use should be considered at the part design stage rather than as a "retrofit" when machining problems become apparent. Manganese sulfide has smaller effects upon dimensional change and strength (Ref 13) and can be used to improve the machinability of existing premixes. The effects of several potential machining agents upon the machinability of P/M steels are described below.

Fig. 6 Dimensional change of F-0008 atomized plus free-machining agents

Fig. 7 Transverse rupture strength of F-0008 atomized plus free-machining steels

References cited in this section

8. H. Chandler, Machining of Powder Metallurgy Materials, Vol 16, Metals Handbook, 9th ed., ASM International, 1989, p 879-892 9. S.A. Kvist, Turning and Drilling of Some Sintered Steels, Powder Metall., Vol 12 (No. 24), 1969 11. D.S. Madan, An Update on the Use of Manganese Sulfide (MnS) Powder in Powder Metallurgy Applications, Advances in Powder Metallurgy, Vol 3, Metal Powder Industries Federation, 1991, p 101 13. J.A. Hamill, R.J. Causton, and S.O. Shah, High Performance Materials for P/M Utilizing High Temperature Sintering, Advances in Powder Metallurgy and Particulate Materials, Vol 5, Metal Powder Industries Federation, 1992, p 193-213 19. I. Sharif and K. Boswell, Prediction and Modelling of Surface Finish in Drilling of P/M Parts, Advances in Powder Metallurgy and Particulate Technology, Vol 2, Metal Powder Industries Federation, 1997, p 15-155 21. R.J. Causton, Machinability of P/M Steels, Advances in Powder Metallurgy and Particulate Materials, Vol 2, Metal Powder Industries Federation, 1995, p 8-149 22. S. Luk, Metal Powder Compositions Containing a Binder Agent for Elevated Temperature Compaction, U.S. Patent 5,154,881, 13 Oct 1992 23. T.M. Cimino and S.H. Luk, Machinability Evaluation of Selected High Green Strength P/M Materials, Advances in Powder Metallurgy and Particulate Materials, Vol 2, Metal Powder Industries Federation, 1995, p 8-129 24. U. Engstrom, Machinability of Sintered Steels, Prog. Powder Metall., Vol 38, 1982, p 417 25. M. Gagne, Sulfur Free Iron Powder Machinable Grade, Advances in Powder Metallurgy, Vol 3, Metal Powder Industries Federation, 1991, p 101 26. L.G. Roy et al., Prealloyed Powders for Improved Machinability in PM Parts, Met. Powder Rep., Feb 1989 27. S. Hironori et al., 250 MSA Resulfurized High Green Strength Steel Powder, Advances in Powder Metallurgy and Particulate Materials, Vol 4, Metal Powder Industries Federation, 1997, p 15-27 28. R.J. Causton, T.M. Cimino, and H.M. Scanlon, Machinability Improvement of P/M Steels, Advances in Powder

Metallurgy and Particulate Materials, Vol 7, Metal Powder Industries Federation, 1994, p 7-169 Machinability of P/M Steels R.J. Causton and T. Cimino, Hoeganaes Corporation

Sulfides Sulfides are probably the most frequently used free-machining agents in both wrought and P/M steels. Powder metallurgy offers more flexibility than wrought metallurgy. Sulfur can be prealloyed in the powder during the primary production process or admixed as sulfur or sulfides during the preparation of a press-ready powder premix. Premixing offers more flexibility in the composition and amount of sulfide formed in the final compact. Prealloy produces a somewhat finer dispersion of sulfides within the powder particles. Sulfur Prealloys. Several powder producers have supplied prealloyed or resulfurized powders, where sulfur is introduced during the primary powder production process. More recently, such powders have become a niche product to meet specific market needs, and admixed manganese sulfide has become a more widely used free-machining agent. It is possible that the powder manufacturing process offers a risk of cross contamination between resulfurized and nonresulfurized grades. Such contamination would introduce undesirable and easily detected sulfide inclusions into other high powder products rendering them unacceptable for high performance applications. Cross contamination can be most easily and efficiently minimized by introducing the sulfur or sulfide as late as possible in the premix stage.

Despite these problems, several powder producers (Ref 26, 27) offer prealloyed or resulfurized sulfur powders. These can offer significant improvements in machinability over nonresulfurized powders, at some loss in compressibility due to the solution-hardening effects of sulfur in iron. By using a resulfurized powder for an F-0008 composition, Ancorsteel 1000M drill life was increased by about 50% compared to a similar composition with no free-machining additives: • • • • •

Ancorsteel 1000M: 128 holes to failure Ancorsteel 1000: 83 holes to failure Compaction: 6.8 g/cm3 Sintered: 1121 °C (2050 °F) Cutting conditions: 0.125 in. HSS drill, 3000 rpm, 0.003 in./revolution

One result (Table 7) shows that controlling both manganese and sulfur content of the prealloyed powder improves the machinability of iron powder indicated by the time required to drill holes and the number of holes drilled before failure. The use of resulfurized powders can be extended to reduced or sponge iron powders with similar beneficial effects upon drill life (Table 8).

Table 7 Machinability of resulfurized iron powders Premix Base powder Time for 25 holes, s Holes to failure Manganese, % Sulfur, %

F-0008 MP 36S 6.2 203 0.38 0.38

MP 37 8.1 82 ... ...

MP 35 14.52 25 0.94 0.236

FC-0208 MP 36S 9.3 40 0.38 0.38

MP 37 9.9 32 ... ...

MP 35 17.7 21 0.94 0.236

Test piece compacted to 6.6 g/cm3. Sintered at 1121 °C (2050 °F). Cutting conditions: 0.25 in. HSS drill, 2300 rpm, 154 pound point loading. Source: Ref 26

Table 8 Effect of 0.5% machining agents on drill life Condition

Sponge Atomized

Drill life, holes to failure No agent Sulfur Manganese sulfide 19 158 33 83 418 890

Molybdenum disulfide 285 718

Compacted to 6.8 g/cm3 and sintered at 1121 °C (2050 °F) Admixed Sulfides. Many different sulfides have been evaluated as free-machining additives for P/M steels. The additives

most frequently used are manganese sulfide, sulfur, and molybdenum disulfide. Manganese sulfide made as an 0.35 to 0.6% addition to a premix is the most widely used. Additives usually take the form of fine high purity powders less than 50 m in particle size as measured by laser particle size analysis (Fig. 8).

Fig. 8 Cumulative particle size distribution of free-machining agents

Effect of Sulfides upon Drill Life. The effect of sulfide free-machining agents upon drill life at an 0.5% addition is

compared for three widely used premixes: F-0008, FC-0208, and FN-0205 in the as-sintered condition. Machinability of F-0008. The machining agents improve drill life, indicated by holes completed before drill failure, when

added to F-0008: • • • •

Powder MH 1024: 45 holes to failure Powder MH 100: 22 holes to failure Compaction: 6.8 g/cm3 Sintered: 1121 °C (2050 °F)

•

Cutting conditions: 0.125 in. HSS drill, 3000 rpm, 0.003 in./revolution

In the mixes made with atomized powder, all the machining additives increase drill life significantly. Manganese sulfide and molybdenum disulfide produce larger increases than sulfur. The improvements are not as large in the materials made with sponge powders where molybdenum disulfide produces the best drill life. Surprisingly, 0.5% manganese sulfide does not perform as well in the sponge test mixes in this sequence of tests. Although drill life is the chosen indicator of performance, the machining agents produce other changes in cutting performance that could influence production use. During the drill test, it was apparent that different agents changed chip form; for example, molybdenum disulfide produces very small chips that are easily removed from the cutting area. The machining agents appear to influence heat transfer to and plastic deformation of the workpiece during the test. As the drill accumulates heat during the test, some materials show local plastic deformation around the cutting area. In extreme cases, the cutting forces push material through the unsupported back of the workpiece, especially in mixes with sulfur and least apparent in molybdenum disulfide. Such extreme behavior is unacceptable in production. The changes in hole diameter that precede it are a significant factor in determining tool life. Machinability of FC-0208. The free-machining additives all improve the machinability of the FC-0208 (Table 9). The

0.5% sulfur addition produces the largest improvement in both atomized and sponge compositions. The drills do not fail before all test material is consumed. Molybdenum disulfide is more effective than manganese sulfide under the chosen test conditions.

Table 9 Effect of 0.5% machining agents on drill life Condition

Sponge Atomized

Drill life, holes to failure No agent Sulfur Manganese sulfide 2 608+ 81 2 608+ 72

Molybdenum disulfide 108 249

Compacted to 6.8 g/cm3 and sintered at 1121 °C (2050 °F) Effect of Sulfides. The drill test results show that the 0.5% sulfide addition improves machinability significantly.

However, the addition of sulfides can change sintered properties, particularly dimensional change. The sulfides are not completely inert during the sintering process and can modify the sintering reactions.

References cited in this section

26. L.G. Roy et al., Prealloyed Powders for Improved Machinability in PM Parts, Met. Powder Rep., Feb 1989 27. S. Hironori et al., 250 MSA Resulfurized High Green Strength Steel Powder, Advances in Powder Metallurgy and Particulate Materials, Vol 4, Metal Powder Industries Federation, 1997, p 15-27 Machinability of P/M Steels R.J. Causton and T. Cimino, Hoeganaes Corporation

Metallography

Free-machining agents have different effects upon the microstructures of the test premixes. These effects depend upon the nature of the base iron (sponge or atomized) and the premix composition (Ref 29) and are illustrated with reference to the F0008 composition in Fig. 6 and 9. Sulfur appears to promote sintering of the test compositions and produces very round pores. A portion of the sulfur dissolves in the iron matrix and diffuses a short distance into the iron matrix. In sponge iron compositions, a portion of sulfur reacts with iron to form iron sulfides. Sulfur also appears to promote pore rounding during sintering.

Fig. 9 Dimensional change of F-0008 sponge plus free-machining agents

Manganese sulfide (Ref 6) is considered to be stable in both iron-graphite and iron-copper-graphite premixes. Metallography indicates that the manganese sulfide occurs in pores or the fine pores remaining at prior particle boundaries. It appears to be almost inert during the sintering process with little evidence of diffusion into the iron matrix in the compositions examined. The effects of molybdenum disulfide appear to be between those of sulfur and manganese sulfide and to depend upon the amount added to the test composition. It appears that a significant portion of the molybdenum disulfide reacts with the sintering atmosphere. At low MoS2 additions, almost all of the addition transforms to molybdenum, which remains within pores and at particle boundaries. Some sulfur evaporates during the sintering process and some dissolves in the iron matrix and can form sulfides on cooling from sintering temperature. At higher sulfur additions, it appears that an equilibrium is reached between the sulfur present as molybdenum disulfide and that in the particles, so that more typical slight pore rounding occurs.

References cited in this section

6. R. Koos and G. Bockstiegel, The Influence of Heat Treatment, Inclusions, and Porosity on the Machinability of Powder Forged Steels, Prog. Powder Metall., Vol 37, 1981, p 145-147 29. D. Madan and A. Fitzgibbon, Shelf Life of MnS Powder and MnS Containing Premixes, Advances in Powder Metallurgy and Particulate Materials, Vol 2, Metal Powder Industries Federation, 1995, p 8-177

Machinability of P/M Steels R.J. Causton and T. Cimino, Hoeganaes Corporation

Stability of Sulfides Sintering. Neither sulfur nor sulfides are completely stable during sintering. Comparison of the sulfur content of the F-0008

premix ingredients to that of the sintered test pieces shows that the sulfur recovery was lower than anticipated (Table 10).

Table 10 Sulfur recovery for sintered F-0008 sponge Addition, % 0 0.25 0.50 0.75 1.00

Sulfur, % Added Measured 0 0.01 0.25 0.17 0.50 0.41 0.75 0.60 1.00 0.76

Manganese sulfide, % Added Measured 0 0.01 0.09 0.11 0.18 0.18 0.28 0.26 0.37 0.35

Molybdenum disulfide, % Added Measured 0 0.01 0.10 0.04 0.20 0.12 0.30 0.18 0.40 0.25

The data show that manganese sulfide is stable under endothermic atmosphere sintering. The measured and predicted sulfur contents of the manganese sulfide mix agree well. As anticipated, a significant portion of the elemental sulfur is lost during sintering. The measured sulfur content in the molybdenum sulfide premix is also less than anticipated confirming the metallographic findings that a portion of the molybdenum disulfide decomposes to release sulfur to the iron matrix and sintering atmosphere. Similar trends were observed in the atomized F-0008 and FC-0208 compositions. These results indicate that a portion of the sulfur can be released to the sintering atmosphere. It is possible that sulfur in the atmosphere could be absorbed by other compositions present in a production sintering furnace. Given the sensitivity of physical properties to sulfur content, it appears that considerable care should be exercised in furnace loading, atmosphere, and scheduling to avoid the possibility of contaminating a furnace or subsequent premix composition. Stability during Handling. Manganese sulfide is hygroscopic. Several authors (Ref 29) have stressed that fine manganese powders should be kept in closed containers to prevent moisture absorption and oxidation of the manganese sulfide. If simple precautions are taken (e.g., closing containers and rapidly consuming all material in an open container), the sulfide is stable and provides very consistent results. Stability during Machining. The potential for reactions between moisture and manganese sulfide can extend to the

machining process. The majority of machining tests of P/M parts are conducted in the dry condition without a cutting fluid or lubricant. In contrast, many high volume machining operations employ a water-based cutting fluid for cooling and removing chips from the cutting area. In machining some lower density P/M components on high-speed transfer lines, reactions between water-based cutting fluids and manganese sulfides occur with detrimental results upon the machining operations (Ref 30).

References cited in this section

29. D. Madan and A. Fitzgibbon, Shelf Life of MnS Powder and MnS Containing Premixes, Advances in Powder Metallurgy and Particulate Materials, Vol 2, Metal Powder Industries Federation, 1995, p 8-177 30. O. Petterson, High Speed Turning and Boring of PM Carbon Steel, Paper 980629, Society of Automotive

Engineers, 1998 Machinability of P/M Steels R.J. Causton and T. Cimino, Hoeganaes Corporation

Effects upon Sintered Properties The experiments indicate that free-machining agents influence sintered microstructure. Thus, they will also influence sintered properties. The effects depend upon the machining agent, premix composition, and the sintering atmosphere. Generally, sulfur and molybdenum disulfide promote growth on sintering. In contrast, manganese sulfide has less effect and could slightly reduce growth. In general, free-machining agents improve the strength of F-0008 compositions but reduce that of FC0208 significantly. F-0008 Atomized Iron Powder. The effects of the machining agents upon the properties of F-0008 made with atomized iron powder are illustrated in Fig. 8 and 9. Increasing sulfur and molybdenum disulfide content tend to increase strength and dimensional change from die size. Increasing manganese sulfide addition from 0 to 1% tends to decrease dimensional change and strength slightly. Sulfur and manganese sulfide have almost no effect upon macrohardness at a density of 6.8 g/cm3 whereas increasing molybdenum disulfide content increases hardness. Both sulfur and molybdenum disulfide promote growth, in contrast, increasing manganese sulfide reduces growth slightly.

The effects of the machining agents upon transverse rupture stress (TRS) are slightly more complex than upon dimensional change. Increasing sulfur contents increases TRS slightly, whereas increasing manganese sulfide contents decreases TRS slightly. Adding molybdenum disulfide initially reduces TRS, but a 1% addition of MoS2 increases the TRS of the F-0008 test premixes. Both sulfur and manganese sulfide increase the hardness of F-0008 test mix slightly (Table 11). Increasing molybdenum disulfide additions increase the hardness of the F-0008 atomized rapidly.

Table 11 Effects of machining agents on hardness (HRB) of F-0008 atomized powder Additions, %

Sulfur, %

0 0.25 0.50 0.75 1.00

54 52 56 53 56

Manganese sulfide, % 53 53 53 53 55

Molybdenum disulfide, % 51 52 58 65 66

Compacted to 6.8 g/cm3 and sintered at 1121 °C (2050 °F), endothermic atmosphere, 30 min F-0008 Sponge Iron Powder. When added to F-0008 made with sponge iron, the free-machining agents had similar

effects to those observed in the F-0008 composition made with atomized powder. Manganese sulfide was relatively neutral, while both sulfur and molybdenum disulfide increased the growth on sintering of the sponge-based F-0008 significantly. Adding manganese sulfide reduced growth from die slightly, the effect of the first 0.25% addition being most significant (Fig. 11). Adding 0.25% of sulfur or molybdenum disulfide increases strength significantly (Table 12). Further increases in sulfur content do not change TRS significantly. Increasing molybdenum disulfide content to 1% increases TRS further. Manganese sulfide has no effect upon strength under the conditions tested. Adding the machining agents to F-0008 sponge causes similar changes in hardness to those observed in TRS. The first 0.25% addition of sulfur increases hardness significantly, but hardness increases slowly with further additions (Table 13). Molybdenum disulfide increases hardness significantly, but an

increase beyond 0.5% causes little increase in hardness. The hardness of the F-0008 increases only slightly with manganese sulfide additions.

Table 12 TRS of F-0008 sponge powder plus free-machining agents Agent addition, % 0 0.25 0.50 0.75 1.00

Transverse, rupture strength, ksi, with addition of: Sulfur MnS MoS 49.3 51.0 50.5 64.0 52.9 57.6 63.6 53.6 68.9 62.4 51.4 68.2 63.1 52.6 72.0

Table 13 Effect of machining agents on hardness (HRB) of F-0008 sponge Addition, %

Sulfur, %

0 0.25 0.50 0.75 1.00

24 36 35 37 40

Manganese sulfide, % 23 24 27 24 26

Molybdenum disulfide, % 22 35 44 32 46

Compacted to 6.8 g/cm3. Sintered at 1121 °C (2050 °F), endothermic atmosphere, 30 min FC-0208 Atomized. For the FC-0208 made with atomized powder, all three agents increase growth significantly (Fig. 10).

The effect of sulfur is slightly greater than manganese sulfide or molybdenum disulfide. The data indicate that additions above 0.5% cause little further increase in growth. All free-machining agents reduce the strength of the FC-0208 composition significantly (Fig. 11). The effect of manganese sulfide was less than that of sulfur and molybdenum disulfide.

Fig. 10 Dimensional change of FC-0208 atomized plus free-machining agents

Fig. 11 Transverse rupture strength of FC-0208 atomized plus free-machining agents

FC-0208 Sponge. All machining agents increase the growth of the FC-0208 premixes made with sponge powder. They

have somewhat less effect on TRS than observed in the FC-0208 premixes made with atomized powder (Table 14).

Table 14 Effect of machining agents on dimensional change of FC-0208 sponge Additive, %

0 0.25 0.50 0.75 1.00

Dimensional change, % Sulfur Manganese Molybdenum sulfide disulfide +0.27 +0.31 +0.31 +0.55 +0.35 +0.54 +0.50 +0.36 +0.44 +0.55 +0.38 +0.44 +0.56 +0.38 +0.41

Compaction of 1% zinc stearate to 6.1 g/cm3. Sintering at 1121 °C (2050 °F) endothermic atmosphere, 30 min

Sulfur additions cause the greatest increase in growth; molybdenum disulfide has a somewhat smaller effect, and manganese sulfide has the least effect. A 0.25% addition of sulfur or molybdenum disulfide causes a significant increase in growth that does not increase with further additions. The growth of FC-0208 sponge tends to increase slowly with increasing manganese sulfide content. The effects of the free-machining agents upon the TRS of the FC-0208 sponge are relatively small (Table 15). Transverse rupture strength tends to increase slightly with increasing molybdenum disulfide. It decreases slightly with increasing sulfur content. Increasing manganese sulfide causes the most significant decrease in TRS at additions above 0.5%.

Table 15 Effect of machining agents on TRS of FC-0208 sponge Transverse rupture strength, ksi Sulfur Manganese sulfide Molybdenum disulfide 84.3 82.0 83.0 78.6 82.5 80.1 80.5 86.0 78.1

80.3 79.9

75.9 73.8

83.0 86.0

Compaction of 1% zinc stearate to 6.1 g/cm3. Sintering at 1121 °C (2050 °F) endothermic atmosphere, 30 min Nonsulfide Machining Agents. In wrought metallurgy, several other free-machining agents, such as lead and selenium, are employed. The P/M industry has made little use of these, possibly due to potential toxicity problems. Powder producers and parts makers (Ref 31, 32, 33) are taking advantage of the premixing operation to introduce free-machining agents that are incompatible with the processing of wrought steels. These agents are nonmetallic solids that are anticipated to assist in chip formation and lubrication of the chip cutting tool interface to reduce wear mechanisms. Unlike sulfide machining agents, these products are proprietary to specific producers and patent protected. Enstatite is a soft mineral that shears easily under stresses but is relatively stable at sintering temperatures. K. Hayashi et al. (Ref 31) describe the use of enstatite in combination with manganese sulfide to enhance the machinability of sintered P/M steels. Boron Nitride. Hexagonal boron nitride is a recognized solid lubricant with microstructure and frictional properties similar

to those of graphite. The use of boron nitride as a free-machining agent in sintered P/M steels has been patented (Ref 32). Published data show that the addition of small quantities of boron nitride to sintered steels enhances machinability significantly in F-0008 and FC-0208 compositions (Table 16).

Table 16 Drill life holes to failure Material Fe-0.3%C F-0005 F-0008 FC-0208

ATOMET 29 17 56 25 2

ATOMET 29M 58 144 87 74

Compaction to 6.7 g/cm3. Sinter at 1120 °C (2048 °F), 90% Ni/10% H atmosphere. Machining by 6.35 mm HSS drill, 4250 rpm. Source: Ref 32 Graphite/Sulfur. Graphite is widely used as a solid lubricant in the production of porous sintered bearings. However, its

use in structural P/M parts is largely confined to that of an alloying agent intended to dissolve during sintering. Prealloyed sulfur has the ability to inhibit graphite solution so that some free graphite remains in the sintered microstructure without adversely reducing mechanical properties. Table 17 illustrates how this combination of prealloyed sulfur and graphite can produce a significant improvement in the machinability of sintered P/M steels (Ref 33).

Table 17 Effect of sulfur modification upon drill life for iron-2% copper Graphite addition, % 0.8 1.0 1.2

Standard 92% of theoretical density) machine like wrought metals. Smearing of self-lubricating porous parts can be a problem. Recommended practice involves the use of sharp tools and light cuts in single-point machining, such as turning or boring. Coolants are preferred in most machining operations. Coolant pickup can be a problem. The rate of pickup is directly related to the amount of porosity. Ideally, all machining except grinding should precede deburring. Retained deburring abrasive can cause excessive tool wear. Ceramic and cubic boron nitride (CBN) inserts are usually run dry; performance is typically better than or at least equal to that obtained with coolants. Material Selection. Powder metallurgy carbon steels are selected primarily for parts with moderate strength and hardness, combined with machinability. Iron-copper and copper steel materials are produced from admixtures of elemental iron powder and elemental copper powder with or without graphite powder (carbon). When secondary machining is required, combined carbon contents of less than 0.5% should be specified. Copper-infiltrated iron and steel materials offer improved machinability because of reductions in interrupted cuts, and machined parts have a smooth surface finish. Among stainless steels, SS-303 is preferred when parts require extensive secondary machining. Brass, bronze, and nickel silver parts usually have good machinability. Additives. Free-machining benefits can be obtained by means of small additions to a standard powder composition. Additives for ferrous powders include lead, sulfur, copper, or graphite; for nonferrous powders, lead is used. The advantages of changing composition in this manner can be at least partially offset by side effects. Additions can cause problems, such as dimensional changes of parts during sintering and deterioration in the properties of parts.

Prealloyed manganese sulfide powders appear to avoid those shortcomings in ferrous alloys. Manganese content is intentionally high to ensure that all sulfur is present in the form of manganese sulfide inclusions. When these inclusions are extensively deformed in the shear plane and in the flow zone adjacent to the tool surface, they contribute to higher cutting speeds, longer tool life, good surface finish on parts, and lower tool forces. In addition, chips are more readily handled than those produced by conventional P/M materials. Oil or resin impregnation of porous P/M parts also improves machinability (see the article "Resin Impregnation of Powder Metallurgy Parts" in this Volume). Design. Certain types of holes, undercuts, and threads are examples of features that cannot be accommodated by the P/M

consolidation (pressing) process and therefore require machining. Holes in the direction of pressing, produced with core rods that extend up through the tools, are readily incorporated in parts, but side holes (those not parallel to the direction of pressing) cannot be made in the same way and are generally produced by secondary machining. Undercuts on the horizontal plane (perpendicular to the die centerline) cannot be produced if they prevent the part from ejecting from the die (Fig. 1). Annular grooves around a part are produced by machining or by making the part in an assembly of two pieces. Likewise, a part with a reverse taper (larger on bottom than on top) cannot be ejected from a die. Because threads in holes and on outside diameters prevent a part from being ejected from a die, they cannot be made with conventional P/M methods; machining is required.

Fig. 1 P/M part design considerations. (a) Undercuts on horizontal plane cannot be produced in P/M process. Machining is required to obtain such features in parts. (b) Example of undercut in flange that is beyond capability of P/M process. (c) Alternative to part in (b) that can be made without secondary machining

Machining of Powder Metallurgy Materials Sigurd Berg, Höganäs AB; Håkan Thoors, Swedish Institute for Metals Research; Bertil Steen, Swedish Institute for Production Engineering Research

Machining Guidelines The machining process is very complex, and tool performance is affected by the properties and condition of the workpiece and the cutting condition. For P/M materials, porosity is a major factor that reduces machinability. The cutting tool configuration (in terms of chip breaker profile, stability, and geometry of tool holder, insert style, etc.) also influences the wear processes that determine tool life. In order to select the right tool and machining parameters, knowledge of the loads on the tool and the properties of the tool material together with an analysis of the wear mechanisms is necessary. The loads associated with the wear process (Ref 3) can be divided into four main groups: • • • •

Mechanical load Thermal load Chemical load Abrasive load

To define cutting parameters, the loads on the tool must be controlled based on the active wear mechanisms. Typically, the loads on the edge of a cutting tool are different at different locations. Consequently, different wear mechanisms are activated and proceed at different rates at the various locations. The processes that influence tool life can be plotted schematically on a wear mechanism map (Fig. 2), which delineates the area's wear and "safe zones" for good tool performance.

Fig. 2 Pressure/feed rate versus cutting speed in a wear mechanism map

The following sections describe machining conditions for common operations (e.g., turning, drilling, tapping, grinding, and milling). In addition, examples are given for machinability evaluations on the turning, drilling, and tapping of various sintered steels. These examples illustrate the influence of chemical composition, tool material, tool geometry, free machining additives, feed rate, cutting velocity, cutting conditions, and surface integrity while forming the basis of guidelines for the machining of sintered steels. Effects from microstructure, carbon content, density, machinability enhancing additives, etc. are examined. Guidelines for optimum machining parameters in turning, drilling, and tapping are stipulated for a wide spectrum of P/M steel conditions based on these examples. Turning Usually, parts with an average hardness of HRB 52 have machining properties similar to those of cast iron. At this hardness level, parts should be held rigid, and weak sections should be supported to prevent distortion. Compressed air is used to cool the tool and maintain swarf clearances. Jets are directed onto tool cutting edges and work surfaces. Liquid coolants cannot be used because parts must be kept dry and clean for subsequent sintering. Carbide tips of ISO designation K10 with a hardness of 92.6 HRA give good results and will accept some interruptions on the cutting surface. Tools must be held rigid, and cutting edges must be sharp with rake angles of up to 3° positive on top and side and frontal clearances of 3 to 5°. Surface speeds of 105 to 120 m/min (350 to 400 sfm) and feeds of 0.050 to 0.10 mm/rev (0.002 to 0.004 in./rev) are satisfactory for form turning, but surface speeds can be increased to 180 to 210 m/min (600 to 700 sfm) in singlepoint turning. Feeds can be increased within the boundaries of economic tool life, the standard of accuracy, and surface finish requirements. In machining fully sintered parts (average hardness, 90 HRB), K10 carbide tips give a satisfactory life with 0° top rake, 7° frontal clearance, and 5° side clearance. Single-point finish turning is used, with a stock removal of 0.125 to 0.20 mm (0.005 to 0.008 in.) of surface depth and 0.050 mm (0.002 in.) of feed per revolution. Surface speeds are 120 to 135 m/min (400 to 450 sfm), and tips require a radius of 0.20 to 0.25 mm (0.008 to 0.010 in.). Form turning is not advisable because of the workhardening characteristics of the material in this state. Abrasive flank wear is the dominating wear mechanism in turning. PVD-TiN coating of the hard metal (HM) inserts reduces the wear rate; CVD coatings (TiN, Al2O3) improves the performance even further. Oil impregnation improves the machinability in general, while cutting fluid is detrimental to the machinability.

The results for the following example show that the axial force is the dominating cutting force component after 0.1 mm flank wear. The micro surface roughness (within the feed marks) is improved by increased density, and addition of MnS improves the macro surface integrity. Tool life is nearly independent on the feed rate in a range of 0.05 to 0.2 mm/rev. Alloy elements in general decrease the machinability. Correlation to hardness is not enough to explain the different machining performance of P/M material. Micro smearing from "soft" phases due to inhomogeneous microstructure is one explanation for P/M materials performance in relation to conventional steel. MnS addition for intermittent cutting has a strong effect on the machinability. Example 1: Machinability Evaluation, Turning of Sintered Steel. Turning tests for a number of P/M materials were performed in a CNC turning lathe. The workpiece geometry was a thickwalled tube with an inner diameter of 35 mm, an outer diameter of 64 mm, and a height of 62 mm. To study intermittent turning a synchronizing hub was used. All turning was performed as a facing operation. The materials were sintered at 1120 °C for 20 min. For materials where carbon is added, an endothermic atmosphere is used, while dissociated ammonia is used for the other materials. The solid wrought reference material was OVAKO 234S (DIN 16MnCr5) 0.5% C (HV 220) with the same workpiece geometry. Initially, evaluation of tool material and cutting conditions was performed. Based on these results, a PVD-coated hard metal insert (ISO code CNMG120408-MF) was selected as standard. The depth of the cut was fixed to 0.5 mm. The force measurements were carried out using a Kistler three component piezo electric table and a digital oscilloscope (Kistler Instrument Group, Amherst, NY). For P/M materials, the main tool wear mechanism during continuous turning is abrasive flank wear. A notch at the depth of cut that can limit the tool life is sometimes formed. The feed force and axial force are also more sensitive to changes in cutting data, wear, and the material type than the main cutting force. When the feed is increased, P/M materials show a smaller increase in these forces than the reference wrought material. Measurements of the three cutting forces at different wear levels reveal that the axial force is the dominating force after 0.15 mm flank wear. The investigations indicate that the wear rate accelerates after 0.1 to 0.15 mm flank wear dependent on grade. The correlation between wear rate and axial force is significant. Key results are summarized below. Surface Integrity. The surface roughness of the machined surfaces in Example 1 were evaluated in terms of Ra and Rz

using a laser measuring station. The Ra value is the arithmetic average of all deviations of the roughness curve from the center line. The Rz value, mean roughness depth, is the mean of the maximum peak-valley distances in five successive lengths, Lc, of the roughness profile. The density influence on Ra is shown in Fig. 3. Increased density will increase the macro Ra value while the micro Ra value is reduced. Looking at Rz it seems as if the value is constant in this density range.

Fig. 3 Density influence on surface integrity for ASC100.29 2% Cu, 0.5% C; density variation; second edge wear,

0.2 mm

Addition of carbon to atomized iron powder with diffusion bonded nickel (4%), copper (1.5%), and molybdenum (0.5%) improves Ra. In the range 0.25 to 0.8% C the micro surface roughness seems to be the same. Addition of MnS free machining additives reduces the macroscopic roughness Ra (Fig. 4). The reduction is even more pronounced when using Rz as a measure of the macroscopic surface.

Fig. 4 Influence of carbon on the surface integrity for Distaloy AE carbon and MnS addition; second edge wear, 0.2 mm

Influence of Feed Rate. For conventional steel, the feed rate has a strong influence on the tool wear, and different wear mechanisms will be critical at different feeds. For P/M material there seems to be no change in the wear mechanism, and the tool life seems to be independent of the feed rate in a certain range. The tool grade and type of coating material determine the maximum feed rate.

For Distaloy AE 0.5% C, the maximum feed rate seems to be in the range of 0.2 mm/rev when using a PVD-TiN coated tool. By switching to a CVD multicoated (TiN, Al2O3) tool, the feed rate can be increased to 0.3 mm/rev (Fig. 5). For ASC100.29 2% Cu, 0.5% C, the maximum feed rate seems to be 0.2 mm/rev with a PVD-TiN or CVD-Al 2O3 coated tool. In general CVD coating improves the performance in tool life by at least 30% compared to a PVD-TiN coated tool. The geometry of the tool has a large influence on the performance. Among the most important factors is a small tool edge radius, which decreases the axial force and prolongs the tool life.

Fig. 5 Influence of feed rate and type of coating on tool life for Distaloy AE 0.5% C and ASC100.29 2% Cu, 0.5%

C. Tool, CNMG 120408

Influence of Carbon Content. Carbon is the most common alloying element in ferrous powder metallurgy. The strengthening effect is due to the increased amount of pearlite in the microstructure. Above 0.8% precipitation of cementite at the grain boundaries will decrease the strength, which influences the machinability strongly as seen in Fig. 6.

Fig. 6 Influence on carbon addition on tool life of Distaloy AE. Tool material: CNMG 120408, GC 1025. Cutting conditions: feed = 0.1 mm/rev; depth of cut = 0.5 mm; criteria, Vb = 0.3 mm; dry

For atomized iron powder with addition of 2% Cu, there is a clear difference between 0% C and 0.25% C. Smearing of workpiece material on the tool edge is believed to be the explanation for this. The surface for this grade is also considered to be rough, both in the macro and micro range, compared to 0.25% C addition. This effect decreases with cutting speed. Influence of Cutting Fluid. The use of cutting fluid is common when machining conventional steel. The pores present in

the microstructure of the P/M materials in combination with water from the cutting fluid can cause oxidation, which is detrimental to the mechanical properties. Oil impregnation is one alternative to enhance the machinability. In our turning studies, oil impregnation has a pronounced positive effect on P/M machinability, while the use of cutting fluid is detrimental. The main reason for the observed deterioration of tool performance with cutting fluid can be traced back to the severe thermomechanical load cycle associated with the test mode. Introduction of cutting fluid can lead to the following: • •

Severe fluctuation in the tool temperature due to the better heat transfer characteristics of the water-based cutting fluid leading to severe thermal cycling Probable reaction between rest products from the coolant and the atmosphere during the interruption period of the machining cycle

The inability of these effects to increase the tool wear when turning conventional steel with cutting fluid indicates that both the maximum tool temperature as well as the temperature difference within a machining cycle is higher when turning P/M materials. In Table 1 the flank wear (mm) after 40 and 90 passes is presented.

Table 1 Influence of liquid coolant and oil impregnation on flank wear Material

Distaloy AE ASC100.29

Composition

0.5% C 2% Cu, 0.5% C

Flank wear, mm Oil impregnated 40 passes 90 passes 0.046 0.098 0.034 0.04425

Cutting fluid 40 passes 90 passes 0.1335 0.739 0.1215 0.39725

Dry 40 passes 0.075 0.04375

90 passes 0.321 0.055

Intermittent Cutting. Initial tests to select the right tool for intermittent cutting revealed that a tougher tool material was needed than for continuous turning. The tool used for continuous cutting was subjected to chipping of the tool edge. The chipping was also reduced by the use of sharper tool geometry when machining P/M parts.

Addition of free machining additives such as MnS gave a clear improvement. The effect was emphasized at increasing cutting speed (Table 2). For conventional steel, it is a common experience that the chip becomes thinner and the forces decrease as the cutting speed is raised. This drop in forces is caused by a decrease in contact area and in shear strength due to higher temperature. This seems also to be the case for P/M material. But it is still not clear if the effect arises from improvement in chip breaking or by reduction of the shear force.

Table 2 Tool life for turning synchronizing hubs Material

Composition

Tool life, min

Distaloy AE

0.5% C

3.36

0.5% C, 0.5% MnS 31.62 Distaloy HP-1 0.5% C

2.79

ASC100.29

...

5.54

ASC100.29

2% Cu, 0.5% C

26.23

Cubic Boron Nitride (CBN) Tools. In cases where the preservation of surface porosity is vital, CBN inserts are used, especially with low-porosity materials. When CBN tools are used, surface speeds can be increased from 600 to 1000 m/min (2000 to 3280 sfm) using the same rake and clearances as those for K10 tools. These tools will also accept some degree of interruption on the cutting surface. With K10, a coolant is necessary to keep the tool cool and to maintain swarf clearance. With CBN, the procedure used depends on the workpiece material.

Cutting speeds with CBN tools can vary widely from 250 to 2000 m/min (820 to 6500 sfm). Feed rates range from 0.050 to 0.075 mm/rev (0.002 to 0.003 in./rev) with depths of cut from 0.13 to 0.40 mm (0.005 to 0.015 in.) for steels with hardnesses in the range of 50 to 400 HB; usually, it is preferable to decrease the cutting speed to below 1000 m/min (3280 sfm) when the hardness is above 250 HB. These conditions are suitable for hot-pressed or cold-pressed ceramic tools. In cases in which uncoated carbide tools are used, cutting speeds should be less than 190 m/min (625 sfm) for machining most ferrous or nonferrous alloys. Drilling In drilling, speeds and feeds are 80 to 85% of those for wrought metals of the same composition. For long tool life, nitrided steel, high-speed steels containing cobalt, and carbide-tipped drills are recommended. Low helix angle drills are not recommended for softer P/M materials because of their poor chip ejection characteristics. Drills with 40° helix angles had twice the tool life of those with 30° helix angles in work performed with soft P/M materials.

Large amounts of coolant are required in drilling medium- or low-density materials; coolant should operate effectively at the drilling point to reduce abrasive wear due to powder particles at the bottom of the hole. A single-nozzle coolant system does not work properly because small or powdered chips do not easily exit through the drill flutes. A ring design system, however, is effective in eliminating the chip-clogging problem. Oil hole drills are the most effective means for removing chips from the cutting zone. Cutting speeds of up to 25 m/min (80 sfm) and feed rates up to 0.25 mm/rev (0.010 in./rev) are recommended for high-speed steel drills. Cutting speed and feed rate could be as high as 120 m/min (390 sfm) and 0.5 mm/rev (0.02 in./rev), respectively, when solid-carbide or carbide-tipped drills are used. Carbide indexable drills are efficient because margins are eliminated. Abrasive margin wear and the welding of powder chips are also eliminated. Holes in planes beyond the capability of the P/M process are best made by drilling when parts are in the presintered (partially sintered) state. In this state, drilling properties are similar to those of cast iron. If it is necessary to qualify the position and size of holes after final sintering, a carbide reamer or carbide-tipped reamer should be used. Roller burnishing can be used to meet accuracy and surface finish requirements. Any size change in the operation is related to the preburnished surface finish and the size of the hole. A change of 0.019 to 0.025 mm (0.00075 to 0.001 in.) in diameter is representative for a hole about 25 mm (1 in.) in diameter. A number of studies have been performed to determine ways of enhancing drill life during P/M machining (Ref 4, 5, 6, 7, 8, 9, and 10). The following example is another. As shown in the following example, drill length is considered as a primary factor for improvements. Additives have strong influence on machinability for high performance materials. Tool life is nearly independent on the feed rate, which ranges from 0.05 to 0.16 mm/rev. Cutting fluid has no significant effect on the productivity. Distribution decreases for coated drills and for short drills. Example 2: Machinability Evaluation, Drilling of Sintered Steel. Blind hole drilling test was performed on a wide spectrum of P/M qualities. Cylindrical blanks with a diameter of 80 mm and a height of 10 mm are used in the test. A survey test is performed in order to select the type of drill and cutting conditions. The main comparisons were made under dry condition using a HSS drill with a diameter of 4 mm and a point angle of 118°. Total breakdown of the drill is chosen as criterion, based on the fact that drilling is commonly used as a bulk removal operation. All materials are investigated in order to divide them into four groups. The classification criterion used is the time required to drill 100 holes. Influence of cutting fluid, density, alloying contents, feed rate, and drill type are evaluated for the selected group representatives. All the tests were carried out in a numeric controlled machining center. Early in the testing procedure, the heat generated in the workpiece during the drilling cycle was recognized to have an effect on the result, and therefore, the time between inserts was increased and temperature measurements were carried out on the workpiece. Key results are noted below. Tool Life. Evaluation for atomized iron with diffusion bonded nickel (4%), copper (1.5%), and molybdenum (0.5%),

addition of 0.8% C regarding influence of length, type of drill, and coating of the drill is shown in Fig. 7. The length of the drill has a large influence on the performance. Evaluation of centered drilling shows improvements in distribution and in performance. It is believed that the first inlet is crucial for the performance. This has its origin from the inhomogeneous microstructure of P/M material.

Fig. 7 Influence of feed rate, tool material, coating, and additives on machinability

Tool material and coating have a strong influence. Geometry changes exist among the three types of drills. Regarding feed rate, the performance is comparable with the result from the turning test presented for HM PVD-TiN tool, which ranges from 0.12 to 0.16 mm/rev. For Distaloy AE 0.5% C using HSS drill, the same tool life exists up to feed rate 0.16 mm/rev. This behavior is unique for P/M material. Effect of Carbon Contents. Addition of carbon above 0.25% C for atomized iron powder or addition of 2% Cu will

decrease the machinability. Micro smearing on the tool is believed to be the reason for the large decrease in machinability (Fig. 8).

Fig. 8 Influence of carbon contents on machinability of ASC100.29 2% Cu. Drill, HSS; point angle, 118°; feed rate, 0.06 mm/rev; D = 4 mm; criteria, total failure

3

Density, Cutting Fluid, and Additives. Influence from density is regarded as small in the range of 6.7 to 7.3 g/cm .

Result from productivity evaluation of atomized iron powder with 2% Cu and 0.5% C at 100 holes reveals 5% decrease for the range. Use of cutting fluid has no significant effect on the productivity. Water as cutting fluid decreases the productivity. For high performance materials, additives have large influence on the machinability. Tapping Conventional tap drill charts should be followed to maintain 65 to 75% depth of thread. Two-flute taps are recommended for diameters up to 8 mm ( in.). Three-flute taps should be used for diameters of 8 to 12.5 mm ( to in.). Spiral-point taps are desirable because they throw the chip out instead of driving it into the pores of the workpiece. Some experimenting in tapping P/M parts may be required to determine which tap is best for a specific metal. As shown in the following example, chip clamping can degrade the performance, but it is improved by the use of cutting fluid. Evaluation of tapping with the use of cutting fluid for the first three holes and after tapping 50 holes is presented in Fig. 9. Tapping under dry condition reveals problems with chip clamping. Cutting fluid improves the performance. Carbon addition decreases the performance and additives like MnS decrease the torque. The distribution during the measuring length still indicates problems with chip clamping. Selection of tap geometry is considered to solve the problem.

Fig. 9 Influence of carbon/MnS addition for the moment during tapping. M5 straight flute tap with 7.2 mm tap length. Bottom hole was used in the initial investigation.

Other Cutting Methods Milling. Slot and side milling cutters are often used for machining P/M materials. Speeds of 70 to 100 m/min (230 to 330

sfm), feed rates of 0.005 to 0.1 mm (0.0002 to 0.004 in.) per tooth, and depths of cut of 0.13 to 0.4 mm (0.005 to 0.015 in.) are recommended in machining ferrous and nonferrous alloys with uncoated carbides. Higher speeds and feeds should be used in machining aluminum. Aluminum P/M alloys have better chip characteristics than their wrought counterparts. Chips are much smaller and are broken more easily, with little or no stringer buildup. In face milling with uncoated carbides, cutting speeds of 90 to 120 m/min (295 to 395 sfm), feed rates of 0.05 to 0. 15 mm (0.002 to 0.006 in.) per tooth, and depths of cut of 0.12 to 0.4 mm (0.005 to 0.015 in.) are recommended for carbon and alloy steels and stainless steels; however, nonferrous materials can be cut at speeds up to 170 m/min (560 sfm) and feed rates as high as 0.1 mm (0.004 in.) per tooth. Speeds of 25 to 50 m/min (80 to 165 sfm) are used in machining P/M iron, steel, stainless steel, copper, and brass with highspeed end mills. On the other hand, cutting speeds in the range of 100 to 200 m/min (330 to 655 sfm) are recommended in

machining soft iron, steels, and aluminum with carbide tools. With harder steels, stainless steels, copper, and brass, speeds should be lowered to the range of 60 to 100 m/min (195 to 330 sfm). Reaming. To control bore accuracy in P/M parts, reaming is sometimes used instead of pin sizing, ball sizing, or burnishing. Standard reamers are satisfactory; left-hand spiral reamers have also proved successful. The cutting edges should have the best possible finish to minimize edge buildup, which results in oversize holes. If the surface finish of the hole is not a factor, the drill should leave a reaming allowance, the amount depending on hole size. Guidelines can be used:

Hole diameter mm in. 6.5 6.5-12.5 12.5-25

0.25 0.25-0.50 0.50-1.0

Allowance mm 0.050

in. 0.002

0.050-0.10 0.10-0.15

0.002-0.004 0.004-0.006

If the surface finish is critical, reaming allowances should be doubled. When possible, reamers should be used in floating holders and run at 7.5 to 15 m/min (25 to 50 sfm). Recommended feeds are:

Hole diameter mm in. 6.5 8-12.5 14-19

0.25 0.30-0.50 0.55-0.75

Feed mm/rev 0.15

in./rev 0.006

0.18 0.25

0.007 0.01

Finishing Burnishing. When the clearance between a shaft and a P/M bearing is ±0.012 mm (±0.0005 in.) or less, burnishing the

bearing bores after they have been installed in the housing is preferred for correcting the bore size. No more than 0.002 mm/mm (0.002 in./in.) of diameter should be displaced, and the smallest amount of displacement that will produce the true diameter is desirable. The type of burnishing tool recommended for this operation is illustrated in Fig. 10.

Fig. 10 Ball broach for burnishing bores in P/M parts. Dimensions given in inches

Given a finished bore diameter of 38 mm, +0.005 mm/-0.0000 mm (1.500 in., +0.0002 in./-0.0000 in.) (B in Fig. 10), the diameter of the starting end of the burnishing tool then becomes 38.10 mm, -0.050 mm (1.5000 in., -0.0020 in.), or 38.05 mm (1.4980 in.), and bearings would be bored to 38.075 mm, +0.0125 mm/-0.0000 mm (1.4990 in., +0.0005/-0.0000 in.). Thus, there would be a minimum clearance of 0.025 mm (0.001 in.) at the entering end of the tool, and the first land would be a line-to-line fit. The tool then becomes progressively larger, and the bearing is expanded. If there were no springback, the operation would be stopped at the fourth or fifth tool land. However, the bearing would ordinarily be burnished to 0.010 mm (0.0004 in.) oversize to allow for springback. Roller burnishing is a cold-working operation that compresses metal rather than removes it. The technique is suitable for sintered (not heat-treated) powder metal materials for which maintenance of open surface porosity is not critical. A significant improvement in surface finish can be obtained using a roller burnishing tool. In addition, the tool is adjustable to match individual product specifications as well as to compensate for wear on the rolls and mandrel. Both through holes and blind holes can be roller burnished. Hole size tolerance depends on the input tolerance of the hole; that is, a prepared tolerance of 0.050 mm (0.002 in.) can be reduced to 0.025 mm (0.001 in.), or a ±0.0025 mm (±0.0001 in.) tolerance can be held if the input tolerance is 0.010 mm (0.0004 in.). Surface finishes of 0.25 m (10 in.) are common after roller burnishing, A lightweight, low-viscosity lubricating oil is recommended for most P/M materials. Honing and Lapping. Holes requiring extreme accuracy can be honed or lapped by normal techniques if retention of porosity is not required. However, size control of holes in P/M parts can usually be obtained more economically by reaming or burnishing.

High-density ferrous metal parts, especially when hardened, have been successfully honed and lapped using conventional procedures. Diamond- and CBN-plated bore finishing tools are recommended for precise hole size control. These tools can be used on standard drilling or honing machines, as well as on multiple-spindle or numerically controlled machines. The use of an adjustable sleeve attached to a mating tapered mandrel increases tool life. The selection of diamond grit size determines the metal removal rate and the surface achieved. The amount of material to be removed from the hole diameter can be determined by: surface finish (start)--surface finish (after honing)/100,000 = required stock removal. If the existing finish is 1.25 m (50 in.) and the desired finish is 0.25 m (10 in.), then 0.010 mm (0.0004 in.) should be removed from the hole diameter.

Honing of infiltrated parts is seldom practical, because the stones become loaded. Neither lapping nor honing is recommended for porous parts, because either of these processes will cause the pores to become filled with abrasive particles. For special applications that require the use of lapping or honing, ultrasonic or solvent cleaning should be performed following grinding. Grinding Grinding of P/M parts can be very complex, especially when materials are low in density because in many cases preservation of surface integrity is essential. Usually, surface porosity decreases during grinding. A large amount of the generated powder chips is forced into pores, and many chips are welded due to the high temperature at the wheel/workpiece interface. When grinding is necessary to achieve dimensional functionality of a part, and surface porosity needs to be preserved, special processes such as ultrasonic or solvent cleaning are applied immediately after grinding. For rough applications, a downfeed of 0.025 to 0.075 mm (0.0010 to 0.003 in.) is recommended, while for finish passes, a maximum of 0.013 mm (0.0005 in.) should be used. Stock removal rates should be either the same as or less than those used in finish turning of cast iron; wheels should be similar. It is important to keep a plentiful supply of coolant (containing an inhibitor) directed onto the wheel and the work to maintain a clean grinding wheel contact. Grinding of P/M Tool Steels. The relative grindability of several conventional and P/M high-speed tool steels is illustrated in Fig. 11. The grinding ratio (volume of metal removal to the volume of wheel worn, as explained in the article "Principles of Grinding" in Machining, Volume 16, ASM Handbook) is clearly superior for the P/M tool steels. As expected, the grinding ratios generally decrease for both the conventional and the P/M tool steels as their alloy and carbon contents increase. The grinding conditions suggested for the CPM tool steels are similar to those recommended for conventional tool steels in the article "Machining of Tool Steels" in Machining, Volume 16, ASM Handbook. Some specific conditions for grinding CPM 10V are given in Table 3.

Table 3 Grinding recommendations for CPM 10V cold-work tool steel

Toolroom grinding (sharpening)(a) • • • • •

Abrasive: very sharp 38A or 32A Grit sizes: 60 to 120 depending on removal and finish requirements Grade: Grade I most effective, but grades as soft as G can work Bond: Vitrified Wheel example: Norton 32A60-I8VBE

Wet surface grinding(a) • •

Grit sizes: 100-150 Wheel example: CBN (Borazon) CB 120TBA

Internal grinding(a) • •

Grit sizes: 100-150 Wheel example: CBN (Borazon) CB 150WBA

Field reports concerning abrasives •

Cubic boron nitride (Borazon) grinders must be rigid, in good condition, and able to mount wheels with

• • •

very good accuracy. Crystolon (silicon carbide) such as 39C60-I8VK are recommended. Use very sharp, very friable aluminum abrasive that remains sharp during grinding, such as 38A. Wet grinding recommended.

Note: Grinding wheel symbols and nomenclature are defined in the article "Grinding Equipment and Processes," Machining, Volume 16, ASM Handbook.

(a)

Based on in-house laboratory testing

Fig. 11 Comparagraph showing the relative grindability of CPM and conventional high-speed tool steels. Source: Crucible Materials Corporation

Tool Steels Rapid solidification of the atomized powders used in the production of wrought P/M tool steels eliminates the segregation present in conventional tool steels and produces a very fine microstructure with a very uniform distribution of small carbides and nonmetallic inclusions. As a result, wrought P/M high-speed tool steels exhibit better machinability, dimensional control, and safety in heat treatment, grindability, and edge toughness during cutting than conventional high-speed tool steels of the same composition. A variety of Anti-segregation process (ASP) and P/M tool steels are available. As with conventional tool steels, P/M tool steels are generally machined in two stages: rough machining of the workpiece with the steel in the annealed condition, followed by finish machining (typically grinding) after heat treatment when the steel is in the hardened-and-tempered condition. Table 4 lists the typical cutting conditions for P/M and conventional AISI highspeed steels of similar composition.

Table 4 Typical machining conditions for P/M and conventional grades of AISI high-speed tool steels Operation

Single-point turning Drilling

Broaching Face milling Cutoff

Tool width or depth of cut mm in. 3.8 0.150 0.64 0.025 6.4

High-speed tooling Speed Feed m/min sfm mm/rev 18 60 0.38 23 75 0.18 12 40 0.08

13

12

40

12 12 3 20 26 14 14 14

40 40 10 65 85 45 45 45

25 50 ... 3.2 0.64 1.6 3.2 6.4

1 2 ... 0.125 0.025 0.062 0.125 0.250

in./rev 0.015 0.007 0.003

Carbide tooling Speed m/min sfm 91 300 111 365 ... ...

Feed mm/rev 0.38 0.18 ...

in./rev 0.015 0.007 ...

0.13

0.005

...

...

...

...

0.23 0.33 0.05 0.20 0.15 0.03 0.03 0.04

0.009 0.013 0.002 0.008 0.006 0.001 0.001 0.0015

... ... ... 78 101 53 53 53

... ... ... 255 330 175 175 175

... ... ... 0.30 0.25 0.05 0.08 0.11

... ... ... 0.012 0.010 0.002 0.003 0.0045

An important advantage of the P/M process relates to the fact that the machinability and grindability of P/M tool steels can be improved by increasing their sulfur content to much higher than conventional levels without sacrificing toughness or cutting performance (see the article "Particle Metallurgy Tool Steels" in this Volume).

References cited in this section

3. A. Thelin, Verschleissmechanismen und Leistungen von Zerspanwerkzeugen, VDI Berichte, No. 762, 1989, p 111-126 4. J.S. Agapiou, G.W. Halldin, and M.F. DeVries, Drillability of 304 Stainless Steel P/M Material: Tool Wear and Life, 1987 Annual Powder Metallurgy Conf. Proc., Vol 43, Metal Powder Industries Federation, 1987, p 181 5. V.V. Podgorkov et al., Finish Machining of Sintered Iron and Copper Base Materials, Sov., Powder Metall. Met. Ceram., Vol 13 (No. 8), 1974, p 674-677 6. S. Suzuki et al., Machinability of 4100 Series Sintered Steel Containing Sulfur, 1987 Annual Powder Metallurgy Conf. Proc., Vol 43, Metal Powder Industries Federation, 1987, p 511 7. U. Engstrom, Machinability of Sintered Steels, Progress in Powder Metallurgy 1982, Vol 38, 1982 National Powder Metallurgy Conf. Proc., Metal Powder Industries Federation, 1982, p 417 8. J.M. Capus and C. Fournel, Tool Wear Measurements in Machining of Sintered Ferrous Alloys, Progress in Powder Metallurgy 1981, Vol 37, 1981 National Powder Metallurgy Conf. Proc., Metal Powder Industries Federation, 1981, p 165 9. Y. Trudel, C. Ciloglu, and S. Tremblay, Selected Additives to Improve Machinability of Ferrous P/M Parts, Modern Developments in Powder Metallurgy, Metal Powder Industries Federation, Vol 15, 1984, p 775 10. A. deRege, G. L'Esperance, L.F. Pease, and L. Roy, Prealloyed MnS Powders for Improved Machinability, Near Net Shape Manufacturing Conf., P.W. Lee and B.L. Ferguson, Ed., ASM International, 1988, p 57-68

Machining of Powder Metallurgy Materials Sigurd Berg, Höganäs AB; Håkan Thoors, Swedish Institute for Metals Research; Bertil Steen, Swedish Institute for Production Engineering Research

References 1. K.H. Roll, Powder Metallurgy at the Turn of the New Century, 1987 Annual Powder Metallurgy Conf. Proc., Metal Powder Industries Federation, 1987 2. J.S. Agapiou and M.F. DeVries, Machinability of Powder Metallurgy Materials, Int. J. Powder Metal., Powder Technol., Vol 34 (No. 1), 1988 3. A. Thelin, Verschleissmechanismen und Leistungen von Zerspanwerkzeugen, VDI Berichte, No. 762, 1989, p 111-126 4. J.S. Agapiou, G.W. Halldin, and M.F. DeVries, Drillability of 304 Stainless Steel P/M Material: Tool Wear and Life, 1987 Annual Powder Metallurgy Conf. Proc., Vol 43, Metal Powder Industries Federation, 1987, p 181 5. V.V. Podgorkov et al., Finish Machining of Sintered Iron and Copper Base Materials, Sov., Powder Metall. Met. Ceram., Vol 13 (No. 8), 1974, p 674-677 6. S. Suzuki et al., Machinability of 4100 Series Sintered Steel Containing Sulfur, 1987 Annual Powder Metallurgy Conf. Proc., Vol 43, Metal Powder Industries Federation, 1987, p 511 7. U. Engstrom, Machinability of Sintered Steels, Progress in Powder Metallurgy 1982, Vol 38, 1982 National Powder Metallurgy Conf. Proc., Metal Powder Industries Federation, 1982, p 417 8. J.M. Capus and C. Fournel, Tool Wear Measurements in Machining of Sintered Ferrous Alloys, Progress in Powder Metallurgy 1981, Vol 37, 1981 National Powder Metallurgy Conf. Proc., Metal Powder Industries Federation, 1981, p 165 9. Y. Trudel, C. Ciloglu, and S. Tremblay, Selected Additives to Improve Machinability of Ferrous P/M Parts, Modern Developments in Powder Metallurgy, Metal Powder Industries Federation, Vol 15, 1984, p 775 10. A. deRege, G. L'Esperance, L.F. Pease, and L. Roy, Prealloyed MnS Powders for Improved Machinability, Near Net Shape Manufacturing Conf., P.W. Lee and B.L. Ferguson, Ed., ASM International, 1988, p 57-68

Resin Impregnation of Powder Metal Parts Charles M. Muisener, Research, Development & Engineering Group, Loctite Corporation

Introduction RESIN IMPREGNATION is a process that eliminates or reduces internal porosity of castings and P/M parts by saturating internal voids with liquid resins. The process has been practiced for many years on castings and P/M parts, and resin impregnation has, to a large extent, eliminated macroporosity (pore diameter >125 m ) in castings. With further process improvements and low-viscosity resins capable of good penetration, impregnation is also capable of significantly reducing microporosity (pore diameter 9.5 mm, or 0.4 in.), the interior remains as ferrite and fine pearlite, experiencing neither shrinkage nor growth. The outer surfaces expand outward, and the inner surfaces shrink inward. This phenomenon also is evident in case-hardened wrought parts. Thus, prediction of exact size change during heat treatment is difficult. Steam Blackening. Sintered parts frequently are treated in steam at 540 to 595 °C (1000 to 1100 °F) for 1 to 4 h to fill the

pores and coat the surface with a hard coating of black iron oxide. The coating causes a uniform growth of 0.0025 to 0.0050 mm (0.0001 to 0.0002 in.) similar to electroplating. The amount of blackening should be controlled, as measured by hardness and destructive break tests; excessive oxide coating thickness may lower impact properties. Thickness of the oxide layers also

can be measured metallographically with a polishing procedure described in the section of this article on density measurement. Evaluation of Dimensional Change in Incoming Powder. New lots of blended or raw powder are checked against

internal standard lots to ensure consistent sintered dimensional change. Transverse-rupture bars 31.8 by 12.7 by 6.4 mm (1.25 by 0.50 by 0.25 in.) are molded at a fixed density or pressure from both the standard and test lot of powder. The two sets of bars are sintered simultaneously in a laboratory or production furnace. Dimensional change in the 31.8 mm (1.25 in.) length are checked against the requirements of American Society for Testing and Materials (ASTM) standard B 610. Although dimensional change from sintering a bar made from the standard powder can differ from previous tests, comparable dimensional changes in the test bar made from incoming powder demonstrate the difference in the performance of the powders. Dimensional change in test and standard lots must agree to within a specified range (±0.1% of the bar length). These bars also can be used to evaluate sintered strength and hardness. Dimensional Control. Table 1 illustrates typical dimensional tolerances of P/M materials. Separate tolerances apply to assintered, as-sized, and as-heat treated conditions. For concentricity between an inside diameter and an outside diameter, a total indicator reading of 0.075 mm (0.003 in.) is permitted. The distance between holes can be as great as 0.075 mm + 0.013 mm/mm (0.003 in. + 0.0005 in./in.). Gears can be molded to American Gear Manufacturers Association (AGMA) class 7, which is limited primarily by the concentricity of the bore to pitch line. If gears are held on the pitch line and bored more concentrically, AGMA class 10 or 11 is achieved.

Table 1 Typical P/M tolerances (other than length) Material

Brass Bronze Aluminum Iron Copper alloy steel Nickel alloy steel Stainless steel

Condition As-sintered mm in. ±0.089 ±0.0035 ±0.089 ±0.0035 ±0.051 ±0.002 ±0.025 ±0.001 ±0.038 ±0.0015 ±0.038 ±0.0015 ±0.025 ±0.001

As-sized mm ±0.013 ±0.013 ±0.013 ±0.013 ±0.025 ±0.025 ±0.013

in. ±0.0005 ±0.0005 ±0.0005 ±0.0005 ±0.001 ±0.001 ±0.0005

As-heat treated mm in. ... ... ... ... ±0.013 ±0.0005 ... ... ±0.038 ±0.0015 ±0.038 ±0.0015 ... ...

Note: Up to 12.7 mm (0.500 in.). Length tolerance, ±0.102 mm (±0.004 in.), unless machined or ground. Source: Ref 1

Other processes, such as P/M hot forging, injection molding, and high-temperature sintering, produce wider tolerances than presented in Table 1. Powder metallurgy forged dimensional tolerances are given in Table 2. High-temperature sintering tolerances are given in Table 3. Injection-molded tolerances range from 0.075 to 0.10 mm/mm (0.003 to 0.004 in./in.), even though parts have experienced 12 to 15% linear shrinkage (Ref 5).

Table 2 Tolerances on P/M forged parts Parameter Outside diameter Outside diameter Inside diameter Thickness Spline Outside diameter Inside diameter Concentricity

Nominal dimension mm in. 50.8 2.00 50.8 2.00 38.1 1.50 25.4 1.00 25.4 1.00 95.25 3.75 63.5 2.50 95.25 3.75

Tolerance mm 0.13 0.25 0.20 0.38 0.23 0.25 0.25 0.10

in. 0.005 0.010 0.008 0.015 0.009 0.010 0.010 0.004

Roundness Thickness Outside diameter Outside diameter Outside diameter Outside diameter Outside diameter Thickness

95.25 15.8 50.8-76.2 25.4-50.8 76.2 50.8 203 25.4

3.75 0.625 2.00-3.00 1.00-2.00 3.00 2.00 8.00 1.00

0.10 0.25 0.13 0.10 0.38 0.13 0.51 0.25-0.634

0.004 0.010 0.005 0.004 0.015 0.005 0.020 0.010-0.025

Source: Ref 5

Table 3 Dimensional tolerances of parts in the as-high-temperature sintered condition Material Composite 3Si-Fe 4600 M-2 Low-alloy steel Stellite

Nominal dimension mm in. 25.4 1.00 19.0 0.75 76.2 3.00 70.3 2.77 22.2 0.88 25.4 1.00

Tolerance mm in. 0.05 0.002(a) ±0.08 ±0.003 0.38 0.015 0.61 0.024 0.08 0.003 0.03 ±0.001(b)

Source: Ref 5

(a) (b)

Roundness. Inside diameter sintered against a mandrel.

References cited in this section

1. P/M Design Guidebook, Metal Powder Industries Federation, 183, p 15 2. "Anchor MH100 Standard Molding Powder," Hoeganaes Corp., Riverton, NJ 3. "Aromet 28, Sintered Properties of P/M Copper Steels," Quebec Metal Powders Ltd., Sorel, Quebec, Canada 4. "Controlled Dimensional Change," SCM Metal Products, Cleveland 5. L. Pease III, "An Assessment of Powder Metallurgy Today, and Its Future Potential," Paper No. 831042, Passenger Car Meeting, Society of Automotive Engineers, Warrendale, PA, 1983 Testing and Evaluation of Powder Metallurgy Parts

Measurement of Density Density is the ratio of mass to volume. For a given material, degree of sintering, and heat treatment, density determines mechanical and physical properties. For example, higher density in sintered steels results in higher tensile strength, elongation, and impact resistance values. As-pressed, or green, density also influences growth or shrinkage that occurs during sintering. With nonuniform green density, parts grow or shrink nonuniformly, as in a thin-walled bronze bearing with a lowdensity region equidistant from the ends. This results in a significantly smaller diameter at midlength than at the ends and necessitates repressing or sizing for close dimensional control. If cubes or right cylinders could be extracted from actual parts, linear dimensions could be measured and volume could be calculated easily. From the weight of a part, density can be easily calculated. This yields a value that, under ideal conditions, differs by 0.04 g/cm3 (0.5%) from a reference (Ref 6). Unless the sintered part is directly molded to an easily measured shape,

such as a transverse-rupture bar (31.8 by 12.7 by 6.4 mm, or 1.25 by 0.50 by 0.25 in.), this method of measuring linear dimensions is used infrequently. Methods Based on Archimedes' Principle. Typical methods of measuring density depend on Archimedes' principle, in which hydrostatic forces in liquids exert buoyant forces proportional to the part volume. This measurement is standardized in ASTM B 328 (Ref 7), MPIF test method 42 (Ref 8), and International Standards Organization test method ISO 2738 (Ref 9). When an object is immersed in a liquid, the liquid exerts an upward buoyant force that is equal to the product of the object volume and the density of the liquid. The difference in weight between an object weighed in air and its weight when suspended in water is equal to the object volume in cubic centimeters times the density of water. Approximating the density of water as unity:

V = Wair - Wwater where V is the volume, cm3; Wair is the weight in air, g; and Wwater is the weight of object suspended in water less the weight of the suspending wire in water (tare), g. Density in g/cm3 is then:

Density = Wair/(Wair - Wwater) For unsintered materials molded with 0.75% lubricant, pores are well sealed, and water cannot penetrate. For such parts, the above calculation is suitable. It is also suitable for materials with pores that are sealed off from the surface (materials close to theoretical density). For most sintered materials that are 70 to 95% dense, water tends to infiltrate the pores during weighing in water. This minimizes the buoyancy effect of the water (that is, the liquid is acting on a smaller volume) and results in an erroneous calculation of low volume. This low volume then causes an erroneously high density value. Infiltration of water into pores usually is accompanied by air bubbles escaping from the part. If the part is blotted to remove surface water and reweighed in air after weighing in water, any weight gain indicates that water has entered the pores. Although not a standard procedure, volume can be approximated as the weight in air after removing the part from the water, minus the weight in water. To prevent infiltration of water, all three standard test methods require that the pores of the part be filled with oil. Oil impregnation is done after the part is weighed in air; this is carried out under vacuum or by immersion in hot oil. Oil prevents the water from entering the pores. The volume of the part is then determined as the part weight in air with oil in the pores, minus the weight of the oiled part suspended in water. Care should be taken to select an oil that is not soluble in water or not soluble in water plus wetting agent. Such oils also must exhibit superior demulsibility. The precision of the ISO method is ±0.25%, regardless of sample density, and assumes a water density of 0.997 g/cm 3. Moyer (Ref 6) has reviewed the literature on precision methods of density determination (Ref 10, 11, 12, 13, 14, 15, 16) and has devised a method that provides accuracy to two or three decimal places, depending on sample porosity. The basic measuring apparatus is shown in Fig. 6. Requirements of precision density measurement include: • • • • • •

Balance capable of measuring to the nearest 0.0001 g Vibration- and draft-free atmosphere Measurement of the density of the immersing liquid (water) by checking the density of a substance of accurately known density (four decimal places) Conversion of all densities back to 20 °C (68 °F) by compensating for thermal expansion of the sintered part Maintenance of liquid level at a constant height on the suspending wire Careful brushing of all bubbles from the test object

Fig. 6 Density measurement apparatus

Using the above procedures, Moyer reports standard deviations of 0.0130 to 0.0005 g/cm3 on 17 g parts with densities ranging from 5.12 to 7.85 g/cm3, respectively. To determine density variation from one point to another in a complex part, the available samples must be considerably 0.8%), iron carbide networks appear in the grain boundaries, and the impact, tensile strength, and elongation are reduced. This carbide network is not to be confused with the divorced eutectoid carbide platelets that will appear occasionally in a grain boundary in the hypoeutectoid steels with less than 0.3% C. This effect is seen in ingot-based steels as well. Sintered iron bearings are fabricated with graphite in solution as well as present as free graphite flakes. The combined carbon is judged by the lever rule, which is important in quality control of iron-graphite bearing materials. Iron-carbon P/M structures are shown in Fig. 53, 54, 55, 56, and 57. Iron-Copper Alloys. Copper is frequently added to iron because it melts and rapidly dissolves, greatly increasing the

strength of the iron. When copper melts, it is drawn by capillary action into the smallest available pores and capillaries. In an atomized iron powder, the copper flows between the particles that are pressed into close contact. It then dissolves in the iron at these points of contact. The copper activates the sintering of the particles that are in contact, resulting in rapid disappearance of particle boundaries and substantial neck growth. The copper may separate the iron particles as it flows among them, causing growth of the part in 1 to 2 min. The subsequent dissolution of the copper and local lattice expansion at points of contact cause later growth. In an undersintered part with 2% added Cu, some of the residual copper may still be visible as a thin line between two iron particles. With sponge irons, the copper can flow into the fine pores inside the particles and thus not cause as much separation of particles. The high surface area also contributes to rapid sintering. These two effects are thought to explain why the sponge iron and copper mixtures do not expand as much on sintering as the mixes based on atomized iron. In conventional sintering of iron-copper alloys (20 to 30 min at 1105 to 1120 °C, or 2025 to 2050 °F), at least 2% Cu disappear into solution in the iron. With 5% or more Cu, some free copper is always present as a copper-rich solid solution with the iron. Depending on the rate of cooling, copper-rich phases precipitate in the iron, and, vice versa. The copper-rich phases precipitate in the iron, darkening the ferrite; slow cooling increases darkening. This effect is limited to the outside of the iron particles, because the melted copper does not readily diffuse to the particle centers under conventional sintering conditions. Picral etching will help to stain the copper precipitate areas for easier identification. Iron-copper P/M structures are shown in Fig. 58, 59, 60, 61, 62. Iron-Copper-Carbon Alloys. The most common of the moderate-strength, as-sintered alloys is Fe-Cu-C with 0.5 to 0.8%

C and 2 to 5% Cu (Fig. 63, 64, 65, 66). It combines the features described above for iron-carbon and iron-copper alloys. The carbon goes rapidly into solution in the iron (perhaps in 5 min at 1040 °C, or 1900 °F) and tends to prevent the sintering expansion prevalent in zero- or low-carbon iron-copper alloys. The combined carbon can be estimated by the lever rule, although the eutectoid may be as low as 0.75% C in this ternary system. Copper-Infiltrated Steels. High-density iron-carbon alloys with 10 or 20% Cu are prepared by infiltrating the copper

alloy into the porous steel matrix. Upon sintering and infiltrating, the copper alloy melts and flows into the iron-carbon matrix with which it is in contact. The copper tends to fill the highest-density, smallest-capillary regions of the matrix first. The coarse-pored lowest-density regions are filled last with whatever liquid copper remains. The structure often appears as islands of ferrite and pearlite with a continuous copper-alloy phase. The alloy of copper may include such elements as manganese and cobalt, which dissolve in iron and alter the alloy content of the steel matrix. Manganese increases the hardenability of the matrix. Elemental nickel contained in the matrix goes into solution in iron and copper, greatly increasing hardenability. Such

materials may exhibit regions of martensite, even as furnace cooled. Copper-infiltrated steel structures are shown in Fig. 67, 68, 69, 70, 71, and 72. Low-alloy steels of the 4600 series type are atomized as low-carbon materials with good compressibility. Because of their alloying elements, they display excellent hardenability and are usually used fully hardened. When viewed in the as-sintered condition, such materials exhibit ferrite and a eutectoid product that does not appear similar to the normal iron-carbon materials. The lamellae are more uniformly spread throughout the structure, and the tendency among the constituents to group into ferrite and pearlite is lessened, which complicates estimating the combined carbon content metallographically. However, this should be possible by devising metallographic photo standards of reference. The powder may contain up to 5% unalloyed iron as a contaminant. In the as-sintered structure, these free-iron particles do not tend to pick up carbon and thus stand out as ferrite (the carbon preferentially dissolves in the prealloyed steel powder particles). Upon quenching, the unalloyed particles are low in carbon and alloy content, do not harden, and are all ferrite or ferrite/pearlite mixtures. Low-alloy steel structures are shown in Fig. 73, 74, 75, 76, 77, 78, 79, and 80. Iron-Phosphorus Alloys. The additions of iron phosphide (Fe 3P) to atomized iron results in the dissolution of phosphorus

in amounts less than 1%. The phosphorus initiates a transient liquid-phase sintering reaction, then goes partly into solution in the iron, resulting in a material with excellent soft magnetic properties. Some of the phosphorus remains visible as a second phase with the ferrite. For magnetic properties, a low carbon content and freedom from pearlite are required. For optimal toughness and strength characteristics, a mixture of up to 1% P and up to 0.3% C is used. The phosphorus also causes pore rounding by virtue of the transient liquid phase, which gives the alloys their toughness and characteristic well-sintered appearance. Iron-phosphorus alloy structures are shown in Fig. 81 and 82. Free-Machining Steels. The machinability of sintered irons and alloys is improved by adding sulfur. Historically, this has

been accomplished by mixing fine sulfur powder (-325 mesh) into sponge iron. More recently, sulfur is dissolved in the liquid melt before atomizing (prealloyed sulfur) to form manganese sulfide (MnS) with carefully controlled amounts of manganese. Manganese sulfide particles have also been mixed with iron for a similar benefit. These additions result in particles of MnS in the pores as a gray phase, or a MnS phase inside the iron particles, if it was prealloyed. The use of high-hydrogen atmospheres at sintering desulfurizes a material to depths of 0.25 to 0.50 mm (0.01 to 0.02 in.), an effect whose analog in carbon is better known. Structures of P/M steels with additions of manganese and sulfur for enhanced machinability are shown in Fig. 83, 84, and 85. Boron nitride powder is also mixed into iron in small amounts to improve machinability. Finally, the most effective enhancer of machinability is impregnation of the pores of the sintered material, using a plastic resin. Nickel Steels. The most common high-strength heat treated materials are the nickel steels. In these mixtures, 2 or 4%

elemental Ni is added to iron, along with 0.4 to 0.8% C and up to 2% Cu (optional). The usual nickel is very finely divided and is often prepared by carbonyl decomposition (production of nickel powder by carbonyl vapor metallurgy processing is discussed in the article "Production of Nickel-Base Powders" in this Volume). The copper is generally added for size control during sintering, because nickel induces shrinkage and copper causes expansion. The copper activates sintering, as noted previously in the section "Iron-Copper Alloys," and promotes the dissolution of nickel in the iron. Nickel-steel structures are shown in Fig. 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, and 98. Nickel-rich regions comprise 20 to 50% of the area of these structures. The regions are extensive because the nickel content of their interiors has been diluted to 12% Ni by inward diffusion of iron. The nickel-rich regions tend to etch lightly. Their interiors often are unetched austenite, and their peripheries contain martensite or bainite with microhardnesses of 40 to 55 HRC, converted from 100 gf Knoop. The pearlite colonies are usually surrounded by a white band that appears similar to ferrite, but never contains eutectoid products. It is probably a higher-alloy diffusion zone. The austenitic cores of the nickelrich regions increase toughness and strength in these alloys and tend to inhibit ductility. The undiffused nickel-rich regions figure significantly in the overall performances of the alloy. These islands with hard phases in the as-sintered condition contribute a degree of wear resistance that would not normally be expected. It is difficult to assess the degree of sintering by studying the nickel-rich areas because copper additions greatly affect their extent and appearance. Sintering is best judged by the disappearance of original particle boundaries and by pore rounding. It is difficult to discern the combined carbon level in the nickel steels because of the presence of the nickel-rich regions, the white diffusion layer, porosity, and the probable lowering of the eutectoid carbon level by the nickel.

Diffusion-alloyed materials, such as Distaloy (Hoeganaes Corp., Riverton, NJ) are powders in which the alloying elements of molybdenum, nickel, and copper are added as finely divided elements or oxides to the iron powder. They are then coreduced with the iron powders at an annealing step, resulting in the firm attachment and partial diffusion of the elements to the iron. This partial alloying increases hardenability compared to elemental mixtures, yet these powders exhibit good compressibility. Bonding of the alloying elements also reduces the tendency toward powder segregation.

The sintered structures exhibit ferrite, pearlite, and nickel-rich regions such as those described previously for the elemental mixes, and the nickel-rich regions have all the benefits noted above. With added copper, additional partial hardening during sintering occurs. In Europe, this is used to advantage by producing medium-carbon alloys that are sold in the pressed, sintered, and sized conditions, but have good strength and impact resistance. This procedure avoids the distortions that can occur during normal heat treating. Diffusion-alloyed structures are shown in Fig. 99, 100, and 101. Sintered stainless steels are available in compositions that approximate AISI designations 303, 304, 316, 409, 410, 430,

and 434. The austenitic materials display austenite grains and annealing twins. The most significant disadvantage may be decoration of the grain boundaries with chromium carbides, indicating loss of chromium from solution and reduction in corrosion resistance. The degree of pore rounding is the most important indication of strength and ductility. The materials are virtually always prepared from prealloyed powders; some variants contain added tin or copper for improved corrosion resistance. The 410 materials are often fabricated with 0.15% graphite mixed with prealloyed powders. This results in such high hardenability that the as-sintered structures are essentially all martensite and require tempering after sintering for optimal properties. Stainless steel P/M structures are shown in Fig. 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, and 113. Powder metallurgy tool steels have long been used for tooling components such as punches and dies. These materials

are produced by hot isostatic pressing of water-atomized, tool steel powders, resulting in a fully dense product with fine grain size and very fine, uniform carbide size. The product displays grindability that is superior to ingot-base tool steels. Such alloys as M2 and T15 are also available in molding grade powders. In addition to hot isostatic pressing, P/M tool steels can be fabricated by pressing to approximately 80% density, followed by vacuum sintering to full density. Tool steel powders of the M2 and T15 compositions can be cold pressed at 550 to 825 MPa (40 to 60 tsi), then liquid phase sintered to full density. For M2, sintering requires 1 h in vacuum at 1240 °C (2260 °F) at 100 to 1000 m nitrogen or argon; T15 takes 1 h at 1260 °C (2300 °F) in the same vacuum. Temperature control within 5 °C (9 °F) may be required for product uniformity. The assintered T15 structures contain retained austenite, because of the high amount of carbon in solution, as well as primary M6C and fine MC (vanadium carbide). The M2 structures contain mainly M6C of varying small sizes against a matrix of retained austenite. The martensite start, Ms, temperature for these materials with the high carbon in solution is below room temperature. Upon annealing, the carbon precipitates out of solution onto the M6C phase, reducing the carbon in the matrix. This structure may then be heat treated at 1150 to 1205 °C (2100 to 2200 °F), but heating and cooling times must be minimized to avoid putting too much carbon back into solution. Upon furnace cooling or air cooling, the matrix then forms martensite with the proper distribution of fine carbides (Ref 12). Powder metallurgy tool steel structures are shown in Fig. 114, 115, and 116 (see also the article "Particle Metallurgy Tool Steels" in this Volume). Nonferrous P/M Materials As discussed in the Section "Metal Powder Production and Characterization" in this Volume, a great many nonferrous metals are also produced in powder form, including: • • • • • • • • •

Copper: by reduction of oxides, atomization, electrolysis, and hydrometallurgical processing Tin: by atomization Aluminum: by atomization Magnesium: by mechanical comminution and atomization Nickel: by carbonyl vapormetallurgy, hydrometallurgy, and atomization Cobalt: by carbonyl vapor metallurgy, hydrometallurgy, reduction of oxides, and atomization Silver: by chemical precipitation, electrolysis, and reduction of oxides Gold, platinum, and palladium: by chemical precipitation Tungsten and molybdenum:by reduction of oxides

• • • • • •

Metal carbides: by carburization, Menstruum process, and exothermic thermite reactions Tantalum: by reduction of potassium tantalum fluoride and a sequence of electron beam melting, hydriding, comminution, and degassing (dehydriding) Niobium: by aluminothermic reduction of oxides Titanium: by reduction of oxides and atomization Beryllium: by reduction of vacuum-melted ingots by comminution Composite powders: by diffusion (alloy coating)

This section reviews copper-, titanium-, and aluminum-base P/M materials. Copper-base alloys include pure copper for high-density electrical applications: 90Cu-10Sn bronzes for bearings and structural parts; brasses with 10, 20, and 30% Zn; and nickel silver (Cu-18Zn-18Ni). The brasses and nickel silvers are used for structural parts that require ductility, moderate strength, corrosion resistance, and decorative value. Copper exhibits a single-phase structure with some annealing twins. The most significant feature is the particle boundaries or their absence. There should be virtual freedom from particle boundaries from the surface to the center of the part. Bronzes should display all -bronze with no gray copper-tin intermetallic compounds. Optimal mechanical properties and machinability dictate a minimum of reddish copper-rich areas and small grain clusters of bronze. Mixes containing admixed graphite show the mottled gray flakes in the pores of the part. Bearings exhibit varying degrees of sintering, depending on the final application. In general, however, a well-sintered bearing results in greater ease of oil impregnation. Bronze P/M structures are shown in Fig. 117, 118, 119, and 120. Brasses and nickel silvers are generally single-phase structures. They should display good pore rounding and almost no original particle boundaries. Some of the materials may contain up to 2% Pb within the particles as an aid to machinability; this will appear as a fine, rounded gray phase (Fig. 121 and 122). Titanium and titanium alloys such as Ti-6Al-4V are produced from metal powders in several ways. The powders may be prealloyed or may be an elemental mix of titanium and a master alloy of vanadium and aluminum. The latter can be pressed and vacuum sintered to an impermeable state, which may then be hot isostatically pressed to full density without a can. The prealloyed materials may be vacuum hot pressed or preformed, canned, and hot isostatically pressed to full density. Titanium alloys can also be consolidated by metal injection molding. Titanium alloy P/M structures are shown in Fig. 123, 124, and 125. Aluminum P/M alloys are pressed and sintered to 90 to 95% density. The common alloys are 201AB and 601AB. The

alloys are prepared using low-alloy aluminum powder with additions of elemental or master alloy copper, magnesium, and silicon. During sintering, the additions cause a liquid phase to form that fluxes away the surface oxides and allows bonding between the aluminum particles. Sintering in nitrogen is performed at approximately 595 or 620 °C (1100 or 1150 °F) at a dew point of -50 °C (-60 °F) to prevent further oxidation of the aluminum. After sintering, the alloys are often solutionized and quenched, then repressed or coined before aging. The repressing densifies the material and establishes close dimensional tolerances. The materials may also be cold forged or rolled to varying reductions in thickness because of their favorable assintered ductility. Aluminum P/M structures are shown in Fig. 126, 127, 128, 129, and 130.

Reference cited in this section

12. M. Svilar, SCM Metal Products, Cleveland, OH, personal communication

Metallography of Powder Metallurgy Materials Leander F. Pease III and Douglas L. Pease, Powder-Tech Associates, Inc.

Representative Micrographs This section includes a discussion of unusual and/or defective structures. Also included are examples of heat treated materials and those subjected to other finishing operations, such as steam blackening. Alternate consolidation processes, such as P/M forging, hot isostatic pressing, metal injection molding, and liquid-phase sintering, are also illustrated. Sintered parts may be undersintered, which is evidenced by the presence of excessive numbers of original particle boundaries. Undersintering is related to the normal pressed-and-sintered structural materials and their mechanical properties as shown in MPIF standard 35 or the various ASTM materials standardized by the B-9 Committee in ASTM Volume 02.05. In general, for ferrous materials, a field of view at 200× would not be expected to show more than approximately five small segments of original particle boundaries. The presence of larger numbers of particle boundaries would necessitate verifying the sintering conditions and the strength of the part. Figures 27, 28, 29, 30, and 31 depict an increasing degree of sintering, as shown by the disappearance of particle boundaries. High-temperature (1290 °C, or 2350 °F) sintered austenitic stainless steel does not exhibit particle boundaries, and the degree of rounding of the pores must be examined to compare sintering (Fig. 102, 103, 104, 105, 106, 107, 108, 109, 110, 111). Injection-molded parts made of fine powders tend to sinter to a closed-pore state with no original particle boundaries (Fig. 131 and 132). Powder metallurgy forgings and hot isostatically pressed parts would not display such boundaries (Fig. 133 and 134). In the etched condition, sintered steels may exhibit carburization or decarburization (Fig. 135 and 136). If parts of nonuniform section are pressed, density may vary, which may be noted and measured metallographically (Ref 15). If. parts are overpressed, the particles will separate, showing microlaminations. Cracks may occur upon ejection, at the change in diameter between two sections of a part, such as between a hub and a flange (Fig. 137 and 138). Even in simple shapes, such as flat tensile bars, improper tool design can cause cracks, which then result in reduced mechanical properties (see Fig. 139). Heat treated ferrous parts vary in structure from nearly all martensite at the surface to a mixture of martensite, ferrite, and 10 to 30% fine pearlite in the interior (Fig. 140, 141, 142). This fine pearlite improves tensile properties (Ref 16). Microhardness testing must be limited to a particular phase when testing with the 100 gf Knoop indenter, for example, martensite. The heat treated structures may display retained austenite, carbides, and subsurface quench cracking (Fig. 143, 144, 145). Most P/M materials do not form a definite shallow case because of penetration of the carburizing gases. At densities above approximately 7.2 g/cm3, a definite case tends to form if the core contains less than 0.2% C, as shown in Fig. 146. Powder metallurgy parts can be finished by steam blackening. The degree of blackening, which should be controlled, generally lowers impact and tensile properties (Ref 17). The gray Fe3O4 layer (Fig. 147, 148) penetrates the pores and increases compressive strength and abrasive wear resistance. The thickness of the oxide layer can be measured metallographically, 10 m (0.0004 in.). Most P/M parts that are to be plated are first impregnated with a resin to prevent the corrosive plating solutions from entering and remaining in the pores. The resin is visible using optical metallography. The various plated layers are also visible, but polishing should be limited to 3 m and 1 m diamond on a short-nap (DAC) cloth to prevent rounding of the plated edge. To examine the original resin in the pores, the part should be mounted in an epoxy resin of contrasting color, for example, red. Also, impregnation with the red epoxy of a specimen that has no existing resin in the pores, allows the epoxy resin to be easily seen. The gray epoxy resins are difficult to image in the light microscope. Often, unfilled pores look dark or filled with debris. An epoxy-filled pore allows one to look through the epoxy at 1000× and see a specular reflection off the bottom.

Powder metallurgy parts can be joined to others by brazing, welding, or adhesive bonding; special precautions are necessary to prevent penetration of the brazing materials. Brazing of sintered steel is done using a special material with a melting point near the normal sintering temperature of iron. When it begins to penetrate iron pores, some iron dissolves in the liquid braze. This raises the solidus or melting point of the liquid, freezing occurs, and penetration proceeds no further into the porous material. Manufacturers of sintered parts have occasion to examine raw materials (powders) metallographically. This is important because different production methods can result in powders with the same nominal chemistry, but disparate properties. Typical powder structures are shown in Fig. 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, and 162. These micrographs are intended to simplify the task of examining a sintered part and attempting to deduce which powder was used to prepare it. Figures 163, 164, 165, 166, 167, 168, 169, and 170 depict various normal and abnormal structures. Among these are undersintered structures (as shown in Fig. 163 and compared with an average sinter in Fig. 164), wear-resistant steels (Ancorwear 500) with high carbon contents (Fig. 165), gravity-sintered bronze filter powders (Fig. 166 and 167), sintered parts that were blistered during heating to the sintering temperature (Fig. 168), and parts that exhibit varying density from die-wall friction and the location in the die (Fig. 169 and 170). Sealed Surfaces. Powder metallurgy parts can have their surface pores closed, either intentionally or inadvertently. Figure

171 shows the outside diameter of a tumbled P/M part in which the surface pores were closed to a depth of 50 m (0.002 in.). Such closure can interfere with intended flow of fluids through the pores. Powder metallurgy parts, for example, gears and sprockets, are deliberately worked on the surface to densify the surfaces and increase the fatigue resistance. Machining P/M parts also is known to close off surface pores. Determination of Crack Origin, Green or Sintered State. If a fracture occurs in the green state and the part is then

sintered, the particles on the fracture surface will be rounded. If a fracture occurs after sintering, the surfaces of the particles show ductile cups, cones, and asperities, as in the scanning electron micrograph (SEM) view in Fig. 34, 36, and 37. If an SEM is not available, one can plate the fracture surface with 25 m (0.001 in.) electroless nickel, Edgemet (Buehler Ltd., Lake Bluff, IL), in about 2 h. The plating is done on a specimen that was sectioned perpendicular to the fracture surface. The specimen is then mounted in epoxy resin, ground, and polished. The plating saves the details of the fracture surface, which is viewed at right angles to the fracture plane. In this way, any asperities and cones can be seen in profile view, at 1000×. It is best to create a deliberate fracture through a sintered region to serve as a comparison with the questioned fracture. Figure 172 shows the fracture that occurred in a green part, and it has rounded particle surfaces. Figure 173 shows a fracture that occurred in a sintered and heat treated part, and the asperities where sinter bonds were broken are shown by the arrows. The plating above keeps epoxy resin from filling pores during sample mounting. This creates some difficulty at polishing, in trying to open all the pores. This is best accomplished at final polishing with 0.05 m alumina on microcloth with a succession of 15 s nital etchings and 2 to 3 min polishing with 1 N force. Several iterations may be required. Smeared pores are most easily viewed at 1000× and are seen as thin gray lines in the unetched structure. Sinter Hardened Steels. The FLC-4608-XXHT alloys contain enough copper, molybdenum, nickel, and carbon to harden

to a largely martensitic structure during typical cooling from the sintering temperature. The structures show martensite with 0 to 40% colonies of dark fine pearlite, as shown in Figures 140 and 141. Such structures can be distinguished from austenitized and oil-quenched structures by the distribution of the fine pearlite. The sinter-hardened structures tend to have the fine pearlite distributed rather uniformly from surface to interior. Oil-quenched parts tend to be all martensite on the surface, and any fine pearlite is mainly formed in the interior. Sinter-hardened parts with more than 70% martensite should be tempered for 1 h at 175 °C (350 °F) in air to relieve brittleness and increase the tensile strength. Such tempering can double tooth strength, as compared with using the part in the as-sinter-hardened condition.

References cited in this section

15. L.F. Pease III, Inspection and Quality Control for P/M Materials, Powder Metallurgy, Vol 7, Metals Handbook,

9th ed., American Society for Metals, 1984, p 483 16. L.F. Pease III, The Mechanical Properties of Sintered Steels and Their Derivation for MPIF Standard 35, Progress in Powder Metallurgy, Vol 37, Metal Powder Industries Federation, 1981 17. L.F. Pease III, J. Collette, and D.A. Pease, Mechanical Properties of Steam Blackened P/M Materials, Modern Developments in Powder Metallurgy, Vol 18-21, Metal Powder Industries Federation, 1988, p 275 Metallography of Powder Metallurgy Materials Leander F. Pease III and Douglas L. Pease, Powder-Tech Associates, Inc.

References 1. "Standard Test Method for Density and Interconnected Porosity of Sintered Powder Metal Structural Parts and Oil-Impregnated Bearings," B 328, Annual Book of ASTM Standards, Vol 2.05, ASTM, 1998 2. "Permeable Sintered Metal Materials, Determination of Density, Open Porosity and Oil Content," ISO 2738, International Organization for Standardization, American National Standards Institute 3. "Standard Method of Sampling Finished Lots of Metal Powders," B 215, Annual Book of ASTM Standards, Vol 2.05, ASTM, 1998 4. G.A. Blann et al., Principles and Practices for Metallographic Preparation for Powder Metallurgy, Microstr. Sci., Vol 22, 1995, p 135 5. L.F. Pease III, Inspection and Quality Control for P/M Materials, Powder Metallurgy, Vol 7, Metals Handbook, 9th ed., American Society for Metals, 1984, p 485-486 6. L.F. Pease III, Metallography and Properties of Sintered Steels, Progress in Powder Metallurgy, Vol 33, Metal Powder Industries Federation, 1977 7. O. Struglics, Hoeganaes Corp., Hoeganaes, Sweden, personal communication 8. S. Coleman and D. Tomkins, A Quantitative Assessment of the Mechanical Properties of Sintered Iron Micrographic Specimens, Powder Metall., No. 2, 1976, p 53 9. "Effective Case Depth of Powder Metallurgy Products, Determination of," Standard 52, Metal Powder Industries Federation 10. S. Kaufmann, Ford Motor Company, Dearborn, MI, personal communication 11. P. Schmey, United States Bronze Powders, Inc., Flemington, NJ, personal communication 12. M. Svilar, SCM Metal Products, Cleveland, OH, personal communication 13. S. Abkowitz, Dynamet Technology, Burlington, MA, personal communication 14. G.F. Millsaps, Alcoa, Pittsburgh, PA, personal communication 15. L.F. Pease III, Inspection and Quality Control for P/M Materials, Powder Metallurgy, Vol 7, Metals Handbook, 9th ed., American Society for Metals, 1984, p 483 16. L.F. Pease III, The Mechanical Properties of Sintered Steels and Their Derivation for MPIF Standard 35, Progress in Powder Metallurgy, Vol 37, Metal Powder Industries Federation, 1981 17. L.F. Pease III, J. Collette, and D.A. Pease, Mechanical Properties of Steam Blackened P/M Materials, Modern Developments in Powder Metallurgy, Vol 18-21, Metal Powder Industries Federation, 1988, p 275 18. J. Hurst, C.I. Hayes, Inc., Cranston, RI, personal communication 19. F. Hanejko, Hoeganaes Corporation, Riverton, NJ, personal communication

Metallography of Powder Metallurgy Materials Leander F. Pease III and Douglas L. Pease, Powder-Tech Associates, Inc.

Selected References • • •

W.J. Hupmann and K. Dalal, Metallographic Atlas of Powder Metallurgy, Verlag Schmid, GmbH, 1986 Metallographic Handbook, Ferrous Powder Metallurgy, Hoeganaes Corporation, 1995 G.F. Vander Voort, Metallography: Principles and Practice, McGraw Hill, 1984

Metallography Materials

of

Powder

Metallurgy

Leander F. Pease III and Douglas L. Pease, Powder-Tech Associates, Inc.

Fig. 12 Pyron 100, hydrogen-reduced sponge iron. A single particle, arrows indicate pores opening into the spongy interior. SEM. 1000×

Fig. 13 Pyron D63, hydrogen-reduced sponge iron, exhibiting high apparent density. SEM. 750×

Fig. 14 MH-100, carbon-reduced iron ore. Arrows indicate one particle with coarse internal porosity. SEM. 750×

Fig. 15 Ancormet 101, carbon-reduced iron ore. One individual particle with coarse and extensive porosity is shown. SEM. 750×

indicate this process can produce iron powder with a fair degree of irregularity or roughness on the surface. SEM. 190×

Fig. 16 Atomet 28 iron powder. Arrows indicate porosity in the spongy regions. SEM. 750× Fig. 19 Ancorsteel 1000, water-atomized and annealed iron powder. Arrows indicate small fines that were agglomerated onto the larger particles. SEM. 190×

Fig. 17 MP35HD iron powder. Arrows indicate porosity in spongy regions. SEM. 750× Fig. 20 Ancorsteel 1000B, water-atomized and double-annealed iron powder. SEM. 190×

Fig.

18

Water-atomized

iron.

Arrows

Fig. 21 Ancorsteel 4600V, water-atomized and annealed prealloyed steel powder. Note that some particles gain surface area and irregularity by agglomeration of fines (see arrow). SEM. 750×

Fig. 22 SCM A283 electrolytic iron powder. Note the flaky shape characteristic of these powders. SEM. 190×

Fig. 23 Type 316, gas-atomized stainless steel powder. Note attached satellites. SEM. 750×

Fig. 24 Type 316L, rotating electrode processed stainless steel powder. Nearly perfect spheres with absence of satellite formation. SEM. 190×

Fig. 25 Ancorsteel 1000 unsintered iron powder. Surface of part, which had been contacted by the upper punch at 275 MPa (20 tsi). Arrow shows the particle boundaries that will disappear during proper sintering. SEM. 750×

Fig. 26 Same as Fig. 25, but showing the view of the surface that was in contact with the die wall. Arrows show the boundary

between particles that must be eliminated during sintering. SEM . 750×

Fig. 27 Distaloy 4600 A (6.7 g/cm3), pressed at 480 MPa, undersintered 5 min in dissociated ammonia in hot zone at 1120 °C (2050 °F). Arrows P: particle boundaries; arrows G: undiffused, gray flakes of graphite in pores. As-polished. 645×

Fig. 28 Same diffusion-alloyed steel as Fig. 27. Arrows P show the many original particle boundaries. Sintering longer will remove these low-strength boundaries. Aspolished. 120×

Fig. 29 Same as Fig. 27, but sintered 15 min. Arrows P indicate persistence of original particle boundaries; arrows R, rounded pores (compare with Fig. 27). Aspolished. 645×

Fig. 30 Same as Fig. 29. Arrows S show segments of original particle boundaries that are shorter and less numerous than those in Fig. 28. Arrows P indicate a row of pores that show how original particles break down into planes of small voids, which coalesce or disappear from diffusion. Aspolished. 180×

Fig. 31 Same as Fig. 27, but sintered approximately 37 min, which is longer than average. Structure still shows a few segments of original particle boundaries (arrow S). Arrows P indicate a row of pores at which a particle boundary is disappearing. As-polished. 180×

Fig. 34 Same as Fig. 33, but sintered 10 min in hot zone (approximately 1 to 3 min at 1120 °C, or 2050 °F). Arrows D show the ductile cup-and-cone fractures that occurred when this particle was torn from the adjacent one above it. Arrows S show the smooth surface of the particle that had not sintered to any adjacent particle. SEM. 750×. (Ref 10) Fig. 32 Fracture surface of Ancorsteel 1000 iron powder (6.4 g/cm3) pressed without lubricant at 275 MPa (20 tsi). Structure shows no evidence of cold welding or bonding of adjacent particles. Arrow indicates a triple particle boundary that will disappear during sintering. SEM. 750×. Source: Ref 10

Fig. 35 Same as Fig. 34, but sintered 10 min in the hot zone at 1120 °C (2050 °F). Arrow D indicates a ductile cup-and-cone fracture where this particle was joined to the one above it. Arrows N show necks forming between two adjacent particles. These necks (solid regions) replace particle boundaries as sintering progresses. SEM. 2850×. (Ref 10) Fig. 33 Fracture surface of Atomet 28 iron powder pressed to 6.6 g/cm3 and sintered 3 min in hot zone at 1120 °C (2050 °F) in dissociated ammonia. Arrows D show where a bond has broken. Arrows S outline the smooth, rounded surface of a particle that did not bond to the adjacent particle above it. SEM. 750×. (Ref 10)

Fig. 36 Same as Fig. 33, but sintered 20 min in the hot zone. Arrows D show the development of ductile cup-and-cone fracture dimples formed when material was torn away from the adjacent particle. Arrows S indicate smooth surfaces where no adjacent particle bonding has occurred. SEM. 750×. (Ref 10)

Fig. 39 Pyron D63 sponge iron (6.2 g/cm3), pressed at 480 MPa (35 tsi) and sintered 30 min at 1120 °C (2050 °F) in dissociated ammonia. Mainly ferrite grain boundaries. Arrow O indicates a small, gray, unreduced oxide particle; arrows C, a few isolated carbide platelets. 2% nital. 645× Fig. 37 Same material (Atomet 28) and processing as described in Fig. 36, but shown at higher magnification. Most of the field of view shows the ductile cup-and-cone fractures that occur when the material is torn apart. SEM. 2850×. (Ref 10)

Fig. 40 Same as Fig. 39, but not etched. Arrows S surround a spongy particle having small, internal pores. Arrow P indicates a much larger pore between powder particles. Very few original particle boundaries are present. As-polished. 180× Fig. 38 Same as Fig. 33, but sintered 40 min in the hot zone. Approximately 50% of the area fraction is occupied by ruptured ductile bonds (arrows D). The remaining area consists of smooth surfaces of particles (arrows S) at which no bonding has occurred. SEM. 750×. (Ref 10)

Fig. 41 Pyron 100 sponge iron (6.2 g/cm3), pressed at 480 MPa (35 tsi) and sintered 30

min at 1120 °C (2050 °F) in dissociated ammonia. Average sinter, no residual particle boundaries. Dark areas are pores. 2% nital. 960×

Fig. 42 Same as Fig. 41, but increased density (6.4 g/cm3). Arrows S surround a spongy particle having small internal pores. Arrow P indicates a larger pore between particles. As-polished. 180×

Fig. 43 MH-100 sponge iron (6.4 g/cm3), pressed at 480 MPa (35 tsi) and sintered 30 min at 1120 °C (2050 °F) in dissociated ammonia. Arrows S indicate the various pore sizes in different particles. The larger pores (arrow P) are between the original particles. As-polished. 180×

Fig. 44 Same as Fig. 43, but at higher magnification. Arrows E show eutectoid (pearlite) indicating 1120 °C, or 2050 °F) and longer sintering times (greater than 20 min at temperature) will, however, influence the number of pores that are present in the microstructure as well as their size and their shape. Pore coalescence occurs as sintering temperature and time are increased; large pores grow at the expense of the smaller pores. Increased time and temperature also lead to an increased amount of pore rounding. Certain applications may require supplemental alloy additions via admixing to enhance the hardenability and mechanical properties of the material. Such materials are termed hybrid alloys. Hybrid Alloy Powders. Hybrid alloys consist of either prealloyed or partially alloyed base powders to which elemental

or ferroalloy additions are made to achieve the desired chemical composition. The practice of making nickel or copper additions to prealloyed powders has become widespread since the introduction of the highly compressible, prealloyed powders with molybdenum as the principal alloy addition (Ref 2, 7). These prealloyed powders are also particularly suitable for the addition of high-carbon ferroalloys to produce chrome-molybdenum, manganese-molybdenum, and chrome-molybdenum-manganese steels (Ref 8). Elemental nickel additions can be made to partially alloyed powders to provide even greater impact properties (Ref 9, 10). The compressibility and hardenability of hybrid alloy materials depends primarily on the compressibility and hardenability of the base powder employed. Hybrid materials generally have a heterogeneous microstructure with nonuniform apparent hardness. Higher sintering temperatures and longer sintering times generally lead to greater diffusion of the alloy additions and to more homogeneous microstructures.

References cited in this section

2. W.B. James, Recent Developments in Ferrous Powder Metallurgy Alloys, Int. J. Powder Metall., Vol 30 (No. 2), 1994, p 157-162 6. P.F. Lindskog and G.F. Bocchini, Development of High Strength P/M Precision Components in Europe, Int. J. Powder Metall., Vol 15 (No. 3), 1979, p 199-230 7. Hoeganaes Technical Data Sheets, Ancorsteel 150 HP, and Ancorsteel 85 HP, Hoeganaes Corporation, Riverton, NJ 8. W.B. James and R.J. Causton, Surface-Hardenable Heat Treated P/M Steels, Advances in Powder Metallurgy & Particulate Materials, Vol 5, J.M. Capus and R.M. German, Ed., Metal Powder Industries Federation, 1992, p 62-92 9. W.B. James, V.C. Potter, and T.F. Murphy, "Steering Column Tilt Lever--P/M Material Development," SAE Technical Paper 900381, SAE Congress and Exposition (Detroit, MI), Feb 1990 10. V.C. Potter, W.B. James, and T.F Murphy, Improved Dimensional Control and Elimination of Heat Treatment for Automotive Parts, Advances in Powder Metallurgy, Vol 3., L.F. Pease III and R.J. Sansoucy, Ed., Metal Powder Industries Federation, 1990, p 33-48 Ferrous Powder Metallurgy Materials W.B. James, Hoeganaes Corporation; G.T. West, National Sintered Alloys Inc.

Ferrous Powder Materials Material code designations for ferrous P/M materials are found in MPIF Standard 35, Materials Standards for P/M Structural Parts. Similar designations have been adopted by ASTM and are summarized in specification B 783 (Ref 11). These standards are subject to periodic review, and users are cautioned to refer to the latest edition. Ferrous material designations have various prefixes: • • • • • • •

F, iron FC, iron-copper or copper steel FD, diffusion alloyed steel (partially alloyed) FL, prealloyed ferrous material except stainless steel FN, iron-nickel or nickel steel FX, copper infiltrated iron or steel SS, stainless steel (prealloyed)

Figure 7 provides examples of material designations. The designation FN-0205-35 (Fig. 7a) refers to a nickel steel with nominally 2 wt%Ni, 0.3 to 0.6 wt% combined carbon, and a guaranteed minimum 0.2% offset yield strength of 240 MPa (35 ksi). The FN prefix designates an admixed nickel steel, and the 02 following the prefix indicates that there is 2 wt% of the major alloy addition (nickel). The 05 refers to the combined carbon content of the material. The code designations for various combined carbon contents are:

Combined carbon range, wt%

Code designation

0-0.3

00

0.3-0.6

05(a)

0.6-0.9

08

(a) The 05 designation for the FL materials (prealloyed) refers to 0.4 to 0.7 wt% combined carbon.

The designation FL-4405-175HT in Fig. 7(b) refers to a quench-hardened and tempered, prealloyed steel with a combined carbon content of 0.4 to 0.7 wt% and a guaranteed minimum ultimate tensile strength of 1210 MPa (175 ksi).

Fig. 7 Examples of ferrous P/M material designations. (a) FN-0205-35 and (b) FL-4405-175HT

For as-sintered materials the material designation suffix refers to the guaranteed minimum 0.2% offset yield strength in thousands of pounds per square inch. For example, FC-0208-50 refers to an admixed copper steel with nominally 2 wt% Cu, a combined carbon content of 0.6 to 0.9 wt%, and a guaranteed minimum 0.2% offset yield strength of 340 MPa (50 ksi).

For heat treated materials the material designation suffix refers to the guaranteed minimum ultimate tensile strength in thousands of pounds per square inch. For example, FN-0205-130HT refers to an admixed nickel steel with nominally 2 wt% nickel, a combined carbon content of 0.3 to 0.6 wt%, and a guaranteed minimum ultimate tensile strength of 900 MPa (130 ksi). Admixed Powders Admixed materials are in widespread use throughout the ferrous P/M industry. They generally contain sufficient graphite to provide the desired carbon content after sintering because the typical base powders used are essentially carbon free. Base powders generally have low carbon contents because carbon forms an interstitial solid solution with iron and reduces the compressibility of the powder significantly. In addition to graphite, the mixes contain a lubricant that aids particle rearrangement during the initial stages of compaction. The principal purpose of the added lubricant is, however, to reduce the friction between the powder particles and the compaction tooling members (die wall, core rods, and punches) and make ejection of the compacted part easier. Lubricant additions vary from 0.5 wt% to sometimes >1 wt%, depending on the part geometry, part density, and the surface area of the compact at the die-wall and core rod interfaces. Lubricant additions of 0.75 wt% are quite common. The lubricants used are generally metallic stearates (e.g., zinc stearate) or ethylene bisstearamides. They are white solids with pycnometric densities of 1 g/cm3. Because of their low density relative to that of iron, while they are added typically in quantities of about 0.75 wt%, they occupy 5.5 vol% in the mixture. This has particular significance when parts are compacted at high pressures (Ref 12, 13, 14). Because the various powders that comprise the admixed alloys have differing particle size and density, admixed alloys are particularly susceptible to dusting and segregation during handling and transfer to the die cavity. Numerous opportunities exist during the P/M process for dusting and segregation to occur (Ref 15). Compositional variations, which result from these demixing phenomena, cause inconsistencies in the green and sintered properties of P/M parts. The tendency for mixes to dust and segregate can be reduced significantly through the use of binder treated mixes (Ref 16, 17, 18, and 19). In addition to reducing dusting and segregation, binder treated mixes improve powder flow and die filling characteristics during part manufacturing. This results in more consistent part sectional densities and improved control of part mass. Improved flow and die fill uniformity result in reduced press cycle times and more consistency throughout the entire part manufacturing process. Improved retention of alloy additions provides an economic advantage in terms of reduced amounts of green scrap, enhanced alloy efficiency, and a more pleasant and easier to maintain work environment. Plant cleanliness is dramatically improved, and the amount of respirable dust in the immediate vicinity of the compaction press is reduced by an order of magnitude (Fig. 8) (Ref 2). Binder treated premixes are produced using a proprietary mixing process that utilizes patented binders (Ref 16, 17, 18, and 19).

Fig. 8 Respirable dust in the work area of a P/M parts making plant. Source: Ref 2

Iron and Carbon Steels. Table 1 summarizes the chemical compositions of the iron and carbon steels listed in MPIF

Standard 35.

Table 1 Chemical composition of iron and carbon steels Material designation

F-0000

F-0005

F-0008

Chemical composition, wt%

Fe

C

Element

97.7

0.0

min.

100.0

0.3

max.

97.4

0.3

min.

99.7

0.6

max.

97.1

0.6

min.

99.4

0.9

max.

Other elements: Total by difference equals 2.0 wt% maximum, which may include other minor elements added for specific purposes.

Unalloyed iron (F-0000) materials are used for lightly loaded structural applications and structural parts requiring self lubrication where strength is not critical. The combined carbon content of these materials can be from 0 to 0.3 wt%. Iron parts that are essentially carbon free are often used, particularly at higher part densities, for their soft magnetic properties (Ref 20, 21, 22). P/M carbon steels with 0.3 to 0.6 wt% combined carbon (F-0005) possess moderate strength and apparent hardness. They are used where such properties combined with machinability are desired. Higher carbon P/M steels (F-0008) have moderately higher strength compared with F-0005 materials, but are more difficult to machine. In the as-sintered condition carbon steels have a ferrite/pearlite microstructure (Fig. 9). Both F-0005 and F-0008 steels can be heat treated to increase tensile strength, improve apparent hardness, and enhance wear resistance. Heat treated carbon steels have a martensitic microstructure (Fig. 10). Table 2 summarizes the mechanical properties of various carbon steels.

Table 2 Mechanical properties of iron and carbon steels Minimum values

Typical values

Material designation code

Minimum strength, ksi

Tensile properties

Yield

Ultimate

Ultimate strength, ksi

Yield strength (0.2%), ksi

F-0000-10

10

...

18

13

F-0000-15

15

...

25

F-0000-20

20

...

F-0005-15

15

F-0005-20

Elongation in 25 mm (1 in.), %

Elastic constants

Unnotched Charpy impact energy, ft · lbf

Transverse rupture strength, ksi

Compressive yield strength (0.1%), ksi

Rockwell hardness

Macro (apparent)

Fatigue limit (90% survival), ksi

Density, g/cm3

Young's modulus, 106 psi

Poisson's ratio

1.5

15.0

0.25

3.0

36

16

40 HRF

N/A

7

6.1

18

2.5

17.5

0.25

6.0

50

18

60 HRF

N/A

10

6.7

38

25

7.0

23.5

0.28

35.0

95

19

80 HRF

N/A

14

7.3

...

24

18

80% theoretical density can be produced. By repressing, with or without resintering, parts of 90% theoretical density can be produced. The density attainable is limited by the size and shape of the compact. Forging of aluminum is a well-established technology. Aluminum also lends itself to the forging of P/M preforms to

produce structural parts. In forging of aluminum preforms, the sintered aluminum part is coated with a graphite lubricant to permit proper metal flow during forging. The part is either hot or cold forged; hot forging at 300 to 450 °C (575 to 850 °F) is recommended for parts requiring critical die fill. Forging pressure usually does not exceed 345 MPa (50 ksi). Forging normally is performed in a confined die so that no flash is produced and only densification and lateral flow result from the forging step. Scrap loss is 99.5% of

theoretical density. Strengths are higher than nonforged P/M parts, and in many ways, are similar to conventional forging. Fatigue endurance limit is doubled over that of nonforged P/M parts. Alloys 601AB, 602AB, 201AB, and 202AB are designed for forgings. Alloy 202AB is especially well suited for cold forging. All of the aluminum powder alloys respond to strain hardening and precipitation hardening, providing a wide range of properties. For example, hot forging of alloy 601 AB-T4 at 425 °C (800 °F) followed by heat treatment gives ultimate tensile strengths of 221 to 262 MPa (32 to 38 ksi), and a yield strength of 138 MPa (20 ksi), with 6 to 16% elongation in 25 mm (1 in.). Alloy 601AB. Heat treated to the T6 condition, 601AB has tensile properties of:

• • •

UTS, 303 to 345 MPa (44 to 50 ksi) Yield strength, 303 to 317 MPa (44 to 46 ksi) Elongation, up to 8%

Forming pressure and percentage of reduction during forging influence final properties. Alloy 201AB. In the T4 condition, alloy 201AB has tensile properties of:

• • •

UTS, 358 to 400 MPa (52 to 58 ksi) Yield strength, 255 to 262 MPa (37 to 38 ksi) Elongation, 8 to 18%

When heat treated to the T6 condition, the tensile strength of 201AB increases from 393 to 434 MPa (57 to 63 ksi). Yield strength for this condition is 386 to 414 MPa (56 to 60 ksi), and elongation ranges from 0.5 to 8%. Properties of cold-formed aluminum P/M alloys are increased by a combination of strain-hardened densification and improved interparticle bonding. Alloy 601AB achieves 257 MPa (37.3 ksi) tensile strength and 241 MPa (34.9 ksi) yield strength after forming to 28% upset. Properties for the T4 and T6 conditions do not change notably between 3 and 28% upset. Alloy 602AB has moderate properties with good elongation. Strain hardening (28% upset) results in 221 MPa (32 ksi) tensile and 203 MPa (29.4 ksi) yield strength. The T6 temper parts achieve 255 MPa (37 ksi) tensile strength and 227 MPa (33 ksi) yield strength. Highest cold-formed properties are achieved by 201AB. In the as-formed condition, yield strength increases from 209 MPa (30.3 ksi) for 92.5% density to 281 MPa (40.7 ksi) for 96.8% density.

Alloy 202AB is best suited for cold forming. Treating to the T2 condition, or as-cold formed, increases the yield strength significantly. In the T8 condition, 202AB develops 280 MPa (40.6 ksi) tensile strength and 250 MPa (36.2 ksi) yield strength, with 3% elongation at the 19% upset level. Conventional Aluminum Powder Metallurgy Alloys

Properties of Sintered Parts Mechanical Properties. Sintered aluminum P/M parts can be produced with strength that equals or exceeds that of iron or

copper P/M parts. Tensile strengths range from 110 to 345 MPa (16 to 50 ksi), depending on composition, density, sintering practice, heat treatment, and repressing procedures. Table 2 lists typical properties of four nitrogen-sintered P/M alloys. Properties of heat-treated, pressed, and sintered grades are provided in Table 3.

Table 3 Typical heat-treated properties of nitrogen-sintered aluminum P/M alloys Heat-treated variables and properties Solution treatment Temperature, °C (°F) Time, min Atmosphere Quench medium Aging Temperature, °C (°F) Time, h Atmosphere Heat-treated (T6) properties(a) Transverse-rupture strength, MPa (ksi) Yield strength, MPa (ksi) Tensile strength, MPa (ksi) Elongation, % Rockwell hardness, HRE Electrical conductivity, %IACS

(a)

Grades MD-22

MD-24

MD-69

MD-76

520 (970) 30 Air H20

500 (930) 60 Air H20

520 (970) 30 Air H20

475 (890) 60 Air H20

150 (300) 18 Air

150 (300) 18 Air

150 (300) 18 Air

125 (257) 18 Air

550 (80) 200 (29) 260 (38) 3 74 36

495 (72) 195 (28) 240 (35) 3 72 32

435 (63) 195 (28) 205 (30) 2 71 39

435 (63) 275 (40) 310 (45) 2 80 25

T6, solution heat treated, quenched, and artificially age hardened

Impact tests are used to provide a measure of toughness of powder metal materials, which are somewhat less ductile than similar wrought compositions. Annealed specimens develop the highest impact strength, whereas fully heat-treated parts have the lowest impact values. Alloy 201AB generally exhibits higher impact resistance than alloy 601AB at the same percent density, and impact strength of 201AB increases with increasing density. A desirable combination of strength and impact resistance is attained in the T4 temper for both alloys. In the T4 temper, 95% density 201AB develops strength and impact properties exceeding those for as-sintered 99Fe-1C alloy, a P/M material frequently employed in applications requiring tensile strengths under 345 MPa (50 ksi). Fatigue is an important design consideration for P/M parts subject to dynamic stresses. Fatigue strengths of pressed and sintered P/M parts may be expected to be about half those of the wrought alloys of corresponding compositions (see comparisons of two P/M alloys with two wrought alloys in Fig. 4). These fatigue-strength levels are suitable for many applications.

Fig. 4 Fatigue curves for (a) P/M 601AB (b) P/M 201AB

Electrical and Thermal Conductivity. Aluminum has higher electrical and thermal conductivities than most other metals. Table 4 compares the conductivities of sintered aluminum alloys with wrought aluminum, brass, bronze, and iron.

Table 4 Electrical and thermal conductivity of sintered aluminum alloys, wrought aluminum, brass, bronze, and iron Materials

Temper

601AB

T4 T6 T61 T4 T6 T61 T4 T6 T61 T4 T6 Hard Annealed Hard Annealed Hot rolled

201AB

602AB

6061 wrought aluminum Brass (35% Zn) Bronze (5% Sn) Iron (wrought plate)

(a) (b)

Electrical conductivity(a) at 20 °C (68 °F), %IACS 38 41 44 32 35 38 44 47 49 40 43 27 27 15 15 16

Thermal conductivity(b) at 20 °C (68 °F), cal/cm·s·°C 0.36 0.38 0.41 0.30 0.32 0.36 0.41 0.44 0.45 0.37 0.40 0.28 0.28 0.17 0.17 0.18

Determined with FM-103 Magnatester. Converted from electrical conductivity values

Machinability. Secondary finishing operations such as drilling, milling, turning, or grinding can be performed easily on

aluminum P/M parts. Aluminum P/M alloys provide excellent chip characteristics; compared to wrought aluminum alloys, P/M chips are much smaller and are broken more easily with little or no stringer buildup. This results in improved tool service life and higher machinability ratings. Conventional Aluminum Powder Metallurgy Alloys

Powder Degassing and Consolidation The water of hydration that forms on aluminum powder surfaces must be removed to prevent porosity in the consolidated product. Although solid-state degassing has been used to reduce the hydrogen content of aluminum P/M wrought products, it is far easier and more effective to remove the moisture from the powder. Degassing is often performed in conjunction with consolidation, and the most commonly used techniques are described below. The various aluminum fabrication schemes are summarized in Fig. 5.

Fig. 5 Aluminum P/M fabrication schemes

Can Vacuum Degassing. This is perhaps the most widely used technique for aluminum degassing because it is relatively

non-capital intensive. Powder is encapsulated in a can, usually aluminum alloys 3003 or 6061, as shown schematically in Fig. 6. A spacer is often useful to increase packing and to avoid safety problems when the can is welded shut. It has been found that packing densities are typically 60% of theoretical density when utilizing this method on mechanically alloyed powders. Care must be used to allow a clear path for evolved gases through the spacer to prevent pressure buildup and explosion.

Fig. 6 Degassing can used for aluminum P/M processing

To increase packing density, the powder is often cold isostatically pressed (CIP) in a reusable polymeric container before insertion into the can. Powder densities in the CIPed compact of 75 to 80% theoretical density are preferred because they have increased packing density with respect to loose-packed powder, yet allow sufficient interconnected porosity for gas removal. At packing densities of about 84% and higher, effective degassing is not possible for several atomized aluminumalloy powders (Ref 2, 3). Furthermore, one must control CIP parameters to avoid inhomogeneous load transfer through the powder, which can lead to excessive density in the outer regions of the cylindrical compact and much lower densities in the center. Such CIP parameters often must be developed for a specific powder and compact diameter. The canned powder is sealed by welding a cap that contains an evacuation tube as shown in Fig. 6. After ensuring that the can contains no leaks, the powder is vacuum degassed while heating to elevated temperatures. The rate in gas evolution as a function of degassing temperature depends on powder size, distribution, and composition. The ultimate degassing temperature should be selected based on powder composition, considering tradeoffs between resulting hydrogen content and microstructural coarsening. For example, an RS-P/M precipitation-hardenable alloy that is to be welded would likely be degassed at a relatively high temperature to minimize hydrogen content (hotter is not always better). Coarsening would not significantly decrease the strength of the resulting product because solution treatment and aging would be subsequently performed and provide most of the strengthening. On the other hand, a mechanically attrited P/M alloy that relies on substructural strengthening and will serve in a mechanically fastened application might be degassed at a lower temperature to reduce the annealing out of dislocations and coarsening of substructure. When a suitable vacuum is achieved (for example, DNi), the effect of interdiffusion is much stronger and depends on whether the component with the higher diffusivity (copper) is the minor or the major component. The faster diffusing species leaves porosity behind (Kirkendall porosity), causing expansion of the compact. An 80%Ni-20%Cu system shows more interdiffusion-induced porosity, while diffusion porosity is less pronounced in a 20%Ni-80%Cu system. Densification. When compacts from a single metal powder are sintered, they undergo dimensional change as sintering

proceeds. Figure 4 shows the dimensional changes that occur during sintering of copper powder compacts. In the initial stage, as the temperature increases, the compacts expand much like solid copper. The entrapped gases in the isolated pores and vaporized lubricant also contribute to this expansion, particularly in compacts pressed to high densities. When sintering begins, the compacts begin to shrink and reach a maximum at the peak sintering temperature. During cooling, the compacts contract like solid copper would. The total result of the three stages is generally a shrinkage and higher density.

Fig. 4 Dimensional change during sintering of compacts from -74+43 mm copper powder. Pressed at 138 MPa (20 ksi) and heated at a rate of 3.9 °C/min (7 °F/min) to 925 °C (1700 °F) and then cooled at the same rate. MPa = 6.8947 ksi. Source: Ref 16

Densification of P/M compacts during sintering depends on several variables, but principal factors are:

• • • •

Sintering temperature Sintering time Powder particle size Green density of the compact, which is primarily a function of compacting pressure

Typical sintering temperatures and times for various alloys are given in Table 10; copper alloys are generally sintered at much lower temperatures than iron- and nickel-base alloys. Typical shrinkage of a copper compact is shown in Fig. 5 for various sintering temperatures and times. The rate of shrinkage is initially high, but then decreases with increasing sintering time. Higher sintering temperatures promote a more rapid shrinkage than do longer sintering times.

Table 10 Typical sintering temperature and time of copper alloys and steels Material Bronze Copper Brass Iron, iron-graphite, etc. Nickel

Temperature °C 760-870 840-900 840-900 1010-1150 1010-1150

°F 1400-1600 1550-1650 1550-1650 1850-2100 1850-2100

Time, min 10-20 12-45 10-45 30-45 30-45

Fig. 5 Linear shrinkage of copper powder compacts from -75+44

m sieve fraction. Source: Ref 17

Another factor affecting densification is the particle size of the powder. In Fig. 6, the sintered density of copper powder compacts pressed at 276 MPa (40 ksi) from two particle size fractions, -105 + 75 and -44 m, and sintered at 865 °C (1590 °F) is plotted as a function of sintering time. Densification of compacts from the finer powder is faster than that of compacts from the coarser powder and, for sintering times above finer powder.

h, the final density is considerably higher for the compacts from the

Fig. 6 Density of compacts from electrolytic copper powder. Source: Ref 18

A final factor affecting the densification of compacts from a single metal powder is the green density of the compact or the pressure at which the compact is pressed, which determines green density. In Fig. 7(a) and (b), dimensional change is plotted versus sintering temperature for various compacting pressures. Higher compacting pressures result in higher green and sintered densities, but the sintering shrinkage, or the change from green to sintered density, is smaller. As previously noted, expansion may also occur when soft metal powders are compacted at higher pressures as shown in Fig. 7(b) for compaction pressures of 550 and 620 MPa (80 and 90 ksi). This effect is associated with expansion of the entrapped gases in the isolated pores.

Fig. 7 Sintering curves for copper powder compacts at various compacting pressures. Source: Ref 19

References cited in this section

12. F.V. Lenel, Powder Metallurgy Principles and Applications, Metal Powder Industries Federation, 1980, p 247 13. Diffusion, American Society for Metals, 1973 14. F. Thummler and R. Oberacker, Introduction to Powder Metallurgy, The Institute of Materials, 1993, p 208 15. Sauerwald, Plansee seminar, 1952, p 201-202 16. Production Sintering Practices, Powder Metallurgy, Vol 7, Metals Handbook, 9th ed., 1985, p 310 17. F.V. Lenel, Powder Metallurgy, Metal Powder Industries Federation, 1980, p 214 18. F.V. Lenel, Powder Metallurgy, Metal Powder Industries Federation, 1980, p 219 19. H.H. Hausner, Handbook of Powder Metallurgy, Chemical Publishing Co., 1973, p 173 Copper Powder Metallurgy Alloys and Composites A. Nadkarni, OMG Americas, Inc.

Pure Copper Pure copper P/M parts are used mainly in electrical and electronic applications because of their high electrical conductivity. It is essential to use very pure copper powders ( 99.95% purity) or to bring about the precipitation of soluble impurities during sintering. As little as 0.023% Fe in solid solution in copper lowers its conductivity to 86% of that of pure copper. Small amounts of iron mechanically mixed with the copper powder lower the conductivity much less, unless the iron dissolves in the copper during sintering. If high-purity copper is used, or if soluble impurities are precipitated during sintering, it is possible to obtain the strength and conductivity values shown in Fig. 8.

Fig. 8 Effect of density on electrical conductivity and tensile strength properties of P/M copper. Source: Ref 5

Conductivity is directly related to porosity; the greater the void content (lower the density), the lower the conductivity. Electrical conductivity of pure copper parts pressed at moderate pressures of 205 to 250 MPa (15 to 18 tsi) and sintered at 800 to 900 °C (1500 to 1650 °F) varies from 80 to 90% IACS on a scale where conductivity of solid annealed copper is 100% IACS. The conductivity of solid copper can be reached or approached in P/M copper parts by sintering the pressed parts at higher temperatures, such as 930 to 1030 °C (1700 to 1900 °F), followed by re-pressing, coining, or forging. Typical applications of pure copper parts in which high electrical conductivity is required include commutator rings, contacts, shading coils, nose cones, and electrical twist-type plugs. Copper powders also are used in copper-graphite compositions that have low contact resistance, high current-carrying capacity, and high thermal conductivity. Typical applications include brushes for motors and generators and moving parts for rheostats, switches, and current-carrying washers.

Reference cited in this section

5. R.W. Stevenson, Powder Metallurgy, Vol 7, Metals Handbook, 9th ed., 1985, p 733-740 Copper Powder Metallurgy Alloys and Composites A. Nadkarni, OMG Americas, Inc.

Bronze Powder metallurgy bronzes typically originate as premixes consisting of elemental copper and tin powders plus 0.5 to 0.75% dry organic lubricants such as stearic acid or zinc stearate. Some structural parts, however, requiring densities >7.0 g/cm3 are fabricated from prealloyed powders. Prealloyed powders have higher yield strengths and work-hardening rates than premixed powders. Therefore, pressing loads required to achieve given green densities in prealloyed powders are higher than the pressures required for elemental powders. Differences in pressing characteristics of premixed and prealloyed powders are compared in Fig. 9.

Fig. 9 Pressing characteristics of premixed and prealloyed 90Cu-10Sn powders. Source: Ref 20

Typical sintering furnace temperatures for bronze range from 815 to 870 °C (1500 to 1600 °F); total sintering time within the hot zone may range from 15 to 30 min, depending on the furnace temperature selected, required dimensional change, and most importantly, the presence of an optimal bronze grain structure. Sintering atmospheres should be protective and reducing to facilitate sintering. Reduction of the copper oxides that may surround each copper powder particle and of tin oxide allow for increased diffusion rates. Typical strength/density data for 90Cu-10Sn sintered bronzes with and without graphite additions are shown in Fig. 10. Control of sintered dimensions in premix systems is achieved by manipulating sintering time and/or temperature.

Fig. 10 Effect of density on the strength of copper-tin and copper-tin-graphite compacts. Source: Ref 21

Generally, copper-tin blends composed of relatively coarser powders sinter to higher growth values than a blend composed of finer powders. After powder blends have been tested and adjusted to provide an approximation of target dimensions, final

adjustments are made during production sintering to obtain dimensional precision. Factors affecting the ultimate dimensional values include physical characteristics of the constituents and compacted density. Bearings. Self-lubricating porous bronze bearings continue to consume the major portion of the copper powder produced

each year. These bearings are made by pressing elemental powder blends of copper and tin, followed by sintering. The most widely used bearing material is 90Cu-10Sn bronze, often with the addition of up to 1.5% graphite. So-called dilute bronze bearings contain various amounts of iron. Dilution with iron reduces the cost of a bearing at the expense of some loss in performance. Compaction pressures for the bronze powders range from about 140 to 415 MPa (20 to 60 ksi). Sintering is typically done in a continuous mesh belt furnace at temperatures between 815 to 870 °C (1500 to 1600 °F) for about 3 to 8 min at temperature. Typical furnace atmospheres are dissociated ammonia or endothermic gas. To obtain reproducible sintering results, it is important to carefully control time and temperature because of their influence upon the kinetics of the homogenization process, which in turn determines the dimensional changes taking place during sintering. Most bearings are sized for improved dimensional accuracy; typical sizing pressures may range from about 200 to 550 MPa (30 to 80 ksi). Bearings are sold either dry or saturated with oil. The pores are filled with oil by a vacuum impregnation process. Most common bearings range in density from 5.8 to 6.6 g/cm3 dry or 6.0 to 6.8 g/cm3 oil impregnated. This range corresponds to about 25 to 35% pore volume. Figure 11 shows an assortment of bronze bearings. The most common shapes are simple or flanged bushings, but some have spherical external surfaces. Sizes range from about 0.8 to 75 mm ( bronze bearings are shown in Table 11.

to 3 in.) in diameter. Typical applications of sintered

Table 11 Applications of self-lubrication sintered bronze bearings (fractional horsepower electric motors)

Automotive components • • • • • • • • •

Starters Light generators Oil and water pumps Windshield wipers Hood and window raisers Heaters Air conditioners Power antennae Power seat adjusters

Home appliance • • • • • • •

Dishwashers Clothes dryers Washing machines Sewing machines Vacuum cleaners Refrigerators Food mixers

Farm and lawn equipment

• • • • • •

Tractors Combines Cotton pickers Lawn mowers String cutters Chain saws

Consumer electronics • • •

Phonographs Record changers Tape recorders

Business machines • • •

Typewriters Computers Copiers

Industrial equipment • • • • • •

Textile machines Packaging machines Electric fans Portable power tools Drills Saws

Fig. 11 Assorted P/M bronze bearings

Filters. Filters constitute one of the major applications for porous P/M parts. The ability to achieve close control of porosity

and pore size is the main reason filters are made from metal powders. Most producers of nonferrous filters prefer atomized

spherical powders with closely controlled particle size to allow production of filters within the desired pore size range. The effective pore size of filters generally ranges from 5 to 125 m. Tin bronze is the most widely used P/M filter material, but nickel silver, stainless steel, Cu-Sn-Ni alloys, and nickel-base alloys also are used. The major advantage of P/M bronze materials over other porous metals is cost. Porous P/M bronze filters can be obtained with tensile strengths ranging from 20 to 140 MPa (3 to 20 ksi) and appreciable ductility, up to 20% elongation. Powder metallurgy bronze also has the same corrosion resistance as cast bronze of the same composition and thus can be used in a wide range of environments. Bronze filters usually are made by gravity sintering of spherical bronze powders, which are generally made by atomization of molten prealloyed bronze. These powders typically contain 90 to 92% Cu and 8 to 10% Sn. Filters made from atomized bronze have sintered densities ranging from 5.0 to 5.2 g/cm3. To produce filters with the highest permeability for a given maximum pore size, powder particles of a uniform particle size must be used. Although not widely used, coarser powders for bronze filters can be obtained by chopping copper wire and tumbling the choppings. Filters made from tin-coated cut copper wire with tin contents ranging from 2.5 to 8% are also used to a lesser extent. Filters made from these materials have sintered densities ranging from 4.6 to 5.0 g/cm3. During sintering the filters shrink slightly--as much as 8%. To avoid excessive shrinkage, filters from powders with fine particle size require lower sintering temperatures in the neighborhood of 815 °C (1500 °F). Because of the shrinkage during sintering, filters must be designed with a slight draft, so they can be removed from the mold. Figure 12 shows an assortment of P/M bronze filters. Such filters are commonly used to filter gases, oils, refrigerants, and chemical solutions. They have been used in fluid systems of space vehicles to remove particles as small as 1 m. Bronze diaphragms can be used to separate air from liquids or mixtures of liquids that are not emulsified. Only liquids capable of wetting the pore surface can pass through the porous metal part.

Fig. 12 Assorted filters made from P/M bronze. Courtesy of Arrow Pneumatics, Inc.

Bronze filter materials can be used as flame arrestors on electrical equipment operating in flammable atmospheres, where the high thermal conductivity of the bronze prevents ignition. They can also be used as vent pipes on tanks containing flammable liquids. In these applications, heat is conducted away rapidly so that the ignition temperature is not reached. Structural Parts. Powder metallurgy bronze parts for structural applications frequently are selected because of corrosion

and wear resistance of bronze. They are generally produced by methods similar to those used for self-lubricating bearings. Typical compositions of bronze structural parts (CT-1000) are included in Table 12, and the typical properties are shown in Table 13.

Table 12 Compositions of copper-base P/M structural materials (brass, bronze, and nickel silver) Material designation CZ-1000 CAP-1002 CAP-2002 CZ-3000 CZP-3002 CNZ-1818 CNZP-1816 CT-1000

Chemical composition, % Cu Zn Pb Sn Ni 88.0 bal . . . . . . ... 91.0 bal . . . . . . ... 88.0 bal 1.0 . . . ... 91.0 bal 2.0 . . . ... 77.0 bal 1.0 . . . ... 80.0 bal 2.0 . . . ... 68.5 bal . . . . . . ... 71.5 bal . . . . . . ... 68.5 bal 1.0 . . . ... 71.5 bal 2.0 . . . ... 62.5 bal . . . . . . 16.5 65.5 bal . . . . . . 19.5 62.5 bal 1.0 . . . 16.5 65.5 bal 2.0 . . . 19.5 87.5 bal . . . 9.5 ... 90.5 bal . . . 10.5 . . .

Element min max min max min max min max min max min max min max min max

Table 13 Properties of copper-base P/M structural materials (brass, bronze, and nickel silver) Mechanical property data derived from laboratory-prepared test specimens sintered under commercial manufacturing conditions Material designation code(a)

CZ-1000-9 CA-1000-10 CZ-1000-11 CZP-1002 CZP-2002-11 CZP-2002-12 CZ-3000-14 CZ-3000-16 CZP-3002-13 CZP-3002-14 CNZ-1818-17 CNZP-1816 CT-1000-13 (repressed)

Minimum yield strength

Typical values Ultimate tensile strength

0.2% yield strength

MPa

ksi

MPa

ksi

MPa

ksi

62 70 75

9 10 11

124 138 159

18.0 20.0 23.0

65 76 83

9.5 11.0 12.0

9.0 10.5 12.0

52 69

106 psi 7.5 10.0

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

75 83 97 110 90 97 117

11 12 14 16 13 14 17

159 207 193 234 186 217 234

23.0 30.0 28.0 34.0 27.0 31.5 34.0

93 110 110 131 103 115 140

13.5 16.0 16.0 19.0 15.0 16.5 20.0

12.0 14.5 14.0 17.0 14.0 16.0 11.0

(b)

(b)

(b)

(b)

(b)

(b)

90

13

152

22.0

110

16.0

Elongation in 25 mm (1 in.), %

Young's modulus

GPa

Unnotched Charpy impact strength J ft · lbf

Density g/cm3

MPa

ksi

270 315 360

39 46 52

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

69 83 62 69 62 69 75

10.0 12.0 9.0 10.0 9.0 10.0 11.0

345 480 425 590 395 490 500

50 70 62 86 57 71 73

(b)

(b)

(b)

(b)

(b)

4.0

38

5.5

310

45

Source: Ref 22

(a) (b)

Transverse rupture strength

Suffix numbers represent minimum yield-strength values in ksi. Additional data in preparation will appear in subsequent editions of MPIF standard 35.

Compressive yield strength

Apparent hardness, HRH

MPa

ksi

7.60 7.90 8.10

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

38 76 31 51.5

28.0 56.0 23.0 38.0

103 110 83 90

15.0 16.0 12.0 13.0

(b)

(b)

(b)

(b)

(b)

(b)

(b)

(b)

32.5

24.0

7.60 8.00 7.60 8.00 7.60 8.00 7.90

172

25.0

75 84 84 92 80 88 90

(b)

(b)

(b)

(b)

(b)

(b)

5.4

4.0

7.20

186

27.0

82

65 72 80

Figure 13 shows an assortment of P/M bronze structural parts. These parts are generally used in automobile clutches, copiers, outboard motors, and paint-spraying equipment.

Fig. 13 Assorted P/M bronze parts. Courtesy of Norddeutsche Affinerie

References cited in this section

20. A. Price and J. Oakley, Powder Metall., Vol 8, 1965, p 201 21. A.K.S. Rowley, E.C.C. Wasser, and M.J. Nash, Powder Metall. Int., Vol 4 (No. 2), 1971, p 71 22. "P/M Materials Standards and Specifications," standard 35, Metal Powder Industries Federation, 1986-1987 Copper Powder Metallurgy Alloys and Composites A. Nadkarni, OMG Americas, Inc.

Brass and Nickel Silver In contrast to bronze structural parts, parts made from brass, leaded brass, and nickel silver are produced from prealloyed atomized powder. Compositions of some common brass and nickel silver alloys used in structural parts are shown in Table 12. The leaded compositions are used whenever secondary machining operations are required. The alloy powders are usually blended with lubricants in amounts from 0.5 to 1.0 wt%. Lithium stearate is the preferred lubricant because of its cleansing and scavenging action during sintering. However, bilubricant systems are common, such as lithium stearate and zinc stearate, which minimize the surface staining attributed to excessive lithium stearate. Lubricated powders are typically compacted to 75% of theoretical density at 207 MPa (30 ksi) and to 85% of theoretical density at 415 MPa (60 ksi). Sintering of brass and nickel silver compacts is normally performed in protective atmospheres. Dissociated ammonia, endothermic gas, and nitrogen-base atmospheres are most common. Temperatures range from 815 to 925 °C (1500 to 1700 °F) depending on alloy composition. To avoid distortion and/or blistering of the compacts, sintering temperatures should not exceed the solidus temperature of the alloy. Through multiple pressing and sintering operations, yield strength and hardness approaching those of the wrought alloy counterparts can be achieved. To minimize zinc losses during sintering, yet allow for adequate lubricant removal, protective-sintering-tray arrangements are used. Table 13 shows the typical properties of common brass and nickel silver P/M parts. Next to bronze bearings, brasses and nickel silvers are the most widely used materials for structural P/M parts. Typical applications include hardware for latch bolts and cylinders for locks; shutter components for cameras; gears, cams, and actuator bars in timing assemblies and in small generator drive assemblies; and decorative trim and medallions. In many of

these applications, corrosion resistance, wear resistance, and aesthetic appearance play important roles. Figure 14 shows some typical P/M brass components.

Fig. 14 P/M brass components. (a) Rack guide for rack-and-pinion steering column. (b) Leaded brass guide for stereomicroscope. (c) Leaded brass guide for microscope. Courtesy of Metal Powder Industries Federation

Copper Powder Metallurgy Alloys and Composites A. Nadkarni, OMG Americas, Inc.

Copper-Nickel Copper-nickel P/M alloys containing 75Cu-25Ni and 90Cu-10Ni have been developed for coinage and corrosion-resistance applications. The 75Cu-25Ni alloy powder pressed at 772 MPa (112 ksi) has a green density 89% of its theoretical density. After sintering at 1090 °C (2000 °F) in disassociated ammonia, elongation is 14%, and apparent hardness is 20 HRB. Repressing at 772 MPa (112 ksi) increases density to 95%. This alloy has the color of stainless steel and can be burnished to a high luster. The 90Cu-10Ni alloy has a final density of 99.4% under similar pressing-and-sintering conditions. It has a bright bronze color and also can be burnished to a high luster.

In one method of producing coins, medals, and medallions, a mixture containing 75% Cu and 25% Ni powders is blended with zinc stearate lubricant and compressed, sintered, coined, and resintered to produce blanks suitable for striking. These blanks are softer than rolled blanks because they are produced from high-purity materials. Therefore, they can be coined at relatively low pressures, and it is possible to achieve greater relief depth with reduced die wear. In another procedure, an organic binder is mixed with copper or copper-nickel powders and rolled into "green" sheets. Individual copper and copper-nickel sheets are pressed together to form a laminate, and blanks are punched from it. Blanks are heated in hydrogen to remove the organic binder and sinter the material. The density of the "green" blanks is low (45% of theoretical), but coining increases density to 97%. After pressing, the blanks are annealed to improve ductility and coinability. Copper Powder Metallurgy Alloys and Composites A. Nadkarni, OMG Americas, Inc.

Copper-Lead Copper and lead, which have limited solubilities in each other, are difficult to alloy by conventional ingot metallurgy. Copper-lead powder mixtures have excellent cold pressing properties; they can be compacted at pressures as low as 76 MPa (11 ksi) to densities as high as 80% of theoretical density. After sintering, they can be re-pressed at pressures as low as 152 MPa (22 ksi) to produce essentially nonporous bearings. Steel-backed copper or Cu-Pb-Sn P/M materials are sometimes used to replace solid bronze bearings. They are produced by spreading the powder in a predetermined thickness on a steel strip, sintering, rolling to theoretical density, re-sintering, and annealing. The end product has a residual porosity of about 0.25%. Blanks of suitable size are cut from the bimetallic strip, formed, and drilled with oil holes or machined to form suitable grooves. These materials include Cu-25Pb-0.5Sn, Cu-25Pb3.5Sn, Cu-10Pb-10Sn, and Cu-50Pb-1.5Sn alloys. Copper Powder Metallurgy Alloys and Composites A. Nadkarni, OMG Americas, Inc.

Copper-Base Friction Materials Sintered-metal-base friction materials are used in applications involving the transmission of motion through friction (clutches) and for deceleration and stopping (brakes). In these applications mechanical energy is converted into frictional heat, which is absorbed and dissipated by the friction material. Copper-base materials are preferred because of their high thermal conductivity; however, lower cost iron-base materials have been developed for moderate to severe duty dry applications. Most friction materials contain copper powders blended with other metal powders, solid lubricants, oxides, and other compounds. These constituents are immiscible in each other and therefore can only be made by powder metallurgy. Compositions of some common copper-base friction materials are shown in Table 14.

Table 14 Compositions of copper-base friction materials for wet and dry applications Country

Use(a)

Composition, wt% Cu Sn Fe 65-80 7-9 4-7 70 9 4 60 10 4 81.5 4.5 ... bal ... ... 60-75 4-10 5-10 52.5 ... ... 72 4.7 3.3 72 7 3 62 7 8 74 3.5 ... bal 3-10 5-10

Pb 5-10 6 5 5 5 ... 7.5 3.5 6 12 ... 1-10

Graphite 3-8 4 4 4 12 3-10 ... 8.7 6 7 16 0.8

MoS2 ... ... ... ... ... 3-12 ... 1.4 ... ... ...

5.1 4-15 5.2

8 5-30 4.5

1.5 ... 1.8

6.2 20-30 6.5

5 ...

Sweden

67.7 bal 68.5

Italy Austria

68.5 68 68

8 5.5 5

4.5 7 8

3 9 1.5

6 6 6.2

6 ...

54.4

0.8

3.7

21.4

19

...

USSR

East Germany United States

United Kingdom West Germany

4

Other 2-4 SiO2 3SiO2, 3 asbestos 9 asbestos, 8 bakelite powder 5 mullite 8 MgO; 5 Ti 2-7 SiO2 5 SiO2; 15 Bi 1.9 SiO2; 0.2 Al2O3 3 SiO2; 4 MoO3 4 sand 2 Sb; 4.5 SiO2 1.5-4 SiO2

W, D W W W W, D D W W, D D D D W

4

2.5 SiO2; 3 Al2O3 3-10 Al2O3 3.3 SiO2; 3 Al2O3

D W W, D

3

4 SiO2 4.5 SiO2 2.5 SiO2; 3 Al2O3

W, D W, D W

0.5 S; 0.04 Mn

D

Source: Ref 23

(a)

W, wet; D, dry.

Mixtures of the appropriate powders are carefully blended to minimize segregation of the constituents. Fine metal powders with high surface area are necessary to provide a strong and thermally conductive matrix. The blended powders are compacted at pressures ranging from 165 to 275 MPa (24 to 40 ksi). Bell-type sintering furnaces are used where the friction facing is bonded to a supporting steel backing plate such as in clutch disks. The green disks are placed on the copper-plated steel plates and stacked. Pressure is applied on the vertical stack of disks. Sintering temperatures range from 550 to 950 °C (1020 to 1740 °F) in a protective atmosphere. Typical sintering times are 30 to 60 min. The sintered parts are typically machined for dimensional accuracy and surface parallelism. The friction elements are usually brazed, welded, riveted, or mechanically fastened to the supporting steel members. They may also be pressure bonded directly to the assembly. The operating conditions encountered by metal-base friction materials can be classified as dry/wet and mild/moderate/severe. Figure 15 shows some typical applications and the corresponding operating conditions. Some examples of copper-base friction elements are shown in Fig. 16.

Fig. 15 Applications of sintered metal friction materials. Source: Ref 24

Fig. 16 Copper-base P/M friction elements. (a) Grooved P/M friction elements for wet applications. (b) Copperbase P/M clutch plates (280 to 500 mm OD) used in power-shift transmissions for tractors. (c) Copper-base P/M friction pad

References cited in this section

23. W. Schatt, Pulvermetallurgie Sinter und Verbundwerkstoffe, VEB Deutscher Verlag fur Grundstoffindustrie, 1979, p 315 24. B.T. Collins, The U.S. Friction Materials Industry, Perspectives in Powder Metallurgy, Vol 4, Plenum Press, 1970, p 3-7 Copper Powder Metallurgy Alloys and Composites A. Nadkarni, OMG Americas, Inc.

Copper-Base Contact Materials Electrical contacts are metal devices that make and break electrical circuits. Arcing, except in applications with low potential or current levels, is a major problem, particularly during opening of the contacts in a live circuit. The arc causes erosion of the contacts by blowing away the molten metal or by vaporizing the material. Welding can occur during closing of the contacts when small areas on the contact surfaces that are molten during arcing fuse together during closure (Ref 25).

Applications involving severe arcing and welding employ contacts made from refractory metals, such as tungsten and molybdenum, which have high melting and boiling points, and excellent resistance to arc erosion. Oxides are often used to prevent welding of the contacts. In both cases, a high conductivity metal such as silver or copper is used in conjunction with the refractory metal or oxide to provide the necessary current-carrying capability. Composites such as these are made by P/M techniques because the individual constituents are immiscible and cannot be made by conventional melt-cast process. Applications where arcing and welding are not severe utilize contacts made of pure metals or alloys. These are generally made by melting and casting followed by suitable metalworking processes. Copper-base materials are used in electrical contacts because of their high electrical and thermal conductivities, low cost, and ease of fabrication. Their main drawbacks are poor resistance to oxidation and corrosion. Therefore, copper-base contacts are used in applications where the voltage drop resulting from the oxide film is acceptable or where it is possible to protect the contact, such as by immersion in oil or by enclosing the contact in a protective gas or vacuum. Common copper alloys used in contacts include yellow brass (C27000), phosphor bronze (C51000), and copper beryllium alloys (C17200 and C17500). These are made by the melt-cast process and are limited to lower current applications where arcing and welding are not severe. Composites of copper with refractory metals or their carbides are used in applications in which limited oxidation of the copper is acceptable or where oxidation is prevented by one of the methods mentioned above. Table 15 presents the compositions, properties, and typical applications for some P/M composite contact materials. The properties of the contacts depend on the manufacturing method used, and therefore the common methods used are also shown in the table. The specific method used depends on the composition of the composite. Generally, materials with 40% or less tungsten or its carbide are manufactured by the conventional pressing, sintering (generally below the melting point of copper), and re-pressing (PSR) technique. Materials containing more than 40% W are generally made by infiltrating (INF) the copper into either loose tungsten powder or pressed-and-sintered tungsten compacts. Their counterparts using tungsten carbide are made by infiltrating the copper into loose powder because the tungsten carbide powder cannot be pressed into compacts.

Table 15 Compositions, properties, and applications of copper-base electrical contacts Nominal Manufacturing composition, method(a) % Tungsten carbide-copper INF 50Cu

Density, g/cm3 Calculated

Typical

11.39

11.0011.27 11.64 12.65

INF INF

11.77 12.78

Tungsten-copper PSR 75Cu-25W

10.37

44Cu 30Cu

Electrical conductivity, % IACS

Hardness

Tensile strength MPa ksi

Modulus of rupture MPa ksi

Data source(b)

Application examples

42-47

90-100 HRF 99 HRF 38 HRC

...

...

1103

160

C, A

... ...

... ...

1241 ...

180 ...

C ...

Arcing contacts in oil switches, wiping shoes in power transformers

...

...

414

60

C, A

...

...

...

...

A

...

...

...

...

A

...

...

...

...

A

...

...

...

...

A

434 ...

63 ...

827 ...

120 ...

C A

43 30

70Cu-30W

...

10.70

9.4510.00 10.45

50-79

65Cu-35W

...

11.06

11.40

72

60Cu-40W

...

11.45

11.76

68

50Cu-50W

INF

12.30

45-63

44Cu-56W 40Cu-60W

INF INF

12.87 13.29

11.9011.96 12.76 12.8012.95

76

55 42-57

35-60 HRB 59-66 HRB 63-69 HRB 69-75 HRB 60-81 HRB 79 HRB 75-86 HRB

Currentcarrying contacts Vacuum interrupter Oil-circuit breakers, arcing tips

Oil-circuit breakers,

30Cu-70W

INF

14.45 14.97

13.8514.18 14.70

26Cu-74W

INF

25Cu-75W

36-51 46

INF

15.11

14.50

33-48

20Cu-80W

INF

15.84

15.20

30-40

15Cu-85W

PSR

16.45

16.0

20

13.4Cu86.6W 10.4Cu89.6W

INF

16.71

16.71

INF

17.22

17.22

86-96 HRB 98 HRB

...

...

1000

145

C, A

621

90

1034

150

...

...

...

...

...

A

758

110

...

...

C

...

...

...

...

M

33

90-100 HRB 95-105 HRB 190 HV(c), 260 HV(d) 20 HRC

621

90

1034

150

C

30

30 HRC

765

111

1138

165

C

tips, contractors Circuit breaker runners, arcing tips, tap change arcing tips Vacuum switches, arcing tips, oil-circuit breakers

Source: Ref 25

(a) (b) (c) (d)

PSR, press-sinter-re-press; INF, press-sinter-infiltrate. A: Advance Metallurgy, Inc., McKeesport, PA. C: Contacts, Materials, Welds, Inc., Indianapolis, IN. M: Metz Degussa, South Plainville, NJ. Annealed. Cold worked.

Reference cited in this section

25. Y.S. Shen, P. Lattari, J. Gardner, and H. Wiegard, Properties and Selection: Nonferrous Alloys and SpecialPurpose Materials, Vol 2, ASM Handbook, ASM International, 1990, p 841-868 Copper Powder Metallurgy Alloys and Composites A. Nadkarni, OMG Americas, Inc.

Copper-Base Brush Materials Brushes are components that transfer electrical current between the stationary and rotating elements in electric motors and generators. Most common brushes are made from composites of graphite and a conductive metal. The graphite provides the required lubrication, and the metal provides the current-carrying capability; copper and silver are preferred metals because of their high electrical conductivity. Table 16 shows the compositions of some typical copper-graphite brush materials. The copper content varies from about 20 to 75%, the balance being graphite. Powder metallurgy is the only way to produce these materials because of the immiscibility of the two components. Copper powder used in the brushes could be made by oxide reduction, electrodeposition, atomization, or flaking.

Table 16 Characteristics of typical graphite-metal brush materials Grade No. 261C 261D FQ 179P 179V GHB GD 22A-S 246 2-S 1-S

Nominal composition

Density, g/cm3

21Cu-79C 35Cu-65C 50Cu-50C 65Cu-35C 75Cu-25C 94metal-6C 97metal-3C 40Ag-60C 65Ag-35C 80Ag-20C 93Ag-7C

2.2 2.5 2.75 3.5 4.0 6.0 6.5 2.7 3.8 4.6 7.0

Specific resistance ·m 0.024 0.016 0.006 0.0016 0.0008 0.0003 0.0001 0.008 0.001 0.0008 0.0001

· in. 0.0006 0.0004 0.00015 0.00004 0.0002 0.0000065 0.0000025 0.0002 0.000025 0.00002 0.0000025

Maximum current density A/m2 A/in.2 125,000 80 125,000 80 130,000 85 190,000 125 235,000 150 235,000 150 235,000 150 150,000 100 190,000 125 235,000 150 270,000 175

Typical voltage, V

Scleroscope hardness

5 offer superior mechanical properties, but form nickel-rich intermetallics, predominantly Ni4W.

If postsinter vacuum heat treatment is not to be performed, alloys with nickel-to-iron ratios of 1.5 to 2 typically give as good properties as higher ratios. For components that can be vacuum heat treated, nickel-to-iron ratios in the range 2.3 to 4 should be chosen, as better mechanical properties are generated. For applications such as kinetic energy penetrators that demand the utmost in mechanical performance, W-Ni-Fe alloys should be resolutionized and quenched. This processing step reduces both the segregation of interstitial species and the presence of any nickel-rich intermetallics and thereby opens the possibility for use of alloys with nickel-to-iron ratios in the range of 6 to 15, which provide noticeably improved mechanical properties. The addition of cobalt to a W-Ni-Fe alloy is a common approach for slight enhancement of both strength and ductility. The presence of cobalt within the alloy provides solid-solution strengthening of the binder and slightly enhanced tungsten-matrix interfacial strength. Cobalt additions of 5 to 15% of the nominal binder weight fraction are most common. Higher levels of addition risk the formation of a mixed -phase ([Fe,Ni]7W6), which would then require high-temperature resolution/quench of the material. For extremely demanding applications, even higher mechanical properties are obtainable from the W-Ni-Co system with nickel-to-cobalt ratios ranging from 2 to 9. Such alloys require resolution/quench, however, due to extensive intermetallic (Co3W and others) formation on cool down from sintering. These alloys not only provide greater quasi-static tensile strength and ductility, but substantial improvement in fracture toughness as well (Table 2). The use of W-Ni-Co alloys is generally limited to defense applications requiring state-of-the-art mechanical properties.

Table 2 Property comparison of unworked alloys Alloy

91W-6.3Ni-2.7Fe 91W-6Ni-3Co

Ultimate tensile strength, MPa 940 960

Elongation, %

Hardness, HRC

Charpy toughness, J

35 40

29 31

300 470+

A number of specialty WHAs are known as well. An example is the W-Mo-Ni-Fe quaternary alloy, which utilizes molybdenum to restrict tungsten dissolution and spheroid growth, resulting in higher strengths (but reduced ductility) in the as-sintered state. There are also a number of alloy systems in various stages of development for kinetic energy penetrators that are intended to provide a WHA that will undergo high deformation rate failure by shear localization in a manner similar to quenched and aged U-0.75Ti for more efficient armor defeat. These alloys to date have not exhibited a property set of interest for industrial applications, however. Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Consolidation WHAs are typically formulated from elemental powders with mean particle sizes in the range 1.5 to 7 m. Blend homogeneity is essential for microstructural uniformity and predictable sintering to full density. Blended powder is sometimes mixed with an organic binder such as 1 to 2 wt% paraffin and granulated. While this procedure is required for even die-cavity filling for some component shapes (especially flat forms), it should be employed only when necessary as the introduction of organics necessitates a debinding operation in the process sequence. It is essential that any organic binder be capable of clean thermal decomposition and that the process be carried to completion as residual carbon in the sintering environment can lead to interstitial embrittlement and in the worst case, WC formation. Unlike WC grade powder, WHA blends tend to be readily pressable without a binder using standard techniques such as uniaxial compaction and isostatic pressing at pressures of 200 MPa. Even without organic binders, the green strength of WHA pressings is generally quite acceptable.

WHAs are most commonly processed to full density by solid-state sintering (SSS) or LPS in H2. The former process is ideal for parts requiring low thermal distortion ("slumping" due to capillary action and/or gravity) and need only minimal mechanical properties. This is often required for precision net-shape parts. Where practical, sintering at a slightly higher temperature allows the formation of an entirely different microstructure capable of significantly higher mechanical properties due to the occurrence of a liquid phase. In the common W-Ni-Fe compositions, liquid-phase conditions are achieved slightly above 1440 °C. Liquid-phase sintering for 90 to 97% W alloys is generally carried out in the range 1470 to 1580 °C, whereas SSS is performed at 1430 to 1440 °C. Linear shrinkages of 20% are typical. Vacuum sintering of WHAs in batch-type furnaces is used only rarely due to the inferior microstructure and mechanical properties obtained. Additionally, if liquid-phase conditions are reached, the high vapor pressure of the transition metals present may cause rapid deterioration of tungsten or molybdenum furnace components in the furnace hot zone. Vacuum sinter furnaces utilizing graphite elements and heat shielding are unsuitable due to excessive carbon transport to the WHA parts, resulting in embrittlement or superficial WC formation. It is essential that the rate of temperature rise experienced by a pressing be gradual, so as to avoid thermal shock or premature surface pore closure. Below 1000 °C, there is significant reduction of metal oxides on transition metal particles and general desorption of O2 and other species. At 1000 °C and slightly higher, tungsten oxides are readily reduced by the hot H2 and significant interdiffusion of metallic species occurs, initially by surface diffusion. As the furnace temperature exceeds 1200 °C, surface closure of interconnected porosity is underway. It is essential that the heating rate is sufficiently slow so as to permit internal deoxidation of the pressing prior to this point. While hydrogen at these temperatures possesses high mobility through the bcc tungsten structure, the H2O formed by reaction relies on interconnected porosity for escape. If surface porosity is closed prematurely, this molecular specie will be trapped, giving rise to residual porosity. In more extreme cases of high oxygen powders, larger voids may form, a process that can be greatly minimized by the use of higher dew point H2. Commercial production of WHA components most commonly utilizes continuous pusher-type furnaces in that such designs offer high throughput rates and good temperature uniformity. Pressings are transported through the furnace on ceramic slabs or partially buried in molybdenum boats containing alumina sand. Stoker furnaces in general produce parts with slightly higher mechanical properties than what is obtainable for batch furnace sintering due to better internal material cleanup by the action of the hot flowing H2, effectively sweeping evolved species from the vicinity of the parts. While hot isostatic pressing (HIP) and rapid consolidation techniques are also available densification options, their use is very rare because the short processing times generally result in microstructures that yield inferior mechanical properties, often at a substantially higher cost as well. Problems with porosity, common in many P/M materials, is virtually nonexistent in WHAs provided LPS is correctly executed using high quality raw materials. Liquid-phase sintering provides a superior means of densifying pressings in which a microstructure consisting of spheroids ( 40 to 60 m in size) of nearly pure tungsten in an austenitic binder phase of transition metals plus dissolved tungsten forms via Ostwald ripening. Such a structure provides better mechanical properties than SSS structures due to less angular grains, lower tungsten-tungsten contiguity for a given tungsten content, and a greater percentage of metastable tungsten in the austenitic solid-solution binder. The amount of tungsten retained for common cooling rates from sintering is substantially greater than equilibrium values and varies with the binder composition as seen in Table 3. The slower the cooling rate from sintering temperature, the lower the percentage of retained metastable tungsten in the binder phase.

Table 3 Metastable tungsten retained in solution Alloy 91W-6.3Ni-2.7Fe (SSS) 91W-6.3Ni-2.7Fe (LPS) 91W-6Ni-3Co (LPS)

% W in solid solution 18 23 40

Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Postsinter Heat Treatment A postsinter heat treatment is generally required for any WHA component that has significant mechanical requirements. The most common is a H2 outgassing treatment employed to virtually eliminate hydrogen embrittlement. During sintering, especially if a liquid phase was allowed to form, WHA parts absorb sufficient hydrogen so as to significantly limit ductility. By subjecting the sintered components to an isothermal treatment in a low pH environment, absorbed hydrogen is given sufficient thermal activation to diffuse to a free surface, recombine into dimer form (H2), and migrate from the furnace environment. This heat treatment is generally performed at 1000 to 1100 °C for several hours, depending on section thickness. While this operation could be performed in flowing N2, vacuum heat treatment is usually the most convenient means of protecting the parts from oxidation during processing and is therefore commonly referred to as vacuum annealing. This procedure typically increases the ultimate tensile strength of an alloy by 5% and the elongation up to 40% as compared to the as-sintered property set. While greatly reducing hydrogen embrittlement, the additional time at moderate temperature increases both the segregation of interstitial species to grain boundaries and the amount of intermetallic precipitation in susceptible alloy systems. While vacuum annealing provides a method of improving low strain rate properties over those of the as-sintered state, dynamic properties such as fracture toughness may in fact be lowered. Resolutionization and quench can be used for dynamic property recovery. This second postsinter step addresses two separate but related embrittlement mechanisms. First, this procedure minimizes segregation-induced embrittlement caused by sulfur, phosphorus, and carbon by providing thermal activation for redistribution of these species into the bulk and preventing reconcentration in the boundary region by imposition of rapid cooling to a temperature at which diffusion is ineffective. This mechanism is operative for all WHAs. Resolution is performed in the vicinity of 1100 °C for 1 to 2 h, followed immediately by a water quench. Secondly, in alloy systems containing intermetallic precipitates, resolution and quench also provides a means of property recovery. The treatment temperature must be increased to 1200 °C for effective resolution of nickel-base compounds. Quenched WHAs typically exhibit another 5% gain in ultimate tensile strength, 10 to 30% gain in elongation, and 300% gain in fracture toughness (as evaluated with unnotched Charpy impact) over vacuum annealed material. Resolution and quench make possible the use of high ductility/toughness alloys such as high nickel-to-iron ratio W-Ni-Fe alloys and all W-Ni-Co alloys. All WHAs benefit from such treatment, though not all part geometries may be suitable for such postsinter processing, however. All applications requiring high dynamic properties should specify quenched WHA. While uncommon in commercial practice, it is possible in modern cold-wall sintering furnaces to combine H2 sintering, vacuum annealing, and quenching into a single energy and time efficient cycle. Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Mechanical Processing For nearly all density- and mass-driven applications, the mechanical properties of unworked WHAs are more than adequate. Yet for many other applications, the strength and hardness of unworked material may be insufficient. Through a combination of mechanical and thermal processing, a wide range of property sets are available. WHAs can be processed for maximum ductility with tensile elongations exceeding 40% or for strength levels exceeding 1650 MPa with a hardness of 50 HRC. As can be seen in Fig. 3, there is a continuous set of mechanical properties available via the trade-off in ductility and toughness for strength and hardness. The greater the initial unworked property set, the greater the range of processing options. The most commonly used means of increasing the strength and hardness of WHAs is by rotary swaging, in which the diameter of a bar is reduced by rapid short-stroke hammering, generally in a four-die configuration. This cold (or occasionally warm) working

results in an increase in bar length and imparts an axial orientation to the microstructure. The extent of swaging possible for a given alloy is a function of its ductility. Table 4 provides a guideline for maximum percentages of reduction in area (RA) via swaging for selected 90 to 93% W alloys, valid for initial bar diameters of 5 cm and smaller.

Table 4 Swaging limitations for various WHAs Alloy family W-Ni-Fe (2.3 Ni/Fe W-Ni-Fe-Co (typical) W-Ni-Co

4)

Maximum % RA swage 25 30 40

Fig. 3 Strength and ductility can be varied over a substantial range by postsinter processing. The higher the initial mechanical properties, the greater the practical range of processing options

Exceeding the listed %RA values often results in microcracking of the material. Such induced microstructural damage results in diminished returns for greater degrees of working. A more successful approach to achieving higher levels of strength and hardness involves taking material worked from 15% RA to the values shown in Table 4 and subjecting the material to an isothermal treatment--typically 1 to 2 h at 350 to 550 °C. This gives rise to significant strain aging of both the binder phase and the tungsten phase. Attempts at aging in the range 700 to 1000 °C may cause embrittlement in some alloys due to intermetallic precipitation. Aging at 1200 °C recovers essentially all of the induced strain hardening and essentially returns the material to its original condition, though some microstructural orientation remains. The original mechanical property set will not be recovered, however. The binder no longer contains the same extent of tungsten supersaturation. As tungsten content of the WHA and/or starting bar diameter increases, the limiting value of %RA must be decreased. There are a number of subtle variations that can only be described generically. These include die geometry and feed rate into the swage as they determine the profile of induced deformation and the extent of work. The effectiveness of such processing is seen in Table 5.

Table 5 Typical properties for 93W-4.9Ni-2.1Fe Condition As-sintered Vacuum annealed Quenched Swaged 15% RA Swaged 25% RA Swaged 25% + aged 500 °C/1 h

Ultimate tensile strength, MPa 860 910 930 1115 1200 1400

Elongation, % 15 25 32 16 14 7

Hardness, HRC 29 29 29 39 40 44

Charpy toughness, J ... ... 230 150 110 80

Although swaging is the most commonly used deformation processing technique for WHAs, alternative methods of working include extrusion (conventional or hydrostatic), forging, rolling, or upsetting for strength and hardness enhancement. Hydrostatic extrusion is capable of successfully achieving very high reduction ratios, producing a fibrous microstructure. Upsetting, a commonly used practice for introducing work into short rods (such as preforms), differs from other techniques in terms of direction of material flow. This difference must be noted for applications in which the Bauschinger effect may prove important. Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Microstructure Considerations In LPS WHAs, the microstructure of most commercial alloys is dual phase and takes the form of spheroids of nearly pure tungsten in an austenitic (face-centered cubic) matrix of the selected transition metals plus dissolved tungsten. In a mature LPS structure of W-Ni-Fe, spheroids are typically 40 to 60 m in size with a degree of connectedness, termed contiguity. For a given composition, mature spheroid size is a function of sintering temperature, higher temperature favoring larger size. Below approximately 91% W, the binder phase is a continuous structure within the solid. Such alloys display noticeably higher ductility and toughness and also are capable of more rapid work hardening. Tungsten spheroids typically contain less than 0.2 wt% Ni and 0.1 wt% Fe in common alloys. In such alloys, three types of interfaces exist: tungsten-tungsten boundaries, binder grain boundaries, and tungsten-matrix interfaces. Of these three structures, the tungsten-tungsten contact regions are the weakest and therefore serve as a principal limiting factor in the overall strength of the alloy. Thus, any microstructural characteristics that affect these features strongly influence mechanical behavior. Segregation of impurities to this interface serves to weaken its strength. Likewise, an increase in contiguity, resulting in a greater fraction of tungstentungsten boundary area, also lowers mechanical properties. It has been demonstrated that the critical flaw size is comparable to the mean spheroid diameter for common alloys. Contiguity is determined by the weight percent of tungsten, the sintering temperature, and the thermochemistry of the binder phase as to tungsten solubility and wetting behavior. High tungsten content alloys exhibit high contiguity due to the shape accommodation that occurred during LPS as tungsten spheroids grew to impingement (Fig. 4). As the microstructure becomes more brittle, the alloy can undergo less strain hardening prior to fracture. Thus, contiguity and interfacial embrittlement serve to limit both the ductility and strength of WHAs.

Fig. 4 High tungsten content WHAs display substantial amounts of tungsten-tungsten contiguity and shape accommodation during Ostwald ripening, as seen in this example of 98% W

Some applications for WHAs demand densities below 17 g/cm3. Such uses, which may often require high bend ductility as well, are best met with SSS material with tungsten contents in the range of 60 to 90%. Solid state sintering provides minimal sintering distortion and near full densification with a more refined microstructure than that available from LPS. A typical SSS microstructure is shown in Fig. 5. The tungsten phase exhibits a mean size of 10 to 15 m in a continuous austenitic binder solid solution. The high volume fraction of ductile binder and the low tungsten-tungsten contiguity of this type result in high deformability. Such material can be easily rolled into thin sheet providing high mechanical properties.

Fig. 5 Typical microstructure of a SSS 70W-Ni-Fe alloy exhibiting fine, polyhedral tungsten grains in a continuous

matrix phase

Whether SSS or LPS, the sintered structure of WHAs can be described as a metal-matrix composite, as each phase presents a distinctly different set of mechanical properties. This structure gives rise to notch sensitivity, and therefore all components requiring high mechanical integrity should avoid grooves, threads, or corners with radii smaller than 0.5 mm. Otherwise, stress concentration will result in reduced mechanical strength. This becomes important also in mechanical property measurement, as specimen preparation, finishing, and marking techniques appropriate for steel specimens may not be suitable for WHA samples. Notch sensitivity is also the reason why WHA toughness is most commonly evaluated using unnotched Charpy specimens per ASTM E 23, as a much higher level of discrimination is possible. Low-toughness WHA typically tests at 8 J notched and 55 J unnotched. The same alloy in a higher toughness condition would test at 10 J notched and 215 J unnotched. In the notched testing mode, statistical noise would mask any difference, whereas unnotched test results are clearly separated. The use of subscale Charpy bars are sometimes used, but results cannot be compared to full-scale values by ratio of areas as the span of the Charpy anvil remains fixed and hence provides a larger relative span for subscale specimens, leading to artificially high values of absorbed energy to failure. WHAs also display strain-rate sensitivity. When tested at medium to high strain rates, WHAs provide a significantly higher dynamic yield strength than suggested by quasi-static values. A 90% WHA exhibits a strain-rate sensitivity exponent, m, of 0.012. During tensile testing, unworked WHAs generate a load-displacement curve with a high-slope linear elastic region, a rather gentle transition into plasticity and work hardening, and minimal rise in engineering strength to failure. The 0.2% offset yield stress is approximately 67% of the ultimate tensile strength. A Poisson ratio of 0.29 is commonly encountered. Work hardening follows power-law behavior with an exponent of n 0.51 for 90% WHA. The rate of work hardening increases slightly with tungsten content. Load-displacement curves of heavily worked WHA exhibit sharp knees with a virtual equivalence of ultimate tensile strength and 0.2% offset yield stress. Toughness of WHA is readily revealed by both the absorbed energy to failure numbers from Charpy testing as well as the resultant fracture morphology when tested in unnotched mode. Optimal mechanical properties are associated with alloy conditions in which the fracture surface will mainly consist of cleaved tungsten spheroids surrounded by binder that has undergone significant elongation, failing in distinct knife edges between the tungsten grains. Embrittled material typically exhibits distinct tungsten-tungsten spheroid contact zones and matrix decohesion. Highly embrittled conditions, such as from intermetallics forming a peritectic reaction layer on the tungsten phase, typically exhibit an almost glassy fracture of the binder phase (Fig. 6).

Fig. 6 Fracture morphology resulting from mechanical testing clearly reveals substantial differences in alloy properties: (A) high toughness material displaying predominant tungsten cleavage and ductile matrix failure, (B) alloy with tungsten-matrix interfacial embrittlement but good matrix ductility, and (C) extremely brittle failure due

to presence of intermetallic precipitates

In addition to the mechanical difference between the two phases, there are significant electrochemical differences as well. While WHAs are generally considered corrosion resistant, corrosion can nevertheless occur due to local galvanic activity between the tungsten phase, and the electrochemically dissimilar austenitic binder phase if a suitable electrolyte is present. Highly alkaline solutions readily attack the tungsten phase whereas the binder is more readily attacked by acidic solutions. Corrosion can occur under ambient conditions in the presence of very high humidity, especially condensing moisture, and in the presence of ionic salt solutions. Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Physical Properties In that the principal phase in WHAs is nearly pure tungsten, it follows that the physical properties of WHAs should be very close to that of elemental tungsten. A comparison of density and elastic stiffness for various W-(3Ni/Fe) alloys and pure tungsten is provided in Table 6. The (longitudinal) sonic velocity of 90 to 95% WHAs is approximately 5150 m/s, slightly lower than that of pure tungsten. In as-sintered material with a bulk hardness of 29 HRC, the microhardness of the tungsten phase is 500 HV (25 gf)--approximately that of pure tungsten. The matrix, however, possesses a hardness of only 300 HV (25 gf).

Table 6 Variation of density and Young's modulus with tungsten content Wt % W 90 91 93 95 97 98 100

Density, g/cm3 17.2 17.4 17.8 18.2 18.6 18.8 19.3

Young's modulus, GPa 345 350 360 370 385 390 410

Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Thermal Properties While primarily comprised of the highest melting point metal, WHAs are limited in service temperature by nature of the transition metal binder phase and the formation of intermetallics. There are other factors as well that serve to limit useful service temperature. Under ambient conditions, oxidation alone may impose a temperature ceiling of 300 to 400 °C. If a WHA bar is in a mechanically worked condition, a temperature as low as 300 °C will result in noticeable strain aging in as short a time as a few hours, resulting in altered properties. Even if a protective atmosphere is employed, service temperatures 600 °C and higher are to be avoided as these composite materials exhibit "hot-short" behavior. The ultimate limit on service temperature is imposed by the threshold of formation of a liquid phase, which can occur in some W-Ni-Fe alloys as low as 1450 °C.

While the coefficient of thermal expansion of pure tungsten is very low at 4.4 ppm/ °C (for 20 to 200 °C), the coefficient of thermal expansion for common WHAs is only slightly higher at 5.0 ppm/ °C. Thermal conductivity for a typical commercial WHA is approximately 120 W/m · K, with a corresponding electrical conductivity of 14% IACS. Specific heat decreases slightly with increasing tungsten content, ranging from approximately 0.037 cal/g · °C for 90 W to 0.032 cal/g · °C for 97% W. While thermal properties are principally determined by the tungsten phase, service temperature limitations are determined by the binder. Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Applications Virtually all of the applications for WHAs depend upon its high density. WHAs are used in an increasingly wide array of industrial, defense, and consumer products. In addition to impressive gravimetric density, WHAs offer a very useful radiographic density that generally exceeds that of lead, but is slightly lower than that of depleted uranium (DU), depending on the type of radiation and its energy. For 1 MeV -radiation, linear absorption by 95W-Ni-Fe is approximately 50% higher than by lead. Additionally, WHAs prove superior for most shielding and collimation applications over lead in that they provide excellent mechanical strength, are readily machined, resist routine handling damage, and pose no toxicity considerations. Consequently, WHAs are widely used as injection syringe shields, bulk radioisotope storage containers, collimator components in cancer therapy devices, and precision shielding for radiation detector assemblies. Some typical configurations are pictured in Fig. 7.

Fig. 7 WHAs can be easily machined into a variety of high density, high tolerance components such as radiation collimators and precision balance weights

A rapidly growing area of use is consumer sporting goods. While WHAs have been employed as high-density dart bodies for decades, large volume consumption is now being realized as these materials are being used for golf club weights and waterfowl shot. In the course of development for premium irons and drivers, club heads are being designed with WHA inserts that provide intelligently concentrated mass so as to increase the "sweet spot" and hence club performance. The remainder of the club head is fashioned from a low density titanium alloy, so as to keep overall club head weight comparable to former all-

steel designs. Some putter designs utilize WHA masses on the extreme ends of the head, thereby increasing the rotational moment of inertia and adding greater directional stability for slightly off-center strikes of the ball. Also of large volume usage is the application of lead-free shot for hunters of wetland birds. A WHA of sintered tungsten-iron provides an excellent density match to lead alloy shot. The tungsten-iron shot does not deform as it passes through the choke of the gun, permitting improved down-range patterning. Currently, shotshells are designed with a special provision for protecting the bore of the gun, thereby overcoming a serious problem present with early tungsten-base shotshells. For several decades WHA has been used in various designs of kinetic energy penetrators for the defeat of armor on account of its high density and favorable mechanical properties. Designs have ranged from small- and medium-caliber (5.56 to 35 mm bore) fin-stabilized short-rod penetrators to high length-to-diameter (L/D 20+) fin-stabilized medium- and large-caliber (35 to 120+ mm bore) rods. These massive projectiles are typically launched at velocities of 1200 to 1650 m/s, with larger designs delivering many megajoules of kinetic energy to the target. Currently, other ordnance concepts require lengths exceeding 1 m and masses greater than 500 kg. Smaller munitions devices in current use employ prescored cases of WHA loaded with high explosive. Upon detonation, the case disintegrates into a controlled dispersion of very high velocity fragments of prescribed size and flight characteristics for both antimateriel and antipersonnel effects. WHA has largely supplanted WC-Co as the material of choice for non-DU penetrators. While political pressure is mounting against the continued used of DU for ordnance applications, its use persists in large caliber due to the 5 to 10% penetration performance advantage commonly observed over WHA. This slight difference is viewed as critical in that it determines the maximum effective engagement distance to enemy tanks. While much development work has been devoted to developing a WHA with a shearable matrix that could simulate the penetration and erosion behavior of DU-0.75Ti rods, satisfactory results to date have not been obtained. All current penetrator designs employ W-Ni-Fe(-Co) alloys with the exception of the highest performance large-caliber rods that are W-Ni-Co. While higher mechanical properties do not necessarily imply better terminal ballistic performance, high properties are nevertheless desirable for WHA penetrator materials. The higher the property set, the less support required for long rods during launch. The reduced parasitic weight of the launch package means a greater percentage of the energy of a given gun system deliverable to the target. In addition to making the strongest possible alloys, controlled embrittlement is useful for addressing another class of ordnance requirements. By carefully controlling the degree of embrittlement of a WHA, a penetrator can be fabricated that dependably withstands chambering and launch stresses, but fragments into a controlled dispersion of fragments on impact with metal or composite sheets, depositing a large amount of energy in a localized volume. Termed frangible rounds, these penetrators are useful in damaging large areas of airframes in anti-aircraft warfare. Such rounds suffer minimal velocity falloff to target and provide a cost-effective means for defeat of spaced light armor and aerostructures. In answer to an opposite concern, WHA sheet by virtue of its high areal density provides a very effective ballistic shielding material against high-velocity particles even when reasonably thin. WHA has an advantage over DU in that it is not pyrophoric, a consideration that is always present when using DU, whether in machining of parts or for the end use. WHA is widely used for high stiffness boring bars, long extension toolholders, and chatter-free tooling. When compared to more standard iron-base toolholder materials, WHA offers a 70% gain in stiffness and WC-10Co a 295% gain, as seen in Table 7. While carbide provides superior stiffness, WHA is much easier to machine to form and offers better vibration damping. The elastic mismatch of the two phases, a problem in other contexts, proves advantageous for increased damping. Additionally, the higher density of the WHA provides better inertial damping than that available from the less dense carbide.

Table 7 Comparison of various toolholder materials Material 4xxx steel 92WHA WC-10Co

Modulus, GPa 210 360 620

Damping Low Moderate Low

Density, g/cm3 8 17.5 14.5

One of the most common applications for WHAs is logically balancing weights and inertial masses. The strength and hardness levels of WHAs permit secure mechanical fastening of components, a prime consideration for such applications as aileron balance weights in commercial aircraft. Sizes and configurations of current WHA balancing weights span the spectrum from fractional gram masses for balancing hard disk drive heads to large keel masses for ships weighing up to a metric ton or more. Race cars, commercial aircraft, and guided missiles all benefit from the combination of high density, good mechanical properties, and low toxicity provided by these materials. WHAs are used for gyroscope rotors and flywheels, as the density of these materials allows much greater mechanical energy storage than from materials such as steel. For applications in which WHA weights are exposed to weather, cleaning fluids, or other harsh environments, corrosion can be effectively prevented by epoxy paint or plating with nickel or chromium, all of which are supplanting cadmium plating as more environmentally acceptable alternatives. Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Machinability A prime advantage of WHAs over pure tungsten is their ability to be machined to complex geometries using common metalworking tools and techniques. WHAs in general machine in a manner similar somewhat to gray cast iron. WHAs tend to form short chips on machining. The exceptions are highly ductile, lower tungsten content alloys that may tend to form continuous chips and require more attention to chip breaking. The machining characteristics of such WHAs resemble those of superalloys. As all WHAs possess high elastic stiffness, cutting forces are higher than for most metalcutting operations, thus requiring carbide tooling, rigid support, and adequate spindle torque. Turning operations are best performed with ISO K10 type inserts. Sharp insert edges are vital for minimizing cutting force. Very rough cuts should be made with negative rake angle inserts though tooling forces will be high. Roughing passes typically utilize depths of cut of 0.7 to 3 mm, feed rates of 0.13 to 0.25 mm/rev, with surface speeds of 60 to 90 m/min. For turning and milling operations, the use of a coolant/lubricant is optional. Finishing cuts are typically made using 0.13 to 0.38 mm depths of cut and feed rates of 0.08 to 0.25 mm/rev. Large nose radius inserts, high spindle speed, and light feed rates contribute to the finest surface finish. Tightest tolerance finishing passes are made with positive rake inserts. WHAs can easily be machined dry. If a cutting fluid is to be used, it should be a water-based, nonalkaline formulation. Multi-insert cutter heads should be employed for milling WHAs. ISO K10 inserts with a positive rake generally prove most useful. Roughing can be performed at surface speeds of 60 to 120 m/min and feed rates of 0.13 to 0.25 mm per insert. Finishing passes are often conducted at 90 to 150 m/min and reduced feeds of 0.08 to 0.25 mm per insert. Grinding is best performed with vitrified-bond alumina or silicon carbide wheels of medium hardness. The ductile binder phase present in WHAs prevents the use of diamond wheels due to loading. Tapping is the most challenging of machining operations for WHAs. Tapping of threads finer than 6-32 may prove impractical. The use of coated cobalt steel taps with two (at most three) heavy flutes and positive rake for increased clearance is highly recommended. Sulfonated oils are effective tapping lubricants. Approaches to minimizing tap/work contact area include the use of large pilot hole sizes and 60% thread area engagement should be considered. Holes -20 and finer may require dropping to 50% engagement due to the high stiffness of the alloy and the correspondingly high torsional stress induced in the flutes and shank. Holes should be tapped to completion without back threading. For larger holes, single-point threading may prove the best option. Designs should always utilize the coarsest thread possible.

Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Joining Mechanical fastening techniques should be employed where possible so as to minimize the uncertainties associated with thermal bonding techniques, most of which relate to the formation of embrittling intermetallics. Additionally, thermal bonding of worked WHA is not feasible due to property alteration that will result as a loss of strain hardening or at lower temperatures, overaging. Due to differences in both chemistry and melting point of the tungsten phase and the binder, WHAs are not readily weldable. Brazing of WHA to dissimilar metals is possible using copper or commercial brazing alloys based on copper, nickel, or silver. In virtually every case, the mechanical properties of the joint are lower due to the formation of intermetallics. Diffusion bonding is also readily performed, but is subject to the same limitations as brazed joints. When it becomes necessary to join sintered WHA bodies together to form larger components, this is readily accomplished by sinter bonding, in which the pieces are placed in contact and heated sufficiently to permit liquid-phase formation. If done properly, this technique yields an imperceptible joint with no local degradation in properties. Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Environmental Considerations As environmental and safety considerations become evermore important drivers in the design of new products and selection of appropriate materials, the fact that WHAs have only a slight percentage of their total weight in slightly toxic metals such as nickel and cobalt provides an incentive for their selection over other common high density materials such as lead and DU. Whereas the latter two materials pose definite toxicity concerns, less than 7% of the total weight of a typical WHA is comprised of metals such as nickel and cobalt. Life cycle considerations are also favorable for WHAs as these materials are readily recyclable by both chemical and oxidation/reduction means. Both WHA machining scrap and spent components serve as raw materials for future alloy production instead of posing a disposal problem. WHAs are therefore an ideal materials choice for applications requiring a combination of high density, good machinability, moderate cost, minimal toxicity, high stiffness, and reasonable thermal stability. Powder Metallurgy Tungsten Heavy Alloys S.G. Caldwell, Teledyne Metalworking Products

Selected References •

A. Bose, D. Sims, and R.M. German, Test Temperature and Strain Rate Effects on the Properties of a Tungsten Heavy Alloy, Metall. Trans. A, Vol 19, 1988, p 487-494

•

S.G. Caldwell, Variation of Ni/Fe Ratio in W-Ni-Fe Heavy Alloys: A Current Perspective, Tungsten & Tungsten Alloys--1992, Metal Powder Industries Federation, 1993, p 89-96 G.B. Dudder and W.E. Gurwell, Hydrogen Embrittlement Effects on Tensile Properties of Tungsten Heavy Alloys, Tungsten and Tungsten Alloys--Recent Advances, A. Crowson and E.S. Chen, Ed., The Minerals, Metals, and Materials Society, 1991, p 161-167 H. Hofmann, "Konstitution des W-Ni-Fe Systems und Mechanische Eigenschaften W-Reicher W-Ni-FeLegierungen," Ph.D. dissertation, Technischen Universitat Berlin, 1983 R.E. Oakes, Jr. and M.W. Moyer, "A Review of Techniques Used to Characterize Cemented Tungsten Alloys for Penetrator Applications," Report Y-2083, Oak Ridge National Laboratory Y-12 Facility, 1977 B.H. Rabin and R.M. German, Microstructure Effects on Tensile Properties of Tungsten-Nickel-Iron Composites, Metall. Trans. A, Vol 19, 1988, p 1523-1532 J. R. Spencer and J.A. Mullendore, "Relationship between Composition, Structure, Properties, ThermoMechanical Processing, and Ballistic Performance of Tungsten Heavy Alloys," Report MTL TR 91-44, U.S. Army Materials Technology Laboratory, Watertown, 1991 S.W.H. Yih and C.T. Wang, Tungsten, Plenum Press, 1979

•

• •

• •

•

Powder Metallurgy Cemented Carbides

Cermets

and

Introduction METAL AND CERAMIC composites include two types of materials known as cermets and cemented carbides. The most outstanding example of the desirable properties obtained from combining metal and ceramic materials involves the hardmetal types made from cemented carbides. Cemented carbides have enjoyed a steady expansion over the past six decades. Over that time, the development of metal-ceramic tool materials has moved away from the early tungsten-base carbides to carbide- and nitride-base compositions of increasing complexity (Table 1). This includes several types of cermet materials.

Table 1 History of cermet product development and marketing Year 1930-1931 1930 1930 1933 1938-1945 1949-1955 1952-1954 1960 1970 1974 1975 1977-1980 1980-1983 1988

Composition WC-Co TiC-Mo2C-(Ni, Mo, Cr) TaC-Ni TiC-TaC-Ni TiC-VC-(Fe, Ni, Co) TiC-(NbC)-(Ni, Co, Cr, Mo, Al) TiC-(Nb, Ta, Ti)C-(Ni, Mo, Co) TiC-(steel, Mo) TiC-(Ni, Mo) Ti(C, N)-(Ni, Mo) (Ti, Mo) (C, N)-(Ni, Mo) TiC-TiN-WC-Mo2C-VC-(Ni, Co) TiC-Mo2C-(Ni, Mo, Al) (Ti, Mo, W) (C, N)-(Ni, Mo, Al) (Ti, Ta, Nb, V, Mo, W) (C, N)-(Ni, Co)-Ti2AlN

Trademark G1 Titanit S Ramet ... ... WZ Kentanium Ferro-TiC ... Experimental alloys Spinodal Alloy KC-3 ... ... TTI, TTI 15

Manufacturer Krupp-Widia Metallwerk Plansee Fansteel Corporation Siemens AG Metallwerk Plansee Metallwerk Plansee Kennametal Sintercast (Chromalloy) Ford Motor Company Technical University Vienna Teledyne Firth Sterling Kyocera Ford Motor Company, Mitsubishi Mitsubishi Krupp-Widia

Source: Ref 4 and Kennametal, Inc.

Acknowledgements The section on cermets was adapted from an article in Volume 2 of the ASM Handbook, 1990, p 978-1007, by John L. Ellis, Consultant, and Claus G. Goetzel, Consultant and Lecturer. The section on cemented carbides was adapted from an article in Volume 2 of the ASM Handbook, 1990, p 951-977, by A.T. Santhanam, P. Tierney, and J.L. Hunt of Kennametal Inc.

Reference

4. R. Kieffer and F. Benesovsky, Hartmetalle, Springer-Verlag, 1965, p 437-489 Powder Metallurgy Cermets and Cemented Carbides

Cermets Cermet is an acronym that is used world wide to designate "a heterogeneous combination of metal(s) or alloy(s) with one or more ceramic phases in which the latter constitutes approximately 15 to 85% by volume and in which there is relatively little solubility between metallic and ceramic phases at the preparation temperature" (Ref 1, 2). A good definition of the term ceramic can be found in the Ceramic Glossary (Ref 3): "Any of a class of inorganic, nonmetallic products which are subject to a high temperature during manufacture or use. Typically, but not exclusively, a ceramic is a metallic oxide, boride, carbide, or a mixture or compound of such materials; that is, they include anions that play important roles in atomic structures and properties." With particular reference to cermets, this definition of the ceramic component could be broadened to include nitrides, carbonitrides, and silicides. Cermets originally were used for cutting tool applications. Some 45 years ago (Ref 2, 4), they began to be considered for use in more taxing applications, such as propulsion systems. The expectations were that the refractory behavior, strength, and corrosion resistance of the ceramic phase could be mated advantageously on a proportional basis with the high ductility and thermal conductivity of the metallic phase, and that some superior new materials would become available for a multitude of high-temperature applications. A significant application of cermets involves cutting tool materials that utilize titanium carbide or titanium carbonitrides as the hard refractory phase. Frequently, molybdenum carbide (Mo2C) and other carbides are also built into these cermet formulations. The cratering and flank wear resistance properties of the titanium carbide and titanium carbonitride cermet tool materials are better than those of the conventional cemented-carbide (that is, cobalt-bond tungsten carbide) tool. In comparison to ceramic cutting tools, these cermets permit heavier cuts, which, at high speed, results in greater amount of metal removal at a comparable level of tool life. Cermets clearly possess characteristics of a cutting tool material that is capable of filling the gap between conventional cemented carbides and ceramics (see Machining, Volume 16 of the ASM Handbook). Classification of Cermets Cermets can be classified according to their hard refractory component. In this system, the principal categories of cermets are determined by the presence of six components: carbides, carbonitrides, nitrides, oxides, borides, and miscellaneous carbonaceous substances.

The metallic binder phase can consist of a variety of elements, alone or in combination, such as nickel, cobalt, iron, chromium, molybdenum, and tungsten; it can also contain other metals, such as stainless steel, superalloys, titanium, zirconium, or some of the lower-melting copper or aluminum alloys. The volume fraction of the binder phase depends entirely upon the intended properties and end use of the material. It can range anywhere from 15 to 85%, but for cutting tool applications it is generally kept at the lower half of the scale (for example, 10 to 15 wt%) (Ref 5). The metallic bond for each cermet is selected in order to produce the desired structure and properties for the specific application. The iron group metals and their alloys dominate as in the cemented tungsten carbide class of hard metals; nickel and, to a lesser extent, cobalt and iron possess a desirable combination of relatively high hardness and good ductility. However, the binder for a cermet can also be chosen from the group of more reactive metals, such as titanium or zirconium, or it can be selected from series of refractory metals that includes chromium, niobium, molybdenum, and tungsten. Lowermelting metals and alloys, primarily those based on copper and aluminum, round out the list of binders at the bottom of the temperature scale. Aluminum, however, is more commonly associated with metal-matrix composites. Carbide-base cermets are by far the largest category of cermets, even if the term is used in its narrower sense and

excludes the broad field of cemented-carbide cutting tools and wear parts based on tungsten carbide (WC). Since the inception of cermet technology, the dominant concept has been that of a material based on TiC as the primary hard and refractory constituent, with the bonding provided by any of a variety of lower-melting ductile metals or alloys (much the same as those used for cemented tungsten carbides). The TiC cermets have found use in tool and wear resistance applications; in selected high-stress, high-temperature systems; and in corrosive environments. Cermets based on SiC and B4C, which generally are classified as metal-matrix composites, have gained considerable industrial significance in wear and corrosion resistance, or antifriction, applications; they are also used in nuclear reactor applications. Cermets with a chromium carbide (Cr3C2) base have been used for a variety of corrosion resistance applications and as gage blocks; however, they have apparently lost much of their industrial usage. Carbonitride-base cermets can be produced with or without additions of various other carbides (of which Mo2C is the most important); they are bonded with the common cemented-carbide binders. At present, these materials are the primary cermets for tool applications. Their enhanced strength, which makes them suitable for high-speed cutting tools, is based on a greatly improved bond between the hard carbide grains and the binder metal. The improved bond is a consequence of a miscibility gap in the quaternary TiC, TiN, MoC, and MoN system that results in a so-called spinodal decomposition into two isostructural phases (Ref 6) with inherently better wettability to the binder (Ref 5). Nitride-base cermets constitute a special class of tool materials. Titanium nitride (TiN) and especially cubic boron nitride (CBN) produce excellent cutting materials if they are combined with a hard binder metal. Titanium nitride and zirconium nitride (ZrN) bonded with their respective metallic elements have been developed for special heat- and corrosion-resistant purposes. Oxide-base cermets constitute a category that includes UO2 or thorium dioxide (ThO2), which are used for a major

fissions component in nuclear reactor fuel elements; Al2O3 or other highly refractory oxides, used for components in liquidmetal manipulation (for example, pouring spouts) and general furnace parts; and SiO2, used for a minor constituent in friction elements. Combinations of Al2O3 with TiC are suitable for hot-machining tools. Boride-base cermets have a boride of one of the transition metals as the dominant phase. These cermets provide excellent

high-temperature corrosion resistance to attack by active metals, such as aluminum, in the molten or vapor state. A combination of ZrB2 and SiC is resistant to erosion from the propulsion gases of chemical rockets. Carbon-containing cermets are materials that contain graphite in varying proportions. They are used for electrical

brushes and contacts or as minor constituents to provide some lubrication in friction elements. Also included in this category are diamond particles within metal matrices that are used in special tools. Cermet Fabrication The methods used for powder preparation, forming, firing or sintering, and post-treatments of cermets generally are similar to conventional ceramic and powder metallurgy (P/M) processing techniques. Figure 1 is a flow chart of the various P/M

techniques applicable to cermets. Table 2 summarizes the relative characteristics of the major forming methods that are practiced in producing carbide-base and most other types of cermets. The principal processes are cold forming and sintering, pressure sintering, and infiltration.

Table 2 Cermet forming techniques

Prismatic shapes without undercuts Simple or complex shapes

Mold or die requirements Hardened steel or carbide dies Rubber molds

Production Rate Labor High Low to moderate Low High

Flat and thin Long pieces with uniform cross section Complex shapes with undercuts Prismatic shapes without undercuts Simple or complex shapes

... Hardened steel or carbide dies Hardened steel or carbide dies Graphite or ceramic molds ...

High High

Low Low

High

Low

Low

High

Medium

Low to high

Long pieces with uniform cross sections Intricate shapes feasible

Alloy steel dies

High

Low

Graphite and ceramic molds

Low

High

Technique

Size capability

Shape capability

Static cold pressing

Limited by press capacity

Hydrostatic cold pressing Powder rolling Warm extrusion

Limited by capacity of pressure vessel Limited width, long length Limited by equipment size

Powder injection molding Static hot pressing

Small pieces

Hot isostatic pressing Hot extrusion

Limited by capacity of pressure vessel Depending on press capacity

Infiltration

Depending on equipment

Limited by press capacity

Production method 1. Presintering 2. Vacuum sintering 3. Canning 4. Hot extrusion 5. Infiltration 6. Warm extrusion 7. Slip casting 8. Powder injection molding 9. Powder rolling 10. Hot pressing

Products Cemented-carbide parts and cermets Steel-bonded carbides (standard pieces) and cermets Steel-bonded carbides (special pieces) Aluminum cermets with moderate amounts of hard-phase additions TiC parts with nickel- or cobalt-base infiltrants and other cermets with about 55-85 vol% hard phase Cemented-carbide rods or other slender cermet parts Cermets with high proportions of hard phase A wide variety of cermet compositions Aluminum, copper, and other nonferrous metals with moderate additions of hard-phase components A wide variety of cermet compositions

Fig. 1 Powder metallurgy production methods for cermet and cemented-carbide products

The cold-pressing process includes static uniaxial and isostatic multiaxial compaction. The powder mixtures are compacted at pressures of 35 to 100 MPa (5 to 14.5 ksi). The predominant method involves pressing dry wax-lubricated powder in hardened steel dies with double-action opposing punches. For long rods or tubes of uniform cross section, these dies are used for the extrusion of a paste in which the powder particles are embedded in suitable organic binder or wax. To form complex or large shapes, the dry powder is placed in a pliable mold and compacted from all sides by hydrostatic pressure inside a sealed, reinforced steel cylinder. Powder Preparation. The first step in the cermet production process is the mixing and milling of the ingredient powders. The mixtures, consisting of the hard-phase substance in powder form and the pure metal or metals in the proper proportions required for the composition of the binder alloy, are milled in ball mills. The balls are made of tungsten carbide or, more frequently, of a highly sintered cermet. The mill can be lined with the same type of cermet material to reduce the possibility of mixture contamination. In addition to conventional ball mills, high-energy vibratory ball mills and attrition mills are used. With the later type of mills, substantial savings in milling time, energy, and floor space can be achieved. During the milling process, the hard phase particles are comminuted and thoroughly coated with binder metal. Organic liquids such as hexane are used in the process to minimize the rise in temperature and prevent oxidation. After the powders have been milled to a particle size of 325 mesh or finer, the mixtures are dried before further processing and use. A lubricant is then added, and those powder mixtures destined for compaction in automatic presses are agglomerated so that they can flow freely from the hopper to the compacting die. Static Cold Pressing. Cold-pressing methods for cermet powder mixtures generally follow the well-known powdercompacting techniques used in conventional powder metallurgy. Small cermet parts needed in reasonably large quantities are compacted in special hard-metal dies by automatic presses with double-action opposing punches (Fig. 2). Whether solid or segmented, the dies are shrunk into a strong, tough, heat-treated steel retainer.

Fig. 2 Static cold pressing with (a) a conventional press and (b) an anvil press. The anvil press has no upper punches; therefore, the misalignment, breakage, and wear problems associated with those punches are eliminated. Courtesy of PTX-Pentronix, Inc.

The automatic compaction cycle consists of filling powder in the die, compacting the powder, ejecting the compact, and removing it. Two methods are used for ejecting the compact from the die. In the first method, the lower punch moves upward and pushes the bottom of the pressed piece to the level of the die table. In the second method, the die is forced down (withdrawn) over the lower punch until the bottom of the compact is on level with the top of the die (Fig. 3). The second method is gaining favor with many specialists because it allows building shorter, less expensive tools, and because it provides better support for the fragile compact during ejection.

Fig. 3 Withdrawal press cycle with controlled die motion (top and bottom pressure). Courtesy of Dorst America

Another interesting variation, particularly for the compaction of fragile cermets, is provided by the so-called anvil press (Fig. 2b). In this method, the powder is pressed against the anvil by the upward motion of a single lower punch. The anvil is then laterally removed to allow ejection of the compact by the continued upward motion of the punch. Because this cold-pressing technique is a single-action pressing method, an anvil press is acceptable only for relatively thin single-level pieces. Yet, where applicable, the anvil press saves tool costs; in addition, it reduces the ejection path to a minimum, which is an important factor for fragile cermets. Medium-to-large rectangular pieces are compacted on hydraulic presses in multiple-section dies held together by powerful steel frames. These frames are capable of counteracting the large internal forces that are exerted on the powder mass and transmitted radially to the die sections. Often, at the completion of the cycle, the dies are opened for the careful removal of the fragile cermet compact. For round pieces it is preferable to use double-action hydraulic presses with a built-in ejection action. These presses can use either one of the two ejection methods: • •

If the die is mounted on a hydraulically activated floating platen, a single-action press will adequately provide the effect of a double-action press; in this case, ejection is accomplished by the withdrawal method. When a full-power double-action hydraulic press is available (obviously a heavier and more expensive machine than the single-action press), the die can be mounted on the stationary platen. The compact is ejected by raising the lower punch to the level of the die table at the completion of the pressing cycle.

Cold compaction of simple cermet shapes gives generally good results when using adequately lubricated powder mixtures, well-designed tooling, and sturdy presses. Even though the compacts are fragile, they should nevertheless be firm, have adequate green strength and well-formed edges, and be free of laminations or other internal defects. Depending on the composition of the cermet, compaction problems can arise that require modifications in the equipment or the process. Generally, the higher the proportion of the hard phase, the greater the difficulties that arise in compaction; these difficulties may not be evident until after sintering. Also, more problems can be expected with iron-, nickel-, or cobalt-base cermets than with the softer, more malleable aluminum-base compositions. Larger pieces, as measured both in diameter and height, intensify the problems associated with composition. Air entrapment, bridging, laminations, intermittent voids, and variable density throughout the compact are just a few of the problems encountered in the cold compaction of cermets (Ref 7). Some of these problems can be overcome by adding more lubricant to the mixture, by increasing the taper of the compacting die, or by slowing down the compaction process. Die wall lubrication between pressings often helps to eliminate compaction or ejection problems. Adequate preloading of the die and die lapping in the pressing direction are other precautions that help to overcome problems. Even for the simplest forms, such as cylindrical or rectangular single-level blanks, not all problems encountered can be solved by the conventional static cold-pressing technique. This is one reason why P/M specialists depend heavily on other, more sophisticated forming processes for the fabrication of a wide variety of products. Cold Hydrostatic Pressing. High-quality cermet compacts require uniform densification throughout. This can be ideally

accomplished by cold hydrostatic pressing. In this method, pressure is applied simultaneously and uniformly from all sides toward the center of gravity of the powder mass while all friction of powders against the die wall is completely eliminated. In order to compact simple or even relatively complex shapes by this method, dry powders are filled into a pliable mold. The powders are settled, and air is removed on a vibrating table; the mold is then sealed and placed into a reinforced steel cylindrical vessel filled with a fluid. After the vessel is closed, hydraulic pressure is built up, thereby compressing the powder contained in the mold. The two hydrostatic pressing methods frequently used for pressing cermets are the wet-bag method and the dry-bag method. The wet-bag method involves placing one or more powder-filled molds inside a hydrostatic pressing vessel (Fig. 4). The

powder-filled mold is pliable and is placed within a perforated container for support. Inside the vessel, the powder-containing mold is completely surrounded by hydraulic fluid. Depending on the size of the vessel and the individual mold, often a number of molds can be placed in the vessels and compacted simultaneously. The entire process--loading the vessel with one or several molds, building up and holding the pressure, releasing the pressure, and reopening and unloading the vessel--is relatively slow. Moreover, filling the mold, assembling the mold, loading the molds into the pressure vessel, and unloading and removing the pressed piece after completion are slow, manual processes that require meticulous attention to detail.

Fig. 4 Schematic of a cold hydrostatic vessel with a wet-bag powder mold. Source: Ref 7

Dry-bag pressing uses a flexible mold that is permanently sealed in the pressure vessel (Fig. 5). After the mold cavity has

been filled with a controlled quantity of powder, the cover plate is closed, and hydraulic pressure is applied. After the pressure is released, the molded piece is removed, and a new cycle commences. Dry-bag pressing is a much faster production process than the wet-bag method and lends itself to automation. Much of this technology has been developed for producing near-net shape ceramic pieces, for example, automotive spark plug bodies. Each dry-mold setup requires special engineering and development (Ref 8).

Fig. 5 Schematic of dry-bag hydrostatic pressing equipment. Courtesy of Olin Energy Systems

Advantages and Disadvantages. Hydrostatic pressing offers the following advantages (Ref 7):

• • • • • • •

Pressed cermet pieces have a uniform density regardless of size and shape. Wet-bag method is well suited for large pieces and often is the only practical method for pressing such pieces. Slender pieces with high ratios of length to cross section are feasible. Mold cost is low compared to that of rigid compacting dies. Low production quantities can thus be economically produced, especially by the wet-bag method. Undercuts and varying cross sections are feasible with either the dry-bag or wet-bag method. Little or no lubricant is required. The process is well suited for research and development work.

The disadvantages of hydrostatic pressing are: • • • • •

Dimensional control of compacts if limited. Mold design must accommodate the radial and axial shrinkage caused by hydrostatic pressing as well as the shrinkage that occurs during subsequent sintering. Surfaces of compacts are less smooth than those of die-pressed pieces. A high liquid-phase sintering step or encapsulation is necessary before hydrostatically pressed cermet pieces can be densified by hot isostatic pressing. Equipment cost is high, and equipment utilization can be low. Labor cost is relatively high.

For difficult-to-press cermet compositions with high loading of the hard phase and/or relatively hard metal and alloy binders, cold hydrostatic pressing often is a convenient production method; sometimes it is the only reliable method for working with certain compositions. Warm Extrusion of Cermet Powder Mixtures. The process of warm extruding cemented ultrafine carbide powder with an admixture of plasticizers has been known for many years. It is successfully used for cermets as well as for forming simple prismatic shapes that have a high ratio of length to cross sections. Cylindrical and triangular shapes and other cross sections can be readily extruded; even tubes are feasible (Ref 9).

Depending on the plasticizer used (for example, polystyrene with an admixture of diphenyl and diphenyl-ether), extrusion requires a temperature somewhere between 160 and 175 °C (320 and 350 °F). Slow and complete debinding under vacuum prior to high sintering is critical in order to avoid distortion, cracking, or microporosity. Screw extruders similar to those used in the plastics industry are adapted to this process (Fig. 6). For the production of a high-quality product, hot isostatic pressing is recommended.

Fig. 6 Warm extrusion of cermet powder mixtures. Courtesy of Dorst America

Power rolling (roll compacting) is a well-known forming process in conventional powder metallurgy that may find application in cermet production. In this process, cermet powder mixtures are fed from a hopper into the gap of a rolling mill and emerge as a continuous strip or sheet. While the horizontal arrangement of the rolls is most convenient for feeding the powder from the hopper into the roll gap (Fig. 7), the vertical arrangement is preferable for feeding the emerging fragile strip horizontally into a series of subsequent operations. The vertical arrangement requires more carefully engineered devices for feeding the powder uniformly into the roll gap (Ref 7).

Fig. 7 Schematic of powder rolling with saturated feed and horizontal roll arrangement. Compression ratio, h0/hg. Source: Ref 7

In contrast to the starting material in slab rolling, the loose powder used in powder rolling has no strength before entering the roll gap and must flow freely or be forced into the gap. During the roll compacting process, the density and physical properties of the powder mixture change. For cermet powder compositions, the feasibility of roll forming a strip of sufficient density and strength depends upon a number of factors, including, but not limited to, the roll diameter and speed, the degree of loading of the cermet mixture with hard-phase substances, the ductility of the metallic phase, and the amount of plasticizer added to the mixture. The presence of the hard component adds to the friction of the powders against the roll and to the internal friction of the powder mixture during the compacting step. This is a favorable characteristic of cermet powders for roll forming; however, it is offset to some extent by the inherently low green strength of the resultant sheet or strip. The sheet thickness that can be compacted with a given diameter roll is quite limited. A ratio of roll diameter to strip thickness between 600 to 1 and 100 to 1 seems to be the range for various metal powders (Ref 7). It is reasonable to assume that the middle-to-lower range applies to cermets. Special devices are required to prevent the powders from flowing laterally out of the roll gap. A uniform flow of powders over the entire width of the roll is essential for obtaining uniform density in the roll-formed strip. Edge cracking can occur, particularly with heavier strips. An optimum strip thickness has to be established experimentally. Thicker strips are too stiff to be coiled, and thinner ones are too fragile. Rolling speed is another variable that can only be optimized through experimentation with a given cermet powder composition. Pure metal powders without hard phase have been roll compacted at speeds of 30 m/h (100 ft/h). It remains to be seen whether an output anywhere near this order of magnitude can be obtained with cermets. A complete powder-rolling line for continuous operation includes debinding and sintering furnaces, rerolling stands, and, if necessary, one or more reannealing furnaces. Up-coiling equipment is needed at the end of the line. This equipment constitutes a major capital investment that is warranted only by a large and continuous demand for the product. Although labor costs for such an operation are low, it may be some time before this line production method finds applications in cermets. A simpler arrangement (Fig. 8) is feasible if, after a debinding step (not shown) and continuous atmosphere

sintering, a product emerges of sufficient strength and ductility to permit upcoiling. Roll-compacting arrangements have been proposed for producing a sandwich-type strip consisting of two layers of different compositions.

Fig. 8 Powder rolling process with strip reeled into individual rolls after first sintering treatment. Source: Ref 7

Slip casting, a method for forming metal powders into a desired shape, follows a technique that has been used for ceramics

for a long time. This method uses an aqueous suspension of cermet powders (the slip) that is poured into a porous plastic mold. The liquid is absorbed by the mold, and the powder is deposited on the mold wall. In the case of hollow shapes, the excess slip is drained off after the deposit reaches the required wall thickness; for solid parts the slip must remain and slowly dry. The water-base slip has low viscosity in order to facilitate pouring, yet it should be stable during standing in order to avoid demixing. Demixing can be a serious problem with cermet powders, particularly those having a substantial difference of specific weight between the hard phase and the binder metal; it can lead to differences in composition and properties from one end of a cermet part to another. Variations in composition can lead to cracks during drying or subsequent sintering. In order to control the viscosity of the slip at the optimum level, it is generally necessary to use a deflocculant and to control the pH. After slow drying, the slip cast part needs a debinding step followed by high sintering. The resultant part has a higher density than the tap density of its original powder mixture. The fine powders that frequently are used to facilitate slip casting can lead to superior properties in the sintered part (Ref 7). Slip casting and mold making together are more art than technology. They require knowledge of parameters such as slip viscosity and suspension stability, wetting agents and deflocculants, and slip-mold interaction and mold release. Other important parameters are wall-building rate and casting crack formation. Slip casting requires only a small investment; however, it is labor intensive and is not well suited for mass production. At the present state of technology, cermet parts of a certain complexity are more likely to be suitable for injection metal molding than for slip casting. The former process is more capital intensive, but it is better suited for a medium-to-large production volume. The P/M injection molding (MIM) process has evoked a great deal of interest since it was first developed in the early 1970s. Commercialization has been slow, mostly because of the long cycle that is required from concept to the point of shipping acceptable parts to a customer. Intensive research and application engineering continues in many laboratories, and more rapid growth is expected in the future (Ref 10, 11).

On a laboratory basis, cermet parts have been made by this process, and its commercialization is underway, particularly in the field of cemented carbides. However, the bulk of current MIM experience is in the area of structural ferrous and nonferrous parts. The powder injection molding process for cermets (Fig. 9) involves mixing and blending the ingredient metal and hard-phase powders with a suitable polymer binder and then granulating the mixture. The granulated product is heated and injection molded under pressure. The polymer imparts viscous flow characteristics to the mixture to aid in forming, mold filling, and uniform packing. After demolding, the binder is removed, and the remaining cermet structure is densified by sintering and, perhaps, by hot isostatic pressing (Ref 10).

Fig. 9 Mold and injection mechanism for the MIM process. Source: Ref 10

Binder compositions and debinding techniques are the main differences among the various MIM processes. There is no universal binder. A primary requirement of the binder is that it allow flow and packing into the mold cavity. It must wet the powder, and it should be designed to minimize debinding time and defects. A multiple-component binder that is not chemically intersoluble allows for progressive extraction in debinding. As one compound is removed and the pores partially opened, the remaining binder holds the particles in place and maintains the shape of the compact. The remainder then vaporizes through the open pores without generating an internal vapor pressure that might cause compact failure. Waxes with additives are most frequently used as binders. The phases of the molding operation are: • • • •

Clamping and filling of the mold Maintaining pressure while the compact becomes solid Retraction of filling mechanism Opening of the mold and ejection of the compact

Mold filling depends on the viscous flow of the feedstock into the mold cavity. The viscosity depends on temperature, shear rate, binder chemistry, powder interfacial chemistry, and loading (Ref 10). Thermal debinding is the most frequently used technique, but capillary wicking and solvent extraction can also be considered as an alternative method. Complete debinding is required before commencing the sintering cycle. Most cermets require a liquid-phase sintering cycle to achieve complete densification of the compact. A modern furnace that combines debinding, high-vacuum sintering, and a final pressure-sintering cycle can accomplish all of these steps economically.

Applications and Advantages of the MIM Process for Cermets. In recent years, major progress has been made in

using the MIM process for the production of heat engine components, military hardware, computers, and aerospace and automotive components. The powder injection molding process offers new opportunities in advanced materials manufacturing (Ref 11), and it offers potential advantages for use in cermet manufacturing technology. General aspects of applying the MIM process in cermet manufacturing include: • •

•

•

• •

In principle, the MIM process is applicable to cermets without modification of the injection molding machines or the typical mold designs (Fig. 9). The production of small- to medium-size complex shapes by the MIM process is feasible, provided that the geometry of the shapes allows for demolding. When this requirement is met, multiple levels, reentrant angles, and undercuts can be accommodated. On small parts, tolerances of ±3 m/mm (0.003 in./in.) after sintering can be obtained on conventional P/M parts with the MIM process. Larger tolerances would be needed on cermets to allow for shrinkage when liquid-phase sintering is needed for complete densification. Small runs (of as few as 2000 parts) are feasible for conventional P/M parts. Because of the higher price level of cermets and the relatively high cost of competitive forming techniques, runs of similar or even smaller size could be economically attractive for cermet parts produced by the MIM process. With proper debinding techniques and a liquid-phase sintering cycle (perhaps followed by hot isostatic pressing), high-quality parts with good physical properties could be produced by the MIM process. Excessive mold wear caused by the hard phase during the injection molding of a cermet composition, particularly one with high loading, does not seem to be a serious problem in the MIM process. Future experience will demonstrate if such a mold wear problem exists and to what extent it affects the economics of using the MIM process for cermets.

Cermet Sintering and Consolidation Sintering. Not all cermets require liquid-phase sintering, but the majority use this process to convert the green compacts

into solid, strong, and dense products. Sintering temperatures depend entirely on the ceramic-metal system involved and on the choice between solid- and liquid-phase sintering. Typical temperatures range from 850 to 1050 °C (1560 to 1920 °F) for products that contain a bronze, silver, or copper metal matrix; 1300 to 1500 °C (2370 to 2730 °F) for cemented carbides and borides; and 1700 to 2200 °C (3100 to 4000 °F), or even higher, for certain ceramic oxide-base cermets. For applications requiring fine machining and grinding, as in many cemented-carbide parts and tools, presintering is performed at 1000 to 1100 °C (1830 to 2010 °F) to bond the metallic contact points and give enough green strength to the body so that it can withstand rough machining. Allowance is made for the substantial shrinkage that occurs during subsequent sintering. Depending on the green density, cermet compacts can shrink during liquid-phase sintering by as much as 18 to 26% linear (45 to 60% by volume). In systems with good sinterability, virtually all porosity is eliminated (Ref 12). During all sintering processes, particularly during the liquid-phase process, many complicated metallurgical phenomena take place that depend on temperature, furnace atmosphere (hydrogen, inert gas, or vacuum), and the dynamics of the particular ceramic-metal system. For example, metals change into alloys; the hard phase partially dissolves in the liquid phase and changes the composition of the latter phase; portions of the liquid phase can diffuse into the hard phase; and reprecipitation of some elements dissolved in the liquid phase can take place during the cooling portion of the cycle. Also, if carbon is present in the furnace atmosphere (perhaps from the furnace furniture), it will reach with oxygen or other elements. The phenomena that occur during the sintering of WC-cobalt systems have been investigated very thoroughly over a long period of time. Ample literature on the basic system and many of its alloy variations is available (Ref 12). Mechanism of Liquid-Phase Sintering. While not strictly a cermet, the liquid-phase sintering of heavy alloys consisting of tungsten-nickel-copper has interesting ramifications that are applicable to the cermet field (Ref 7). This liquidphase sintering process includes these principal features:

• • • • •

The hard phase is partially soluble in the liquid phase during sintering. At temperature, the liquid phase is limited so that the compacts keep their shape. The sintering temperature must be high enough so that an appreciable amount of liquid phase is present. When these conditions are met, densification takes place. The finer the particle size of the hard phase is, the more rapid and complete the densification of the compact is. Final density is independent of the compacting pressure. To reach theoretical density, compacts pressed at low pressure will shrink correspondingly more than those pressed at high pressure. The microstructure will show grain growth when compared to the particle size of the original hard-phase powder. This grain growth can be appreciable and is dependent on the sintering time and temperature.

When the original hard-phase particles are angular (for example, titanium carbide), they can become rounded during the sintering process. However, this is not always the case because some angular hard substances (for example, tungsten carbide) seem to possess shape memory. During the reprecipitation of dissolved elements from the liquid phase during cooling, angular contours reappear on the hard-phase particles of these substances. Furnaces. Continuous high-temperature sintering furnaces have been used in the cemented-carbide and refractory metal industries for many years. They are equipped with a hydrogen or protective atmosphere with a low dew point to reduce residual oxygen in the compacts and to prevent oxidation. Continuous pusher-type furnaces equipped with silicon carbide or molybdenum heating elements have been particularly successful and are used for sintering a large volume of small parts. Batch-type vacuum furnaces have become very popular in the last 30 to 40 years. When there is a choice between pusher-type continuous furnaces and batch-type furnaces, the equipment and operating costs favor the former.

In a typical operation, the parts are laid out without packing on graphite plates that are stacked with spacers within the furnace. When direct contact between the graphite and the compacts is undesirable, the plates are lined with an inert ceramic. It is essential to use vacuum equipment when highly reactive powder mixtures are sintered. The optimum level of vacuum to be reached during liquid-phase sintering varies greatly depending on the hard-phase binder system being treated. Several advanced furnace designs provide for initial operation using a hydrogen atmosphere, with a switch to vacuum at a later stage in the sintering cycle. Others provide a pulsating cycle of hydrogen pressure alternating with vacuum. Static Hot Pressing. Hot pressing is a cermet production method in which the pressure and temperature are applied

simultaneously. The powder mixtures are either compacted directly in the hot press mold or prepressed cold in dies and then transferred to the hot press tools of the pressure-sintering furnace. Pressures are considerably lower than for the cold press method. They can range from deadweight loads up to 3 MPa (500 psi) for pressure sintering (of friction elements, for example), or from 10 to 35 MPa (1500 to 5000 psi) for hot pressing; the lower end of the hot-pressing range applies to liquidphase systems. Sintering temperatures are reached by induction or resistance heating of the mold, or by direct induction or resistance heating of the powder compact. In the former case, the mold material consists of graphite. This is the more practical process because usually no controlled atmosphere supply is required. The latter method requires ceramic molds, which are sensitive to thermal shock, break easily on product removal, and are costly to produce accurately to the dimensions of the mold opening. The advantage of direct compact heating--that the tooling and surrounding area can remain cool--can be offset by a temperature gradient and the resulting microstructure segregation effects in the product. For most systems with readily oxidizing metal matrices, a controlled atmosphere is required. The densification effect of conventional static hot pressing is more pronounced than the effect that can be achieved by cold pressing and subsequent sintering. Heating the cermet powder mixture increases its plasticity and produces larger areas of inter-particle contact. The surface shearing action that occurs during the process mechanically disrupts surface oxide films and generates clean bonding surfaces (Ref 13). Shape limitations are similar to those for static cold pressure. Prismatic singlelevel pieces that have no undercuts or re-entrant angles are preferred; however, shallow details in the punch faces are acceptable. Very large pieces are well suited for this process.

A typical graphite induction-heated vacuum furnace (Fig. 10) with graphite tooling is capable of double-action hot compacting or repressing of a 125 mm (5 in.) diam billet at pressures up to 90 Mg (100 tons) and temperatures up to 2300 °C (4200 °F) in a high-vacuum or controlled atmosphere. The fully pressed compact is ejected from the die while still hot to reduce cooling time, minimize sticking, and prolong mold life. Items produced in this type of furnace include WC-Co draw dies and friction elements, Al-B4C billets, stable oxide components, and some boride-base cermets.

Fig. 10 Production-scale 225 Mg (250 tons) vacuum hot press. Courtesy of Vacuum Industries, Inc.

Among the various cermet production processes, static hot pressing in such a furnace is the only reliable single-step method for producing a fully dense, high-quality, near-net shape compact from a cermet mixture. However, graphite die life is limited, and cermet compositions that do not react with graphite are preferred. In order to avoid a cermet-graphite reaction, ceramic molds can be used for hot pressing, although they are more fragile and costly than cermet molds. The complete hot press setup (including a vacuum system or atmosphere generator power system, hydraulic system, controls, and instrumentation) is expensive. In addition, static hot pressing is labor intensive because products are pressed one at a time. Therefore, hot isostatic pressing may be more appropriate for many sensitive high-temperature consolidation tasks involving small- and medium-size pieces. For large and very large pieces, vacuum hot pressing is often used because large equipment is readily available (Ref 14). For example, Fig. 10 shows a 225 Mg (250 ton) vacuum hot press with double-action bottom rams and a 1070 mm (42 in.) square platen; the press has a maximum operating temperature of 1315 °C (2400 °F). Hot isostatic pressing (HIP) has become increasingly popular as a means for producing carbide-base and other cermets of very high and uniform density. Internal flaws and micro- or macroporosity are virtually eliminated in the resultant product. Isostatic pressing is a batch process accomplished in water-cooled pressure vessels capable of withstanding internal pressures of up to 210 MPa (30 ksi). Heating up to a temperature of 1600 °C (3000 °F) is achieved with a high-frequency or resistance furnace mounted inside the pressure vessel. The pressure medium is an inert gas, usually argon. The pressure medium can also be helium (Ref 8), which at the pressure employed (100 to 150 MPa, or 15 to 20 ksi) has a density close to that of water.

Hot isostatic pressing was originally developed for use in gas pressure-assisted diffusion bonding processes such as the encapsulation of nuclear fuel elements (uranium oxide, for example) in a zircaloy sheath. This was soon followed by applications such as powder consolidation and densification of difficult-to-sinter substances and cermet composition (Ref 13). Hot isostatic pressing is most successfully applied in the cemented-carbide industry and in the manufacture of steel-bonded titanium carbide. Notwithstanding the ease with which these cermets sinter to high density, they often have slight localized porosity in the range up to 50 m; on occasion they have voids in the range from 0.25 to 2.5 mm (0.01 to 0.1 in.) caused by random or accidental contamination.

Hot isostatic pressing is an improvement over static hot pressing in that it eliminates the need for costly and highly perishable molds. However, before compacts are submitted to isostatic pressing, they need to have a sufficiently dense structure (at least at the skin) to inhibit gas penetration. Compacts with lower density and interconnecting pores require a gas-tight encapsulation of some sort before being treated. Three methods are in common use to accomplish this encapsulation (Ref 13). In the first method, compacts are formed and sintered to 95% or more of theoretical density, and the resultant continuous, dense surface structure acts as an impenetrable envelope to the high-pressure gas. Alternatively, the density is raised to a sufficiently high level that no interconnecting porosity remains within the compact. This method is used primarily for smalland medium-size pieces. The second method involves a steel can that is prepared in accordance with the desired form of the compact. The can is filled with a powder mixture that is densely packed by vibration or pressing. (A cold-compacted or hydrostatically pressed piece could be encapsulated in the steel can instead of the powder mixture.) After loading, the can is closed tightly by welding and then evacuated. This method is most commonly used for medium- and large-size pieces, in situations where the additional cost of a perishable can is economically bearable. The third method is the Ugine-Sejournet process, in which vitrified glass is used to encapsulate the compact. This method may be more economical than the steel can method. None of the methods for preventing the high-pressure gas from penetrating the compact are inexpensive. Fortunately, the HIP process is flexible enough to allow for the simultaneous hot isostatic pressing of a number of freestanding or encapsulated compacts. The process cost can be apportioned according to the volume occupied by each piece in the available furnace space. The cemented-carbide and P/M products that undergo HIP processing are of substantially higher quality than those produced by any other process. The higher quality is a result of the near theoretical densities produced by HIP processing. Pieces with near theoretical densities have high strength levels and reliable physical properties. Hot isostatic pressing is a capitalintensive batch process; when the encapsulation method is used, it is labor intensive as well. Hot extrusion of cermet billets is unique among the various cermet processes because it is essentially a solid-state process. All of the other previously discussed densification processes involve a liquid phase. During hot extrusion of powdered material, large hydrostatic compression forces occur. A unidirectional force component first compresses the powder material to full density and then forces the material through the die. Depending on the configuration of the front surface of the extrusion die, a large shear component can absorb as much as one-half of the total energy needed for extrusion. The total amount of one-step deformation (ratio of the cross section of the billet to the cross section of the die opening) is much larger than in any other cermet hot-working process.

Hot extrusion is an attractive forming and densification process for cermets. Unfortunately, its application in this broad field has serious limitations caused by the loading of the nonmetallic material, the choice of the metallic component, and the degree of interaction between the two substances. Precise limitations on the volumetric amount of nonmetallic substance have not been established, but when this substance exceeds about 18 to 25 vol%, the composite material exhibits hot shortness to such an extent that the problems of edge cracking, internal lamination, and distortion of the extruded product become intolerable. When the nonmetallic phase occurs in the form of very fine powders, the hot shortness problems are aggravated; they are also aggravated when the nonmetallic substance interacts with the metallic component at the extrusion temperature to form new phases or eutectoids. How to deal with these complex problems for each particular pairing of cermet components is beyond the scope of this review. A careful study of the constitution diagram of the binder metal and the hard phase before undertaking any serious work is recommended. Like metal alloy billets, cermet billets are much easier to extrude when aluminum or aluminum alloys are used for the metallic components. Billet densification and extrusion occur at lower temperatures and pressures. The expensive canning and decanning process can generally be avoided, as can the use of a protective atmosphere. Moreover, straightening and finishing of the extruded product often can be performed at room temperatures. Defects due to hot shortness in extruded aluminum-base products occur only at higher levels of loading with hard-phase substances. Compared with cermets containing aluminum or aluminum alloys, cermets with iron, nickel, cobalt, or alloys of these metals as the metallic phase are more technically demanding and more costly to extrude. For example, billet heating and hot

compacting require a controlled atmosphere, billet canning is practically unavoidable, and a glass process is required for lubrication and reduction of die wear. Postextrusion finishing, such as the removal of can material, straightening, finish rolling, and so on, also requires more costly processes. The extruded cermet product can be expected to be harder, stiffer, less malleable, and more brittle than the metal binder component. Because of the higher extrusion temperature, metallurgical interaction between the matrix metal or alloy and the hard-phase cermet component is far more likely to occur, particularly if the hard phase cermet component is a nonmetallic compound of the carbide or boride group. Oxide-base hard-phase components are less likely to interact with the metallic component. The undesirable interaction between the hard substance and the metallic component of the cermet is easier to control with the solid-state extrusion process than with any densification process involving a liquid phase. Also, in spite of the aforementioned limitations and problems, hot extrusion continues to attract the attention of product development engineers as a possible production method for certain cermets, particularly those with an aluminum base. It has the potential to be a relatively low labor cost, yet capital-intensive, mass-production process for high-technology rod or strip material. Combination Sintering-Compacting. Combination debinding, sintering, and pressure consolidation furnaces have been developed in an attempt to simplify the manufacturing process for cermets and similar products. As stated before, a debinding step is essential before sintering green products that contain admixed lubricants, organic binders, or plasticizers. These additives are needed in varying proportions for static cold pressing, warm extrusion, powder roll compacting, slip casting, and injection metal molding of cermets. After debinding and during liquid-phase sintering, the green compacts shrink to nearly complete density. When densification progresses to the point that pressurized gas can no longer penetrate into the compact, hot isostatic compacting occurs. Gas compacting the already-sintered dense cermet at high gas pressures and at a temperature near that of liquid-phase formation improves the product quality by eliminating all residual porosity, internal flaws, and defects.

Recent experience with WC-Co compacts has shown that using a lower isostatic pressure of only 2.7 MPa (390 psig) can produce compacts with strength and densification nearly equal to those of compacts produced by the high-pressure HIP process (Table 3). Based on these findings, multimode single-chamber pressure furnaces have been developed that are capable of operating in vacuum, with partial pressure, and with positive gas pressure up to 10 MPa (1500 psig); the furnaces operate at temperatures between 1450 and 2200 °C (2640 and 3390 °F) (Ref 15).

Table 3 Comparison of transverse rupture strength for various cemented carbides after hot isostatic pressing and pressure sintering Processing method 6% cobalt Vacuum sintering Vacuum sintering and HIP Pressure sintering 9% cobalt Vacuum sintering Vacuum sintering and HIP Pressure sintering 12% cobalt Vacuum sintering Vacuum sintering and HIP Pressure sintering

Density g/cm3 lb./in.3

Hardness, HRA

Transverse rupture strength Mpa ksi

14.87 14.90 14.89

0.537 0.538 0.538

91.6 91.9 92.0

2180 2645 2480

316 384 360

14.59 14.58 14.63

0.527 0.527 0.529

91.2 91.2 91.2

2170 2380 2843

315 345 412

14.09 14.07 14.11

0.509 0.508 0.510

89.9 90.0 90.8

2140 2515 2565

310 365 372

This new furnace concept of combining three operations in one cycle offers several advantages over separate sintering and hot isostatic pressing operations: •

Debinding, sintering, and densification under pressure take place in one cycle and in a single vessel.

• • • •

During the various stages of the process, the compacts do not come in contact with air. Controlling the cycle with an electronic microprocessor ensures automatic operation and a high degree of program reproducibility (Fig. 11). Transfer of parts from one process step to another is avoided, saving labor cost and process time. Combining processes saves energy.

Fig. 11 Schematic cycle diagram for low-pressure dewaxing and overpressure sintering. Source: Ref 16

Infiltration is a process that is similar to liquid-phase sintering, except that the solid phase is first formed into a porous

skeleton body, and the liquid-metal phase is introduced during sintering from the outside and allowed to penetrate the pore system. Excessive shrinkage associated with in situ liquid-phase sintering is avoided, and dimensional stability of the product is obtained, except for about 1% growth that is due to a thin surface film formed by the liquid metal. This technique is used for systems of two or more components that have widely differing melting temperatures. Aside from hot pressing, infiltration is the only powder processing method that can obtain essentially full density of a near-net shape. All of the other densification processes involve substantial shrinkage and thus destroy shape and dimensional accuracy. By machining, hydrostatic pressing, or powder injection molding the skeletal preforms prior to infiltration, complexities in part design, such as undercuts, re-entrant angles, and multiple levels, can be realized to an extent not possible in parts of comparable high density that are made by extrusion or hot pressing. The other unique feature of infiltration is that--under suitable conditions of low contact angles and limited solubilities between the high- and low-melting phases-- systems of completely intertwined continuous networks can be obtained. This is of considerable importance for making products that must combine high thermal or electrical conductivity with acceptable levels of strength and abrasion or erosion resistance. The procedure used for TiC cermets involves two steps (Ref 17). First, an approximately 60% dense carbide skeleton body of near-net shape is formed by mixing the TiC powder with a small percentage of nickel binder and wax, cold pressing the mixture at about 35 MPa (5000 psi) into a slab, vacuum sintering the slab at about 1300 °C (2370 °F), and then machining the contour (for example, a turbine blade). The second step consists of inserting the skeleton shape into a mold assembly that contains the metal in a ceramic tundish on top and provides for the gravity feeding of the liquid to the skeleton at the preferred contact faces (Ref 18).

The mold assembly is made of graphite, and its cavity is lined with a refractory ceramic in powder form that interfaces with the TiC skeleton. The ceramic liner is chosen so that it does not react with the titanium carbide up to infiltration temperature and also so that it shrinks at a controlled rate, permitting the formation of a uniform gap all around. The mold assembly is heated in a vacuum furnace to about 1400 to 1500 °C (2550 to 2730 °F), well above the melting temperature of infiltrating alloys, such as 80Ni-20Cr and 70Co-24Cr-6Mo. During infiltration, the liquid metal first fills the gap between the liner and the skeleton exterior by capillary forces and then penetrates the interior of the porous TiC part. After furnace cooling, the fully infiltrated product can be readily extracted by fragmenting the sintered ceramic liner without degrading the graphite mold assembly, which can be reused. Graded cermet parts can be produced by varying the density of the TiC skeleton through the use of a special die filling and multiple-step pressing. For example, a turbine blade can be made that has a high concentration of titanium carbide and, therefore, high strength at the center of the foil and in the transition to the root. The turbine blade also has a metal-rich jacket around the foil and especially at the mechanical shock-sensitive blade edges, as well as around the serrated root needed for blade attachment to the turbine disk. The infiltration process has been successfully applied to other cermet systems; it has been especially successful when used with interfacial reactions. An example of such an application is the production of complex ceramics that are reinforced by microscopic-size platelets of another compound and bonded by a third species (Ref 19). The preformed ceramic is a metalloidal carbide, the binder is a relatively high-melting reactive metal, and the platelets are the reaction product of the carbide and binder. The unique composite microstructure of such a platelet-reinforced ceramic is obtained by gravity infiltration of the molten metal into a porous preform or bed of the carbide. In the case of a B4C ceramic and a zirconium metal infiltrant, controlled oxidation at the contact faces produces a new phase, ZrB2, that precipitates abundantly in the form of platelets that reinforce the ceramic. A similar result is achieved with SiC preforms or fillers that are infiltrated with molten aluminum metal under oxidizing conditions; the reaction product in this case is Al2O3 (Ref 19). The container or mold used in the infiltration process is made from graphite, and it is shaped in accordance with the configurations of the desired product; the process is conducted in an argon atmosphere. The infiltration and reaction temperature depends on the melting point and liquidity of the metal. It can be as high as 2000 °C (3630 °F) in the system involving zirconium. The time to complete penetration and reaction is in the 1 to 2 h range, and the end product typically contains 5 to 15% residual binder metal. The platelet-reinforced infiltrated ceramic system of the type ZrB 2/ZrCx/Zr exhibits a good combination of high strength, high fracture toughness, and high thermal conductivity. This combination makes it an interesting candidate material for rocket engine components and wear parts (Ref 20). Other systems that have been successfully produced or are potentially workable by the infiltration process include TiB2 ceramics combined with nickel as second phase (Ref 21), TiC with steel (Ref 22), WC with cobalt (Ref 23), AlN with aluminum (Ref 24), and Al2O3 with aluminum (Ref 25). Infiltration processing of boron carbide and boride-reactive metal cermets has also been used successfully in the development of high-strength, hard, and lightweight products that offer an interesting combination of toughness with high thermal and electrical conductivity (Ref 26). The process involves the infiltration of molten reactive metals, particularly aluminum, into chemically treated boron carbide, or it can use metal-boride starting constituents, such as powders or low aspect ratio fibers, that have been consolidated into a porous ceramic precursor sponge. This process is an alternative to the infiltration of the molten aluminum into thermally modified precursor sponges. Conventional or colloidal chemistry is used in the chemical reaction-controlled casting and infiltration procedures. The potential also exists for consolidating the precursor sponges by injection molding in either a single-step or a two-step process. The key to the process lies in controlling the surface chemistry of the starting constituents. In the two-step process, the first step is the production of highly configured geometries that can be molded by using chemically pretreated binders. In the second step, the binder is volatilized from the precursor, leaving it as a skeleton ready for infiltration.

References cited in this section

1. ASTM Committee C-21, "Report of Task Group B on Cermets," American Society for Testing and Materials, 1955 2. J.R. Tinklepaugh and W.B. Crandall, Chap. 1 in Cermets, Reinhold, 1960

3. 4. 5. 6.

E.C. Van Schoick, Ed., Ceramic Glossary, The Ceramic Society, 1963 R. Kieffer and F. Benesovsky, Hartmetalle, Springer-Verlag, 1965, p 437-489 P. Ettmayer and W. Lengauer, The Story of Cermets, Powder Metall. Int., Vol 21 (No. 2), 1989, p 37-38 E. Rudy, Boundary Phase Stability and Critical Phenomena in Higher Order Solid Solution Systems, J. LessCommon Met., Vol 33, 1973, p 43-70 7. F.V. Lenel, Powder Metallurgy, Principles and Applications, Metal Powder Industry Federation, 1980 8. P. Popper, Isostatic Pressing, British Ceramic Research Association, Heyden & Sons Ltd., 1976 9. R. Kieffer and P. Schwarzkopf, Hartstoffe and Hartmetalle, Springer-Verlag, 1953 10. R.M. German, Molding Metal Injection, Powder Injection Molding, Metal Powder Industries Federation, 1989 11. L.F. Pease III, Present Status in PM Injection Molding (MIM): An Overview, Progress in Powder Metallurgy, Vol 43, Metal Powder Industries Federation, 1987 12. K.J.A. Brookes, World Directory and Handbook of Hardmetals, 2nd ed., Engineer's Digest Publications, 1979 13. E. Lardner, Metallurgical Applications of Isostatic Hot Pressing, Chap. 10 in High Pressure Technology, Marcel Dekker, 1977 14. Vacuum Hot Press Furnaces for Powder Compaction, Met. Powder Rep., Vol 37 (No. 11), 1982 15. S.W. Kennedy, "Development in Combination Debinder/Pressure Consolidation Furnace," Technical Note, Vacuum Industries Inc., 1989 16. R.E. Bauer, Sinter-HIP Furnaces Sintering and Compacting in a Combined Cycle, Modern Developments in Powder Metallurgy, Metal Powder Industries Federation, 1988 17. C.G. Goetzel, Infiltration Process, Cermets, Reinhold, 1960, p 73-81 18. H.W. Lavendel and C.G. Goetzel, Recent Advances in Infiltrated Titanium Carbides, High Temperature Materials, R.F. Heheman and G.M. Ault, Ed., John Wiley & Sons, 1959, p 140-154 19. W.B. Johnson, T.D. Claar, and G.H. Schiroky, Preparation and Processing of Platelet Reinforced Ceramics by the Directed Reaction of Zirconium with Boron Carbide, Ceram. Eng. Sci. Proc., Vol 10 (No. 7/8), 1989 20. T.D. Claar, W.B. Johnson, C.A. Anderson, and G.H. Schiroky, Microstructure and Properties of Platelet Reinforced Ceramics Formed by the Directed Reaction of Zirconium with Boron Carbide, Ceram. Eng. Sci. Proc., Vol 10 (No. 7/8), 1989 21. V.J. Tennery, C.B. Finch, C.S. Yust, and G.W. Clark, Structure-Property Correlations for TiB2-Based Ceramics Densified Using Active Liquid Metals, Proc. of the Int. Conf. on the Science of Hard Materials, Plenum, 1983 22. C.G. Goetzel and L.P. Skolnick, Some Properties of a Recently Developed Hard Metal Produced by Infiltration, Sintered High-Temperature and Corrosion-Resistant Materials, F. Benesovsky, Ed., Pergamon Press, 1956, p 92-98 23. R. Kieffer and F. Benesovsky, The Production and Properties of Novel Sintered Alloys (Infiltrated Alloys), Berg Hüttenmänn. Monatsh., Vol 94 (No. 8/9), 1949, p 284-294 24. D.K. Creber, S.D. Poste, M.K. Aghajanian, and T.D. Claar, AlN Composite Growth by Nitridation of Aluminum Alloys, Ceram. Eng. Sci. Proc., Vol 9 (No. 7/8), 1988, p 975 25. M.S. Newkirk, H.D. Lesher, D.R. White, C.R. Kennedy, A.W. Urquhart, and T.D. Claar, Preparation of Lanxide Ceramic Matrix Composites: Matrix Formation by the Directed Oxidation of Molten Metals, Ceram. Eng. Sci. Proc., Vol 8 (No. 7/8), 1987, p 879-882 26. D.C. Halverson, A.J. Pyzik, I.A. Aksay, and W.E. Snowden, Processing of Boron Carbide- Aluminum Composites, Advanced Ceramic Materials, Preprint UCRL-93862, Lawrence Livermore National Laboratory, 1986

Powder Metallurgy Cermets and Cemented Carbides

Cemented Carbides Cemented carbides belong to a class of hard, wear-resistant, refractory materials in which the hard carbide particles are bound together, or cemented, by a soft and ductile metal binder. These materials were first developed in Germany in the early 1920s in response to demands for a die material having sufficient wear resistance for drawing tungsten incandescent filament wires to replace the expensive diamond dies then in use. The first cemented carbide to be produced was tungsten carbide (WC) with a cobalt binder. Although the term cemented carbide is widely used in the United States, these materials are better known internationally as hard metals. Tungsten carbide was first synthesized by the French chemist Henri Moissan in the 1890s. There are two types of tungsten carbide: WC, which directly decomposes at 2800 °C (5070 °F), and W2C, which melts at 2750 °C (4980 °F). Early attempts to produce drawing dies from a eutectic alloy WC and W2C were unsuccessful, because the material had many flaws and fractured easily. The use of powder metallurgy techniques by Schroeter in 1923 paved the way for obtaining a fully consolidated product. Schroeter blended fine WC powders with a small amount of iron, nickel, or cobalt powders and pressed the powders into compacts, which were then sintered at approximately 1300 °C (2400 °F). Cobalt was soon found to be the best bonding material. Over the years, the basic WC-Co material has been modified to produce a variety of cemented carbides, which are used in a wide range of applications, including metal cutting, mining, construction, rock drilling, metal forming, structural components, and wear parts. Approximately 50% of all carbide production is used for metal cutting applications. This section discusses the manufacture and composition of cemented carbides and their microstructure, classifications, physical and mechanical properties, and applications. The current status of cemented carbides in nonmetal cutting applications will also be covered. Manufacture Cemented carbides are manufactured by a powder metallurgy process consisting of a sequence of steps in which each step must be carefully controlled to obtain a final product with the desired properties, microstructure, and performance. The steps include: • • • • • •

Processing of the ore and the preparation of the tungsten carbide powder Preparation of the other carbide powders Production of the grade powders Compacting or powder consolidation Sintering Postsinter forming

The sintered product can be directly used or can be ground, polished, and coated to suit a given application. Preparation of Tungsten Carbide Powder. There are two methods by which tungsten carbide powders are produced

from the tungsten-bearing ores. Traditionally, tungsten ore is chemically processed to ammonium paratungstate and tungsten oxides. These compounds are then hydrogen-reduced to tungsten metal powder. The fine tungsten powders are blended with carbon and heated in a hydrogen atmosphere between 1400 and 1500 °C (2500 and 2700 °F) to produce tungsten carbide particles with sizes varying from 0.5 to 30 m. Each particle is composed of numerous tungsten carbide crystals. Small amounts of vanadium, chromium, or tantalum are sometimes added to tungsten and carbon powders before carburization to produce very fine (1500 >1500 96 14 50 >1500 355

8 9 4 9 9 7 6 5 10 8

PRE, Pitting resistance equivalent (%Cr + 3.3% Mo + 16% N) in wt%; NSS1, time to first sign of corrosion, salt-spray test in 5% NaCl; NSS2, rust rating after 1500 h of testing where 10 = no corrosion and 0 = surface half covered with corrosion products.

(a) (b)

0.1%Cl-, pH 5, 30 °C, 5 mV/min. 0.1% Cl-, pH 5, 30 °C, 25 mV/8h.

(c) (d) (e)

18.3% Cr, 18.3% Ni, 5.6% Mo, 1.7% Cu, 1.3% Sn, 0.78% Si, 0.23% Mn, bal Fe. 20% Cr, 17.0% Ni, 5.0% Mo, 0.75% Si, >0.15% Mn, bal Fe. 16.3% Cr, 24.3% Ni, 7.7% Mo, 0.81% Si, 0.25% Si, bal Fe.

Mechanical properties of 316L stainless steel as a function of some of the more important processing parameters are presented in Fig. 28, 29, and 30 (Ref 23). As these figures reveal, nitrogen-containing atmospheres result in the absorption of considerable amounts of nitrogen, which increases strength, decreases ductility, and, as is seen in the following paragraphs, influences P/M alloy corrosion resistance. Mechanical properties for a few P/M stainless steel alloys sintered in dissociated ammonia are presented in Table 13 (Ref 1).

Fig. 28 Effect of sintering temperature on tensile and yield strengths and apparent hardness of type 316L stainless steel. Parts (density: 6.85 g/cm3) were sintered for 30 min in various atmospheres. Source: Ref 23

Fig. 29 Effect of sintering temperature on elongation and dimensional change during sintering of type 316L stainless steel. Parts (density: 6.85 g/cm3) were sintered for 30 min in various atmospheres. Source: Ref 23

Fig. 30 Effect of sintering time on tensile and yield strengths of type 316L stainless steel. Parts were pressed to 6.85 g/cm3 and sintered at various temperatures in dissociated NH3. Source: Ref 23

Ferritic and Martensitic Stainless Steels. The 400 series alloys are typically less heavily alloyed than the austenitic grades and, as a result, they usually exhibit inferior corrosion resistance. In addition to the lower pitting and crevice-corrosion resistance resulting from lower concentrations of passivity enhancing elements, ferritic stainless steels are also more susceptible to sensitization and intergranular corrosion. Ferritic stainless steels exhibit a greater affinity for sensitization than austenitic stainless steels because the solubility limit of carbon in the austenite phase is greater than in the ferrite phase. Hence, the precipitation of carbides is more prevalent in ferritic microstructures (Ref 18).

References cited in this section

1. E. Klar and P.K. Samal, Powder Metals, Corrosion Tests and Standards Manual, R. Baboian, Ed., ASTM, 1995, p 551-557 14. E. Maahn, S.K. Jensen, R.M. Larsen, and T. Mathiesen, Factors Affecting the Corrosion Resistance of Sintered Stainless Steel, Advances in Powder Metallurgy and Particulate Materials, Vol 7, C. Lall, Ed., Plenum Press, 1994, p 253-271 18. M. Baran, "An Evaluation of a P/M Ferritic Stainless Steel Automotive Exhaust Flange," Master's dissertation, The Pennsylvania State University, 1997 20. A.J. Sedriks, Effects of Alloy Composition and Microstructure on the Passivity of Stainless Steels, Corrosion, Vol 42 (No. 7), 1986, p 376-388 23. E. Klar, Corrosion of Powder Metallurgy Materials, Corrosion, Vol 13, Metals Handbook, ASM International, p 824-845 24. A.J. Sedriks, Corrosion of Stainless Steels, 2nd ed., John Wiley & Sons, 1996 Corrosion-Resistant Powder Metallurgy Alloys Barbara Shaw, Penn State University

Influence of Processing Parameters on the Corrosion Resistance of P/M Stainless Steels Influence of Iron or Steel Contamination on Corrosion Resistance. It should come as no surprise that the

corrosion resistance of P/M stainless steels is seriously degraded if iron or steel particles become incorporated into the alloy. The potential difference between iron or steel and stainless steel is typically on the order of several hundred millivolts and easily results in the establishment of galvanic or dissimilar metal corrosion within the contaminated component. There are numerous possible contamination sources: contamination of the initial powder at the supplier; inadvertent introduction during mixing/blending, feeding, or pressing operations; incorporation of airborne particles during processing or storage; and inadequate furnace cleaning. Cleanliness is of the utmost importance and separate or dedicated equipment is often used for the production of stainless components. Figure 31 shows an example of the appearance of iron-contaminated sintered 316L stainless steel after exposure to a 5% NaCl solution. Rusting became apparent within minutes of exposure to the chloridecontaining solution.

Fig. 31 Small circles of rust around iron particles embedded in the surface of sintered type 316L stainless steel after testing in 5% aqueous NaCl. 35×

Corrosion resulting from iron or steel contamination is perhaps the worst and, ironically, most avoidable corrosion problem encountered with P/M stainless steels. A concentrated copper sulfate solution can be used to easily detect iron, or an iron alloy, present in a stainless steel powder or on the surface of a sintered part. Dissolved copper from a copper sulfate solution readily plates out on the anodic (lower potential) iron sites, making them easy to see at low magnification. Influence of Lubricant and Carbon. As Fig. 32 suggests (Ref 24), carbide formation (especially, chromium carbide formation) with concomitant sensitization is an issue when the carbon content of an austenitic stainless steel exceeds 0.03%. In order to resist intergranular corrosion, water-atomized stainless steel powders have carbon contents greater than 0.03%. Unfortunately, other sources of carbon are associated with processing sintered stainless steels. These sources include the carbon resulting from inadequate organic lubricant dissipation and carbon contamination (soot) from insufficiently cleaned furnaces. Microstructures from two sintered 316L stainless steels, one below and one above the critical 0.03% concentration, are shown in Fig. 33. Thin, undecorated grain boundaries are observed in the low-carbon stainless steel, whereas, heavily decorated grain boundaries are observed for the high-carbon stainless steel. In insufficiently cleaned furnaces, loose, adherent soot can fall onto the surface of stainless steel parts or moisture from the sintering atmosphere can react with soot and form carbon monoxide and carburize the stainless steel. If care is not taken to limit the uptake of carbon, sensitization of the microstructure can occur and severely compromise the overall corrosion resistance of the alloy. The influence of carbon content on the pitting potential for a number of different sintered stainless steels is shown in Fig. 34 (Ref 14). Sensitization can be minimized with proper lubricant dissipation, a clean furnace, and low initial carbon concentration in the powder. It should be noted that when optimal sintering conditions are used, differences in corrosion resistance have not been noted as a function of lubricant type, as Table 15 reveals (Ref 25).

Table 15 Effect of binder/lubricant on the corrosion resistance of sintered 316L stainless steel in deaerated 1000 ppm Cl buffered with acetate at 30 °C (pH=5) The dew point of the gas atmospheres in the furnace was approximately -30 °C. Binder Acraw Metlb Acraw Metlb Acraw Metlb Acraw Metlb Acraw Metlb Acraw Metlb

Sintering, °C/min 1120/20

Atmosphere DA

1160/45

H2

1250/30

H2

1120/30

Vacuum

1200/50

Vacuum

1295/30

Vacuum

Ipass, A/cm2 45 18 3 2 16 16 4 4 8 5 2 2

Epit, mV SCE 65 230 455 400 230 230 390 380 475 425 405 500

Salt spray, h 20 24 30 620 24 24 500 500 560 560 240 330

Acraw, Acrawax; DA, dissociated ammonia; Metlb, Metallub. Source: Ref 25

Fig. 32 Solid solubility of carbon in an austenitic stainless steel. Source: Ref 24

Fig. 33 Microstructures of type 316L stainless steel sintered in hydrogen at 1150 °C (2100 °F). (a) Low carbon content. (b) Excessive carbon content. Both 400×

Fig. 34 Influence of carbon concentration on the pitting potential for a number of different materials. Source: Ref 14

Carbon contents in excess of 0.03% can be of benefit when stainless steels are vacuum sintered. In vacuum sintering, the excessive carbon is used for the reduction of some oxides on the water-atomized stainless steel, improving strength, ductility, and corrosion resistance. Microstructures of vacuum-sintered 430L stainless steel with and without the addition of 0.2% graphite are shown in Fig. 35. The graphite-containing stainless steel exhibited clean grain boundaries, while the alloy without graphite had grain boundaries containing carbides.

Fig. 35 Cross sections of vacuum-sintered (30 min at 1330 °C, or 2430 °F) type 430L stainless steel. (a) No oxides are present in grain boundaries after addition of 0.2% graphite. (b) Small, gray, rounded oxide particles in

grain boundaries (no graphite added)

Influence of Nitrogen and Sintering Atmosphere. Dissociated ammonia is a commonly used sintering atmosphere

because it costs less than other sintering atmospheres. However, sintering in dissociated ammonia usually leads to the pickup of nitrogen by the stainless steel--a factor that can enhance susceptibility to corrosion in a manner analogous to that observed with chromium carbide formation. For wrought stainless steels, enhanced passivity is observed with increased nitrogen content. However, if chromium nitrides precipitate, the sensitized stainless steel is susceptible to intergranular corrosion. Equilibrium solubilities for nitrogen in austenitic stainless steels with different chromium contents are presented in Fig. 36 (Ref 23). The concentration of dissolved nitrogen depends on the amount of nitrogen in the atmosphere, the sintering temperature, and the cooling rate of the sintered alloy. An example of the influence of nitrogen concentration in the sintering atmosphere on the corrosion resistance of 316L stainless steel is shown in Table 16 (Ref 14). These results reveal that up to the point where supersaturation associated with sensitization occurs, no significant difference in corrosion behavior was noted. In addition, these results show that the EPR test is very sensitive to the identification of nitride formation. Reference 26 reported that chromium nitrides were not present in 316L when its nitrogen content was lower than 0.4 wt%. This point is supported by the weight loss data after exposure to 10% HNO3 (as a function of absorbed nitrogen content) for sintered 316L stainless steel shown in Fig. 37 (Ref 3). The data presented in this figure were obtained using several sintering atmospheres. An example of the influence of temperature on dissolved nitrogen concentration is shown in Fig. 38 (Ref 13). The data in this figure were obtained by continuously measuring weight gain during heating in dry nitrogen (nitrogen containing 0.01% water and 1% water). In the 700 to 1000 °C temperature range a large absorption of nitrogen occurs. A maximum of 9 mg per gram of stainless steel was absorbed--24 mg/g would be required to convert all of the chromium in the alloy to Cr2N (Ref 13). An example of the influence of cooling rate, and hence the dissolved nitrogen concentration, on the corrosion rate of 316L stainless steel is shown in Table 17 (Ref 13). While the rate of heating had no influence on corrosion of the alloy, cooling rates in excess of 100 °C/s inhibited or eliminated corrosion. Chromium nitride sensitization with concomitant loss of corrosion resistance is not limited to 316L stainless steel. Other stainless steels (both 300 and 400 series alloys) are subject to loss of corrosion resistance when nitrogen-containing sintering atmospheres are used. Table 2 showed weight-loss data for sintered 304L and 316L stainless steels as a function of sintering atmosphere and revealed that nitride formation lowered corrosion resistance (weight loss after 5% NaCl exposure).

Table 16 Corrosion properties of 316L stainless steel sintered in hydrogen or nitrogen/hydrogen mixtures Sintering, °C/min 1120/30

1250/120

(a) (b) (c)

Nitrogen content in atm, % 0 5 10 25 0 5 10 25

Nitrogen, ppm

ipeak(a), A/cm2

Ipass(a), A/cm2

Epit(a), mV SCE

Estp(b), mV SCE

NSS1, h

NSS2

Ir/Ia(c) × 1000

360 1710 2100 5670 20 1350 1850 7180

10 11 10 330 8 11 10 400

11 11 11 34 10 9 12 160

375 475 525 325 600 550 600 -25

100 350 350 225 375 300 375 -25

>1500 >1500 >1500 24 990 864 240 24

8 9 9 4 8 6 3 2

0.0 0.7 7.2 28.9 0.0 0.0 0.1 390

0.1% Cl-, pH 5, 30 °C, 5 mV/min. 0.1% Cl-, pH 5, 30 °C, 25 mV/8h. EPR test in 0.5 M H2SO4 + 0.01 M KSCN, 30 °C.

Table 17 Influence of heating and cooling rates on the corrosion resistance of 316L stainless steel specimens sintered at 1150 °C in dissociated ammonia Heating rate, °C/min

Cooling rate, °C/min

22 67 22 5 22 200

22 67 100 200 200 200

Weight increase, mg/g 3.3 2.5 3 2.9 3 2.3

Result of corrosion test in 5% NaCl solution Corroded in 1 day Corroded in 1 day Slight attack in 4 days No attack in 5 days(a) No attack in 5 days No attack in 5 days

Source: Ref 13

(a)

Test continued to 12 days with no attack.

Fig. 36 Solubility of nitrogen in austenitic stainless steel in equilibrium with gaseous or Cr2N. Source: Ref 23

Fig. 37 Effect of the absorbed nitrogen content during sintering on the corrosion resistance of 316L stainless steel sintered at 1160 °C in several atmospheres; corrosion rate is given in terms of weight loss resulting from immersion in 10% HNO3. Source: Ref 3

Fig. 38 Increase in weight of specimens heated in nitrogen containing various amounts of water vapor. Source: Ref 13

Influence of Oxygen and Water Vapor/Dew Point on Corrosion Resistance. The influence of oxygen on the

corrosion resistance of sintered stainless steels can, perhaps, best be understood by visualizing the structure of the as-sintered material. The as-received powders contain oxygen, much of which resides on the surface of the powder. Reduction of these oxides in industrial furnaces is not always complete, and the grain boundary oxides within the as-sintered structure provide paths for easier corrosion of the alloy. The role these grain-boundary oxides play in the corrosion of sintered metals is likely similar to the role that such oxides play in the degradation of thermal sprayed coatings (Ref 27, 28). Upon cooling, high oxygen affinity elements oxidize when they reach the metal-oxide equilibrium temperature, and the water content (dew point) of the atmosphere determines the stability of the oxides according to Fig. 39 (Ref 23). This figure reveals that the oxides are more easily formed at lower temperatures. The negative effect of high dew point on the corrosion resistance of sintered 316L stainless steel is further shown by the data in Table 18.

Table 18 Corrosion properties of 316L stainless steel sintered in hydrogen with a dew point of -35 or -70 °C at different combinations of time and temperature Sintering, °C/min 1120/30 1250/30 1120/120 1250/120

ipeak(a), A/cm2 at dew point: -35 °C -70 °C 150 10 105 7 120 10 83 4

ipass(a), A/cm2, at dew point: -35 °C -70 °C 29 11 20 12 25 10 19 9

Epit(a), mV SCE at dew point: -35 °C -70 °C 250 375 325 325 325 375 325 500

NSS1, at dew point

NSS2, at dewpoint

-35 °C 36 288 48 24

-35 °C 5 4 5 1

-70 °C >1500 1260 1272 96

-70 °C 9 8 7 7

0.1% Cl-, pH 5, 30 °C, 5 mV/min.

(a)

Fig. 39 Redox curves for chromium and silicon alone and in solution. Source: Ref 23

A number of investigators have observed decreased corrosion resistance with increasing oxygen content (Ref 14, 18, 25, 29). Figure 40 reveals that the pitting potential of P/M 316L stainless steel is found to decrease with increasing oxygen content (Ref 25). Immersion data in a 5% NaCl solution shows an identical trend, as seen in Fig. 41 (Ref 23). High dew points (greater than -34 °C) resulted in high oxygen concentrations within sintered alloys leading to a reduction in mechanical properties and corrosion resistance. This is not surprising in light of the type of microstructure that is attained in high dew point sintering atmospheres and illustrated in Fig. 42 (Ref 22). The P/M 316L stainless steel specimen shown in Fig. 42 exhibited a lack of interparticle bonding, resulting from the high grain boundary oxides, that lead to poor mechanical and corrosion properties. Immersion data for sintered 316L stainless steel in a 5% NaCl solution, as a function of water vapor content of the hydrogen sintering atmosphere, is shown in Table 19 (Ref 30). When the water vapor content was 45 ppm or lower, no corrosion was noted after eight days of exposure.

Table 19 Effect of water-vapor content on the corrosion resistance of stainless steel specimens sintered at 1150 °C in hydrogen Water-vapor content, ppm 30 45 90 110 150

Result of corrosion test in 5% NaCl solution No attack in 8 days No attack in 8 days(a) Rusted after 3 days Stained in 3 h Rusted in 1

h

Source: Ref 13

(a)

Test continued to 14 days with no attack.

Fig. 40 Influence of the oxygen content on the pitting potential for a number of different materials of sintered AISI 316L. Source: Ref 25

Fig. 41 Effect of oxygen content on corrosion resistance of sintered type 316L and tin-modified type 316L (sintered density: 6.65 g/cm3; cooling rate: 75 °C/min, or 135 °F/min). Parenthetical values are sintering temperature (°C), dew point (°C), and nitrogen content (ppm), respectively. Time indicates when 50% of specimens showed first sign of corrosion in 5% aqueous NaCl. Source: Ref 23, 30

Fig. 42 Microstructure of type 316L stainless steel sintered in a high-dew-point atmosphere. Oxygen content: 5100 ppm; sintered density: 7.5 g/cm3. Etched with Marble's reagent. 200×

Influence of Sintering Temperature, Sintering Time, and Cooling Rate on Corrosion Resistance. Improved

electrochemical corrosion resistance has been noted for sintered 316L stainless steel with increased sintering time, as Table 20 reveals (Ref 25). These improvements in corrosion resistance were attributed to the reduced nitrogen, oxygen, and carbon levels observed for the specimens after the longer sintering time. Carbon, oxygen, and CO, H2O, and CH4 concentrations in a hydrogen sintering gas as a function of time at temperature are presented in Fig. 43. Earlier, the significant influence of cooling rate on corrosion resistance is shown in Table 17. When there is sufficient water vapor to cause corrosion with slow cooling, it appears that a fast cooling rate (200 °C/min) retards corrosion, as shown in Table 21 (Ref 13).

Table 20 Effect of sintering temperature and time on the corrosion resistance of sintered 316L stainless steel The materials were sintered in hydrogen with a dew point of -70 °C. Sintering, °C/min 1120/30 1250/30 1120/240 1250/240

Open pores, % 17.4 16.3 15.8 13.8

Nitrogen, %

Oxygen, %

Carbon, %

Ipass, A/cm2

Epit, mV SCE

380 110 70 20

2230 1980 1640 450

250 130 130 70

3.4 3.1 3.0 1.8

383 357 508 561

Salt spray, h >1500 >1500 >1500 260

Source: Ref 25

Table 21 Effect of cooling rate on the corrosion resistance of 316L stainless steel specimens sintered at 1150 °C in a hydrogen atmosphere containing 100 ppm water vapor Heating rate, °C/min 5 200 200

Source: Ref 13

Cooling rate, °C/min 22 67 200

Result of corrosion test in 5% NaCl solution Corroded in 2 days Attack started after 1 day; severe attack after 8 days No attack after 3 days; slight staining after 5 days

Fig. 43 Gas composition and progress of reactions for a sintering experiment performed with pure hydrogen. Source: Ref 14

Influence of Porosity/Alloy Density. Sintered stainless steels are used in low-density forms (e.g., in filters) and in a

wide variety of forms requiring higher-density alloys. The literature on P/M stainless steels in acid solutions reveals that corrosion resistance improves with increasing density, as Fig. 44 shows (Ref 31). In saline solutions the situation is not as clear--some researchers have reported that increasing density is beneficial, while others have reported it to be detrimental. These discrepancies are believed to be a result of differences, from study to study, in pore morphology and alloy density. This point is illustrated in Fig. 45, which shows the corrosion resistance of sintered 316L stainless steels as a function of density (Ref 1, 32). At low sintered densities, the network of pores, including boundary oxides and particle boundaries, is rather open and discourages the formation of the occluded cell environment associated with crevice-corrosion initiation in stainless steels. At relatively high sintered densities, this network is tighter and encourages both the establishment of an aggressive environment within the crevice and a high potential drop down the crevice. At very high sintered densities, crevice-corrosion susceptibility decreases as the porous network is closed off with increasing alloy density. Data showing the percentage of open pores as a function of sintered density and the resulting time to first rust during salt spray exposure are presented in Table 20 (Ref 25).

Fig. 44 Relationship between sintered density and weight loss of three austenitic stainless steels in 40% HNO3 solution. Source: Ref 31

Fig. 45 Effect of sintered density upon corrosion resistance of sintered 316 type alloys. Source: Ref 1, 32

In the corrosion literature for wrought alloys, it is well recognized that the aspect ratio of a crevice (the ratio of its width to its length) is a critical parameter in the establishment of crevice corrosion. Figure 46 illustrates that for narrow crevice gaps, crevice corrosion initiates at shallow depths, whereas, for wider crevice gaps, attack initiates deeper within the crevice (Ref 24). As a result, narrow and/or long crevices are likely initiation sites for crevice corrosion. In sintered stainless steels the inherent porous nature of the material provides a narrow tortuous electrolyte path that both encourages crevice-corrosion initiation and sustains its propagation.

Fig. 46 Effect of crevice gap and depth on the initiation of crevice corrosion in various stainless steels and alloy 625. The gaps and depths below and to the right of the curve for each material define crevice geometries where initiation of crevice corrosion is predicted by the mathematical model of T.S. Lee and R.M. Kain, NACE Corrosion 83 Conference proceeding paper 69, 1983. Source: Ref 24

Several changes occur in the occluded cell environment of a crevice during the initiation stages of crevice corrosion: oxygen depletion within the crevice that establishes a separation of anodic and cathodic sites (where the cathode is largely outside the crevice and the anode is inside the crevice), a lowering of the pH of the solution within the crevice by hydrolysis reactions, and migration of chloride into the crevice to maintain charge neutrality. For some of the less heavily alloyed stainless steels, the reduction in Ep resulting from the increased chloride concentration is enough to initiate crevice corrosion, as the schematic polarization curve and mixed potential analysis in Fig. 47 reveal (Ref 33). The crossover point in the mixed potential analysis indicates dissolution of the metal within the crevice via a pitting type of attack. Changes in the anodic polarization behavior for the stainless steel within the crevice, resulting from acidification and increase in chloride ion concentration of the crevice solution, are illustrated in Fig. 48 (Ref 33). As the aggressive nature of the crevice solution increases, icrit, Epp, and ip increase, while Ep decreases. A mixed potential analysis of crevice-corrosion initiation (using the cathodic polarization behavior for the stainless steel in the environment outside of the crevice and the anodic polarization behavior for the stainless steel in the environment inside the crevice) is depicted in Fig. 49 (Ref 33). This analysis reveals that when the potential drop associated with the tortuous electrolyte path of a crevice exceeds a certain value, IR* in this illustration, crevice corrosion is initiated by active dissolution of the metal in the crevice. While the mechanisms just described were proposed for wrought materials containing intentional or unintentional crevice formers, such as O-rings or gaskets, they are equally applicable to P/M materials containing inherent porosity. In fact, the extremely tortuous path provided by the pores, oxides, and particle boundaries in a sintered stainless steel provide what could be viewed as the ultimate geometry for establishing and, perhaps, even studying crevice corrosion. Clearly, by gaining a better understanding

of the influence that processing parameters play in establishing pore morphologies susceptible to crevice corrosion, it will be possible to alter pore morphology to discourage crevice-corrosion initiation.

Fig. 47 Crevice-corrosion initiation resulting from lowering of Ep with increasing chloride concentration. Source: Ref 33

Fig. 48 Changes in anodic polarization behavior that occur as the environment in the crevice becomes increasingly aggressive. Source: Ref 33

Fig. 49 Crevice-corrosion initiation by an IR-induced mechanism. When the IR drop down the crevice exceeds IR*, crevice corrosion initiates. Source: Ref 33

Another means of altering the susceptibility of sintered stainless steels to crevice attack is to alter alloy composition. By altering alloy composition the hydrolysis reactions responsible for lowering the pH within the crevice can be influenced, and these changes can be used to discourage crevice-corrosion initiation, as Fig. 50 reveals (Ref 32). The enhanced crevicecorrosion resistance of the 317L and SS100 alloys is attributed to their higher molybdenum concentrations. It should be noted that the addition of molybdenum does not make these alloys immune to crevice corrosion because it is possible to initiate crevice corrosion in neutral pH crevice environments without the presence of chloride when the aspect ratio of the crevice is severe enough; it simply makes initiation more difficult.

Fig. 50 Effect of density on corrosion resistance of sintered austenitic stainless steel. Source: Ref 32

Approaches Used to Improve the Corrosion Resistance of Sintered Stainless Steels. A variety of means have

been employed to improve the corrosion resistance of sintered stainless steels--some of which simply alter the number of open pores, others are aimed at both reducing the number of open pores and enhancing passivity of the alloy. These approaches include the following finishing processes: tumbling, grinding and shot blasting, passivating treatments, liquid phase sintering, double pressing and sintering, and the addition of alloying elements, such as copper, tin, and noble metals (Ref 34). One group of researchers evaluated the influence of several finishing processes on the anodic polarization behavior of P/M 316L stainless steel in a 0.1 NNaCl/0.4 N NaClO4 solution (ASTM B 627). As shown in Fig. 51, the investigation revealed the following results: tumbling likely smears the pores and is ineffective at improving corrosion resistance; coining/sizing introduces residual stresses in the surface of the alloy that may increase corrosion; grinding, turning, and shot blasting can seal surface pores and improve corrosion resistance; and thermal and chemical passivation processes can alter the thickness and/or composition of the passive film, thus enhancing corrosion resistance (Ref 3). In another investigation, thermal passivation in the temperature range of 325 to 500 °C for 30 min also showed promise for enhancing passivity, and thus corrosion resistance, of sintered 316L stainless steel exposed to a 1 N H2SO4 solution, as Fig. 52 reveals (Ref 35).

Fig. 51 Anodic potentiodynamic polarization curves for 316L stainless in 0.1 N NaCl/0.4 N NaClO4 as a function of surface finishing. Source: Ref 3

Fig. 52 Rest (open-circuit) potential measurements for sintered 316L thermally prepassivated at temperatures between 325 and 500 °C. Source: Ref 35

Closing of porosity through operations such as double pressing and double sintering (DPDS), while expensive, has been found to significantly reduce the amount of open porosity and yield anodic polarization behavior in aggressive acid solutions similar to that of wrought materials. Figure 53 shows anodic polarization data for wrought and hot pressed and hot sintered 316 stainless steel in 1 N H2SO4 (Ref 29). When a DPDS operation is used, a P/M 316 stainless steel alloy of nominal composition exhibits a degree of passivity almost identical to that of wrought 316 (Ref 36). Another, more thorough, means for reducing or eliminating open porosity in P/M materials is liquid phase sintering. Sintering additives such as boron, NiB, BN, and CrB have been found to be effective in producing dense microstructures with enhanced salt-spray corrosion resistance, as Table 22 reveals (Ref 14). Unfortunately, sensitized microstructures, as evidenced by the high Ir/Ia ratio listed in Table 22 are a by-product of liquid phase sintering (Ref 14). Careful development of liquid phase sintering additives for stainless steels will be needed in order to produce dense alloys that are not susceptible to intergranular corrosion. Injection molding also shows promise for producing dense P/M alloys with enhanced corrosion resistance. Figure 54 compares weightloss values for injection-molded and wrought 14-4PH stainless steel as a function of exposure time in either full strength chlorine bleach or a 10% FeCl3 solution (Ref 37). The injection-molded P/M compacts had densities of 96 to 97% of the wrought density and exhibited weight losses comparable to those of the wrought 17-4PH stainless steel.

Table 22 Corrosion properties of liquid phase sintered 316L steels with the addition of boron-base sintering additives All stainless steels were sintered at 1250 °C for 60 to 120 min in pure hydrogen. Additive None 0.2% B (-38 m) 1% BN (-63 m) 1% NiB (-38 m) 1% CrB (-38 m)

Density, g/cm3 6.86 7.83 7.61 7.67 7.64

Open pores, % 8.2 0.1 0.2 0.1 0.1

Estp(a) mV SCE 150 500 400 525 550

NSS1, h 96 >1500 762 >1500 >1500

NSS2 7 10 9 10 10

Ir/Ia ×103 0.0 4.4 2.5 3.8 2.5

Source: Ref 14

(a)

0.1% Cl-, pH 5, 30 °C, 25 mV/8 h.

Fig. 53 Potentiodynamic polarization curves of samples tested in H2SO4, scanning rate 2 mV/S (1 N H2SO4). Source: Ref 29

Fig. 54 Comparison of the corrosion weight loss versus exposure time in (a) ferric chloride and (b) chlorine bleach tests at room temperature. P-W, polymer wax binder, 69% paraffin wax, 20% polypropylene, 10% carnuba wax, 1% stearic acid; SPS, 73% acetanilide, 18% polystyrene, and 9% stearic acid. Source: Ref 37

A number of alloying additions or infiltrants have been explored as means for enhancing passivity in P/M stainless steels. Among the most popular are copper and tin. Infiltration of P/M 316L and 304L alloys with either copper or bronze have been observed to increase corrosion resistance in boiling and room-temperature acid solutions (see Fig. 55). As this figure reveals, similar results were noted for the bronze and copper in the room-temperature acid, while the bronze infiltrant gave the best results in the boiling acid. Enhanced corrosion resistance was attributed to the elimination of open-connected porosity and the effect of the alloying. It should be noted that in both cases the infiltrants diffused either partially (bronze at all infiltration percentages and copper at the 6% infiltration level) or totally (copper at 4 and 6% infiltration levels) into the matrix. Table 1 illustrates the beneficial influence of copper alloying additions on the corrosion resistance of 316L stainless steel. Additional data supporting the beneficial influence of copper alloying additions on hot-pressed-and-sintered 316 stainless steel are provided in Fig. 38 (Ref 35). The improved passivity of the copper-containing alloys is attributed to its depolarizing influence on the oxygen reduction reaction. Depolarization of the cathodic, oxygen-reduction reaction enables the alloys to passivate more easily. Figure 56 shows the benefits of tin additions to the polarization behavior of 316 stainless steel (Ref 39). The addition of tin decreases the passive and critical current densities of the hot-pressed specimens by two orders of magnitude. Improvements in corrosion resistance, such as the ones just described, led to the development of copper- and tin-modified stainless steel powder chemistries, which are currently being marketed under the trademarks of Ultra 303L, Ultra 304, and

Ultra 316. The enhanced corrosion resistance of these alloys, when compared to those of nonmodified 303L, 304L, and 316L, is evident in Fig. 55 and 56. When both copper and tin are prealloyed into the powder, the alloys are marketed with an "LSC" designation. The improved corrosion resistance for 316LSC in comparison to that of nonmodified 316L is presented in Fig. 57 (Ref 23).

Fig. 55 Potentiodynamic polarization curves of as-received samples tested in 1 N H2SO4, scanning rate 1 mV/s. (a) A, wrought plate stainless steel of type 316; B, hot-pressed sintered stainless steel; C, hot-pressed-and-sintered stainless steel containing 0.25% Cu. (b) A Hot-pressed-and-sintered stainless steel containing 1 wt% Cu; B, hotpressed-and-sintered stainless steel containing 3 wt%; C, hot-pressed-and-sintered stainless steel containing 5 wt% Cu. Source: Ref 38

Fig. 56 Potentiodynamic polarization curves of samples tested in 5 N H2SO4, scanning rate 2 mV/s. (a) A, Wrought plate stainless steel of type 316; B, 2 wt% Sn; C, HPHS 316. (b) Hot-pressed-and-sintered sample containing A, 3 wt% Sn; B, 2 wt% Sn; C, 0.5 wt% Sn. Source: Ref 39

Fig. 57 Typical corrosion behavior of regular and copper-tin modified (type 316LSC) sintered type 316L stainless sintered in dissociated NH3 under various conditions of cooling and contamination. B rating indicates that 150 to 200 °C/min (>270 to 360 °F/min) 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. Use low oxygen content powder, preferably 200 °C/min (360 °F/min) For nitrogen-containing atmospheres, use dew point of -37 to -45 °C (-35 to -50 °F) in cooling zone. For sintering in H2, ensure that water vapor content of atmosphere is below 50 ppm.

Use lower density to increase pore size and circulation of

sintered part

crevice corrosion

corrodent. In acidic environments, corrosion resistance improves with increasing density due to a decrease of specific surface area.

Source: Ref 23

References cited in this section

1. E. Klar and P.K. Samal, Powder Metals, Corrosion Tests and Standards Manual, R. Baboian, Ed., ASTM, 1995, p 551-557 3. C. Molins, J.A. Bas, J. Planas, and S.A. Ames, P/M Stainless Steel: Types and Their Characteristics and Applications, Advances in Powder Metallurgy and Particulate Materials, Vol 5, J.M. Capus, Ed., Plenum Press, 1992, p 345-357 13. R.L. Sands, G.F. Bidmead, and D.A. Oliver, The Corrosion Resistance of Sintered Austenitic Stainless Steel, Modern Developments in Powder Metallurgy, Vol 2, H.H. Hausner, Ed., Plenum Press, 1966, p 73-83 14. E. Maahn, S.K. Jensen, R.M. Larsen, and T. Mathiesen, Factors Affecting the Corrosion Resistance of Sintered Stainless Steel, Advances in Powder Metallurgy and Particulate Materials, Vol 7, C. Lall, Ed., Plenum Press, 1994, p 253-271 18. M. Baran, "An Evaluation of a P/M Ferritic Stainless Steel Automotive Exhaust Flange," Master's dissertation, The Pennsylvania State University, 1997 22. M.A. Streicher, Corrosion Tests and Standards Manual, R. Baboian, Ed., ASTM, 1995, p 197-217 23. E. Klar, Corrosion of Powder Metallurgy Materials, Corrosion, Vol 13, Metals Handbook, ASM International, p 824-845 24. A.J. Sedriks, Corrosion of Stainless Steels, 2nd ed., John Wiley & Sons, 1996 25. E. Maahn and T. Mathiesen, Corrosion Properties of Sintered Stainless Steel, UK Corrosion Proc. (Manchester, UK), Vol 2, Institute of Corrosion, 1991, p 1-15 26. K. Frisk, A. Johansson, and C. Lindberg, Nitrogen Pick Up During Sintering Stainless Steel, Advances in Powder Metallurgy and Particulate Materials, Vol 3, J.M. Capus and R.M. German, Ed., Metal Powder Industries Federation, 1992, p 167 27. B.A. Shaw and G.S. Morton, "Thermal Spray Coatings--Marine Performance and Mechanisms," ASM International Thermal Spray Technology Proc. of the National Thermal Spray Conf. (Cincinnati, Ohio), 24-27 Oct 1988, ASM International 28. B.A. Shaw, "Characterization of the Corrosion Behavior of Aluminum, Zinc and Zinc-Aluminum Thermal Spray Coatings," Master's dissertation, Johns Hopkins University, 1985 29. D. Itzhak and E. Aghion, Corrosion Behavior of Hot-Pressed Austenitic Stainless Steel in H 2SO4 Solutions at Room Temperature, Corros. Sci., Vol 23 (No. 10), 1983, p 1085-1094 30. E. Klar, M. Svilara, C. Lall, and H. Tews, Corrosion Resistance of Austenitic Stainless Steels Sintered in Commercial Furnaces, Advances in Powder Metallurgy and Particulate Materials--1992, Vol 5, J.M. Campus and R.M. German, Ed., Metal Powder Industries Federation, American Powder Metallurgy Institute, p 411-426 31. F.M.F. Jones, The Effect of Process Variables on the Properties of Type 316L Powder Compacts, Progress in Powder Metallurgy, Vol 30, Metal Powder Industries Federation, 1970, p 25-50 32. E. Klar and P.K. Samal, Optimization of Vacuum Sintering Parameters for Improved Corrosion Resistance of P/M Stainless Steels, Advances in Powder Metallurgy and Particulate Materials, Vol 7, C. Lall, Ed., Plenum Press, 1994, p 239-251 33. B.A. Shaw, "Crevice Corrosion of a Ni-Cr-Mo-Fe Alloy in Natural and Chlorinated Seawater," Ph.D. dissertation, Johns Hopkins University, 1988 34. P. Peled and D. Itzhak, The Effect of Noble Alloying Elements Ag, Pt, and Au on the Corrosion Behavior of

Sintered Stainless Steel in an H2SO4 Environment, Corros. Sci., Vol 28 (No. 10), 1988, p 1019-1028 35. 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 36. P. Peled and D. Itzhak, The Corrosion Behavior of Double Pressed, Double Sintered Stainless Steel Containing Noble Alloying Elements, Corros. Sci., Vol 30 (No. 1), 1990, p 59-65 37. R.M. German and D. Kubish, Evaluation of Injection Molded 17-4 PH Stainless Steel Using Water Atomized Powder, Int. J. Powder Metall., Vol 29 (No. 1), 1993, p 47-62 38. D. Itzhak and P. Peled, The Effect of Cu Addition on the Corrosion Behavior of Sintered Stainless Steel in H2SO4 Environment, Corros. Sci., Vol 26 (No. 1), 1986, p 49-54 39. D. Itzhak and S. Harush, The Effect of Sn Addition on the Corrosion Behavior of Sintered Stainless Steel in H2SO4, Corros. Sci., Vol 25 (No. 10), 1985, p 883-888 Corrosion-Resistant Powder Metallurgy Alloys Barbara Shaw, Penn State University

P/M Superalloys Development of P/M superalloys began in the 1960s with the search by the aerospace industry (and later the electric power industry) for stronger high-temperature alloys in order to operate engines at higher temperatures and thus improve fuel efficiency. Initially, lower production costs were a major objective in exploring the P/M approach. Later, specific advantages linked to the P/M approach, such as the use of more complex and greater volume fractions of dispersoids, reduced segregation, improved workability, and reducing the cost of consolidating superalloy powders, particularly of oxide-dispersionstrengthened (ODS) superalloys, through the development of suitable forging techniques (Ref 40). Efforts are also underway to exploit the advantages of microcrystallinity and extended solid solutions of rapid solidification technology. Uses and State of Commercialization. Table 24 summarizes the uses of P/M superalloys in terms of components, engine use, and reasons for using P/M technology. Other uses of superalloys include nuclear reactors, heat exchangers, furnaces, sour gas well equipment and other high-temperature applications.

Table 24 Aerospace applications of P/M superalloys P/M superalloy

Component

Engine

Aircraft/ manufacturer

IN-100

Turbine disks, seals, spacers

F-100

René 95

Turbine disks, cooling plate Turbine disks, compressor shaft Vane High-pressure turbine blade retainer, disks, forward outer seals High-pressure turbine disks

T-700 F-404 F-404 F-101

Pratt & Whitney Helicopter/G.E F-18 Fighter G.E. ...

... X ... X

... ... ... ...

...

X

...

... F-18 Fighter ...

X ... ...

X X X

Astroloy Merl 76 Inconel MA-754

Turbine disks Turbine nozzle vane High- and low-pressure turbine vanes

JTSD-17R Turbofan Turbofan F-404 Selected

Reason for using P/M technology Cost Improved reduction properties X X

Stellite 31 Inconel MA6000E

engines TF 30-P100 TFE 731

Turbine blade dampers Turbine blades

USAF F-111F ...

X ...

... X

Source: Ref 41 Manufacturing of P/M Superalloys. An important prerequisite for making P/M superalloys that possess reliable dynamic properties is the use of clean powders. Years of intensive work were spent in identifying and controlling the problems related to unclean powders. Today, argon and vacuum (also known as soluble gas process) atomization, as well as atomization by the rotating electrode process, are known to be suitable for producing powders with the required low oxygen content and low degree of contamination. The so-called prior particle boundary (PPB) problem, that is, the presence of carbides segregated at PPBs, was solved through the development of low-carbon alloys. Special equipment is used for removing ceramic particles and particles containing entrapped argon. Some of these problems are minimized or avoided in ODS alloys made by mechanical alloying. In mechanical alloying, elemental and master alloy powders as well as refractory compounds are mechanically alloyed by high-energy milling (Ref 42, 43).

Two established powder consolidation techniques for P/M superalloys are hot isostatic pressing (HIP) and isothermal forging. Both P/M methods permit the manufacture of so-called near-net shape parts with attendant improved material use and reduced machining costs. Powder metallurgy forging exploits the improved forgeability deriving from the higher incipient melting temperature and reduced grain size of P/M material. Hot compaction by extrusion leads to very fine grain size, improved hot ductility, and superplasticity. Depending on the application of a superalloy part, the powder consolidation process can be controlled in order to yield either a fine or a coarse grain size. Fine grain size is preferred for intermediate temperatures (up to about 700 °C, or 1290 °F) because of its higher strength and ductility at these temperatures. For high-temperature blade and vane applications, however, a large grain size (ASTM 1 to 2) provides superior creep strength due to reduced grain-boundary sliding. Grain coarsening of ODS alloys is achieved through special heat treatments after consolidation (Ref 40). Compositions and Properties. Table 25 shows the compositions of the best known P/M superalloys. Many have the

same compositions as cast alloys but are manufactured similarly to wrought alloys. The important P/M superalloys--IN-100, René 95, and Astroloy--were adapted in the P/M process by reducing their carbon content and by adding stable carbide formers to eliminate the problem of PPB carbides. To facilitate HIP, alloy compositions were modified to increase the temperature gap between the ' solvus (above which HIP has to be carried out for increasing grain size) and the solidus temperature.

Table 25 Nominal compositions of several P/M superalloys Alloy IN-100 René 95 MERL 76 AF 115 PA101 Low-carbon Astroloy MA 754 MA 956 MA 6000

Composition, % C Cr Mo 0.07 12.5 3.2 0.07 13.0 3.5 0.02 12.4 3.2 0.05 10.5 2.8 0.1 12.5 . . . 0.04 15.0 5.0 0.05 20.0 . . . ... 20.0 . . . 0.05 15.0 2.0

W ... 3.5 ... 6.0 4.0 ... ... ... 4.0

Ta ... ... ... ... 4.0 ... ... ... 2.0

Ti 4.3 2.5 4.3 3.9 4.0 3.5 0.5 0.5 2.5

Nb ... 3.5 1.4 1.7 ... ... ... ... ...

Co 18.5 8.0 18.5 15.0 9.0 17.0 ... ... ...

Al 5.0 3.5 5.0 3.8 3.5 4.0 0.3 4.5 4.5

Hf ... ... 0.4 2.0 1.0 ... ... ... ...

Zr 0.04 0.05 0.06 ... ... 0.4 ... ... 0.15

B 0.02 0.01 0.02 ... ... 0.025 ... ... 0.01

Ni rem rem rem rem rem rem rem ... rem

Fe ... ... ... ... ... ... ... rem ...

V 0.75 ... ... ... ... ... ... ... ...

Y2O3 ... ... ... ... ... ... 0.6 0.5 1.1

Oxidation. Nickel-, cobalt-, and iron-base superalloys use the selective oxidation of aluminum or chromium to develop

oxidation resistance (Ref 44). These alloys are therefore often referred to as Al2O3 or Cr2O3 formers depending on the composition of the oxide scale that provides protection. Alloy composition, surface conditions, gas environment, and

cracking of the oxide scale affect the selective-oxidation process (Ref 44). Figure 58 shows the development of superalloys in terms of the progress achieved against high-temperature oxidation.

Fig. 58 Advancing steps in the protection of superalloys against oxidation of high temperatures showing life (in h) to 0.25 mm (10 mils) penetration of 980 °C (1800 °F). Source: Ref 45

Cyclic oxidation causes protective scales of Al2O3 and Cr2O3 to crack and spall. Regeneration of the scales will eventually result in the complete depletion of chromium and aluminum. The length of time for which superalloys are Al2O3 or Cr2O3 formers under given conditions is very important because of the subsequent appearance of less protective oxides. The importance of chromium for imparting oxidation resistance is demonstrated in Table 26, which lists fatigue crack growth rates for different alloys.

Table 26 Relative increase in fatigue crack growth rates after 15 min at 650 °C (1200 °F) K = 30 MPa

(27 ksi

Alloy René 95 after HIP + forge IN-100 HIP-consolidated MERL 76 NASA II B-7 HIP-consolidated Astroloy Waspaloy Astroloy after HIP + forge

Source: Ref 46

) Relative increase in crack growth rates 242 43.4 41.5 335 3.3 4.0 3.5

Grain size, m 50-70 4-6 15-20 4-6 50-70 40-150 50-100

Chromium content, % 12.8 12.0 12.0 8.9 15.1 19.3 14.7

Figures 59 and 60 show comparisons of the oxidation resistances of superalloys with and without oxide dispersions. Many studies have confirmed the beneficial effect of dispersed oxides on oxidation resistance. The lower oxidation rates of ODS alloys have been attributed to the reduced time required to form a continuous Cr2O3 scale due to the presence of dispersed oxides, which act as nuclei for oxidation (Ref 47). Based on marker studies with platinum, one investigation attributed the beneficial effect of oxide dispersions to the predominant, inward diffusion of oxygen ion (O2-) and a slowdown of the chromic ion (Cr3+) diffusion (Ref 48). The latter may be caused by the blocking of the dispersions in the Cr2O3 scale. With the dispersed oxides becoming dissolved in the scale, it also appears possible that trivalent ions, such as yttrium (Y3+) and lanthanum (L3+), will reduce the number of vacant cation sites, thus lowering the diffusivity of Cr3+.

Fig. 59 Comparison of the oxidation resistance of ODS alloys MA 956, MA 754, and MA 6000 with that of other superalloys. Testing conditions: 504 h at 1100 °C (2010 °F) in air containing 5% H2O. Temperature was cycled between test temperature and room temperature every 24 h. Source: Ref 46

Fig. 60 Cyclic oxidation of ODS alloys MA 956, MA 8077, MA 953, and TD-NiCr compared to that of coated alloy MM 200. Testing conditions: held at 1100 °C (2010 °F) for 1 h and cooled by a 3 min air blast. Source: Ref 46

Dispersed oxides may also improve scale adhesion because of the thinner scale or because of increased porosity or smaller grain size in the oxide scale (Ref 46). It was reported that at 1300 °C (2370 °'F) the outer regions of the Al2O3 film of MA 956 (Fe-20Cr-4.5Al-0.5Ti-0.5Y2O3) became enriched with titanium, giving rise to a continuous layer of titanium-rich oxide (Ref 49). Pegging of the oxide by titanium carbide particles and the irregular metal/oxide interface is said to contribute to the good spalling resistance of the alloy. In oxidation tests in air and in an inert atmosphere at 1260 °C (2300 °F) for MA 754 (Ni-20Cr-0.5Ti-0.5Y2O3-0.3Al-0.05C), Ni-20Cr (cast/wrought), and an ODS nickel-chromium alloy, subsurface porosity was attributed to the oxidation of chromium and aluminum (Kirkendall porosity), and thermally induced porosity was excluded as a cause (Ref 50). This type of porosity decreases with improving oxidation resistance of the alloy. Results of cyclic oxidation tests at 1100 °C (2010 °F) for MA 956, TD-NiCr, and Hastelloy X are given in Table 27. The superior resistance of MA 956 is attributed to a very stable Al 2O3 film and parabolic oxidation for more than 500 h. Tables 28 and 29 list sulfidation and carburization resistance data for MA 956. As in the case of oxidation, the alloy shows marked superiority to the other alloys tested. Tables 30 and 31 provide similar data for alloy MA 6000E (Ni-15Cr-4.5Al-4W-2Mo2Ta-2.5Ti-1.1Y2O3) (Ref 51). The functions of the various alloying elements are as follows: • • • • •

Aluminum, titanium, and tantalum for ' hardening Y2O3 for high-temperature strength and stability Aluminum and chromium for oxidation resistance Titanium, tantalum, chromium, and tungsten for sulfidation resistance Tungsten and molybdenum for solid-solution strengthening

Table 27 Cyclic oxidation resistance of superalloys at 1100 °C (2010 °F) 504 h test in atmosphere of air containing 5% H2O; temperature cycled between 1100 °C (2010 °F) and room temperature every 24 h Alloy

MA 956 TD-NiCr Hastelloy X

Weight mg/cm2 Undescaled 0.99 -4.66 -11.81

change, Descaled -1.57 -12.52 -20.62

Metal loss, m

Maximum attack, m

2 20 50

15 33 256

Table 28 Sulfidation resistance of superalloys at 925 °C (1700 °F) 312 h test in burner rig with air-to-fuel ratio of 30 to 1; fuel contained 0.3% S and 5 ppm seawater. Temperature cycle: 58 min at temperature, followed by 2 min cool to room temperature Alloy

MA 956 TD-NiCr Hastelloy X

Weight mg/cm2 Undescaled 1.04 -1.69 -2.25

change, Descaled -0.17 -11.57 -6.83

Metal loss, m

Maximum attack, m

5 25 33

18 129 132

Table 29 Carburization resistance of superalloys at 1095 °C (2000 °F) 100 h test in atmosphere of hydrogen containing 2% methane Alloy

MA 956 Incoloy 800 Alloy 814

Weight mg/cm2 Undescaled 0.07 33.74 0.82

change, Descaled -0.42 29.89 -0.73

Metal loss, m

Maximum attack, m

10 132 13

10 7615 363

Source: Ref 51

Table 30 Sulfidation resistance of superalloys at 925 °C (1700 °F) Tested in burner rig with air-to-fuel ratio of 30 to 1; fuel contained 0.3% S and 5 ppm seawater. Temperature cycle 58 min at temperature, followed by 2 min cool to room temperature Alloy MA 6000E IN-100 Alloy 713LC IN-738C

Exposure time, h 312 48 168 312

Descaled weight loss, mg/cm2 -11.11 -367.36 -488.63 -9.73

Maximum attack, m 24 169 328 28

Source: Ref 51

Table 31 Cyclic oxidation resistance of superalloys at 1100 °C (2010 °F) 504 h test in air containing 5% H2O; temperature cycled from 1100 °C (2010 °F) to room temperature every 24 h Alloy MA 6000E IN-100 Alloy 713LC IN-738C

Descaled weight change, mg/cm2 -14.12 -7.27 -22.08 -49.51

Source: Ref 51 Hot Corrosion. The requirement of hot corrosion resistance of superalloys derives from the use of sodium- and sulfurcontaining fuels and the presence of salt in the air necessary for combustion. Under such conditions, combustion gases often leave deposits of sulfates or chlorides with metallic constituents of sodium, calcium, magnesium, or potassium on the surfaces of superalloys. The resulting corrosion problems are particularly severe if these condensed phases are liquid. Hot corrosion may occur in gas turbines, boiler tubes, and incinerators. Typically (and similar to what happens in pure oxidation), hot corrosion of superalloys occurs in two stages: a slow rate initiation stage, followed by a propagation stage of rapid degradation. The difference, compared to oxidation, is that the conditions causing hot corrosion simply shorten the time in which superalloys form protective Al2O3 or Cr2O3 scales by selective oxidation (Ref 44). Factors affecting the length of the initiation stage (at the end of the initiation stage, the superalloy must be removed from service because of the start of excessive corrosion) include alloy composition, alloy fabrication conditions, gas composition and velocity, deposit composition and its physical state, amount of deposit, temperature, temperature cycles, erosion, and specimen geometry.

When a protective scale dissolves into a liquid deposit, so-called fluxing reactions can occur with the appearance of other basic or acidic nonprotective reaction products. Propagation may also be caused by components from the deposit that can accumulate in the deposit or the alloy and thus cause a nonprotective scale to form. Chlorine and sulfur produce such effects, and hot corrosion caused by the latter is known as sulfidation. Figure 61 shows the temperature ranges over which the various hot corrosion propagation modes are important.

Fig. 61 Schematic showing the temperature regimes over which different propagation modes are most prevalent.

Some superalloys corrode in several modes. For example. hot corrosion of IN-738 proceeds by alloy-induced acidic fluxing, but is preceded by other propagation modes, including a basic fluxing mode. The higher chromium content alloys IN-738 and IN-939 were developed to improve the hot corrosion resistance of land-based gas turbines. Carbide stabilization through tungsten and tantalum and delay of M23C6 formation in service were expected to allow the large chromium content to impart improved hot corrosion resistance. Increasing the chromium and decreasing the Al2O3, however, lowered ' solution temperatures and strength, which necessitated the use of coatings. The use of coatings led to the current use of enhanced aluminum, that is, carefully balanced coating alloys (based on nickel, iron, or cobalt with chromium, aluminum, and other active elements). Generally, all superalloy load-bearing parts used at very high temperatures under dynamic conditions are coated (Ref 45). Nevertheless, coatings generally last longer on more corrosion-resistant base materials. In a model study, IN-738 was used to demonstrate the effect of grain size and Y2O3 dispersions on hot corrosion behavior (Ref 52). Under gas turbine simulated hot gas corrosion test conditions at 850 and 950 °C (1560 and 1740 °F) (Fig. 62), the presence of a Y2O3 dispersion lowered the corrosion rate. At 950 °C (1740 °F), a finer grain size further reduced the corrosion rate, which was thought to be mainly due to a higher diffusion rate of chromium and aluminum. The effect of the dispersion was predominant at 850 °C (1560 °F). Reduced sulfate formation at 850 °C (1560 °F) was attributed to the likely formation of yttrium oxysulfide.

Fig. 62 Comparison resistance of alloy IN-738LC in hot (850 °C, or 1560 °F) gases. A, IN-738LC; B, IN-738LC with Y2O3 dispersion, annealed at 1270 °C (2320 °F); C, IN-738LC with Y 2O3 dispersion, annealed at 1100 °C (2010 °F). Source: Ref 40

In a study of the oxidation and hot corrosion resistance of P/M LC Astroloy and IN-100, isostatically pressed samples were found to be moderately attacked in a sulfate-chloride environment and heavily corroded by pure sodium sulfate (Na2SO4) (Ref 53). Heat treatment and the use of coarse powder (62 to 150 m for Astroloy and 88 to 200 m for IN-100) lowered the susceptibility to catastrophic corrosion. Additions of yttrium to IN-100 improved the corrosion resistance in pure sulfate, but were detrimental when NaCl was present. Therefore, yttrium additions to IN-100 cannot be recommended for marine turbines. It was concluded that in many cases impregnation coatings must be considered for components made of IN-100 alloys. As a part of an evaluation of improved alloys for use in oil and gas drilling at depths of 6100 m (20,000 ft), HIP nickel-base alloy Inconel 625 was studied in a simulated deep, hot, sour cell environment (Ref 54). The alloy demonstrated resistance to pitting and crevice corrosion, sulfide stress cracking, chloride stress-corrosion cracking (SCC), and elevated-temperature anodic stress cracking. Hot isostatically pressed Inconel 625 exhibited essentially the same corrosion resistance as wrought Inconel 625. Fatigue and Creep Crack Growth. Fatigue crack growth rates of nickel-base superalloys measured at frequencies above

0.1 Hz, at intermediate temperatures, and at an intermediate stress intensity range, K, were found to be several times higher than those measured in inert atmospheres (Ref 55). The buildup of corrosion products with decreasing K, however, was thought to enhance crack closure, thus reducing the effective stress intensity range and leading to fatigue thresholds higher than those in inert environments. Table 32 shows creep crack growth rates of IN-750 with various grain-boundary carbide microstructures (Ref 46, 55). Aggressive environments (helium + 3% SO2 and air) produce order of magnitude increases over the rates in inert gas. In general, the reaction of both oxide dispersoid-free P/M superalloys and cast and wrought superalloys to aggressive environments is similar. This suggests that crack growth is governed mainly by microstructure and alloy chemistry.

Table 32 Dependence on carbide microstructure of creep crack growth rates of alloy IN-750 in four environments Grain-boundary carbide microstructure Blocky Cellular None

Crack growth rate (da/dt), mm/min(a) Helium Air Helium + 4% methane 7 × 10-4 1.95 × 10-2 7 × 10-4 -3 1.15 × 10 2.6 × 10-3 3.1 × 10-3 -4 -2 7 × 10 6.3 × 10 7 × 10-4

Helium + 3% SO2 1.5 × 10-1 1.3 × 10-1 1.6 × 10-1

Source: Ref 46, 55

(a)

K = 35 MPa

(32 ksi

).

Oxide-Dispersion-Strengthened Alloys. Mechanical alloying has removed many constraints on the development of

new superalloys. Many new ODS alloys were designed specifically for corrosion resistance because alloying requirements for precipitation strengthening can be greatly reduced. Superior creep, corrosion, and erosion resistance at high temperatures have been claimed to enable the use of lower-grade fuels (Ref 56). Figures 63, 64, 65, and 66 show the corrosion resistances of ODS alloys MA 956, MA 6000, and MA 754 compared to several conventional superalloys.

Fig. 63 Comparison of the corrosion resistance of MA 6000 and MA 956 with that of other superalloys. Tested in a burner rig for 312 h using a 30-to-1 air-to-fuel ratio. Fuel contained 0.3% S and 5 ppm seawater, and specimens were held at temperature for 58 min of each hour, then cooled 2 min an air blast. Source: Ref 46

Fig. 64 Corrosion rate (a) and temperature capability (b) of MA 6000, MA 754, and non-ODS superalloys as a function of chromium content. A stress of MPa (29 ksi) was applied to the specimens during the 10,000 h test. Source: Ref 46

Fig. 65 Temperature capability as a function of corrosion rate for various superalloys. Same data as in Fig. 64. Source: Ref 46

Fig. 66 Hot corrosion of alloys MA 953, HDA 8077, and MA 956 compared to that of some non-ODS alloys. Test conditions: 900 °C (1650 °F), 1 h, followed by a 3 min air blast, 5 ppm sea salt. Source: Ref 46

Coatings for ODS Alloys. As mentioned above, for extended high-temperature service, superalloys require additional protection through coatings. The use of aluminide coatings appears to be unsatisfactory due to the development of subsurface Kirkendall porosity and early spalling of the protective scale. Kirkendall porosity decreases with increasing aluminum content of the substrate alloy as well as with decreasing grain size (Ref 56). Only limited information exists on the properties of Cr-Al-Y coatings (Ref 46) and on diffusion barrier coatings (Ref 57, 58, 59, 60).

References cited in this section

40. G.H. Gessinger, Recent Developments in Powder Metallurgy of Superalloys, Powder Metall. Int., Vol 13 (No 2), 1981, p 93-101 41. R.F. Singer, Recent Developments and Trends in High Strength P/M Materials, Powder Metall. Int., Vol 17 (No. 6), 1985, p 284-288 42. L.R. Curwick, The Mechanical Alloying Process:Powder to Mill Product, Frontiers of High-Temperature Materials, J.S. Benjamin, Ed., Proc. International Conference on Oxide Dispersion Strengthened Superalloys by Mechanical Alloying, Inco Alloy Products, 1981, p 3-10 43. J.S. Benjamin and T.E. Volin, The Mechanism of Mechanical Alloying, Metall. Trans., Vol 5, 1974, p 19291934 44. F.S. Pettit and G.H. Meier, Oxidation and Hot Corrosion of Superalloys, Superalloys, 1984, M. Gell et al., Ed., Proc. the Fifth International Symposium on Superalloys, The Metallurgical Society, 1984, p 651-687 45. C.T. Sims, A History of Superalloy Metallurgy for Superalloy Metallurgists, Superalloys, 1984, M. Gell et al., Ed., Proc. the Fifth International Symposium on Superalloys, The Metallurgical Society, 1984, p 309-419 46. G.H. Gessinger, Powder Metallurgy of Superalloys, Butterworths, 1984 47. J. Stringer, B.A. Wilcox, and P.I. Jaffee, Oxid. Met., Vol 5, 1972, p 11 48. C.S. Giggins and F.S. Pettit, The Oxidation of TD Ni Cr (Ni-20 Cr-2 Vol. pct ThO 2) Between 900 and 1200 C, Metall. Trans., Vol 2, 1971, p 1071-1078 49. F. Perry, Oxide-Dispersion-Strengthened P/M Alloys Produced for Severe Service Applications, Ind. Heat., Vol 49 (No. 5), May 1982, p 22-25 50. 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 51. J.H. Weber, High Temperature Oxide Dispersion Strengthened Alloys, Proc. 25th National SAMPE Symposium and Exhibition Society for the Advancement of Material and Process Engineering, 1980, p 752763 52. G.H. Gessinger, High Temperature Alloys for Gas Turbines, D. Coutsouradis et al., Ed., Applied Science, 1978, p 817 53. P.L. Antona, A. Bennani, P. Cavalloti, and O. Ducati, Heat Treatments and Oxidation Behavior of Some P/M Ni-Base Superalloys, European Symposium on Powder Metallurgy, Vol 1, Jernkontoret Activity Group B, 1978, p 137-142 54. 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 55. S. Floreen, Effects of Environment on Intermediate Temperature Crack Growth in Superalloys, Proc. AIME Symposium (Louisville, KY), American Society of Mining, Metallurgical, and Petroleum Engineers, 1981 56. 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 57. D.H. Boone, D.A. Crane, and D.P. Whittle, Thin Solid Films, Vol 84, 1981, p 39 58. F.R. Wermuth and A.R. Stetson, Report NASA CR-120852, National Aeronautics and Space Administration, 1971 59. M.A. Gedwill, T.K. Glasgow, and LS. Levine, "A New Diffusion Inhibited Oxidation Resistant Coating for Superalloys," NASA TM 82687, National Aeronautics and Space Administration, 1981 60. 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 Corrosion-Resistant Powder Metallurgy Alloys Barbara Shaw, Penn State University

References 1. E. Klar and P.K. Samal, Powder Metals, Corrosion Tests and Standards Manual, R. Baboian, Ed., ASTM, 1995, p 551-557 2. E. Klar and P.K. Samal, Effect of Density and Sintering Variables on the Corrosion Resistance of Austenitic Stainless Steels, Advances in Powder Metallurgy and Particulate Materials, Vol 3, M. Phillips and J. Porter, Ed., Metal Powder Industries Federation, American Powder Metallurgy Institute, 1995, p 3-17 3. C. Molins, J.A. Bas, J. Planas, and S.A. Ames, P/M Stainless Steel: Types and Their Characteristics and Applications, Advances in Powder Metallurgy and Particulate Materials, Vol 5, J.M. Capus, Ed., Plenum Press, 1992, p 345-357 4. J.A. Bas, J. Peñafiel, A. Bolarin, and M.P. Latre, Determination Methods of the PM Stainless Steels Corrosion Resistance, Advances in Powder Metallurgy and Particulate Materials, Plenum Press, p 47-59 5. D. Itzhak and E. Aghion, An Anodic Behavior Study of an Analogical Sintered System of Austenitic Stainless Steel in H2SO4 Solution, Corros. Sci., Vol 24 (No. 2), 1984, p 145-152 6. G. Lei, R.M. German, and H.S. Nayar, Corrosion Control in Sintered Austenitic Stainless Steels, reprinted from Progress in Powder Metallurgy, Vol 39, Metal Powder Industries Federation, 1984, p 55-74 7. A. Tremblay and R. Angers, Corrosion Resistance of 316L P/M Stainless Steel, Advances in Powder

Metallurgy and Particulate Materials--1994, Vol 7, C. Lall and A.J. Neupaver, Ed., Metal Powder Industries Federation, American Powder Metallurgy Institute, p 225-237 8. D.A. Jones, Principles and Prevention of Corrosion, 2nd ed., Prentice Hall, 1996 9. G. Scavino, E. Angelini, M. Rosso, and F. Rosalbino, Comparison of Corrosion Resistance Properties of PM Steels, Computer Methods and Experimental Measurements for-Surface Treatment Effects, Computational Mechanics Publications, 1993, p 337-346 10. M. Baran, A.E. Segall, B.A. Shaw, et al., "Evaluation of P/M Ferritic Stainless Steel Alloys for Automotive Exhaust Applications," presented at the P/M Tec `97 Conf. (Chicago, IL), Metal Powder Industries Fe deration, June 1997 11. D.W. Yuan, J.R. Spirko, and H.I. Sanderow, Colorimetric Corrosion Testing of P/M Stainless Steel, Int. J. Powder Metall., Vol 33 (No. 2), 1997 12. D.W. Yuan, T. Prucher, and H.I. Sanderow, "An Evaluation of the Relative Corrosion Resistance of P/M Stainless Steel Alloys," Technical Paper No. 950391, Society of Automotive Engineers, 1995 13. R.L. Sands, G.F. Bidmead, and D.A. Oliver, The Corrosion Resistance of Sintered Austenitic Stainless Steel, Modern Developments in Powder Metallurgy, Vol 2, H.H. Hausner, Ed., Plenum Press, 1966, p 73-83 14. E. Maahn, S.K. Jensen, R.M. Larsen, and T. Mathiesen, Factors Affecting the Corrosion Resistance of Sintered Stainless Steel, Advances in Powder Metallurgy and Particulate Materials, Vol 7, C. Lall, Ed., Plenum Press, 1994, p 253-271 15. T. Raghu, S.N. Malhotra, and P. Ramakrishnan, Corrosion Behavior of Porous Sintered Type 316L Austenitic Stainless Steel in 3% NaCl Solution, Corrosion, Vol 45 (No. 9), 1989, p 698-704 16. D.H. Ro and E. Klar, Corrosion Behavior of P/M Austenitic Stainless Steels, Modern Developments in Powder Metallurgy, Vol 2, H.H. Hausner, Ed., Plenum Press, 1980, p 247-287 17. P. Peled and D. Itzhak, The Surface Composition of Sintered Stainless Steel Containing Novle Alloying Elements Exposed to a H2SO4 Environment, Corros. Sci., Vol 32 (No. 1), 1991, p 83-90 18. M. Baran, "An Evaluation of a P/M Ferritic Stainless Steel Automotive Exhaust Flange," Master's dissertation, The Pennsylvania State University, 1997 19. Laboratory Testing, Corrosion, Vol 13, ASM Handbook, ASM International, 1987, p 212-228 20. A.J. Sedriks, Effects of Alloy Composition and Microstructure on the Passivity of Stainless Steels, Corrosion, Vol 42 (No. 7), 1986, p 376-388 21. L. Campbell, "Corrosion Behavior of an Alumina-Reinforced Aluminum Metal Matrix Composite," Baccalaureate thesis, The Pennsylvania State University, 1996 22. M.A. Streicher, Corrosion Tests and Standards Manual, R. Baboian, Ed., ASTM, 1995, p 197-217 23. E. Klar, Corrosion of Powder Metallurgy Materials, Corrosion, Vol 13, Metals Handbook, ASM International, p 824-845 24. A.J. Sedriks, Corrosion of Stainless Steels, 2nd ed., John Wiley & Sons, 1996 25. E. Maahn and T. Mathiesen, Corrosion Properties of Sintered Stainless Steel, UK Corrosion Proc. (Manchester, UK), Vol 2, Institute of Corrosion, 1991, p 1-15 26. K. Frisk, A. Johansson, and C. Lindberg, Nitrogen Pick Up During Sintering Stainless Steel, Advances in Powder Metallurgy and Particulate Materials, Vol 3, J.M. Capus and R.M. German, Ed., Metal Powder Industries Federation, 1992, p 167 27. B.A. Shaw and G.S. Morton, "Thermal Spray Coatings--Marine Performance and Mechanisms," ASM International Thermal Spray Technology Proc. of the National Thermal Spray Conf. (Cincinnati, Ohio), 24-27 Oct 1988, ASM International 28. B.A. Shaw, "Characterization of the Corrosion Behavior of Aluminum, Zinc and Zinc-Aluminum Thermal Spray Coatings," Master's dissertation, Johns Hopkins University, 1985 29. D. Itzhak and E. Aghion, Corrosion Behavior of Hot-Pressed Austenitic Stainless Steel in H 2SO4 Solutions at

Room Temperature, Corros. Sci., Vol 23 (No. 10), 1983, p 1085-1094 30. E. Klar, M. Svilara, C. Lall, and H. Tews, Corrosion Resistance of Austenitic Stainless Steels Sintered in Commercial Furnaces, Advances in Powder Metallurgy and Particulate Materials--1992, Vol 5, J.M. Campus and R.M. German, Ed., Metal Powder Industries Federation, American Powder Metallurgy Institute, p 411-426 31. F.M.F. Jones, The Effect of Process Variables on the Properties of Type 316L Powder Compacts, Progress in Powder Metallurgy, Vol 30, Metal Powder Industries Federation, 1970, p 25-50 32. E. Klar and P.K. Samal, Optimization of Vacuum Sintering Parameters for Improved Corrosion Resistance o f P/M Stainless Steels, Advances in Powder Metallurgy and Particulate Materials, Vol 7, C. Lall, Ed., Plenum Press, 1994, p 239-251 33. B.A. Shaw, "Crevice Corrosion of a Ni-Cr-Mo-Fe Alloy in Natural and Chlorinated Seawater," Ph.D. dissertation, Johns Hopkins University, 1988 34. P. Peled and D. Itzhak, The Effect of Noble Alloying Elements Ag, Pt, and Au on the Corrosion Behavior of Sintered Stainless Steel in an H2SO4 Environment, Corros. Sci., Vol 28 (No. 10), 1988, p 1019-1028 35. 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 36. P. Peled and D. Itzhak, The Corrosion Behavior of Double Pressed, Double Sintered Stainless Steel Containing Noble Alloying Elements, Corros. Sci., Vol 30 (No. 1), 1990, p 59-65 37. R.M. German and D. Kubish, Evaluation of Injection Molded 17-4 PH Stainless Steel Using Water Atomized Powder, Int. J. Powder Metall., Vol 29 (No. 1), 1993, p 47-62 38. D. Itzhak and P. Peled, The Effect of Cu Addition on the Corrosion Behavior of Sintered Stainless Steel in H2SO4 Environment, Corros. Sci., Vol 26 (No. 1), 1986, p 49-54 39. D. Itzhak and S. Harush, The Effect of Sn Addition on the Corrosion Behavior of Sintered Stainless Steel in H2SO4, Corros. Sci., Vol 25 (No. 10), 1985, p 883-888 40. G.H. Gessinger, Recent Developments in Powder Metallurgy of Superalloys, Powder Metall. Int., Vol 13 (No 2), 1981, p 93-101 41. R.F. Singer, Recent Developments and Trends in High Strength P/M Materials, Powder Metall. Int., Vol 17 (No. 6), 1985, p 284-288 42. L.R. Curwick, The Mechanical Alloying Process:Powder to Mill Product, Frontiers of High-Temperature Materials, J.S. Benjamin, Ed., Proc. International Conference on Oxide Dispersion Strengthened Superalloys by Mechanical Alloying, Inco Alloy Products, 1981, p 3-10 43. J.S. Benjamin and T.E. Volin, The Mechanism of Mechanical Alloying, Metall. Trans., Vol 5, 1974, p 19291934 44. F.S. Pettit and G.H. Meier, Oxidation and Hot Corrosion of Superalloys, Superalloys, 1984, M. Gell et al., Ed., Proc. the Fifth International Symposium on Superalloys, The Metallurgical Society, 1984, p 651-687 45. C.T. Sims, A History of Superalloy Metallurgy for Superalloy Metallurgists, Superalloys, 1984, M. Gell et al., Ed., Proc. the Fifth International Symposium on Superalloys, The Metallurgical Society, 1984, p 309-419 46. G.H. Gessinger, Powder Metallurgy of Superalloys, Butterworths, 1984 47. J. Stringer, B.A. Wilcox, and P.I. Jaffee, Oxid. Met., Vol 5, 1972, p 11 48. C.S. Giggins and F.S. Pettit, The Oxidation of TD Ni Cr (Ni-20 Cr-2 Vol. pct ThO 2) Between 900 and 1200 C, Metall. Trans., Vol 2, 1971, p 1071-1078 49. F. Perry, Oxide-Dispersion-Strengthened P/M Alloys Produced for Severe Service Applications, Ind. Heat., Vol 49 (No. 5), May 1982, p 22-25 50. 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 51. J.H. Weber, High Temperature Oxide Dispersion Strengthened Alloys, Proc. 25th National SAMPE Symposium and Exhibition Society for the Advancement of Material and Process Engineering, 1980, p 752-

763 52. G.H. Gessinger, High Temperature Alloys for Gas Turbines, D. Coutsouradis et al., Ed., Applied Science, 1978, p 817 53. P.L. Antona, A. Bennani, P. Cavalloti, and O. Ducati, Heat Treatments and Oxidation Behavior of Some P/M Ni-Base Superalloys, European Symposium on Powder Metallurgy, Vol 1, Jernkontoret Activity Group B, 1978, p 137-142 54. 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 55. S. Floreen, Effects of Environment on Intermediate Temperature Crack Growth in Superalloys, Proc. AIME Symposium (Louisville, KY), American Society of Mining, Metallurgical, and Petroleum Engineers, 1981 56. 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 57. D.H. Boone, D.A. Crane, and D.P. Whittle, Thin Solid Films, Vol 84, 1981, p 39 58. F.R. Wermuth and A.R. Stetson, Report NASA CR-120852, National Aeronautics and Space Administration, 1971 59. M.A. Gedwill, T.K. Glasgow, and LS. Levine, "A New Diffusion Inhibited Oxidation Resistant Coating for Superalloys," NASA TM 82687, National Aeronautics and Space Administration, 1981 60. 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 Corrosion-Resistant Powder Metallurgy Alloys Barbara Shaw, Penn State University

Selected References • • • • • • • •

•

L. Fedrizzi, A. Molinari, F. Deflorian, A. Tiziani, and P.L. Bonora, Corrosion Study of Industrially Sintered Copper Alloyed 316L Austenitic Stainless Steel, Br. Corros. J., Vol 26 (No. 1), 1991, p 46-50 M. Hanada, Y. Takeda, N. Amano, et al., Development of a PM Sensor Ring for Use in an Antilock Brake System, Met. Powder Rep., Vol 44 (No. 10), 1989, p 695-698 C. Lall and M. Svilar, The Corrosion Resistance of P/M Stainless Steels and Selected Alloys in Methanol-Based Fuels, P/M Steels, J.M. Capus, Ed., Vol 5, Plenum Press, 1992, p 427-435 G. Lei and R.M. German, Corrosion of Sintered Stainless Steels in a Sodium Chloride Solution, Modern Developments in Powder Metallurgy, E.N. Aqua, Ed., Vol 16, Plenum Press, p 261-275 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 S.N. Malhotra and P. Ramakrishnan, Corrosion Behavior of PM Stainless Steel Filters, Met. Powder Rep., 1991, p 48-51 T. Mathiesen and E. Maahn, Evaluation of Sensitization Phenomena in Sintered Stainless Steels, Powder Metallurgy World Congress (PM), Paris, Vol 3, Les Editions de Physique, 1994, p 2089-2092 H.S. Nayar, R.M. German, and W.R. Johnson, The Effect of Sintering on the Corrosion Resistance of 316L Stainless Steel, Progress in Powder Metallurgy, Vol 37; Metal Powder Industries Federation, p 17 P. Peled, S. Harush, and D. Itzhak, The Effect of Ni Addition on the Corrosion Behavior of Sintered

• •

•

• •

Stainless Steel in H2SO4, Corros. Sci., Vol 28 (No. 4), 1988, p 327-332 J.R. Pickens, Techniques for Assessing the Corrosion Resistance of Aluminum Powder Metallurgy Alloys, Rapidly Solidified Powder Aluminum Alloys, ASTM, 1986 p 381-409 J.H. Reinshagen and T.J. Bockius, Stainless Steel Based P/M Alloys with Improved Corrosion Resistance, Advances in Powder Metallurgy and Particular Materials, Metal Powder Industries Federation, Vol 3, 1995, p 19-30 J.H. Reinshagen and G.D. Flick, Improved Corrosion Resistant Stainless Steel Based P/M Alloys, Advances in Powder Metallurgy and Particulate Materials--1996, Metal Powder Industries Federation, Vol 5, 1995, p 61-71 J.H. Reinshagen and A.J. Neupaver, Fundamentals of P/M Stainless Steels, Advances in Powder Metallurgy, Vol 2, Plenum Press, 1989, p 283-295 F. Velasco, J.R. Ibars, J.M. Ruiz-Roman, et al., Improving the Corrosion Resistance of Power Metallurgy Austenitic Stainless Steels through Infiltration, Corrosion, Vol 52 (No. 1), 1996, p 47-52

Magnetic Materials and Properties for Powder Metallurgy Part Applications Kenneth H. Moyer, Magna-Tech P/M Labs

Introduction MAGNETIC POWDER METALLURGY PARTS are produced predominantly from "soft" (or temporary) magnetic materials, which generate a magnetic force only in the presence of an externally applied magnetic field. Commercial soft magnetic P/M alloys are commonly produced from high-purity iron or various ferrous alloy types such as Fe-2Ni, Fe-3Si, Fe-0.45P, Fe-0.6P, and 50Ni-50Fe. The use of P/M techniques to make magnetically soft components are effective for applications in which complicated magnetic parts would otherwise require considerable machining. In some applications, P/M fabrication eliminates all machining operations and saves a substantial amount of the total cost, compared to conventional manufacturing. Powder metallurgy and metal injection molding (MIM) also have other advantages for fabrication of magnetic materials. One major advantage includes economical blending of small-lot specialty alloys. Powder metallurgy also has the advantage of processing materials that would be difficult for fabrication as wrought metal forms. For example, ironphosphorus alloys cannot be made in wrought form because of "hot shortness," while Fe-P alloys can be processed effectively by P/M methods for soft magnetic applications. Powder metallurgy also is the method used for the relatively new permanent magnetic materials based on combinations of neodymium, iron, and boron (introduced in 1983). The high reactivity of neodymium or other rare earths and their alloys requires the suppression of contamination during alloy preparation and processing. In particular, oxidation by O2 or H2O or both must be kept to a minimum. Any oxidation of the alloy occurring during processing depletes the alloy of the rare-earth components. This usually results in the production of an alloy having unfavorable magnetic qualities. To obtain powder compacts with maximum magnetization, the powder is magnetically aligned and pressed. Magnetic Materials and Properties for Powder Metallurgy Part Applications Kenneth H. Moyer, Magna-Tech P/M Labs

Basic Magnetism In simple terms, magnetism is a field force that can be produced either by permanent ("hard") magnetic materials or by an electric current generated in a coil of wire that surrounds a section of iron. The magnetic field is defined quantitatively in terms of the magnetic field strength (or the magnetizing force), designated as H, with the unit of oersteds (Oe) or amps/meter (A/m). The magnetic response of a material is characterized by the induced magnetic field that is caused by an applied magnetic field, as shown in Fig. 1. The current in the primary winding produces an external magnetic field, which then induces a magnetic field in the toroid and the secondary circuit. This method characterizes the magnetic response in terms of magnetic induction (B) versus the applied field (H) as shown in Fig. 1(b). These basic characteristics of magnetic response are briefly summarized in this introductory section.

Fig. 1 Standard toroid test of magnetic response. (a) Test setup. (b) Hysteresis response. Source: Ref 1

Intensity of Magnetization and Magnetic Induction. The intensity of magnetization (I) is defined in terms of the number of unit north poles and south poles, m, within a given cross-sectional area. A rectangular bar of length l and crosssectional area a with m unit north and south poles at each end would have an intensity of magnetization defined as m/a = I. The intensity of magnetization also can be likened to the flow of electrons through a wire where the concentration and direction of the electrons are known at any point.

In a magnetic field, flux lines pass from a magnetized material into the air at the north pole, re-enter at the south pole, and then pass through the material to the north pole to form a closed loop. The total number of lines crossing a given area at right angles is defined as the flux in the area. The flux per unit area is defined as the flux density or the magnetic induction, B. Both the magnetizing force or the applied field, H, and the intensity of magnetization, I, contribute to the lines of induction, but in magnetic materials, the intensity of magnetization, I, is generally much larger. The magnetic induction is therefore defined by the formula: B = H + 4 I. The cgs unit of magnetic induction is gauss (G), the SI unit is tesla (T). A tesla is equivalent to 104 G. Magnetism and Permeability. When a piece of iron is brought near a permanent magnet or a magnetic field of an

electrical current, the magnetization induced in the iron by the applied field is described by a magnetization curve obtained by plotting either the intensity of magnetization or the magnetic induction, B, as a function of the applied field, H. A typical magnetization curve is shown in Fig. 2(a). However, if the direction of magnetization is reversed, the magnetization curve does not retrace itself but has a property known as hysteresis. Owing to this hysteresis, a loop rather than a curve is generated (Fig. 2b). However, unless losses are realized, the magnetization curve should be reproduced repeatedly.

Fig. 2 Schematic of (a) magnetization curve (b) hysteresis loop. Source: Ref 1, 2

The behavior of any magnetic material can be defined by the hysteresis loop and the permeability. The ratio B/H is defined as the permeability, . This value indicates the relative increase in flux or magnetic induction caused by the presence of a magnetic material. Of the many useful permeabilities, the initial permeability, 0, and the maximum permeability, max, are of most importance for P/M applications. The initial permeability, 0, is simply the slope of the initial portion of the magnetization curve, say to 0.01 T (100 G). The maximum permeability, max, is the slope to the tangent of the magnetization curve (Fig. 2). After max is achieved, the permeability falls off rapidly and finally reaches a value of 1. At this point, the magnet is said to be saturated. In other words, even though the magnetization curve will continue to rise, the contribution to magnetization will derive only from the applied field. The magnetic induction at a permeability of 1 is defined as the saturation induction (Bsat or Bm in Fig. 2b). Each magnetic material at a temperature of 20 °C will have a specific constant value for the magnetic saturation. In the case of iron, the value is 2.15 T (21,500 G). Hysteresis Loop. When the applied magnetic field (H) is cycled over time, the magnetic induction lags the applied

field and defines a hysteresis loop (Fig. 2b). There are two magnetic properties of interest that occur as a result of the hysteresis. These properties include the remanence or residual induction, Br, and the coercive field, Hc. The remanent

magnetization is the specific value of the magnetic induction when the applied field has been reduced to 0. In other words, the remanent magnetization is the magnetic induction at an applied field of 0, or the intercept of the B-axis. The coercive field, Hc, is the applied field required to reduce the magnetic induction to a value of 0. This value is the intercept of the H-axis. Both of these properties, including magnetic induction, are important to the performance of any magnetic material. These properties are identified in Fig. 2b as specific points on the hysteresis curve. Demagnetizing Fields, Air Gaps. Most magnetic devices are not solid but contain discontinuities, such as air gaps. When a discontinuity or an air gap is present, it creates a magnetic field in a direction opposite to the applied field. The net effect is to reduce or subtract from the magnetic induction that results from the application of the applied field. Thus, the true applied field causing the magnetic induction is somewhat lower than the applied field, owing to any demagnetizing forces that are present, such as air gaps. The true applied field can therefore be expressed as follows: H = Ha - H. The demagnetizing field is therefore approximately proportional to the intensity of magnetization. In other words: H = NI, where N is defined as the demagnetizing factor, and where:

The effect of air gaps is to shear the magnetization curve. In other words, air gaps counteract the positive effects of purifying materials and growing grains (for example, by the use of high-temperature sintering during P/M processing). Ferromagnetism and Factors Affecting Magnetic Quality. The magnetic properties of magnetic materials

depend on the chemical composition, cold deformation, and heat treatment. There are essentially three elements and their respective alloys that are truly magnetic. These ferromagnetic materials are iron, nickel, and cobalt. Iron and cobalt have high magnetic saturations (Table 1). Because of the expense, scarcity, and difficulty of sintering, cobalt and its alloys are not generally selected, except for special applications. Iron, on the other hand, is inexpensive and quite common. Nickel has a magnetic saturation of only about 0.6 T (6 kG). Owing to its much lower magnetic saturation with respect to iron and its higher cost, nickel and its alloys are used mainly when quick response is required.

Table 1 Room-temperature magnetic properties of selected materials Material/condition

Density, g/cm3

Maximum induction, Bmax T kG

Wrought metals 1.87(a) Unalloyed cobalt 8.8 7.87 2.15(a) Unalloyed iron 0.616(a) Unalloyed nickel 8.9 Wrought alloys 7.65 2.0 Fe-3%Si 8.2 1.6 Fe-50%Ni 7.62 1.2 430F stainless Soft magnetic P/M Iron (ASTM A 811) 6.6 0.9 H = 15 Oe 6.9 1.0 H = 15 Oe 7.2 1.2 H = 15 Oe Pure P/M iron(b) 7.0 1.09 At H = 15 Oe 7.0 1.15 At H = 25 Oe 7.0 1.35 At H = 100 Oe P/M Fe-0.45%P(b) 7.0 1.19 At H = 15 Oe 7.0 1.24 At H = 25 Oe 7.0 1.43 At H = 100 Oe P/M Fe-50%Ni(b) 7.2 0.99 At H = 15 Oe 7.2 1.01 At H = 25 Oe 7.2 1.14 At H = 100 Oe

Residual induction, Br T kG

Maximum permeability,

18.7(a) 21.5(a) 6.16(a)

0.49 1.18 0.3

4.9 11.8 3.0

245 5,000 1,240

708 80 167

8.9 1.0 2.1

20 16 12

... 0.8 0.6

... 8.0 6.0

8,000 70,000 2,000

56 4.0 160

0.7 0.05 2.0

9 10 12

0.8 0.9 1.1

8 9 11

1,800 2,100 2,500

175 ... 160

2.2 ... 2.0

10.9 11.5 13.5

0.97 1.0 1.1

9.7 10.0 11.0

2,400 2,400 2,400

135 135 135

1.7 1.7 1.7

11.9 12.4 14.3

1.12 1.14 1.24

11.2 11.4 12.4

3,300 3,300 3,300

127 127 127

1.6 1.6 1.6

9.7 10.1 11.4

0.81 0.84 0.88

8.1 8.4 8.8

12,300 12,300 12,300

22 22 22

0.28 0.28 0.28

Source: ASM Handbook, Vol 2, 1990, Ref 1, 3

max

Coercive force, Hc A/m Oe

Magnetic saturation (Bsat) values.

(a) (b)

Samples sintered at 1250 °C (2280 °F), 45 min in dissociated ammonia.

In crystalline materials, the basis for ferromagnetism lies in the alignment of magnetic moments from noncompensated electron spins in the 3d shell of the transition series elements, such as iron, nickel, and cobalt (Table 2). In ferromagnetic materials that are below their Curie temperature, the magnetic moments of adjacent atoms are "coupled" parallel to each other. For a small volume of material, all of the individual magnetic moments are aligned in one direction. This small volume is magnetized to saturation and is known as a magnetic domain. An adjacent volume of material, also magnetized to saturation, can have the summation of its magnetic moments point in another direction. Where two such volumes meet with differing alignments, a domain boundary wall must exist. The total magnetization of a sample of material is the net vector summation of all the individual component domain magnetization vectors.

Table 2 Magnetic moments of neutral atoms of 3d transition elements Unpaired 3d electrons 3

Atom V

Number of electrons 23

Electronic 3d orbitals

configuration

a/D ratio(a) 2

5

Cr

24

1

5

Mn

25

2

4

Fe

26

2

3

Co

27

2

2

Ni

28

2

0

Cu

29

1

Source: Ref 4

(a)

The ratio of atomic spacing, a, and diameter, D, of the 3d orbitals determines the occurrence of positive exchange energy for ferromagnetism.

The parallel "coupling" of magnetic dipoles in iron, cobalt, and nickel occurs from an "isotropic exchange energy," which is the mechanism responsible for ferromagnetism that was a complete mystery until the advent of quantum mechanics by Heisenberg. In very simple terms, parallel "coupling" of magnetic dipoles in iron, cobalt, and nickel is due to a positive exchange energy that occurs when the ratio of atomic spacing (a) to the diameter (D) of the third orbital is between 1.4 and 4.7 (Ref 4). Thus iron, cobalt, and nickel are ferromagnetic, but manganese and chromium are not (Table 2). Factors Affecting Permeability and Hysteresis Losses. Permeability, coercive field, and hysteresis loss are

affected by impurities within the alloy, by cold deformation such as sizing, and by any heat treatment that is provided. Impurities that are most harmful to these alloys include carbon, nitrogen, oxygen, and sulfur. All P/M parts also contain a lubricant to facilitate compaction. If this lubricant is not properly removed during the presinter portion of the sintering cycle, the structure-sensitive properties (permeability, coercive field, and hysteresis loss) may be significantly degraded. Sintering furnace condition and choice of protective atmosphere are also important. Hydrogen or vacuum is preferred. However, hydrogen/nitrogen mixtures, exothermic gas, endothermic gas, and dissociated ammonia can be selected as the protective atmosphere. Any of the latter can cause nitriding or carburizing, which will degrade the structure-sensitive magnetic properties. Because most powders that are used are water atomized and therefore contain surface oxides, these surface oxides must be reduced during the sintering cycle or else degradation of structure-sensitive magnetic properties will occur. Knowledge of phase diagrams is important to the development of alloys for magnetic applications. Alloys most commonly selected are single-phase alloys. For instance, more than 0.008 wt% C or more than 0.006 wt% N causes carbide or nitride formation in -iron. The more carbon contamination or the more nitriding that occurs, the greater the degradation of the structure-sensitive magnetic properties. In phosphorus irons, additions of >1.5 wt% P result in formation of phosphorus iron intermetallics. In silicon irons, >3 wt% Si makes the alloys difficult to compact; however,

as much as 6 wt% Si has been added to P/M silicon iron alloys. Soft magnets should be processed from single-phase compositions of high purity. If second phases form, the structure-sensitive magnetic properties can be affected over time from aging. At the elevated temperatures required to sinter the parts, iron can dissolve considerable carbon or nitrogen into the structure, which consists of austenite ( -phase). After sintering and cooling, the structure-sensitive properties can appear to be unaffected. However, over time in service, the properties degrade as a result of the nitrogen or carbon precipitating from solution. Some carbon or nitrogen can also be trapped in the iron, or a reversible action can occur. The net result is a degradation of the structure-sensitive magnetic properties over a period of time (aging). If the alloy is cooled rapidly after sintering, the carbon or nitrogen is trapped in solution in the iron, which in turn strains the iron lattice, causing degradation of the structure-sensitive magnetic properties. This latter effect is small, but it occurs nonetheless.

References cited in this section

1. R. German, Powder Metallurgy of Iron and Steel, John Wiley & Sons, 1998 2. C. Lall, Fundamentals of Magnetism, Chapter I, Soft Magnetism, Metal Powder Industries Federation, 1992, p 1-27 3. Magnetic Properties, Vol 03.04, Annual Book of ASTM Standards, ASTM 4. W.F. Smith, Magnetic Materials, Chapter 11, Principles of Material Science and Engineering, McGraw-Hill, 1986, p 604 Magnetic Materials and Properties for Powder Metallurgy Part Applications Kenneth H. Moyer, Magna-Tech P/M Labs

Magnetic P/M Materials Most P/M materials are designed for soft magnetic properties with little or no retentivity--that is, if they are magnetized in a magnetic field and then are removed from that field, they lose most, if not all, of the magnetism they exhibited while in the field. Magnetically soft materials that are produced in large quantities include high-purity iron, low-carbon steels, silicon steels, iron-nickel alloys, iron-cobalt alloys, and soft-magnetic ferrites. In contrast, permanent magnets are normally used in a single magnetic state and have stable magnetic fields. Permanent magnet materials include a variety of metals, intermetallics, and ceramics. Commonly included are certain steels, Alnico, CuNiFe, Fe-Co alloys containing V or Mo, Pt-Co, hard ferrites, and cobalt-rare earth alloys. Only a few P/M alloys are useful as hard magnetic materials. In this regard, the major alloy is Fe-Nd-B rapidly solidified powders. This section briefly describes the common soft magnetic P/M alloys including high-purity iron Fe-2Ni, Fe-3Si, Fe-0.45P, Fe-0.6P, and Fe-50Ni. Table 3 compares the typical magnetic properties of P/M compositions. Values in Table 3 are only typical as considerable variability occurs in these properties, especially the magnetic force needed to demagnetize an alloy. Part of this variation is due to differences in sintered density, but impurity control is also a major factor. Like strength, the magnetic characteristics are sensitive to contaminants.

Table 3 Typical Magnetic properties of P/M alloys Alloy

Resistivity, · cm

410L 430L 434L Fe

90 93 104 20

Maximum magnetization (Bmax), T 1.2 1.1 1.1 1.6

Remanence magnetization (Br,) T 4.1 4.8 5.7 1.2

Coercive force, Oe 3.5 2.7 2.3 1.4

Maximum permeability 960 630 1,300 4,000

Fe-49Co-2V Fe-50Co Fe-5Mo Fe-2Ni Fe-50Ni Fe-0.45P Fe-0.8P Fe-3Si Fe-6.5Si

40 60 ... 15 40 21 23 45 81

2.0 1.7 1.6 1.5 1.3 1.4 1.8 1.4 1.3

5.6 0.56 1.3 1.2 0.9 1.2 1.3 1.3 1.2

3.0 2.0 1.8 1.0 0.3 1.1 0.4 0.9 0.3

3,000 2,000 2,900 ... 25,000 4,000 6,100 4,500 4,000

Except for 400 series stainless steels at approximately 6.9 g/cm3, all others are high-density products over 7.4 g/cm3. Source: Ref 1 Pure Iron. Iron powder can be produced in many forms. All are magnetic, some more so than others. In early applications, sponge iron was used to produce magnets. The magnetic properties of these early magnets were not as good as those of current magnets because the sponge iron magnets contained as much as 2 wt% insolubles, generally in the form of oxides. Besides being impure, the sponge iron powder was also harder and therefore could not be compacted to densities above 6.2 g/cm3. Therefore, magnetic induction was limited, owing to the low sintered density. The structuresensitive properties--permeability and coercive field--were limited by the impurities. If improved magnetic properties were required, either carbonyl iron or electrolytic iron was selected, at a great increase in cost of the powder.

Today, inexpensive water-atomized iron powder is readily available. This powder has 1260 °C (>2300 °F), and they are held at temperature for long periods of time, maximum permeabilities of 350,000 and coercive fields as low as 0.8 A/m (0.01 Oe) are possible. The control is dependent on both grain growth and impurities within the material. Iron is selected to produce parts because of the dimensional control that is realized when iron is sintered at 1120 °C (2050 °F) in standard belt furnaces for 30 min. These parts do not have the best magnetic properties but are adequate as inexpensive flux paths for magnetic devices. In reality, the P/M iron parts serve as inexpensive replacements for carbon steel parts. Properties are limited at these sintering temperatures because grain growth and sintered densities are limited by the limited diffusion at these sintering temperatures. Phosphorus Irons. Although the attributes of alloying iron with phosphorus formerly were well known, wrought

products, when hot rolled, became embrittled when phosphorus was alloyed with the iron in significant amounts. However, because phosphorus is normally admixed with water-atomized iron powder as 10 m Fe2P or Fe3P powder, P/M technology can take advantage of phosphorus additions to iron. When compacted, the compressibility of the soft iron is realized without loss of green density because the hard, small 10 m Fe2P or Fe3P particles rearrange themselves among the larger iron particles as compaction occurs. When sintered, the phosphorus iron intermetallic compounds melt at a temperature of 1050 °C (1920 °F) and diffuse into the iron, forming a solid solution of phosphorus in iron (Fig. 3). The liquid phase enhances diffusion rates and assists in the rearrangement of the pores and particle boundaries, thereby further densifying the iron. If a normal belt furnace sinter of 1120 °C (2050 °F) for 30 min is used, then sintered densities as high as 7.2 g/cm3 can be realized (Table 3). If the sintering temperature is increased to 1260 °C (2300 °F), and the parts are sintered for 1 h in a partial pressure of hydrogen, a sintered density as high as 7.6 g/cm3 can be realized. Because magnetic induction is a linear function of the sintered density, the magnetic induction realized from the phosphorus iron alloys is higher than that realized from pure iron parts. However, the drawback to the use of phosphorus iron parts is that, owing to the liquid phase sintering that occurs, dimensional control is more difficult. When sintered at 1120 °C (2050 °F) in belt furnaces, dimensional control is not too difficult because the normal phosphorus addition of 0.45 wt% provides what is commonly known as a zero-growth part. However, if the phosphorus content is increased to 0.8 wt% or if hightemperature sintering is specified to improve density and structure-sensitive magnetic properties, shrinkage becomes

greater and dimensions are more difficult to control. The problems with dimensional control and added cost of hightemperature sintering generally cause a compromise to be made between achievable magnetic properties and commercial manufacturing practice.

Fig. 3 Iron-phosphorous phase diagram. Source: Ref 5

From the iron-phosphorus binary phase diagram in Fig. 3, it can be seen that the maximum solubility of phosphorus in iron is 1.5 wt%. Theoretically, phosphorus can be alloyed to this wt%; however in P/M technology, phosphorus additions to 2 wt% have been investigated (Ref 6). Increased alloying of phosphorus in iron results in several advantages to improve magnetic properties. Resistivity is one physical property that is improved by alloying phosphorus with iron. If 0.45 wt% P is alloyed, resistivity is approximately doubled. If 0.8 wt% P is alloyed, the resistivity is approximately 24 m · cm. Higher resistivity is required to minimize eddy current losses at low levels so that heat generation will not destroy the magnetic device. The low resistivities of the phosphorus irons do not enable these alloys to be used for most alternating-current applications. However, if 0.8 wt% P is used, then the phosphorus irons can be considered for some pulsed direct-current applications.

It is therefore readily understandable that phosphorus irons have become the workhorse for specification of material for magnetic parts. Depending on impurity content, sintering conditions, and heat treatment, these alloys can compete with most other P/M magnetic materials for demanding devices that require improved performance for direct current, and for many devices that require a pulsed direct current for application. Owing to low resistivity and the inability to produce thin sheet for laminations, the alloys are not normally considered for alternating-current applications. Typical magnetic properties of the two phosphorus iron alloys considered for current commercial applications are shown in Table 4.

Table 4 Typical-direct-current magnetic properties of phosphorous iron alloys Property

Sintered density, g/cm3 Induction, for a field of 1200 A/m(15 Oe), T (kG) Residual induction, T (kG) Maximum relative permeability Coercive force, A/m (Oe)

Compacting pressure 0.45% P compacted at pressures of: 410 MPa 550 MPa 685 MPa (30 tsi) (40 tsi) (50 tsi) 6.80 7.10 7.23 1.06 1.19 1.27 (10.6) (11.9) (12.7) 0.87 0.99 1.08 (8.7) (9.9) (10.8) 2400 2800 3100 135 135 127 (1.7) (1.7) (1.6)

0.80% P compacted at pressures of: 410 MPa 550 MPa 685 MPa (30 tsi) (40 tsi) (50 tsi) 6.88 7.12 7.28 1.12 1.23 1.31 (11.2) (12.3) (13.1) 1.01 1.13 1.20 (10.1) (11.3) (12.0) 3680 4240 4640 112 112 112 (1.4) (1.4) (1.4)

Source: Ref 3

It is possible to enhance structure-sensitive magnetic properties (permeability and coercive field). If a 0.8 wt% alloy is sintered at 1260 °C (2300 °F) for 1 h in a partial pressure of hydrogen, maximum permeabilities of 15,000 and coercive fields as low as 55 A/m (0.7 Oe), are possible (see Moyer, 1990 in "Selected References" ). These values are comparable to those of commercial 3 wt% silicon iron. In reality, even more improved structure-sensitive properties are possible if impurity products, such as carbon, nitrogen, oxygen, and sulfur are maintained at low levels, and heat treatment is provided to grow grains. However, practical difficulties (cost, time, dimensional control) prevent industrial application. Ferritic Stainless Steels. The automotive market consumes approximately three-fourths of the P/M parts that are fabricated today. This statement is also true for direct-current magnetic parts. As newer applications emerge, magnetic part applications can involve higher operating temperatures and require improved corrosion resistance. The optimal material selection for these applications are the ferritic stainless steels. Although corrosion resistance is enhanced by this selection, the magnetic properties are sacrificed at the expense of this requirement.

Four ferritic stainless steels (409L, 410L, 434L, and 430L), are commercially available for magnetic applications. The 410L powder nominally contains 12 wt% Cr as an alloying addition. The 434L powder contains nominally 17 wt% Cr and 2 wt% Mo. The latter alloy is more corrosion resistant but more expensive. In addition, because magnetic induction depends on the quantity of iron present in the alloy, the magnetic induction of the ferritic stainless steels is lower than that of either iron or the phosphorus irons. Most stainless steel powders are generally water atomized, although gas-atomized powders are used for parts made by the MIM process. Because of the solid-solution strengthening of chromium in the iron matrix, the powder particles are harder than iron powder, which limits compressibility. Therefore, sintered densities >7.2 g/cm3 are not common. The lower densities also limit the magnetic induction of these alloys. Present state-of-the-art production generally requires a high-temperature sinter for these alloys to attain their best level of performance magnetically. For highest magnetic properties, the sintering temperature is normally 1260 °C (2300 °F) or greater for at least 1 h in either vacuum or hydrogen (Table 5). This sintering temperature is required because chromium oxides present on the surfaces of these water-atomized powders do not readily reduce at temperatures lower than 1260 °C (2300 °F).

Table 5 Magnetic data on ferritic stainless steels Steel type

Wrought 430F, Bm at H = 796 A/m, Bs = 1.42 T Powder metal 410L, H = 1990 A/m: 1260 °C, vacuum 45 min 1288 °C, H2 430L H = 1194 A/m, 1121 °C, H2, 30 min H = 1194 A/m, 1121 °C DA, 30 min

H = 1194 A/m, 1260 °C, H2 H = 1194 A/m (vacuum + backfill of listed gas): 1121 °C, H2 1232 °C, H2 1121 °C, H2 + N2 1232 °C, H2 + N2 434L H = 1990 A/m: 1260 °C, vacuum 45 min 1288 °C, H2 H = 1194 A/m: 1288 °C, DA, 30 min 1288 °C, H2, 30 min H = 1194 A/m: 1121 °C, DA, 30 min

1288 °C H = 1194 A/m, 1232 °C, H2

Density, kg/m3

Maximum induction (Bm), T

Residual induction (Br), T

Relative permeability,

7620

1.2

0.6

2000

159

7100 7100

1.09 5.6

0.94 ...

2200 320

169 590

6450 6670 5810 6130 6420 7250

0.73 0.79 0.034 0.041 0.045 1.05

0.47 0.51 0.003 0.005 0.005 0.80

1000 1043 11 13 14 1900

182 185 231 294 279 159

6690 7110 6490 6990

0.81 0.98 0.016 0.038

0.74 0.89 0.009 0.024

1200 1800 ... ...

207 167 740 669

7000 7100

1.01 6.5

0.84 ...

1700 450

159 220

6430 6650 5830 6030 6290 7100 7350

0.728 0.791 0.053 0.063 0.079 0.89 0.97

0.463 0.483 0.009 0.013 0.019 0.44 0.77

1092 1165 416.8 719.7 424.9 1275 1600

160 151 414 477 533 119 143

r

Coercive field (Hc),A/m

DA, dissociated ammonia. Source: Ref 6

High-temperature sintering presents some problems in addition to increased cost. If ferritic stainless steel parts are hightemperature sintered, shrinkage of at least 1% is realized. These high shrinkage rates again make dimensional control difficult. Some parts fabricators therefore elect to sinter at either 1205 °C (2200 °F) or at 1230 °C (2250 °F) to reduce the problem of dimensional control. More recently, parts fabricators supplying parts to the automotive industry have even used belt furnaces to sinter ferritic stainless steels at 1150 °C (2100 °F). In all cases, lower temperatures result in lower sintered densities and higher oxygen contamination of the parts. This being the case, magnetic properties and corrosion resistance are sacrificed. In addition to compromises with respect to sintering temperature, many parts fabricators elect to sinter in dissociated ammonia, nitrogen/hydrogen gas mixtures, nitrogen, exothermic, or endothermic gas. These atmospheres are rich in nitrogen, carbon, or both. Chromium in the ferritic stainless steel readily reacts with the components of these atmospheres to form carbides or nitrides in grain boundaries, thereby depleting the chromium protective film. These carbides or nitrides lower structure-sensitive magnetic properties and corrosion resistance and also decrease ductility. 50 Nickel/50 Iron. Nickel irons are usually the alloys selected when exceptional structure-sensitive properties are

required. The alloy composition generally selected is 50 wt% Ni combined in solid solution with 50 wt% Fe. Although this alloy is most predominantly selected, nickel/iron combinations of all ranges have been selected for magnetic applications. Nickel irons are the most expensive of all the commonly used magnetic alloys today. The nickel irons are selected when fast response is required in magnetic properties. Fe-50Ni has a high maximum permeability (Table 3).

Because structure-sensitive magnetic properties are generally required, these alloys are normally high-temperature sintered at 1260 °C (2300 °F) or higher, in either vacuum or hydrogen, for at least 1 h (Table 6). These more demanding sintering conditions are necessary to promote diffusion of the substitional elements (iron, nickel) and to aid grain growth and purification (removal of carbon, nitrogen, oxygen, and sulfur). If the above conditions are met, relative maximum permeability as high as 10,000, and a coercive field as low as 15 A/m (0.2 Oe) are possible in the as-sintered condition.

Table 6 Magnetic data on 50 nickel/50 iron Density, kg/m3

Treatment

Wrought ... Powder metal H = 1990 A/m, 1260 °C, DA, 45 min(a)

H = 1990 A/m, 1260 °C, vacuum At temperature of: 1121 °C 1288 °C H = 1194 A/m: 1246 °C, vacuum 2 h 1246 °C, vacuum 2 h

(a)

Maximum induction (Bm),T

Residual induction (Br),T

Relative permeability,

8200

1.60

0.80

70,000

3.98

6800 7100 7500 7400

0.93 1.09 1.27 1.08

0.71 0.80 0.94 0.86

... ... 21,000 10,600

20.7 19.9 19.1 23.9

7100 7300

1.15 1.29

0.85 0.90

11,000 16,000

31.4 23.9

7300 7500

1.12 1.23

0.70 0.75

... ...

23.9 23.9

Resistivity 0.78, 0.69, 0.60 dissociated ammonia.

Coercive field (Hc),A/m

r

·m at 6800, 7100, and 7500 kg/m3, respectively. DA,

These alloys, however, are sensitive to heat treatment. Generally, if annealing is required, the wrought practice of heating to a temperature of 1120 to 1175 °C (2050 to 2150 °F) in dry hydrogen for 4 h is employed. Using this practice, the relative maximum permeability can be increased to 21,000; however, the coercive field is not changed. It must be stressed at this point that further work on heat treatment is needed to improve the structure-sensitive magnetic properties of the P/M alloy and make them comparable to the wrought product. Typical magnetic properties of the 50 nickel/50 iron alloy are shown in Table 6. Silicon Irons. Wrought silicon irons have been selected for magnetic applications since the 1940s. These alloys

normally contain 1.5 to 3 wt% Si. They are used primarily for motors, generators, and relays. Alloys containing from 3 to 5 wt% have found service in high-efficiency motors and in transformers. Barrett, Brown, and Hadfield in 1900 published a paper which stated that 2 to 2.5 wt% Si decreased the coercive field of the alloy to half that of iron. Silicon steels were first produced in Germany for commercial applications in 1903. Commercial silicon iron was also produced in the United States in the same year. Its features included increased permeability, lower hysteresis loss, lower eddy current losses owing to higher resistivity, and no aging. The solubility of silicon in ferrite extends to 16 mol% (Fig. 4). Although the solubility is high, providing an advantage of high resistivity, most wrought alloys are limited to 3 wt% Si because of the difficulty of deforming alloys containing greater additions of silicon.

Fig. 4 Iron-silicon phase diagram. Source: Ref 5

Although P/M magnetic parts had been fabricated commercially from 3 wt% silicon iron since the 1950s, only a few parts fabricators considered producing these parts. First, the silicon irons are made by admixing hard intermetallic compounds with a soft base iron matrix. In the case of the silicon irons, sufficient master alloy is added to yield 3 wt% Si upon sintering. Master alloys containing 17% Si can cause significant die wear. In addition to an increased die wear problem, problems in sintering these alloys are also encountered. Silicon is even more readily oxidized than chromium. Therefore, high-temperature sintering is a must. If sintering temperatures below 1205 °C (2200 °F) are employed, the oxide film encasing the silicon master alloy is not reduced, and the master alloy particles do not diffuse into the ferrite, but they remain as small spheres within the basic ferrite matrix (Ref 8). A liquid phase develops (Fig. 4) below 1260 °C (2300 °F); therefore if a hydrogen atmosphere is employed as the protective atmosphere, the oxide film is reduced, and a proper sinter can be provided between 1260 and 1315 °C (2300 and 2400 °F). Because there is a large concentration of the master alloy present in the parts, sintering is rapid, and there is a danger of melting occurring if the sintering temperature selected approaches 1315 °C (2400 °F). Owing to the large amount of liquid phase that occurs, dimensional control is also more difficult and sizing may be required. Provided that the parts fabricator is aware of the practical difficulties of processing silicon irons, enhanced magnetic properties can be realized from these alloys at reasonable cost. In addition to the above-mentioned disadvantages, the magnetic saturation of the silicon irons is lower than that of iron. Iron has a saturation magnetization of 2.15 T (21,500 G)

compared to 1.95 T (19,500 G) for Fe-3 wt% Si. Therefore, iron or phosphorus iron is a better choice for applications such as flux paths. However, the resistivity of a 3 wt% silicon iron is 65 cm, which is five times that of pure iron. High resistivity lowers hysteresis and eddy current losses. Therefore, if a pulsed direct current or a 60 Hz application is specified, the 3 wt% silicon iron is a logical choice for this application, provided that the part is not too thick. The only commercial P/M material with better structure-sensitive magnetic properties is 50 nickel/50 iron, but this is much more expensive. Typical magnetic properties of the silicon-iron alloy are shown in Fig. 5. Work continues to develop alloys with increasing silicon content for commercial applications.

Fig. 5 Magnetic properties of Fe-3%Si as a function of sintering temperature in dissociated ammonia. Source: Ref 7

References cited in this section

1. R. German, Powder Metallurgy of Iron and Steel, John Wiley & Sons, 1998 3. Magnetic Properties, Vol 03.04, Annual Book of ASTM Standards, ASTM 5. D.T. Hawkins and R. Hultgren, Constitution of Binary Alloys, Metallography, Structures and Phase Diagrams, Vol 8, Metals Handbook, 8th ed., American Society for Metals, 1973, p 251-376 6. B. Weglinski and J. Kaczmar, Effect of Fe3P Addition on Magnetic Properties and Structure of Sintered Iron, Powder Metall., Vol 23 (No. 4), 1980, p 210-216 7. C. Lall, Soft Magnetic Properties of Selected Alloys, Chapter III, Soft Magnetism, Metal Powder Industries Federation, 1992, p 49-87 8. L.W. Baum, Jr., "Theoretical and Practical Considerations for P/M Production of Magnetic Parts," presented at P/M Technical Conference (Philadelphia), Hoeganaes Corporation, Oct 1978

Magnetic Materials and Properties for Powder Metallurgy Part Applications Kenneth H. Moyer, Magna-Tech P/M Labs

Metal Injection Molding Metal injection molding (MIM) is a process that has recently been developed to produce parts with complex geometry. Gas-atomized powders are normally selected to fabricate these parts because they are purer and finer. Iron and nickel powders produced by the carbonyl process are also widely selected. Gas-atomized and carbonyl powders are considerably more expensive than water-atomized powder. To date, the expense has restricted parts made by the injection molding process to those of a size less than that of a golf ball. However, as developments in this technology progress, there should be less restriction on the size of parts that can be produced using MIM. One part already being developed using MIM technology is a stainless steel magnetic fuel injector. There is no restriction of the materials that can be selected to be metal injection molded, provided they are magnetic (alloys containing as base iron, nickel, or cobalt). Therefore, parts for magnetic applications have been made from all of the alloy systems discussed above. In addition, there is no restriction that limits development of new alloy systems for magnetic applications. Normally, magnetic parts are produced from fine (10 m) pure gas-atomized powders. These fine powders are mixed with a system of polymers to form a mixture with the texture of toothpaste when heated. The feedstock is then heated and injected under pressure into a die cavity that is of the geometry of the desired part. The molded part is heated to remove the polymers and then sintered. Some processing techniques may use a solvent to extract the binder prior to sintering. Nonetheless, MIM is not an easy process, and more detailed information is contained in the article "Powder Injection Molding" in this Volume. For magnetic parts, much is to be gained by selecting MIM as the processing technology, provided that the increased cost can be justified. Because fine, pure powders and high-temperature sintering are normally considered in conjunction with MIM processing, densities approaching theoretical density are possible. Therefore, magnetic inductions close to saturation magnetic induction are possible. This means that magnetic inductions equivalent to those of wrought alloys are possible for alloys of similar composition. In addition, new alloys can be developed to take advantage of additional alloying elements that cannot be considered using wrought fabrication technology. Although metal injection molded parts shrink much more than conventional P/M parts, dimensional control becomes less of a problem because the shrinkage is more uniform. The uniform shrinkage results because the part has uniform stress after injection and, if binder removal is done correctly, rearrangement of the particles occurs. Coupled with enhanced diffusion across pure reduced particle surfaces, this rearrangement results in a high-density part of more uniform dimensions. The purity of the fine gas-atomized powders that are employed is also helpful in maintaining improved structure-sensitive magnetic properties. If the binder is removed properly and if the parts are sintered correctly, the relative maximum permeability and coercive field should equal those of the best processed conventional P/M parts. An added benefit is the magnetic induction should be equivalent to that of wrought parts, because the part density is so high. A sampling of magnetic properties that are possible from MIM technology is shown in Table 7.

Table 7 Magnetic data on metal injection molded alloys Alloy

Density, kg/m3

Fe-2% Ni, 1316 °C Fe-50% Ni, 1316 °C Fe-3% Si, 1316 °C Then, H2 annealed Fe-3% Si, 1232 °C Fe-6% Si, 1316 °C Then, H2 annealed Fe, 1371 °C, DA Fe, 1371 °C, DA Fe-3% Si, 1371 °C, DA Fe-3% Si, 1371 °C, DA Fe-6% Si, 1371 °C, DA Fe-50% Ni, 1371 °C, DA 430L stainless steel, 1371 °C, vacuum Fe-3% Si

7670, 97% 7660, 94% 7550, 98% 7550, 98% 7540, 98% 7540, 99% 7410, 98% 7600, 97% 7550, 96% 7550, 98% 7550, 98% 7420, 99% 7660, 93% 7400, 95% ...

Fe-50% Ni

...

430L stainless steel

...

Maximum induction (Bm), T 1.51 1.27 1.50 1.47 1.44 1.32 1.37 1.53 1.55 1.45 1.50 1.37 1.27 1.15 1.48 1.46 1.22 1.04 0.85

Residual induction (Br), T 1.29 0.42 1.21 1.14 0.62 1.12 1.22 1.37 1.34 1.07 1.21 1.21 0.42 0.54 0.88 1.10 0.53 0.64 0.22

Coercive field (Hc), A/m 82 16 45 47 60 46 46 294 183 57 51 40 16 202 62 56 16 19 119

DA, dissociated ammonia. Source: Ref 7

Reference cited in this section

7. C. Lall, Soft Magnetic Properties of Selected Alloys, Chapter III, Soft Magnetism, Metal Powder Industries Federation, 1992, p 49-87 Magnetic Materials and Properties for Powder Metallurgy Part Applications Kenneth H. Moyer, Magna-Tech P/M Labs

Optimizing Magnetic Properties Powders. Most magnetic parts are fabricated from water-atomized metal powders. The major exception to this is metal

injection molded parts, where gas-atomized powders are selected, as a rule. Because purity is important, carbon, nitrogen, and sulfur are maintained at very low levels if structure-sensitive magnetic properties are to be optimized. The major contaminant in the powders is oxygen in the form of surface oxides. In the case of ferritic stainless steel powders, as much as 2000 ppm of oxygen is concentrated on the surface of the water-atomized powders. Base iron powder normally contains 1000 ppm or less of oxygen. Gas-atomized powders contain even less oxygen. Nevertheless, these surface oxides must be readily reduced to provide a good sinter of the part. The water-atomized powders are generally classified as -100 mesh powders and have a normal distribution. The gasatomized powders, or the admixed additions, are normally 10 m in size. All additions must also be as free as possible of carbon, nitrogen, oxygen, and sulfur. Ferritic stainless steels and the 50 nickel/50 iron powders are prealloyed. In the case of the phosphorus irons and the silicon steels, alloying is accomplished by admixing the alloying components. For the phosphorus irons, generally a 10 m Fe3P powder is admixed with the base iron powder, although a Fe2P powder can also be selected as the alloying addition. Some practitioners have found, however, that Fe2P powder is more abrasive and contributes to increased die wear. For the silicon irons, a master alloy containing 17 wt% Si is generally selected as the alloying addition.

Compaction. Generally, parts are compacted to the highest green density possible because the magnetic induction is

dependent on the sintered density. This is a linear function; the higher the density, the higher the magnetic induction. For flux paths, the higher the magnetic induction, the greater the amount of flux that can be carried in a given cross section. The limitation that normally controls the green density that can be achieved is the compaction pressure. Normally, most parts fabricators do not wish to use compaction pressures 620 MPa (>45 tsi) because tool breakage becomes a problem. High compacting pressures (1000 MPa) do not improve the magnetic properties of Fe-Si alloys (Ref 9). For certain applications with Fe-Si alloys, it is recommended to combine low compacting pressures and long sintering time (24 h) or medium compacting pressures (600 MPa) and short sintering time (2 h) (Ref 10). The density of complex or long parts also can vary appreciably within the cross section of the part. It is therefore important that the parts fabricator and the user agree on the critical density of the section of the part that is to be fabricated. Generally, green densities of 6.8 to 7.2 g/cm3 are desired. Sintering. Of all the variables to consider in selection of the sintering conditions, the selection and control of the

sintering atmosphere is the most critical. For example, dissociated ammonia is often used instead of hydrogen for cost reasons. This may introduce a contamination problem from the nitrogen in dissociated ammonia. When parts are sintered in dissociated ammonia, structure-sensitive magnetic properties can be seriously degraded (Table 8). In this example, rings fabricated from 0.45 wt% phosphorus iron were sintered in hydrogen, and companion rings were sintered in dissociated ammonia, at 1120 °C (2050 °F). As can be seen, contamination levels of both sets of parts are equivalent, except for nitrogen and oxygen contamination. Note that the sintered density of the parts sintered in hydrogen was higher. As a result, the magnetic induction of the parts sintered in hydrogen was also higher. The major changes in properties are seen in the structure-sensitive magnetic properties, relative maximum permeability, and coercive field. The relative maximum permeability of the parts sintered in hydrogen was doubled, and the coercive field was significantly lower. Minimization of carbon, nitrogen, oxygen, and sulfur is a must if structure-sensitive magnetic properties are to be achieved. Even if the proper sintering temperature is selected, magnetic properties will be severely degraded if the sintering atmosphere selection is incorrect.

Table 8 Effect of sintering atmosphere on magnetic properties Property

Carbon, % Nitrogen, % Oxygen, % Sulfur, % Density, g/cm3 Induction at 15 Oe, kG Residual induction, kG Maximum permeability

0.45 mol % phosphorus iron 2050 °F 75H2/25N2 H2 800 °C, or 1470 °F) to fill the airbag. Besides cooling and diffusing the gas flow, the filter prevents hot particles as small as 1 m (0.00004 in.) in diameter, and other undesirable respirables, from entering the airbag. The sintered media can be sheared, rolled, pleated, and/or welded without cracking, splitting, or destroying its porous properties. Extremely fine metallic fibers fabricated through a bundle drawing process provide an integral, controlledporosity medium. The stainless steel fibers have diameters as fine as 4 m (0.00016 in.) in a sintered web matrix combined with a stainless wire mesh. The sintered media strips are provided in a preformed, crimped condition. This facilitates handling by the customer, who flanges and roll-forms the strips to various diameters, and then welds the overlap seam. The part is made by Memtec America Corp. 1997 Parts Competition Warm-Compacted Turbine Hub. The first automotive powertrain application fabricated through warm compaction

technology is a steel torque converter turbine hub (Fig. 14). The hub is formed by means of the EI-Temp warm compaction system from powder in an 825 ton (7.3 MN) compacting press.

Fig. 14 1997 P/M Part-of-the-Year Competition winners. Left to right, foreground: Latchbolt and seven gun parts. Left to right, background: Exhaust manifold flanges, turbine hub, and a powder tool ratchet wheel set, which includes the wheel, flange, and thrust ring

The part has a minimum density of 7.2 g/cm3 (0.26 lb/in.3), a tensile strength of 807 MPa (117 ksi), a yield strength of 428 MPa (62 ksi), a fatigue limit of 242 MPa (35 ksi), and a 17 HRC apparent hardness.

Subjected to severe durability testing, it must withstand an internal spline torque of 1210 N · m (890 ft · lb) for one million cycles. Powder metallurgy replaced a machined and heat-treated forged part. It provided a cost savings of more than 30% and an assembly savings of more than 50%. The part is made by Chicago Powdered Metal Products. 3

3

Brass latchbolt (Fig. 14) is made to a density of 7.9 g/cm (0.28 lb/in. ) and has an ultimate tensile strength of 210

MPa (30 ksi). Its yield strength is 110 MPa (16 ksi), elongation is 15%, and hardness is 84 HRH. The latchbolt is assembled with mating parts by crimping the tail section, requiring strength and ductility in the thinwalled areas. Secondary operations are limited to drilling a 2.4 mm (0.94 in.) hole, tumbling, and resin impregnation. A contour tolerance of 0.15 mm (0.006 in.) must be held. The art must pass a stringent pendulum ram test where ram blows are applied to the lock. The part is made by ASCO Sintering Company. Automotive exhaust manifold flanges (Fig. 14) are made of stainless steel alloy 434. The manifold flange

connects the manifold to the exhaust pipe that leads to the catalytic converter. The exhaust converter outlet flange connects the exhaust pipe leading from the catalytic converter to a flange welded to the manifold. The three-level manifold flange (6.9 g/cm3, or 0.24 lb/in.3) required special tooling to form the "tulip" shape. Powder metallurgy replaced a two-piece stamped and welded assembly, which leaked exhaust fumes. The P/M part dramatically reduces exhaust leakage at the manifold/exhaust system junction. Alloy 434 P/M stainless steel has tensile properties superior to wrought 409 stainless steel at elevated temperatures, as well as improved corrosion and oxidation properties. Formed to a minimum density of 7.0 g/cm3 (0.25 lb/in.3), the exhaust converter outlet flange is supplied with two cold-headed wrought bolts, which are press-fitted into it and shipped as an assembly. Deburring is the only secondary operation performed on both parts. The components are made by SSI Technologies Inc. Complex MIM gun parts (Fig. 14) are used for seven parts that go into .22 and .38 caliber revolvers. The frame,

weighing 240 g (8.46 oz), is made from nickel-steel and has a minimum hardness of 25 to 30 HRC. The other parts are made from 4140 steel. All parts have a density of 7.6 g/cm3 (0.27 lb/in.3). The barrel must withstand heavy shock and wear. Metal injection molding technology replaced investment casting and forging to achieve better finishes and closer tolerances at more than 40% cost savings. Secondary operations include heat treating and tapping, threading and rifling the barrel, and applying a black oxide or chrome finish. Testing of the parts includes firing 1000 overload rounds. The parts are made by Taurus International Manufacturing Inc. Warm compacted ratchet-wheel set for a new angle grinder (Fig. 14) is made to a density of more than 7.0 g/cm

3

(0.25 lb/in.3), has an ultimate tensile strength of 760 MPa (110 ksi), impact strength of 22 J (16 ft · lb) and a fatigue strength of 300 MPa (40 ksi). The warm compacted thrust ring and flange are formed to a density exceeding 7.2 g/cm3 (0.26 lb/in.3) and have an ultimate tensile strength of 1100 MPa (160 ksi), an impact strength of 24 J (18 ft · lb), and a fatigue strength of 414 MPa (60 ksi). Powder metallurgy provided almost 50% cost savings and very high precision. Secondary operations include heat treating and minor machining on the wheel; the thrust ring is machined, and the flange is heat treated. The part is made by Porite Taiwan Co. Ltd. Clutch Plates. P/M steel notch plate and pocket plate (Fig. 15) is for a new one-way clutch that revolutionizes

powertrain/drivetrain designs. The P/M plates offer significant cost savings over alternative manufacturing processes. The new clutch is more economical than conventional "sprag and roller ramp" designs, and is less sensitive to vibration, unbalanced loading, and centrifugal force.

Fig. 15 1997 Award of distinction winners. Left to right, foreground: Five stainless steel mortise lock parts; three opto holders; and metal injection molded latch door, gear, and racks. Center: Metal injection molded link, receiver, and clevis; and padlock body. Background: Notch plate and pocket plate

The two-level notch plate has a minimum density of 7.0 g/cm3 (0.25 lb/in.3), a minimum ultimate tensile strength of 720 MPa (105 ksi), and a minimum apparent hardness in the range 20 to 25 HRC. The three-level pocket plate has a minimum density of 6.9 g/cm3 (0.24 lb/in.3) and a minimum ultimate tensile strength of 900 MPa (130 ksi). Both parts are heat treated, machined, and tumbled. A Stroker load cycle and overrun test are performed on the assembly. The impact load specification is 20 cycles at 1580 J (1200 ft · lb), and the Stroker requirement is 200,000 cycles at 790 J (600 ft · lb) and 14,000 cycles at 1580 J (1200 ft · lb). The P/M clutch also withstood a maximum ultimate torque of 3950 J (3000 ft · lb), more than six times the torque that the more conventional roller ramp clutch could withstand. The P/M notch and pocket plates are assembled with four struts, four springs, and a threaded weir to form the one-way clutch. The design is patented by Epilogics Inc., and the part is made by Metal Powder Components. Brass optical sensing part (Fig. 15) made by FMS Corp. holds a photoelectric emitter and receiver in place. The part

has thin walls and four pyramid-shaped "locking ribs" located on the center bridge sections. These protrusions allow the part to be staked into a plastic molded frame. The part is repressed and sized after sintering, then resin-impregnated, deburred, and tin plated. It has a minimum yield strength of 80 MPa (12 ksi) and a density of 8.0 g/cm3 (0.30 lb/in.3). The 0.635 mm (0.025 in.) wall thickness required high-precision tooling for compacting and sizing. A process capability index, Cpk, of 1.33 is required on critical dimensions, which include 1.3 mm ±0.13 mm (0.050 ±0.005 in.) on bridge thickness and 2 mm +0.10 mm/-0.05 mm (0.080 + 0.004/-0.002 in.) on the distance from peak-to-peak on the gripping ribs. 3

3

Mortise lock parts (Fig. 15) are 304 stainless steel with a minimum density of 6.4 g/cm (0.23 lb/in. ). The stainless

steel parts have a 61 HRB hardness, a yield strength of 207 MPa (30 ksi), and an ultimate tensile strength of 296 MPa (43 ksi). Brass forging, stainless casting, and zinc die casting were also considered, but could not approach the economies of powder metallurgy. The part also provides high hardness, wear resistance, and net shape. The new mortise lock design can change function and handling in the field without disassembly and extra parts, which means the lock can be installed on a door regardless of the way it is swung. The lock also contains five copper-infiltrated P/M steel parts. The parts are made by Ceromet. Three intricate 17-4PH stainless steel parts (Fig. 15) are metal injection molded parts for a new endoscopic

surgical stapler. The thin-walled parts are produced very close to their final shape. The heat-treated parts have an ultimate tensile strength of 1180 MPa (170 ksi), a yield strength of 1100 MPa (160 ksi), a 36 HRC hardness, a 3% elongation, and a 7.5 g/cm3 (0.26 lb/in.3) density. The receiver has a 0.7 mm (0.028 in.) wall section, which forms a slot held to between 8.8 and 8.9 mm (0.34 to 0.35 in.) over the entire length. The clevis holds a roundness of 11.7 mm ±0.05 mm (0.460 in. ±0.002 in.) on a 0.864 mm (0.034 in.) wall section. The link holes are held to 1.19 mm ±0.019 mm (0.047 ±0.00075 in.). The endoscopic linear cutter combines a 45° bilateral articulation with a 360°

rotation, providing excellent maneuverability and tissue access for endoscopic procedures, such as appendectomies and lung resections. The parts are made by PC Advanced Forming Technology. MIM Parts for Locking Mechanism. A gear, rack, and latch door (Fig. 15) are used as a locking door unit gear mechanism for closing small battery doors in a portable terminal. The gear and latch door are made from a P/M nickel steel to a density of 7.75 g/cm3 (0.27 lb/in.3). They have a tensile strength range of 414 to 448 MPa (60 to 65 ksi), a yield strength of 276 to 310 MPa (40 to 45 ksi), and a 65 to 75 HRB hardness range.

The rack is made from 316L stainless steel to a density of 7.70 g/cm3 (0.27 lb/in.3) and has a tensile strength range of 470 to 515 MPa (68 to 75 ksi), a yield strength range of 150 to 172 MPa (22 to 25 ksi), and a 65 HRB hardness. Originally, the parts were produced from plastic and aluminum, but they failed. Metal injection molding provided a more reliable design and a cost savings of up to 50% compared to machining. The parts are made by Metalor 2000. A net-shape steel padlock body

(Fig. 15) is used in a gear-shift lever locking mechanism. Produced by Gino Olivares s.r.l., Gessate-Milan, Italy, to a density range of 6.4 to 6.6 g/cm 3 (0.23 to 0.24 lb/in.3), the part required complex tooling. It has a 65 HRB hardness along with good ductility. Close tolerances are required in the hole and height of the part. The padlock body is

Abbreviations and Symbols

o •

Abbreviations and Symbols

A

•

•

ampere

•

angstrom

•

ABC

•

ac

•

activity-based costing

•

alternating current

•

artificial intelligence

•

AI

•

AMS

•

ANSI

• •

Aerospace Material Specification American National Standards Institute

•

ASTM

•

at.%

• •

•

atm

•

b

•

bal

•

bcc

•

bct

atomic percent

•

atmosphere (pressure)

•

Burgers vector

•

balance

•

body-centered cubic

•

body-centered tetragonal

•

Btu

•

C

•

C-C

•

British thermal unit

•

Coulomb; heat capacity

•

•

CAD

•

CAE

•

CAM

• •

carbon-carbon computer-aided design computer-aided engineering

•

computer-aided manufacturing

•

candela

•

cd

•

CIP •

cold isostatic pressing

•

centimeter

•

cm

•

CMC

•

CNC

•

American Society for Testing and Materials

• •

ceramic-matrix composite computer numerical control

cpm •

cycles per minute

•

cps

•

CPS

•

CSA

• •

cycles per second creative problem solving

•

Canadian Standards Association

•

centiStokes

•

cSt

•

CTE •

coefficient of thermal expansion

•

CTOD

•

CVD

•

CVI

• • •

•

CVN

•

dB

crack tip opening displacement chemical vapor deposition chemical vapor infiltration (impregnation)

•

Charpy V-notch (impact test or specimen)

•

decibel

•

DBTT

•

dc

•

dhcp

•

ductile-brittle transition temperature

•

direct current

•

•

diam

•

DIN

• •

double hexagonal close-packed diameter Deutsche Institut für Normung

•

DTA

•

DWTT

•

e

• •

drop weight transition temperature

•

charge of an electron; natural log base, 2.71828

•

modulus of elasticity; Young's modulus; potential

•

secant modulus

•

E

•

Es

•

EDM •

•

emf

•

EMI

•

electromotive force electromagnetic interference

•

equation

Eq

•

ESC

•

esu

•

environmental stress cracking

•

electrostatic units

•

et al.

•

eV f

electrical discharge machining

•

•

•

differential thermal analysis

•

and others

•

electron volt

•

f

•

fn

•

F

•

fiber

•

focal length; frequency

•

normal load

•

Faraday constant; force

•

face-centered cubic

•

face-centered tetragonal

•

fcc

•

fct

•

FEA •

•

FEM

•

Fig.

• •

finite-element analysis finite-element modeling figure

•

FM

•

FMEA

•

ft

• •

failure modes and effects analysis

•

foot

•

FTA

•

g

•

G

•

G

•

GIc

•

fault tree analysis

•

gram

•

gauss

•

mean grain size; shear modulus

•

interlaminar fracture toughness (mode I, peel; mode II, shear; mode III, scissor shear)

•

gallon

•

gal

•

GPa •

•

gpd

•

gr

•

h

•

h

•

H

•

H

figure of merit

gigapascal

•

grains per denier

•

grain

•

hour

•

thickness

•

henry

•

change in height; degree of homogenization; enthalpy; hardness; height; magnetic field

•

HAZ

•

HB

•

heat-affected zone

•

Brinell hardness

•

hcp

•

HDT

• •

hexagonal close-packed heat-deflection temperature

•

HERF

•

HIP

•

HK

• •

horsepower

HR

•

•

Knoop hardness

hp •

•

HV

•

Hz

•

i I

•

IC

Vickers hardness

•

hertz

•

current (measure of number of electrons)

•

current; emergent intensity

•

integrated circuit

•

inside diameter

•

inch

ID

•

in.

•

IPD •

•

IPTS

•

IR

International Practical Temperature Scale

•

infrared (radiation)

ISO

•

ITS

•

IV

•

•

JIS

•

k

•

k

•

K

•

integrated product development

•

•

J

Rockwell hardness (requires a scale designation, such as HRC for Rockwell C hardness)

•

•

•

hot isostatic pressing

•

•

•

high-energy-rate forging

International Organization for Standardization

•

International Temperature Scale

•

intrinsic viscosity

•

joule

•

Japanese Industrial Standard

•

karat

•

Boltzmann constant; notch sensitivity factor; thermal conductivity; wave number

•

Kelvin

•

bulk modulus of elasticity; coefficient of thermal conductivity; empirical constant; interface reaction; mean integrated thermal conductivity

•

stress-intensity factor

•

mode I critical stress-intensity factor; plane-strain fracture toughness

•

dynamic fracture toughness

K

•

KI

•

KIc

•

KId

•

KIscc

•

threshold stress intensity for stress-corrosion cracking

•

Kc

•

Kf

•

Kt

•

Kth

•

kg

•

plane-stress fracture toughness

•

stress-concentration factor

•

stress-concentration factor

•

threshold crack tip stress-intensity factor

•

kilogram

•

kilometer

•

kilonewton

•

km

•

kN

•

kPa

•

ksi

•

kV

•

kilopascal

•

kips (1000 lb) per square inch

•

kilovolt

•

kW

•

l

•

L

•

L

•

lb

•

kilowatt

•

mean free path; length

•

liter; longitudinal direction; liquid

•

length

•

pound

•

pound-force

•

lbf

•

LCA

•

L/D

• •

life cycle analysis (or assessment) length-to-diameter ratio

•

LED

•

LEFM

•

ln

•

•

log

•

m

•

m

•

M

•

M

•

linear-elastic fracture mechanics

•

natural logarithm (base e)

•

common logarithm (base 10)

•

matrix; meter

•

ion mass; Weibull modulus

•

metal atom

•

molecular weight

•

mA

•

MeV

•

• •

mg

light-emitting diode

milliampere megaelectronvolt

•

milligram

•

megagram

•

Mg

•

MIM

•

min

•

metal injection molding

•

minimum; minute

•

megajoule

•

MJ

•

mL

•

MLT

• •

•

mm

•

MMC

• •

•

mod

•

mol%

•

MOR

• • •

•

mPa

•

MPa

• •

•

mpg

•

mph

•

ms

•

milliliter marketing lead time millimeter metal-matrix composite modified mole percent modulus of resilience; modulus of rupture millipascal megapascal miles per gallon

mT

millisecond

•

•

MS

miles per hour

•

•

• •

megasiemens millitesla

•

MTBF

•

MTTF

•

mV

• • •

•

MV

•

MW

•

n

•

•

n

•

N

mean time between failures mean time to failure millivolt megavolt

•

molecular weight

•

growth exponent

•

integral number

•

Newton

•

fatigue life (number of cycles)

•

N

•

NASA •

National Aeronautics and Space Administration

•

NBS

•

NC

•

NDE

•

National Bureau of Standards

•

numerical control

•

•

NDI

•

NDT

• •

•

NIST

•

nm

•

No.

nondestructive evaluation nondestructive inspection nondestructive testing

•

National Institute of Standards and Technology

•

nanometer

•

number

•

NTC

•

OD

•

negative temperature coefficient

•

outside diameter

•

oersted

•

Oe

•

OEM

•

OHA

• •

original equipment manufacturer operating hazards analysis

•

ORNL

•

OSHA

•

•

oz

•

p

•

p

•

Occupational Safety and Health Administration

•

ounce

•

page

•

pressure

•

applied load; power; pressure

•

pascal

•

P

•

Pa

•

PCE •

•

PDS

•

PDT

•

pyrometric cone equivalent product design specification

•

product design team

•

powder forging

•

negative logarithm of hydrogen-ion activity

•

P/F

•

pH

•

P/M •

•

ppb

•

ppba

•

Oak Ridge National Laboratory

• •

ppm

powder metallurgy parts per billion parts per billion atomic

•

parts per million

•

parts per trillion

•

pounds per square inch

•

ppt

•

psi

•

psia •

•

psid

•

psig

•

PTC

• • •

•

PVD

•

QC

• •

•

QFD

•

QLF

•

r

•

•

r

•

R

•

pounds per square inch absolute pounds per square inch differential pounds per square inch gage positive temperature coefficient physical vapor deposition quality control quality function deployment

•

quality loss function

•

radius vector in a plane normal to the axis

•

particle radius; radius of curvature; rate of reaction; reflectivity; spherical particle radius

•

roentgen

•

average particle radius; gas constant; radius; ratio of the minimum stress to the maximum stress; resistance; reliability

•

surface roughness in terms of arithmetic average

•

reduction of area

R

•

Ra

•

RA

•

rad

•

RCF

•

RE

•

absorbed radiation dose; radian

•

rolling contact fatigue

•

rare earth

•

reference

•

Ref

•

rem •

remainder

•

rf, RF

•

RH

•

rms, RMS

• • •

•

ROM

•

rpm

•

RT

•

•

radio frequency relative humidity root mean square rough order of magnitude; rule of mixtures

•

revolutions per minute

•

room temperature

RTI

•

s

•

S

•

S

•

relative thermal index

•

second

•

siemens

•

distance traveled

•

SAE

•

scfm

•

SEM

• • •

•

sfm

•

SG •

standard grade

•

Système International d'Unités

•

sineh

•

S/N

• •

•

S-N

•

SPC

• •

•

SPF

•

sp gr

•

SRIM

• •

•

•

STM

•

Sv

•

t

•

Tm

•

tan

statistical process control superplastic forming specific gravity

standard

STEM

Tg

stress-number of cycles

•

•

•

signal-to-noise (ratio)

structural reaction injection molding

std

T

sine hyperbolic

•

•

•

scanning electron microscopy surface feet per minute

SI

T

standard cubic foot per minute

•

•

•

Society of Automotive Engineers

scanning transmission electron microscopy

•

scanning tunneling microscopy

•

sievert

•

thickness; time

•

Tesla; transverse direction

•

absolute temperature; temperature; tenacity; total dispersion; transmittance

•

glass transition temperature

•

melt temperature

•

equal to ratio of the loss modulus to the storage modulus

•

TBC

•

TEM

• •

thermal barrier coating transmission electron microscopy

•

TGA

•

TMA

•

TP

• •

thermomechanical analysis

•

thermoplastic

•

TQM

•

TS

•

total quality management

•

thermoset

•

tons per square inch

•

tsi

•

TTT

•

UL

•

time-temperature-transformation

•

Underwriters' Laboratories

•

UNS

•

UTS

•

•

UV

•

v

•

V

•

Vf

•

Vm

•

Vv

•

VI

•

vol

ultimate tensile strength

•

ultraviolet

•

workpiece velocity

•

volt

•

volume fraction of fiber

•

volume fraction of matrix

•

volume fraction of void content

•

viscosity index

•

volume

vol%

•

w

•

W

•

volume percent

•

whisker

•

watt

•

total radiation; width

•

W

•

wt% x

•

X

•

Y

•

weight percent

•

axial distance

•

neck diameter

•

scale of microstructural segregation

•

YAG

•

yr

•

z

Unified Numbering System

•

•

•

thermogravimetric analysis

•

yttrium aluminum garnet

•

year

•

Z

•

°

•

°C

•

°F

•

0°

•

90°

• • • •

b

s

• • •

• • •

•

ion charge

•

atomic number; impedance

•

angular measure; degree

•

degree Celsius (centigrade)

•

degree Fahrenheit

•

fiber direction

•

perpendicular to fiber direction

•

coefficient of thermal expansion

•

shear strain; surface energy; surface tension

•

interfacial energy

•

surface energy

•

an increment; a range; change in quantity; grain boundary width; thickness of liquid boundary

•

an increment; a range; change in quantity

•

temperature difference or change

•

strain

•

strain rate

•

loss coefficient; viscosity

•

angle; geometrical constant

•

angle of incidence

•

angle of refraction

•

dielectric constant

•

thermal conductivity

•

friction coefficient; linear attenuation coefficient; magnetic permeability; the mean (or average) of a distribution

T

•

•

•

i r

• • • •

in.

•

m

• • • • •

•

microinch

•

micrometer (micron)

•

microsecond

•

Poisson's ratio; velocity

•

pi (3.141592)

•

density; resistivity

s

•

a

•

f

• • • • • • • • •

•

•

•

applied axial load

•

summation of

•

shear stress

•

angle of internal friction; dihedral angle; porosity; power

•

reaction rate (kinetic function); volume concentration of impurity ions

•

damping; geometrical constant

•

frequency

•

atomic volume; electrical conductivity; molecular volume of solute; ohm

•

partial derivative

•

kinetic function

•

direction of reaction

•

divided by

•

equals

•

approximately equals

•

not equal to

•

identical with

•

greater than

•

much greater than

•

greater than or equal to

•

infinity

•

is proportional to; varies as

•

integral of

•

less than

•

much less than

•

less than or equal to

•

maximum deviation

•

minus; negative ion charge

•

diameters (magnification); multiplied by

>

• • • • •