EDITOR IN CHIEF Rudy J. M. Konings European Commission, Joint Research Centre, Institute for Transuranium Elements, Karlsruhe, Germany
SECTION EDITORS Todd R. Allen Department of Engineering Physics, University of Wisconsin, Madison, WI, USA Roger E. Stoller Materials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA Shinsuke Yamanaka Division of Sustainable Energy and Environmental Engineering, Graduate School of Engineering, Osaka University, Osaka, Japan
Elsevier Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK 225 Wyman Street, Waltham, MA 02451, USA Copyright © 2012 Elsevier Ltd. All rights reserved The following articles are US Government works in the public domain and not subject to copyright: Radiation Effects in UO2 TRISO-Coated Particle Fuel Performance Composite Fuel (cermet, cercer) Metal Fuel-Cladding Interaction No part of this publication 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 prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (þ44) (0) 1865 843830; fax (þ44) (0) 1865 853333; email:
[email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein, Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Catalog Number: 2011929343 ISBN (print): 978-0-08-056027-4 For information on all Elsevier publications visit our website at books.elsevier.com Cover image courtesy of Professor David Sedmidubsky´, The Institute of Chemical Technology, Prague Printed and bound in Spain 12 13 14 15 16 10 9 8 7 6 5 4 3 2 1
Editorial : Gemma Mattingley Production: Nicky Carter
EDITORS BIOGRAPHIES Rudy Konings is currently head of the Materials Research Unit in the Institute for Transuranium Elements (ITU) of the Joint Research Centre of the European Commission. His research interests are nuclear reactor fuels and actinide materials, with particular emphasis on high temperature chemistry and thermodynamics. Before joining ITU, he worked on nuclear fuel-related issues at ECN (the Energy Research Centre of the Netherlands) and NRG (Nuclear Research and Consultancy Group) in the Netherlands. Rudy is editor of Journal of Nuclear Materials and is professor at the Delft University of Technology (Netherlands), where he holds the chair of ‘Chemistry of the nuclear fuel cycle.’
Roger Stoller is currently a Distinguished Research Staff Member in the Materials Science and Technology Division of the Oak Ridge National Laboratory and serves as the ORNL Program Manager for Fusion Reactor Materials for ORNL. He joined ORNL in 1984 and is actively involved in research on the effects of radiation on structural materials and fuels for nuclear energy systems. His primary expertise is in the area of computational modeling and simulation. He has authored or coauthored more than 100 publications and reports on the effects of radiation on materials, as well as edited the proceedings of several international conferences.
Todd Allen is an Associate Professor in the Department of Engineering Physics at the University of Wisconsin – Madison since 2003. Todd’s research expertise is in the area of materials-related issues in nuclear reactors, specifically radiation damage and corrosion. He is also the Scientific Director for the Advanced Test Reactor National Scientific User Facility as well as the Director for the Center for Material Science of Nuclear Fuel at the Idaho National Laboratory, positions he holds in conjunction with his faculty position at the University of Wisconsin.
v
vi
Editors Biographies
Shinsuke Yamanaka is a professor in Division of Sustainable Energy and Environmental Engineering, Graduate School of Engineering, Osaka University since 1998. He has studied the thermophysics and thermochemistry of nuclear fuel and materials. His research for the hydrogen behavior in LWR fuel cladding is notable among his achievements and he received the Young Scientist Awards (1980) and the Best Paper Awards (2004) from Japan Atomic Energy Society. Shinsuke is the program officer of Japan Science and Technology Agency since 2005 and the visiting professor of Fukui University since 2009, and he is also the associate dean of Graduate School of Engineering, Osaka University since 2011.
PREFACE There are essentially three primary energy sources for the billions of people living on the earth’s surface: the sun, radioactivity, and gravitation. The sun, an enormous nuclear fusion reactor, has transmitted energy to the earth for billions of years, sustaining photosynthesis, which in turn produces wood and other combustible resources (biomass), and the fossil fuels like coal, oil, and natural gas. The sun also provides the energy that steers the climate, the atmospheric circulations, and thus ‘fuelling’ wind mills, and it is at the origin of photovoltaic processes used to produce electricity. Radioactive decay of primarily uranium and thorium heats the earth underneath us and is the origin of geothermal energy. Hot springs have been used as a source of energy from the early days of humanity, although it took until the twentieth century for the potential of radioactivity by fission to be discovered. Gravitation, a non-nuclear source, has been long used to generate energy, primarily in hydropower and tidal power applications. Although nuclear processes are thus omnipresent, nuclear technology is relatively young. But from the moment scientists unraveled the secrets of the atom and its nucleus during the twentieth century, aided by developments in quantum mechanics, and obtained a fundamental understanding of nuclear fission and fusion, humanity has considered these nuclear processes as sources of almost unlimited (peaceful) energy. The first fission reactor was designed and constructed by Enrico Fermi in 1942 in Chicago, the CP1, based on the fission of uranium by neutron capture. After World War II, a rapid exploration of fission technology took place in the United States and the Union of Soviet Socialist Republics, and after the Atoms for Peace speech by Eisenhower at the United Nations Congress in 1954, also in Europe and Japan. A variety of nuclear fission reactors were explored for electricity generation and with them the fuel cycle. Moreover, the possibility of controlled fusion reactions has gained interest as a technology for producing energy from one of the most abundant elements on earth, hydrogen. The environment to which materials in nuclear reactors are exposed is one of extremes with respect to temperature and radiation. Fuel pins for nuclear reactors operate at temperatures above 1000 C in the center of the pellets, in fast reactor oxide fuels even above 2000 C, whereas the effects of the radiation (neutrons, alpha particles, recoil atoms, fission fragments) continuously damage the material. The cladding of the fuel and the structural and functional materials in the fission reactor core also operate in a strong radiation field, often in a dynamic corrosive environment of the coolant at elevated temperatures. Materials in fusion reactors are exposed to the fusion plasma and the highly energetic particles escaping from it. Furthermore, in this technology, the reactor core structures operate at high temperatures. Materials science for nuclear systems has, therefore, been strongly focussed on the development of radiation tolerant materials that can operate in a wide range of temperatures and in different chemical environments such as aqueous solutions, liquid metals, molten salts, or gases. The lifetime of the plant components is critical in many respects and thus strongly affects the safety as well as the economics of the technologies. With the need for efficiency and competitiveness in modern society, there is a strong incentive to improve reactor components or to deploy advanced materials that are continuously developed for improved performance. There are many examples of excellent achievements in this respect. For example, with the increase of the burnup of the fuel for fission reactors, motivated by improved economics and a more efficient use of resources, the Zircaloy cladding (a Zr–Sn alloy) of the fuel pins showed increased susceptibility to coolant corrosion, but within a relatively short period, a different zirconium-based alloy was developed, tested, qualified, and employed, which allowed reliable operation in the high burnup range.
xxi
xxii
Preface
Nuclear technologies also produce waste. It is the moral obligation of the generations consuming the energy to implement an acceptable waste treatment and disposal strategy. The inherent complication of radioactivity, the decay that can span hundreds of thousands of years, amplifies the importance of extreme time periods in the issue of corrosion and radiation stability. The search for storage concepts that can guarantee the safe storage and isolation of radioactive waste is, therefore, another challenging task for materials science, requiring a close examination of natural (geological) materials and processes. The more than 50 years of research and development of fission and fusion reactors have undoubtedly demonstrated that the statement ‘technologies are enabled by materials’ is particularly true for nuclear technology. Although the nuclear field is typically known for its incremental progress, the challenges posed by the next generation of fission reactors (Generation IV) as well as the demonstration of fusion reactors will need breakthroughs to achieve their ambitious goals. This is being accompanied by an important change in materials science, with a shift of discovery through experiments to discovery through simulation. The progress in numerical simulation of the material evolution on a scientific and engineering scale is growing rapidly. Simulation techniques at the atomistic or meso scale (e.g., electronic structure calculations, molecular dynamics, kinetic Monte Carlo) are increasingly helping to unravel the complex processes occurring in materials under extreme conditions and to provide an insight into the causes and thus helping to design remedies. In this context, Comprehensive Nuclear Materials aims to provide fundamental information on the vast variety of materials employed in the broad field of nuclear technology. But to do justice to the comprehensiveness of the work, fundamental issues are also addressed in detail, as well as the basics of the emerging numerical simulation techniques. R.J.M. Konings European Commission, Joint Research Centre, Institute for Transuranium Elements, Karlsruhe, Germany T.R. Allen Department of Engineering Physics, Wisconsin University, Madison, WI, USA R. Stoller Materials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA S. Yamanaka Division of Sustainable Energy and Environmental Engineering, Graduate School of Engineering, Osaka University, Osaka, Japan
FOREWORD ‘Nuclear materials’ denotes a field of great breadth and depth, whose topics address applications and facilities that depend upon nuclear reactions. The major topics within the field are devoted to the materials science and engineering surrounding fission and fusion reactions in energy conversion reactors. Most of the rest of the field is formed of the closely related materials science needed for the effects of energetic particles on the targets and other radiation areas of charged particle accelerators and plasma devices. A more complete but also more cumbersome descriptor thus would be ‘the science and engineering of materials for fission reactors, fusion reactors, and closely related topics.’ In these areas, the very existence of such technologies turns upon our capabilities to understand the physical behavior of materials. Performance of facilities and components to the demanding limits required is dictated by the capabilities of materials to withstand unique and aggressive environments. The unifying concept that runs through all aspects is the effect of radiation on materials. In this way, the main feature is somewhat analogous to the unifying concept of elevated temperature in that part of materials science and engineering termed ‘high-temperature materials.’ Nuclear materials came into existence in the 1950s and began to grow as an internationally recognized field of endeavor late in that decade. The beginning in this field has been attributed to presentations and discussions that occurred at the First and Second International Conferences on the Peaceful Uses of Atomic Energy, held in Geneva in 1955 and 1958. Journal of Nuclear Materials, which is the home journal for this area of materials science, was founded in 1959. The development of nuclear materials science and engineering took place in the same rapid growth time period as the parent field of materials science and engineering. And similarly to the parent field, nuclear materials draws together the formerly separate disciplines of metallurgy, solid-state physics, ceramics, and materials chemistry that were early devoted to nuclear applications. The small priesthood of first researchers in half a dozen countries has now grown to a cohort of thousands, whose home institutions are anchored in more than 40 nations. The prodigious work, ‘Comprehensive Nuclear Materials,’ captures the essence and the extensive scope of the field. It provides authoritative chapters that review the full range of endeavor. In the present day of glance and click ‘reading’ of short snippets from the internet, this is an old-fashioned book in the best sense of the word, which will be available in both electronic and printed form. All of the main segments of the field are covered, as well as most of the specialized areas and subtopics. With well over 100 chapters, the reader finds thorough coverage on topics ranging from fundamentals of atom movements after displacement by energetic particles to testing and engineering analysis methods of large components. All the materials classes that have main application in nuclear technologies are visited, and the most important of them are covered in exhaustive fashion. Authors of the chapters are practitioners who are at the highest level of achievement and knowledge in their respective areas. Many of these authors not only have lived through a substantial part of the history sketched above, but they themselves are the architects. Without those represented here in the author list, the field would certainly be a weaker reflection of itself. It is no small feat that so many of my distinguished colleagues could have been persuaded to join this collective endeavor and to make the real sacrifices entailed in such time-consuming work. I congratulate the Editor, Rudy Konings, and
xxiii
xxiv
Foreword
the Associate Editors, Roger Stoller, Todd Allen, and Shinsuke Yamanaka. This book will be an important asset to young researchers entering the field as well as a valuable resource to workers engaged in the enterprise at present. Dr. Louis K. Mansur Oak Ridge, Tennessee, USA
Permission Acknowledgments The following material is reproduced with kind permission of Cambridge University Press Figure 15 of Oxide Dispersion Strengthened Steels Figure 15 of Minerals and Natural Analogues Table 10 of Spent Fuel as Waste Material Figure 21b of Radiation-Induced Effects on Microstructure www.cambridge.org The following material is reproduced with kind permission of American Chemical Society Figure 2 of Molten Salt Reactor Fuel and Coolant Figure 22 of Molten Salt Reactor Fuel and Coolant Table 9 of Molten Salt Reactor Fuel and Coolant Figure 6 of Thermodynamic and Thermophysical Properties of the Actinide Nitrides www.acs.org The following material is reproduced with kind permission of Wiley Table 3 of Properties and Characteristics of SiC and SiC/SiC Composites Table 4 of Properties and Characteristics of SiC and SiC/SiC Composites Table 5 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 5 of Advanced Concepts in TRISO Fuel Figure 6 of Advanced Concepts in TRISO Fuel Figure 30 of Material Performance in Supercritical Water Figure 32 of Material Performance in Supercritical Water Figure 19 of Tritium Barriers and Tritium Diffusion in Fusion Reactors Figure 9 of Waste Containers Figure 13 of Waste Containers Figure 21 of Waste Containers Figure 11 of Carbide Fuel Figure 12 of Carbide Fuel Figure 13 of Carbide Fuel Figure 4 of Thermodynamic and Thermophysical Properties of the Actinide Nitrides Figure 2 of The U–F system Figure 18 of Fundamental Point Defect Properties in Ceramics Table 1 of Fundamental Point Defect Properties in Ceramics Figure 17 of Radiation Effects in SiC and SiC-SiC Figure 21 of Radiation Effects in SiC and SiC-SiC Figure 6 of Radiation Damage in Austenitic Steels Figure 7 of Radiation Damage in Austenitic Steels Figure 17 of Ceramic Breeder Materials Figure 33a of Carbon as a Fusion Plasma-Facing Material Figure 34 of Carbon as a Fusion Plasma-Facing Material i
ii
Permission Acknowledgments
Figure 39 of Carbon as a Fusion Plasma-Facing Material Figure 40 of Carbon as a Fusion Plasma-Facing Material Table 5 of Carbon as a Fusion Plasma-Facing Material www.wiley.com The following material is reproduced with kind permission of Springer Figure 4 of Neutron Reflector Materials (Be, Hydrides) Figure 6 of Neutron Reflector Materials (Be, Hydrides) Figure 1 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 3 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 4 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 5 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 6 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 7 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 8 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 9 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 10 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 11 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 12 of Properties and Characteristics of SiC and SiC/SiC Composites Figure 22d of Fission Product Chemistry in Oxide Fuels Figure 3 of Behavior of LWR Fuel During Loss-of-Coolant Accidents Figure 14a of Irradiation Assisted Stress Corrosion Cracking Figure 14b of Irradiation Assisted Stress Corrosion Cracking Figure 14c of Irradiation Assisted Stress Corrosion Cracking Figure 25a of Irradiation Assisted Stress Corrosion Cracking Figure 25b of Irradiation Assisted Stress Corrosion Cracking Figure 1 of Properties of Liquid Metal Coolants Figure 5b of Fast Spectrum Control Rod Materials Figure 3 of Oxide Fuel Performance Modeling and Simulations Figure 8 of Oxide Fuel Performance Modeling and Simulations Figure 10 of Oxide Fuel Performance Modeling and Simulations Figure 11 of Oxide Fuel Performance Modeling and Simulations Figure 14 of Oxide Fuel Performance Modeling and Simulations Figure 5 of Thermodynamic and Thermophysical Properties of the Actinide Nitrides Figure 51 of Phase Diagrams of Actinide Alloys Figure 6 of Thermodynamic and Thermophysical Properties of the Actinide Oxides Figure 7b of Thermodynamic and Thermophysical Properties of the Actinide Oxides Figure 9b of Thermodynamic and Thermophysical Properties of the Actinide Oxides Figure 35 of Thermodynamic and Thermophysical Properties of the Actinide Oxides Table 11 of Thermodynamic and Thermophysical Properties of the Actinide Oxides Table 13 of Thermodynamic and Thermophysical Properties of the Actinide Oxides Table 17 of Thermodynamic and Thermophysical Properties of the Actinide Oxides Figure 18 of Radiation Damage of Reactor Pressure Vessel Steels Figure 7 of Radiation Damage Using Ion Beams Figure 9b of Radiation Damage Using Ion Beams Figure 28 of Radiation Damage Using Ion Beams Figure 34 of Radiation Damage Using Ion Beams Figure 35 of Radiation Damage Using Ion Beams Figure 36d of Radiation Damage Using Ion Beams Figure 37 of Radiation Damage Using Ion Beams Table 3 of Radiation Damage Using Ion Beams
Permission Acknowledgments
Figure 5 of Radiation Effects in UO2 Figure 9a of Ab Initio Electronic Structure Calculations for Nuclear Materials Figure 9b of Ab Initio Electronic Structure Calculations for Nuclear Materials Figure 9c of Ab Initio Electronic Structure Calculations for Nuclear Materials Figure 10a of Ab Initio Electronic Structure Calculations for Nuclear Materials Figure 23 of Thermodynamic and Thermophysical Properties of the Actinide Carbides Figure 25 of Thermodynamic and Thermophysical Properties of the Actinide Carbides Figure 26 of Thermodynamic and Thermophysical Properties of the Actinide Carbides Figure 27 of Thermodynamic and Thermophysical Properties of the Actinide Carbides Figure 28a of Thermodynamic and Thermophysical Properties of the Actinide Carbides Figure 28b of Thermodynamic and Thermophysical Properties of the Actinide Carbides Figure 2 of Physical and Mechanical Properties of Copper and Copper Alloys Figure 5 of Physical and Mechanical Properties of Copper and Copper Alloys Figure 6 of The Actinides Elements: Properties and Characteristics Figure 10 of The Actinides Elements: Properties and Characteristics Figure 11 of The Actinides Elements: Properties and Characteristics Figure 12 of The Actinides Elements: Properties and Characteristics Figure 15 of The Actinides Elements: Properties and Characteristics Table 1 of The Actinides Elements: Properties and Characteristics Table 6 of The Actinides Elements: Properties and Characteristics Figure 25 of Fundamental Properties of Defects in Metals Table 1 of Fundamental Properties of Defects in Metals Table 7 of Fundamental Properties of Defects in Metals Table 8 of Fundamental Properties of Defects in Metals www.springer.com The following material is reproduced with kind permission of Taylor & Francis Figure 9 of Radiation-Induced Segregation Figure 6 of Radiation Effects in Zirconium Alloys Figure 1 of Dislocation Dynamics Figure 25 of Radiation Damage Using Ion Beams Figure 26 of Radiation Damage Using Ion Beams Figure 27 of Radiation Damage Using Ion Beams Figure 4 of Radiation-Induced Effects on Material Properties of Ceramics (Mechanical and Dimensional) Figure 7 of The Actinides Elements: Properties and Characteristics Figure 20 of The Actinides Elements: Properties and Characteristics Figure 18a of Primary Radiation Damage Formation Figure 18b of Primary Radiation Damage Formation Figure 18c of Primary Radiation Damage Formation Figure 18d of Primary Radiation Damage Formation Figure 18e of Primary Radiation Damage Formation Figure 18f of Primary Radiation Damage Formation Figure 1 of Radiation-Induced Effects on Microstructure Figure 27 of Radiation-Induced Effects on Microstructure Figure 5 of Performance of Aluminum in Research Reactors Figure 2 of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 3 of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 5 of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 10a of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 10b of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 10c of Atomic-Level Dislocation Dynamics in Irradiated Metals
iii
iv
Permission Acknowledgments
Figure 10d of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 12a of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 12b of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 12c of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 12d of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 16a of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 16b of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 16c of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 16d of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 16e of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 17a of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 17b of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 17c of Atomic-Level Dislocation Dynamics in Irradiated Metals Figure 17d of Atomic-Level Dislocation Dynamics in Irradiated Metals www.taylorandfrancisgroup.com
5.01
Corrosion and Compatibility
S. Lillard Los Alamos National Laboratory, Los Alamos, NM, USA
ß 2012 Elsevier Ltd. All rights reserved.
2
5.01.1
Theory
5.01.1.1 5.01.1.2 5.01.1.3 5.01.1.4 5.01.1.5 5.01.1.6 5.01.2 5.01.2.1 5.01.2.2 5.01.2.3 5.01.2.4 References
Introduction Half Cell Reactions Cell Potentials and the Nernst Equation Reference Electrodes and Their Application to Nuclear Systems The Thermodynamics of Corrosion from Room Temperature to the PWR Kinetics of Dissolution and Passive Film Formation Analytical Methods Introduction Potentiodynamic Polarization Electrochemical Impedance Spectroscopy Mott–Schottky Analysis
Abbreviations BWR CNLS
Boiling water reactor Complex nonlinear least squares fitting of the data EC Electrical equivalent circuit EIS Electrochemical impedance spectroscopy EPBRE External pressure-balanced reference electrode FFTF Fast Flux Test Facility HIC Hydrogen-induced cracking HIFER Hi-Flux Isotope Reactor IG Intergranular LBE Lead–bismuth eutectic PWR Pressurized water reactor SCC Stress corrosion cracking SS Stainless steel
Symbols A ai C ci CR E EW Ecorr f
Surface area Activity of species i Capacitance Concentration of species i Corrosion rate Potential Equivalent weight Corrosion potential Mass fraction
ƒ F i icorr ji k L M MM n ND Q r R Rp RV S t ti T Vo z Z Z0 Z00 jZj b ba bc d
2 2 3 4 6 8 10 10 11 12 14 16
Fugacity Faraday’s constant Current density Corrosion current density Square root of 1 Rate constant Oxide thickness Molecular weight Metal cation Number of electrons Donor concentration Reaction quotient Rate of reaction Gas constant Polarization resistance Solution resistance Entropy of transport Time Transport number of species i Temperature Oxygen vacancy Charge Impedance Real part of the impedance Imaginary part of the impedance Magnitude of the impedance Symmetry factor Anodic Tafel slope Cathodic Tafel slope Double layer thickness
1
2
Corrosion and Compatibility 0
DCp Change in standard partial molar heat capacity DE0 Standard reduction potential DG Change in Gibbs energy DG0 Standard Gibbs energy DS0 Standard entropy change « Electronic charge Permittivity of space «0 f Applied potential h Overpotential u Phase angle r Material density v Frequency
5.01.1 Theory 5.01.1.1
Introduction
Mars Fontana identified eight forms of corrosion in his book Corrosion Engineering1 and it is quite easy to find examples of almost all of these in nuclear reactors in both the primary and secondary cooling water systems. For example, galvanic corrosion in zirconiumstainless steel couples,2,3 crevice corrosion in tube sheets4 and former baffle bolts,5 and pitting corrosion in alloy 600 steam generator tubes.6,7 Perhaps the most infamous form of corrosion observed in nuclear reactors is stress corrosion cracking (SCC), or environmental fracture, as we shall refer to it here, which has numerous examples in the literature. Environmental fracture includes both intergranular SCC (IG), such as that which occurs in austenitic stainless steel, and hydrogen-induced cracking (HIC), frequently observed in nickel base alloys. Failure by one of these mechanisms results from an interplay between stress, microstructure, and the environment (e.g., the electrochemical interface). The goal of this chapter is not to address each of the corrosion mechanisms outlined by Fontana individually, that will be accomplished in the following chapters. Rather, this chapter is meant to provide the reader with the fundamental electrochemical theory necessary to critically evaluate the data and discussions in the corrosion chapters that follow. In this section, we will review the fundamental theory of the electrochemical interface. In the first three subsections, we review Half Cell Reaction, Cell Potentials and the Nernst Equation, and Reference Electrodes in Nuclear Systems. In these sections, we develop the theory necessary to understand the role of electrochemical potential in environmental fracture
and corrosion mechanisms. For example, intergranular stress corrosion cracking (IGSCC) is only observed at potentials more positive than a critical value while HIC is only observed at potentials more negative than a critical value. In the remaining two sections, we review the Thermodynamics from Room Temperature to the pressurized water reactor (PWR) and Kinetics of Dissolution and Passive Film Formation. These sections should help the reader to understand the role of the passive film in the corrosion mechanism and the competition that occurs between film formation and metal dissolution rate. As the fundamental role of irradiation in corrosion and environmental fracture mechanism is far from well established, in each section, we incorporate empirical irradiation data as examples and discuss concepts that are more broadly important to nuclear systems. 5.01.1.2
Half Cell Reactions
The electrochemical interface is characterized by an electrode (in this case a metal such as a cooling pipe) and an electrolyte (e.g., the cooling water in a reactor). While the bulk electrolyte contributes to variables such as solution chemistry and ohmic drop (solution resistance is discussed later in this chapter), it is the first nanometer of electrolyte that plays the most important role in electrochemistry. In this short distance, referred to as the electrochemical double layer, a separation of charge occurs. It is this separation of charge that provides the driving force (potential drop) for corrosion reactions. For example, a 100 mV-applied potential across a typical double layer will result in an electric field on the order of 106 V cm2. In the model proposed by Helmholtz,8 the double layer may be thought of as capacitor, with positive charge on the metal electrode and the adsorption of negatively charged cations on the solution side (Figure 1). The capacitance of the double layer is equal to that in its electrical analog e0D/d, where e0 is the permittivity of space, D is the dielectric, and d is the thickness of the layer. For most electrochemical double layers, C is on the order of 106 F cm2. Electrochemical reactions that take place in the double layer are reactions in which a transfer of charge (electrons) occurs. There are two different types of cells in which electrochemical reactions may occur9: Electrolytic cells in which work, in the form of electrical energy, is required to bring about a nonspontaneous reaction.
Corrosion and Compatibility
Metal
Because the system cannot store charge, the electrons produced during the anodic reaction must be used. This occurs at the cathode where typical reactions may include oxygen reduction:
Excess negative charge
Excess positive charge
Bulk solution
+
-
+
-
+
-
+
-
+
-
+
+
-
+
Acid: O2 þ 4Hþ þ 4e ) 2H2 O
-
Base: O2 þ 2H2 O þ 4e ) 4OH +
-
3
½II ½III
or hydrogen reduction: 2Hþ þ 2e ) H2
fmetal
½IV
From eqns [I] and [III], the general corrosion of an Fe surface in basic solution may then be written as: 2Fe þ O2 þ 2H2 O ) 2FeðOHÞ2
fsolution Double layer Figure 1 A diagram depicting the separation of charge at the electrochemical double layer and the associated potential drop (f).
H2O 2H+
ClH2
H2O H+
H2O
ClH2O
Cl-
H+
Oxide Fe
For any chemical reaction the driving force, the Gibbs energy, may be written as10:
Figure 2 Diagram of what the anodic and cathodic reactions may look like on an iron surface depicting the separation of reactions and ionic conduction.
DG ¼ DG 0 þ RT lnQ
Voltaic cells in which a spontaneous reaction occurs resulting in work in the form of electrical energy. Electrolytic cells cover a fairly large number of electrochemical reactions but may generally be thought of as ‘plating’ or ‘electrolysis’ type reactions and will not be treated here. Corrosion reactions are voltaic cells and will be the focus of this chapter. As in an electrolytic cell, voltaic cells are characterized by two separate electrodes, an anode and a cathode. In corrosion, reactions at the anode take the form of metal dissolution, the formation of a soluble metal cation: Fe ) Feþ2 þ 2e
where Fe(OH)2 is the corrosion product. An example of what the anodic and cathodic reactions on Fe electrode might look like is presented in Figure 2. Though the anodic and cathodic reactions occur at physically separate locations, as shown in this figure, the reactions must be connected via an electrolyte (aqueous solution). Figure 2 also suggests that corrosion reactions are controlled by variables such as mass transport (diffusion, convection, migration), concentration, and ohmic drop (resistivity of the electrolyte). These variables will be considered in our discussion of corrosion kinetics. 5.01.1.3 Cell Potentials and the Nernst Equation
Fe2+
e-
½V
½I
½1
where DG0 is the standard Gibbs energy, Q is the reaction quotient equal to the product of the activities (assumed to obey Raoult’s Law for dilute solutions and, thus, equal to the concentration) of the products divided by the reactants, R is the gas constant, and T is temperature. The electrical potential, E, is related to the Gibbs energy of a cell by the relationship: nFE ¼ DG
½2
where n is the number of electrons participating in the reaction and F is Faraday’s constant. For the reduction of hydrogen on platinum: 1 þ þ e ðPtÞ , H2ðgÞ Haq 2
½VI
4
Corrosion and Compatibility
The reaction quotient, Q (starting conditions) becomes: Q ¼
½fH2 1=2 ½Hþ
½3
where fH2 is the fugacity of hydrogen gas. Substituting eqns [2] and [3] into eqn [1], we find for the reduction of hydrogen on platinum that: " # RT ½fH2 1=2 0 ½4 ln E ¼ DE þ ½Hþ F where F is Faraday’s constant and DE0 is the standard reduction potential for the reaction in eqn [VI]. Equation [10] is commonly referred to as the Nernst equation and defines the equilibrium reduction potential of the half cell and is pH dependent. The Nernst equation is commonly expressed in its generalized form as: E ¼ DE 0 þ
RT ln½Q F
½5
5.01.1.4 Reference Electrodes and Their Application to Nuclear Systems In Equation [4], all of the parameters are easily calculated with one exception, DE0. Therefore, we define DE0 ¼ 0 in eqn [4] for a set of specific parameters and refer to this cell as the standard hydrogen electrode (SHE): H2 pressure of 1 atm, a pH ¼ 0, and a temperature of 25 C. This provides a reference from which we can calculate the standard potentials for all other reduction reactions using eqn [5]. These are referred to as standard reduction potentials and a few examples are provided in Table 1. While the SHE is the accepted standard, from a practical standpoint, this reference electrode is difficult to construct and maintain. As such, experimentalists have taken advantage of a number of other reduction reactions to construct reference electrodes for laboratory use. The reaction selected typically depends on the application. One common reference electrode is the silver–silver chloride electrode (Ag/AgCl) which is based on the reduction of Agþ in a solution of potassium chloride: Agþ þ e , AgðsÞ
½VII
Agþ þ Cl , AgClðsÞ
½VIII
and the overall reaction being: AgðsÞ þ Cl , AgClðsÞ þ e
½XI
Table 1 Standard reduction potentials for several reactions important to the nuclear power industry Reduction reaction
Standard reduction potential (V)
Au3þ þ 3e ⇄ Au Cl2 þ 2e ⇄ 2Cl O2 þ 4Hþ þ 4e ⇄ 2H2O Agþ þ e ⇄ Ag Fe3þ þ e ⇄ Fe2þ O2 þ 2H2O þ 4e ⇄ 4OH AgCl þ e ⇄ Ag þ Cl 2Hþ þ 2e ⇄ H2 (NHE) Ni2þ þ 2e ⇄ Ni Fe2þ þ 2e ⇄ Fe Cr3þ þ 3e ⇄ Cr Zr4þ þ 4e ⇄ Zr Al3þ þ 3e ⇄ Al Liþ þ e ⇄ Li
1.52 1.36 1.23 0.80 0.77 0.4 0.22 0.0 0.25 0.44 0.74 1.53 1.66 3.04
The Nernst equation for eqn [XI] is equal to: ½aAgCl RT 0 E ¼ DEAg=AgCl þ ln ½aAg ½aCl F 0 ¼ DEAg=AgCl ln½aCl
½6
where aCl is the activity of chloride and for which the concentration (mCl ) in molal (mol kg1) is frequently substituted. In the corrosion lab, the reference electrode is constructed by electrochemically depositing an AgCl layer onto a silver wire. This wire is then placed in a glass capillary filled with a solution of potassium chloride the concentration of which then defines the cell potential (aCl in eqn [6]). One end of the capillary is sealed using a porous frit (typically a porous polymer) that acts as a junction between the solution of the reference electrode and the environment of the corrosion experiment. While the Ag/AgCl reference electrode construction described above is straightforward for the lab, there are several obstacles that must be overcome before it can be used in a nuclear power plant setting, namely, radiation flux, pressure, and temperature. As it turns out, the primary impact of ionizing radiation on laboratory reference electrodes relates to damage of the cotton wadding and polymer frits used in their construction and no change in cell potential occurs.11 As such, two approaches based on the Ag/AgCl reference electrode have been used to measure electrode potential in nuclear power reactors. In the first approach, an internal reference electrode operates in the same high-temperature environment as the reactor. In this case, one must consider the solubility of
Corrosion and Compatibility
Sapphire lid AgCl pellet
5
Rulon adapter Compression fitting
Sapphire container
Pt cap Ni wire Alumina insulators
Ceramic to metal braze Restrainer
Ag/AgCl
Kovar TIG weld 304SS
Seal
Coaxial cable Figure 3 Diagram of an internal Ag/AgCl reference electrode used in BWRs. Top end is inserted into the cooling loop, while the coaxial cable provides electrical connection. Reprinted from Indig, M. E. In 12th International Corrosion Congress, Corrosion Control for Low-Cost Reliability; NACE International: Houston, TX, 1993; p 4224, with permission from NACE International.
Ag/Cl complexes that form as a function of temperature eqn [6].12,13 That is, reactions in addition to eqns [VIII] and [XI] must be considered. From a construction viewpoint, the internal reference electrode consists of a silver chloride pellet on a platinum foil (Figure 3).14 External potential measurement is made via contact with a nickel wire which is connected to an electrometer via a coaxial cable. The electrode is housed in a sapphire tube that is sealed via a porous sapphire cap. In this configuration, there is no internal electrolyte per se. Upon placing the electrode in a boiling water reactor (BWR), the porous cap allows the cooling water to penetrate the electrode and the potential is determined from eqn [6] and the solubility of AgCl in high purity water as a function of temperature.15 In the second approach, an external pressurebalanced reference electrode is used (EPBRE). In the EPBRE, the reference electrode is maintained at room temperature and pressure and the corresponding constants are used in eqn [6]. The reference is connected to the high-temperature environment via a nonisothermal salt bridge sealed with a porous zirconia plug (Figure 4).16 As a result of this configuration, the EPBRE is not susceptible to
Compression fitting 1/4 NPT Pure water or 0.01 M KCl Glass wick 1/4 OD SS tube
Heat-shrinkable PTFE tube
SS nut Rulon sleeve Zirconia plug
Figure 4 A diagram of a pressure-balanced reference electrode is used in BWRs. Bottom of figure is sealed into pressure vessel via compression fitting while Ag/AgCl electrode (top) remains at room temperature and pressure. Reprinted from Oh, S. H.; Bahn, C. B.; Hwang, I. S. J. Electrochem. Soc. 2003, 150, E321, with permission from The Electrochemical Society.
potential deviations owing to the solubility of AgCl and its complexes as a function of temperature. However, the temperature gradient between the reactor and the reference electrode results in a junction potential that must be subtracted from eqn [6]. The corresponding thermal liquid junction potential (ETLJ)17 is given by: ð 1 T2 tMþ SMþ tCl SCl dT ½7 ETLJ ¼ þ zMþ zCl F T1 where t, S, and z represent the transport number, the entropy of transport, and the charge on the cation, respectively. The symbol M in eqn [7] represents the metal in the chloride salt, MCl, and is commonly Li, Na, or K. In addition to ETLJ, there is also the isothermal liquid junction potential, EILJ, which arises due to the differences in cation and anion mobilities through the porous frits and the fact that the electrolyte in the external reference (typically KCl) is vastly different from the reactor cooling water in which it is immersed17: ð RT T2 ti EILJ ¼ dln½ai ½8 F T1 zi
6
Corrosion and Compatibility
where the subscript i denotes a species that may be transported through the zirconia plug and for a nuclear power reactor may include species ions as Agþ, Cl, Hþ, OH, Kþ, and B(OH4). It has been shown that both ETLJ and EILJ each increase by as much as 0.15 V over the temperature range of 25–350 C. The result is a decrease in the measured potential of 0.30 V at 350 C. While these junction potentials can be calculated and used to correct eqn [7], it has been shown that there is some deviation at higher temperatures (>200 C) and an experimental fitting procedure is the preferred method for calibration of the reference electrode.
reactions with two soluble species
3þ Fe ½11 Fe2þ ¼ Fe3þ þ e E 0 ¼ 0:771 þ 0:059log Fe2þ
solubility of iron and its oxides Fe ¼ Fe2þ þ 2e E 0 ¼ 0:440 þ 0:0295logðFe2þ Þ
2Fe2þ þ 3H2 O ¼ Fe2 O3 þ 6Hþ þ 2e E 0 ¼ 0:728 0:177pH 0:059logðFe2þ Þ
An atlas of electrochemical equilibria has been created by M. Pourbaix for metals in aqueous solution at room temperature.18 This atlas contains potential– pH diagrams, so-called Pourbaix diagrams, which define three equilibrium thermodynamic domains for metals in aqueous solutions: immunity, passivity, and corrosion. Immunity is defined as the state where the base metal is stable while corrosion is defined as the formation of soluble metal cations and passivity the formation of a stable oxide film. Pourbaix’s derivation requires that the values of the standard chemical potential, m0, for all of the reacting substances are known for the standard state at the temperature and pressure of interest. For chemical reactions at room temperature, the equilibrium conditions are defined by the relationship18: P 0 nm ½9 logK ¼ 5708 and for electrochemical reactions at room temperature (Table 1) equilibrium is defined by: P 0 nm 0 ½10 E ¼ 96485n where K is the equilibrium constant for the reaction, m0 is in Joules per mole, v is the stoichiometric coefficient for the species, n is the number of electrons, 5708 is a conversion constant equal to RT/(log10e) where T is temperature (298.15 K) and R the ideal gas constant (8.314472 J (K mol)1), and 96 485 is Faraday’s constant in J (mol V)1. As an example of these diagrams, consider the iron–water system and the solid substances Fe, Fe3O4, and Fe2O3. Pourbaix18 defined the relevant equations for this system as:
½13
þ Fe þ 2H2 O ¼ HFeO 2 þ 3H þ 2e
E 0 ¼ 0:493 0:089pH þ 0:0295log½HFeO 2
5.01.1.5 The Thermodynamics of Corrosion from Room Temperature to the PWR
½12
½14
þ 3HFeO 2 þ H ¼ Fe3 O4 þ 2H2 O þ 2e
E 0 ¼ 1:819 þ 0:029pH 0:088log½HFeO 2
½15
reaction of two solid substances 3Fe þ 4H2 O ¼ Fe3 O4 þ 8Hþ þ 8e E 0 ¼ 0:085 0:059pH
½16
2Fe3 O4 þ H2 O ¼ 3Fe2 O3 þ Hþ þ 2e E 0 ¼ 0:221 0:059pH
½17
An example of a simplified Pourbaix diagram for Fe at room temperature based on the reactions in eqns [11]–[17] is presented in Figure 5, where Eq. [12] corresponds to figure line 23, [13] to line 28, [14] to line 24, [15] to line 27, [16] to line 13 and [17] to line 17. Note that Eq. [11] is the boundary between Fe2þ and Fe3þ and was not drawn in the original figure. In addition to the lines separating the domains for Fe, Pourbaix diagrams will typically include the domains associated with water stability (oxidation and reduction) represented by the dashed lines marked a and b in Figure 5. Upon inspection of this diagram one would conclude what is know from experience with Fe: that iron is passive in alkaline solutions and at higher applied potentials owing to oxide film formation while at more acidic solutions Fe is susceptible to corrosion owing to Fe2þ. It is worth noting again that these potential–pH domains are defined solely by the thermodynamic stability of the species within them and these diagrams do not consider kinetics which will be addressed later in this chapter. This is important as while a species/reaction may be thermodynamically stable it may be kinetically hindered. While the use of Pourbaix diagrams to characterize room temperature corrosion reactions is
Corrosion and Compatibility
1.5
Fe3+
1.5
Fe3+ 20
7
20
1.0
1.0 b
Fe2O3
Fe2+
0
EH2 (200 C) (V)
28
2
EH (25 C) (V)
0.5
a 26
-0.5
17 Fe3O4
23
0.5
0
Fe2+
26
13
-1.0
HFeO-2
17
Fe3O4
23 27 24
Fe
Fe2O3 a
-0.5
13
-1.0
b
28
Fe
HFeO22-
-1.5
-1.5 0
5
10
15
pH
widespread, these diagrams and the method for generating them as presented thus far cannot be used at the higher temperatures associated with nuclear power reactors. This is due to the lack of standard potentials at elevated temperature as required by eqns [9] and [10] (e.g., the application of Table 1 to higher temperature). In the absence of these high-temperature thermodynamic data, Townsend19 used an extrapolation method introduced by Criss and Coble (the correspondence principle). The method allows for empirical entropy data of ionic species at 25 C to be extrapolated to higher temperatures. In this method, the standard Gibbs free energy is calculated from the relationship: ð
T
ðT 250
0
T
250
DC p ðT ÞdlnT
0
DC p ðT ÞdT ½18 0
where DS is the standard entropy change and DC p is the change in standard partial molar heat capacity. The potential–pH diagram for the Fe–H2O system and the solid substances Fe, Fe3O4, and Fe2O3 at 200 C calculated by Townsend is presented in Figure 6. In comparison with the diagram at 25 C (Figure 5), the Fe2O3 and Fe3O4 regions are extended to lower pH and potentials. As a result the area associated with corrosion at lower solution pH is decreased. However, 0
24
5
10
15
pH
Figure 5 A simplified potential–pH diagram for the Fe–H2O system and the solid substances Fe, Fe3O4, and Fe2O3 at 25 C based on the reactions in eqns [11]–[17]. Reprinted from Townsend, H. E. Corrosion Sci. 1970, 10, 343, with permission from Elsevier.
DðDG 0 Þ ¼ DT DS 0 ð250 Þ þ
0
29
27
Figure 6 The potential–pH diagram for the Fe–H2O system and the solid substances Fe, Fe3O4, and Fe2O3 at 200 C. Most dramatic influence of increased temperature is the presence of a large region of soluble species (corrosion) at high pH. Reprinted from Townsend, H. E. Corrosion Sci. 1970, 10, 343, with permission from Elsevier.
the most notable change in the diagram is at high pH where the area associated with corrosion owing to the soluble HFeO 2 has increased dramatically. The Criss and Coble method is limited, however, to the 150–200 C range and, to extend the Pourbaix to the temperatures of power reactors, Beverskog used a Helgeson–Kirkham–Flowers model to extend the heat capacity data to 300 C.20 Thus far, we have described a method for generating electrochemical equilibria diagrams and regions of passivity, corrosion, and immunity for pure metals from 25 to 300 C. From an engineering standpoint, we would like to know this information for structural alloys such as austenitic stainless steels and super nickel alloys. At temperatures near 25 C, the predominant oxide responsible for passivity is Cr2O3 and it is sufficient to rely only on the Cr potential–pH diagram for alloys with a high Cr content. However, at higher temperatures other oxides form such as Fe(Fe,Cr)2O4, (Cr,Fe)2O3, (Cr,Fe,Ni)3O4, and (Cr,Fe,Ni)2O3, and it is desirable to know the thermodynamic stability of the alloy. Beverskog has developed the ternary potential– pH diagrams for the Fe–Cr–Ni–H2O–H2 system for temperatures up to 300 C using heat capacitance data and the revised Helgeson–Kirkham–Flowers model described above.21 However, Fe–Cr–Ni phases lack thermodynamic data and the ternary oxides were,
8
Corrosion and Compatibility
npH 2
H2CrO4(aq) HCrO4-
Potential (VSHE)
1
CrO42Cr(OH)2+ 0 NiCr2O4(cr) Cr2+
FeCr2O4(cr)
-1
Cr2O3(cr) Cr(cr) -2
0
2
4
6 pH300 C
8
10
Figure 7 Potential–pH diagram for chromium species in Fe–Cr–Ni at 300 C. Concentration of aqueous species is 106 molal. Reprinted from Beverskog, B.; Puigdomenech, I. Corrosion 1999, 55, 1077, with permission from NACE International.
thus, not considered. The diagrams assumed that the metallic elements in the alloy had unit activity, that is, equal amounts of iron, chromium, and nickel. An example of the potential–pH diagram for chromium species in Fe–Cr–Ni at 300 C and aqueous species with a concentration of 106 molal is presented in Figure 7. Unlike the Fe diagram, where the presence of soluble HFeO2 2 species increased with temperature (Figure 6), the diagram for Cr in Fe–Cr–Ni is dominated by passive region where the stable oxides of Cr2O3, FeCr2O4, and NiCr2O4 are formed. 5.01.1.6 Kinetics of Dissolution and Passive Film Formation The study of dissolution kinetics, corrosion rate, attempts to answer the question: ‘‘What are the relationships that govern the flow of current across a corroding interface and how is this current flow related to applied potential?’’ Consider the anodic dissolution of a metal with an activation barrier equal to G1a ¼ nFE (eqn [2]). If we increase the driving force (potential) from its equilibrium condition, E0, to a new value, f, the new barrier is given by the relationship22:
decreasing the barrier, that is, not all of the applied potential is dropped across the electrochemical double layer. The rate (ra) of this reaction is expressed in the same, Arrhenius, form as for chemical reactions: DGa ½20 ra ¼ ia =nF ¼ ka co exp RT where ia is the anodic current density, ka is the rate constant, co is the concentration of oxidized species, and DGa is the change in free energy for the anodic reaction. Substituting G2a G1a in eqn [20] for DGa in eqn [19], we express the anodic reaction rate as22: ð1 bÞnF ½21 ia ¼ nFka cR exp RT where (the overpotential) represents a departure from equilibrium and is equal to f E0. We can derive a similar expression for the cathodic reaction22: bnF ½22 ic ¼ nFkc co exp RT where ic is the cathodic current density, kc is the rate constant, and co is concentration of oxidized species. Combining eqns [21] and [22] and rearranging them, we can write an expression for the total current, i: ð1 bÞnF bnF exp ½23 i ¼ io exp RT RT where io is the exchange current density and is equal to b b nFkcc1b o ka cR . This expression is commonly referred to as the Butler–Volmer equation. To apply eqn [23] to corrosion reactions, we need to be able to relate to the corrosion potential, Ecorr , that is, as it stands the Butler–Volmer equation is derived for equilibrium conditions. Returning to our definition of the overpotential ¼ f E0, by both subtracting and adding Ecorr from the right side of this definition, inserting the resulting expression back into eqn [23] and rearranging we find22: ð1 bÞF ðEcorr Ea Þ icorr ¼ ia exp RT bF ¼ ic exp ðEcorr Ec Þ ½25 RT
½19
For small applied potentials around Ecorr , the Stern– Geary approximation of eqn [24] is used23: ba þ bc ðf Ecorr Þ ½26 i ¼ 2:303icorr ba bc
where b is the symmetry factor and reflects the fact that not all of the increase in potential goes to
where ba and bc are defined as the anodic and cathodic Tafel slopes (discussed in Section
Ga2 ¼ Ga1 ð1 bÞnF ðf E0 Þ
9
0.4
0.1
0.35
0 Current density (A m-2)
Potential (V vs. SCE)
Corrosion and Compatibility
0.3 0.25 0.2 Beam on at 100 nA ~540 s
0.15 0.1
0
500
1000 Time (s)
1500
2000
5.01.2.2) having units of volts and are empirical factors related to the symmetry factor by the relationships22: RT ð1 bÞnF
bc ¼ 2:303
RT bnF
-0.1 -0.2 Increasing radiation flux Increasing cathodic reaction rate
-0.3 -0.4 -0.5
Figure 8 Influence of proton irradiation on the Ecorr of a SS 304L electrode in dilute sulfuric acid, pH ¼ 1.6. The increase is caused by the production of water radiolysis products.
ba ¼ 2:303
Beam = 35 na Beam = 62 na Beam = 100 na
½27 ½28
As it relates to the nuclear power industry, eqn [24] not only relates the corrosion rate (icorr) to the applied potential, f, but it can also help us to rationalize other processes such as the influence of water radiolysis products on corrosion rate. For example, it is generally observed that ionizing radiation (g, neutron, proton, etc.) increases Ecorr potential (Figure 8) and corrosion rate at Ecorr .24 The flux of ionizing radiation on the cooling water results in radiolysis, the breaking of chemical bonds. During the course of water radiolysis, a wide variety of intermediate products are formed, such as O2, eaq, and the OH radical.25 The vast majority of these species have very fast reaction rates so that the end result is a handful of stable species. These stable products are typically oxidants, such as O2, H2, and H2O2. That is, these products readily consume electrons (eqns [II]–[IV]) and increase cathodic reaction rate (Figure 9). From eqn [25] we see that an increase in the cathodic reaction rate, ic, necessarily results in an increase in corrosion rate, icorr , consistent with the observation described. While the development of dissolution kinetics is straightforward, the kinetics associated with passive
-0.6 -0.2
-0.1
0 0.1 0.2 0.3 Potential (V vs. SCE)
0.4
0.5
Figure 9 Influence of proton irradiation on the cathodic reactions on a Au electrode in dilute sulfuric acid, pH ¼ 1.6. The increase is caused by the production of water radiolysis products.
film formation and breakdown are less well understood yet equally as important to our understanding of corrosion mechanisms. One such example is the case of localized corrosion where the probability for a pit to transition from a metastable to a stable state is governed by the ability of the active surface to repassivate. Another example is the initiation of SCC where passive film rupture results in very high dissolution rates and, correspondingly, crack advance rate which is controlled by the activation kinetics described above as the bare metal dissolves.26–28 During the propagation stage of SCC, the crack tip must propagate faster than (1) the oxide film can repassivate the surface and (2) the corrosion rate on the unstrained crack sides so that dissolution of the walls does not result in blunt notch. To evaluate the role of repassivation kinetics in SCC and corrosion mechanisms in general, investigators set about measuring three critical experimental parameters, namely film: ductility,29,30 bare surface dissolution rates,31–33 and passive film formation rates34,35 for various alloys. Each of these techniques involves the depassivation of a metal electrode using a tensile frame or a nano-indenter (in the case of ductility studies) or scratching/breaking an electrode (bares surface current density and repassivation studies) and measuring the resulting current transient as a function of time. An example of a current transient for SS 304L in chloride solution is presented in Figure 10. The surface was under potentiostatic control and was bared using a diamond scribe. The data were collected using a high-speed oscilloscope.
10
Corrosion and Compatibility
7.0 10−4 6.0 10−4
Current (A)
5.0 10−4
td
tr
0
0.002 0.004 0.006 0.008 Time (s)
4.0 10−4 3.0 10−4 2.0 10−4 1.0 10−4 0.0 0.01
0.012
Figure 10 Scratch test current transient from a SS 304L electrode in 0.1 M NaCl. The transient is characterized by a growth period, td, and a repassivation period, tr.
The transient is characterized by two separate processes, anodic dissolution and repassivation represented by td and tr in Figure 10. To analyze the repassivation rates, the period tr is typically fit to an expression and evaluated as a function of solution pH or electrode potential. The most prolific work in this field is probably on the alloy SS 304L. For this alloy, it has been proposed that the kinetics of film growth are controlled by ion migration under high electric field.36–38 The kinetics of high-field film growth were first proposed by Cabrera and Mott39 to obey the kinetic relationship: BV ½29 i ¼ Aexp L where i is the current density, V is the voltage, L is the oxide thickness, and A and B are constants.
5.01.2 Analytical Methods 5.01.2.1
Introduction
In this section, we will review the principle analytical methods used to probe the electrochemical interface. In the Section 5.01.2.2 Potentiodynamic Polarization, we discuss linear polarization resistance and the practical application of corrosion kinetics, eqns [25] and [26]. In that section, we also describe the salient points of the anodic polarization curve. In the Section 5.01.2.3 Electrochemical Impedance Spectroscopy, we introduce an ac method for interrogating the electrochemical interface. This technique is
probably the most versatile experimental method available to scientists and researchers. As it relates to nuclear reactors, this technique has the ability to subtract out the contribution of the solution resistance to polarization resistance measurements which, if not accounted for in highly resistive cooling water measurements will result in nonconservative corrosion rates. In the final section, we introduce a more seldom used technique, Mott–Schottky analysis. While this is by no means a common experimental method, it provides a conduit for the reader to become familiar with defects in the oxide film, their transport and ways to quantify it. This has particular interest here as ionizing radiation may promote corrosion rates by increasing transport of these defects through the passive oxide film. Regretfully, the scope of this chapter is limited, and we are not able to discuss the step-by-step details of the experimental methods that are used to make corrosion measurements. A comprehensive guide to experimental methods in corrosion has been published by Kelly et al.40 as well as Marcus and Mansfeld41 while a more broad description of electrochemical methods has been published by Bard and Faulkner.42 The reader is also encouraged to become familiar with the equipment that is used to make electrochemical measurements and a good introductory chapter on this topic has been presented by Schiller.43 The most important instrument is, no doubt, the potentiostat. While this instrument is the cornerstone of corrosion science, it does have its pitfalls including bandwidth limitations and the potential for ground loop circuits when used in conjunction with other equipment such as load frames, autoclaves and cooling loops. The latter can be overcome using proper instrumentation such as potentiostat with a floating ground, or isolation amplifiers. To investigate the influence of the neutron flux on corrosion rates and mechanisms, real-time in-situ corrosion measurements are often made ‘in-reactor’ or at neutron facilities such as Oak Ridge National Lab’s Hi-Flux Isotope Reactor (HIFER) and Argonne National Lab’s Fast Flux Test Facility (FFTF). Alternately, neutron damage can be simulated using ion beams. As it relates to ion irradiation, this method provides opportunities for studying the interaction of the components of reactor environments (radiation, stress, temperature, aggressive media) that are not possible with in-reactor or neutron irradiation facilities. For a full discussion of this topic, see Chapter 1.07, Radiation Damage Using Ion Beams. To summarize these experiments, controlled
11
Corrosion and Compatibility
environmental cells are coupled to accelerator beamlines to study the interaction of the environment and irradiation on structural materials. Corrosion at the substrate–environment interface is studied in real time by numerous electrochemical techniques including those described below. With respect to dose, light ions can be used to reach doses up to 10 dpa in several days. However, the depth of penetration is low (tens or micrometers), which puts unrealistic limitations on electrochemical cell construction. Heavy ion irradiation can reach several hundred displacements per atom in a matter of days but the penetration depth is much less. 5.01.2.2
Potentiodynamic Polarization
During our development of Butler–Volmer reaction kinetics, we introduced the Stern–Geary approximation and defined Tafel slopes within the context of the symmetry factor without much further explanation. To understand the empirical source of the Tafel slopes, we rearrange eqn [26] to define the polarization resistance Rp in units of O cm2: Rp ¼
ba bc DE ¼ 2:303icorr ðba þ bc Þ Di
½30
The Tafel slopes can be obtained from a plot of potential as a function of the logarithm of current density as shown for the cathodic curve in Figure 11 where bc has units of volts. A similar anodic plot can be generated to obtain ba. It is important to realize that these slopes are frequently not equal as they are related to separate mechanisms on the electrode surface. From a plot of both the anodic and cathodic Tafel slopes, we can also obtain icorr (Figure 11). With knowledge of icorr , ba, and bc, the polarization iOH2 EH
+
+
H
2
20
10 Slope =
Applied current curve
Ecorr +
M
20
20
iapp(cathodic)
icorr
E iOM
h(mv)
40 H+
/H2
resistance can then be determined from eqn [30]. For small voltage perturbations, the slope of a plot of applied potential (DE ) versus the change in current density (Di ) is equal to Rp (Figure 12), often referred to as the linear polarization resistance as the slope is only linear for small voltage perturbations around Ecorr . Errors not accounted for in eqn [30] include uncompensated solution resistance (RO) and choosing a scan rate that is too fast. From a reactor standpoint, the value of RO in the cooling water or a simulant where the resistivity is high may be a significant contribution to the measured Rp and, therefore, result in a nonconservative corrosion rate, that is, the calculated corrosion rate will be too low. For a complete standardized method of conducting potentiodynamic polarization resistance measurements and data analysis, the reader is referred to the appropriate ASTM standards G3, G59, and G102.44–46 Potentiodynamic polarization curves can also be used to assess corrosion rate as well as determine if a material is passive, active, or susceptible to pitting corrosion in a given environment. Consider the curve in Figure 13. It plots the applied potential as a function of the log of the absolute value of the current density. The corrosion potential and corrosion current density are shown at the intersection of the cathodic and anodic curves. Mass loss (m, in grams) for a given period of exposure can then be
DE Diapp
40 iapp(anodic)
-10
Tafel region
M
- EM/M+ log iapplied
Figure 11 Diagram depicting the cathodic polarization of an electrode near Ecorr. Tafel extrapolation showing the determination of icorr is also presented. Reprinted from Fontana, M. G. Corrosion Engineering; McGraw Hill: New York, 1986, with permission from McGraw Hill.
-20 h(mv) Figure 12 Diagram depicting the linear polarization for an electrode about Ecorr. Slope, DE/Di, is equal to the polarization resistance. Reprinted from Fontana, M. G. Corrosion Engineering; McGraw Hill: New York, 1986, with permission from McGraw Hill.
12
Corrosion and Compatibility
positive hysteresis in the reverse portion of the curve that is not observed in the case of solution oxidation or transpassivity.
Epit Potential
Erepass
5.01.2.3 Electrochemical Impedance Spectroscopy Eflade
ba ipass
icorr
bc
Ecorr
Log current density Figure 13 Diagram depicting the potentiodynamic polarization of an electrode far from Ecorr. Relevant potentials and currents are defined in the text.
determined from the corresponding icorr using Faraday’s Law, for a pure metal40: icorr t EW ½31 FA where t is time in seconds, icorr has units of A (cm2), A is surface area, and EW is equivalent weight and is equal to molecular weight (M) divided by the number of electrons in the reaction. Similarly, for an alloy: m¼
EW ¼ P
1 n i i fi =Mi
½32
where the subscript i denotes the alloying element of interest and f is the mass fraction of that element in the alloy. From Faraday’s law, it is also possible to calculate corrosion rate, the penetration depth owing to corrosion over a period of time in units of mm year1 (CR): CRmm year1 ¼
3:27 103 icorr EW r
½33
where r is the material density in g cm3. Other critical parameters in Figure 13 include the Flade potential (EFlade) which marks the critical potential necessary for passive film formation, the passive current density (ipass), the pitting potential (Epit), and the repassivation potential (Erepass). With respect to EFlade, this active to passive transition is not observed for materials that are spontaneously passive in a given environment such as SS in a BWR. In such a case, the current would be limited by the film dissolution current (ipass). As it relates to Epit, the onset of localized corrosion is characterized by a sharp increase in current at a given potential. As such, materials that are more susceptible to pitting have lower Epit values. Pitting is also characterized by a
While potentiodynamic polarization is typically considered a destructive technique, that is, it alters the surface of the corrosion sample, electrochemical impedance spectroscopy (EIS) is a powerful nondestructive technique for obtaining a wealth of data including Rp.47 Further, this technique has the ability to subtract out the uncompensated solution resistance (RO) from the measurement which, for highly resistive reactor cooling water environments, is a significant advantage. In EIS, a small sinusoidal voltage perturbation (10 mV) is applied across the electrode interface over a broad frequency range (mHz to MHz). By measuring the transfer function of the applied voltage to the system current, the system impedance may be obtained. For corrosion systems, the impedance (Z) is a complex number and may be represented in Cartesian coordinates by the relationship: ZðoÞ ¼ Z0 þ Z00
½34
where o is the applied frequency in radians, Z0 is the real part of the impedance, and Z00 is the imaginary part of the impedance, and the magnitude of the impedance jZj is given by: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ½35 jZj ¼ ðZ0 Þ2 þ ðZ00 Þ2 In its simplest form, the electrochemical interface can be thought of as an electrical equivalent circuit (EC): a resistor (R) with an impedance Z(o) ¼ R and a capacitor (C) with Z(o) ¼ 1/joC where j is the square root of 1.48 Thus, the impedance of a resistor is purely real and independent of frequency while the impedance of a capacitor is purely imaginary and inversely proportional to frequency. With respect to the electrochemical interface, the polarization resistance is in parallel with the double layer capacitance (Cdl, owing to adsorption of charged anions/cations in the electrolyte). These two components act in series with the solution resistance, RO as seen in the EC in Figure 14. This circuit is referred to as a simple Randles circuit and represents an ideal interface. Commonly, however, Cdl does not behave as an ideal capacitor and its impedance is better represented by the expression
13
Corrosion and Compatibility
-6 104
Rp RW
-5 104 -4 104
Figure 14 Simplified Randles equivalent circuit of an electrochemical interface where Rp is the polarization resistance, Cdl is the double layer capacitance, and RO is the geometric resistance associated with the solution resistance.
Z (W)
Cd1
-3 104
-2 104 -1 104
0
105
0
1 104 2 104 3 104 4 104 5 104 6 104 Z (W)
Rp + RW
Figure 16 Nyquist format for data in Figure 15 where Rp ¼ 5 104 O, Cdl ¼ 4 106, and RO ¼ 200 O.
104 |Z| (W)
ZðoÞ ¼ 1=Cð j oÞa 103
RW 102 10-3 10-2 10-1 100 101 102 (a) Frequency (Hz)
103
104
105
0 -10
Q ()
-30 -40 -50
-70
wmax
-80 -90 10-3 10-2 10-1 100 101 102 (b) Frequency (Hz)
where a is typically found to be between 0.5 and 1. The element that represents this behavior is known as a constant phase element (Ccpe).49 The response of a simple Randle’s circuit as a function of frequency is shown in Figure 15(a) and 15(b). These plots are referred to as the Bode magnitude plot (eqn [35]) and the Bode phase plot50 where the phase angle, y, is equal to: 00 Z ½37 y ¼ tan1 Z0 This phase angle is a result of the double layer capacitance where the current leads the applied ac voltage perturbation as is the case for pure capacitors. The parameters Rp and RO may be determined graphically from the Bode magnitude plot as shown in Figure 15(a) while Cdl is determined from the graphical parameter omax (converted to radians, 2pf) in Figure 15(b) and the relationship40: 1 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 þ Rp =RO ½38 Cdl ¼ omax Rp
-20
-60
½36
103
104
105
Figure 15 (a) Bode magnitude and (b) Bode phase plots. Data were generated from an electrical equivalent of a simplified Randles circuit (Figure 13) where Rp ¼ 5 104 O, Cdl ¼ 4 106, and RO ¼ 200 O.
Alternately, the data may be presented using the Nyquist format which plots the imaginary impedance as a function of the real impedance as seen in Figure 16 (sometimes referred to as a Cole–Cole plot). As graphical analysis is somewhat imprecise, commercially available complex nonlinear least squares fitting of the data (CNLS) is commonly used to obtain these parameters (Figure 17).
14
Corrosion and Compatibility
|Z| exper. |Z| fit
102
ROX
0
RW
-10
|Z| (W m2)
-30
Q exper. Q fit
-40 -50
100
Q (degrees)
-20 101
-60
COX Figure 19 Equivalent circuit of an oxide-covered metal in liquid PbBi eutectic (LBE) where Rox is the resistance of the passive film, Cox is the double layer capacitance of the oxide, and RO is the geometric resistance associated with the LBE.
-70 10-1 10-3
10-2
10-1
100
101
102
-80 103
Frequency (Hz) Figure 17 Bode magnitude and Bode phase plots for SS 304L during proton irradiation in a pressurized deionized cooling water loop at 125 C showing experimental data and complex nonlinear least squares fit of equivalent circuit in Figure 14.
Corrosion rate (mm year −1)
6.0
5.0
4.0
3.0 Photons Neutrons 2.0
1.0
Protons
0
20
100 120 40 60 80 Flux (particles m–2 per proton)
140
Figure 18 Corrosion rate as a function of particle flux for a SS 304L electrode during proton irradiation in a pressurized deionized cooling water loop at 125 C.
EIS has been used successfully to investigate the passive films on Zr alloys,51–53 SS 304L,54 and nickel base alloys55 in environments that simulate reactor cooling water systems. In addition, it has also been used to measure the real-time corrosion rates of materials during irradiation. Lillard et al.56,57 measured the corrosion rate of materials as a function of immersion time in a deionized water cooling loop during proton irradiation. In that work, radiationhardened probes made from Alloy 718, SS 304L, and Al 60 601 were exposed to a proton beam at
various current densities. The impingement of the beam on the probes resulted in a mixed neutron, photon, and proton flux. It was shown that corrosion rate was almost linear with photon and neutron flux as compared to proton flux where anomalies existed at intermediate fluxes (Figure 18). The data were ultimately used to extrapolate lifetimes for accelerator materials. In other studies, EIS has also been applied to investigate passive films on metals in sodium (Na)cooled and lead–bismuth (LBE) systems that simulate reactor environments. The equivalent circuit (EC) used to model the data is similar to the simplified Randles circuit; however, in this case, there is no electrochemical double layer, only the impedance and capacitance associated with the oxide (Figure 19). In this EC, Rox is the dc resistance of the oxide, Cox is the oxide capacitance, and RO is the geometric resistance associated with the liquid metal (Na or LBE). In one such study, Isaacs investigated the capacitance of anodized films on Zr in liquid Na. In that study, it was shown that both Rox and Cox were a function of Na temperature between 50 and 400 C. Lillard et al.58 reported similar trends for HT-9 in LBE. At constant temperature, Rox and Cox for HT-9 in LBE were a function of immersion time (Figure 20). The data were used to calculate oxide thickness as a function of time. In addition, Rox was related to ionic transport through the film and corrosion rates were calculated using Wagner’s oxidation theory. Upon irradiation in a proton beam, this rate fell even further.59 Additional information about leadeutectic coolants may be found in Chapter 5.09, Material Performance in Lead and Lead-bismuth Alloy. 5.01.2.4
Mott–Schottky Analysis
The formation and growth of passive oxide films is driven, fundamentally, by interfacial reactions and defect, electron and ion transport processes. Yet the
Corrosion and Compatibility
105
ROX
15
5.5 1010
101
5.0 1010
104
Prior to irradiation
Fit
Beam on
Fit
10–1
101
1 C-2 (cm4 F-2)
ROX (W cm−2)
Rp = 0.9 ´ t1.7 102
COX (nF cm–2)
100 103
4.5 1010 4.0 1010 3.5 1010
10–2 COX
100 10–1 –50
0
50 100 Immersion time (h)
150
3.0 1010 10–3 200
2.5 1010 -0.2
0
0.2
0.4
0.6
0.8
1
Potential (V vs. SCE)
Figure 20 Impedance capacitance for the oxide on an HT-9 steel as a function of immersion in liquid PbBi eutectic. Reprinted from Lillard, R. S.; Valot, C.; Hanrahan, R. J. Corrosion 2004, 60, 1134, with permission from the author.
Figure 21 Mott–Schottky plots for a SS 304L electrode in dilute sulfuric acid, pH ¼ 1.6. Before and after proton irradiation. Reprinted from Lillard, R. S.; Vasquez, G. J. Electrochem. Soc. 2008, 155, C162, with permission from the author.
nature and relative importance of these processes are still far from being understood. Key to oxide growth, and, therefore, passivation, is the mobility of these defects, specifically vacancies. Under irradiation, however, the defect properties of the oxide are undoubtedly changed, and the extent of corrosion is related both to the microstructure and transport properties of defects. One way to probe the transport properties of the oxide film is by using Mott–Schottky theory. Returning to our discussion of EIS above, the oxide capacitance may be obtained at high frequency from the relationship
depends on what type of semiconductor the oxide is (p vs. n) and this effects the sign of the slope. Lillard used Mott–Schottky analysis to examine the influence of proton irradiation on defect generation and transport in the oxide film on SS 304L.61 The passive film on SS 304L is an n-type semiconducting oxide and the major defect is the oxygen vacancy which acts as an electron donor. According to the Point Defect Model62 for oxide growth, oxygen vacancies are produced at the metal–film interface by the injection of a metal atom into the oxide lattice:
Z00 ðoÞ ¼ 1=oC or C ¼ 1=oZ00 ðolim Þ
½39
By measuring Z00 as a function of applied dc voltage at the high frequency limit (olim) and calculating the film capacitance from eqn [39], it is possible to evaluate donor concentration (ND) in the oxide film from the well-known Mott–Schottky relationship: 1 kT ½40 ¼ 2=ee N U U 0 D fb C2 q where e0 is the permittivity of space and is equal to 8.854 1014 F cm1, e is the electronic charge and is equal to 1.602 1019 C, U is the applied potential in V, Ufb is the flatband potential in V, kT/q is equal to 25 mV at 25 C, and ND is the donor concentration (oxygen vacancies) in cm3.60 ND is the slope of a plot of 1/C2 versus applied potential. The type of defect
m ! MM þ ðx=2ÞVo þ xe
½41
where m is a metal atom, MM is a metal cation, Vo is a oxygen vacancy, e is an electron, and x is the stoichiometric coefficient. In pH 1.6 H2SO4, it was found that the cation vacancy concentration in the oxide increased from 2.94 1021 in the absence of irradiation to 3.41 1021 during irradiation (Figure 21). It was proposed that the film on SS 304L is composed of an inner Cr-rich p-type semiconductor and an outer Fe-rich n-type semiconductor. This bilayer film results in a p–n junction where the two layers meet. On the p side of the junction, there is a surplus of holes, while on the n side of the junction, a surplus of electrons exists. The energy bands are such that it is ‘uphill’ (an increase in energy) for electrons moving across the junction (from the n side to the p side). On a related topic, a discussion of defects in bulk oxides can be found in Chapter 1.02, Fundamental Point Defect Properties in Ceramics.
16
Corrosion and Compatibility
References 1. 2. 3. 4. 5.
6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.
28. 29. 30. 31. 32.
Fontana, M. G. Corrosion Engineering; McGraw Hill: New York, 1986. Cox, B. J. Nucl. Mater. 2005, 336, 331. Lysel, G.; Nystrand, A. C.; Ullberg, M. J. ASTM Int. 2005, 2, 355. Abella, J.; Balachov, I.; Macdonald, D. D.; Millet, P. J. Corrosion Sci. 2002, 44, 191. Scott, P. M.; Meunier, M. C.; Deydier, D.; Silvestre, S.; Trenty, A. In American Society for Testing and Materials Special Technical Publication; Kane, R. D., Ed.; ASTM: Conshohocken, PA, 2000; p 210. Hur, D. H.; Choi, M. S.; Lee, D. H.; Song, M. H.; Han, J. H. Corrosion 2006, 62, 905. Hwang, S. S.; Kim, H. P.; Kim, J. S. Corrosion 2003, 59, 821. Helmholtz, H. Pogg. Ann. 1853, LXXXIX, 211. Whitten, K. W.; Gailey, K. D. General Chemistry; Saunders College: Philadelphia, PA, 1981. Castellan, G. W. Physical Chemistry; Addison-Wesley: Reading, MA, 1983. Danielson, M. J. Corrosion 1995, 51, 450. Macdonald, D. D. Corrosion 1978, 34, 75. Oijerholm, J.; Forsberg, S.; Hermansson, H. P.; Ullberg, M. J. Electrochem. Soc. 2009, 156, P56. Indig, M. E. In 12th International Corrosion Congress, Corrosion Control for Low-Cost Reliability; NACE International: Houston, TX, 1993; p 4224. Indig, M. E.; Nelson, J. L. Corrosion 1991, 47, 202. Oh, S. H.; Bahn, C. B.; Hwang, I. S. J. Electrochem. Soc. 2003, 150, E321. Lvov, S. N.; Macdonald, D. D. J. Electroanal. Chem. 1995, 403, 25. Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions; NACE/Cebelcor: Houston, TX, 1974. Townsend, H. E. Corrosion Sci. 1970, 10, 343. Beverskog, B.; Puigdomenech, I. Corrosion Sci. 1996, 38, 2121. Beverskog, B.; Puigdomenech, I. Corrosion 1999, 55, 1077. Prentice, G. Electrochemical Engineering Principles; Prentice Hall International Series; Prentice-Hall: Upper Sadddle River, NJ, 1991. Stern, M.; Geary, A. L. J. Electrochem. Soc. 1957, 104, 56. Marsh, G. P.; Taylor, K. J.; Bryan, G.; Worthington, S. E. Corrosion Sci. 1986, 26. Lewis, M. B.; Hunn, J. D. J. Nucl. Mater. 1999, 265, 325. Hoar, T. P.; Hines, J. G. J. Iron Inst. 1956, 182, 124. Diegle, R. B.; Boyd, W. K. In Stress Corrosion Cracking – The Slow Strain-Rate Technique; Ugiansky, G. M., Payer, J. H., Eds.; ASTM: West Conshohocken, PA, 1979; p 26. Ford, F. P. Corrosion 1996, 52, 375. Bubar, S. F.; Vermilyea, D. A. J. Electrochem. Soc. 1966, 113, 892. Bubar, S. F.; Vermilyea, D. A. J. Electrochem. Soc. 1967, 114, 882. Hoar, T. P.; Scully, J. C. J. Electrochem. Soc. 1964, 111, 348. Hoar, T. P.; West, J. M. Nature 1958, 181, 835.
33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56.
57. 58. 59. 60. 61. 62.
Frankel, G. S.; Jahnes, C. V.; Brusic, V. V.; Davenport, A. J. J. Electrochem. Soc. 1995, 142, 2290. Ford, F. P.; Burstein, G. T.; Hoar, T. P. J. Electrochem. Soc. 1980, 127, 1325. Marshal, P.; Burstein, G. T. Corrosion Sci. 1983, 23, 1219. Burstein, G. T.; Davenport, A. J. J. Electrochem. Soc. 1989, 136, 936. Burstein, G. T.; Marshall, P. I. Corosion Sci. 1984, 24, 449. Davenport, A. J.; Burstein, G. T. J. Electrochem. Soc. 1990, 137, 1496. Cabrera, N.; Mott, N. F. Rep. Prog. Phys. 1948, 12, 163. Kelly, R. G.; Scully, J. R.; Shoesmith, D. W.; Buchheit, R. G. Electrochemical Techniques in Corrosion Science and Engineering; Marcel Dekker: New York, 2003. Marcus, P.; Mansfeld, F. Analytical Methods in Corrosion Science and Engineering; Taylor & Francis: New York, 2006. Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundementals and Applications; Wiley: New York, 2001. Schiller, C. A. In Analytical Methods in Corrosion Science and Engineering; Marcus, P., Mansfeld, F., Eds.; Taylor & Francis: New York, 2006; p 361. ASTM, G 3. Standard practice for conventions applicable to electrochemical measurements in corrosion testing; ASTM International; 1989. ASTM, G 102. Standard practice for calculation of corrosion rates and related information from electrochemical measurements; ASTM International; 1989. ASTM, G 59. Standard test method for conducting potentiodynamic polarization resistance measurements; ASTM International; 1997. MacDonald, J. R. Impedance Spectrocopy; Wiley: New York, 1987. Mansfeld, F. In Analytical Methods in Corrosion Science and Engineering; Marcus, P., Mansfeld, F., Eds.; Taylor & Francis: Boca Raton, FL, 2006; p 463. Hsu, C. H.; Mansfeld, F. Corrosion 2001, 57, 747. Mansfeld, F. Corrosion 1988, 44, 558. Ai, J.; Yingzi, C.; Uriquidi-Macdonald, M.; Macdonald, D. D. J. Nucl. Mater. 2008, 379, 162. Barberis, P.; Frichet, A. J. Nucl. Mater. 1999, 273, 182. Nagy, G.; Kerner, Z.; Battistig, G.; Csordas, A. P.; Balogh, J. J. Nucl. Mater. 2001, 297, 62. Kim, Y. J. Corrosion 2000, 56, 389. Fulger, M.; Ohai, D.; Mihalache, M.; Pantiru, M.; Malinovschi, V. J. Nucl. Mater. 2009, 385, 288. Lillard, R. S.; Gac, F.; Paciotti, M.; et al. In Effects Radiation on Materials; American Society for Testing and Materials Special Technical Publication #1045; Rosinski, S. T., Grossbeck, M. L., Allen, T. R., Kumar, A. S., Eds.; ASTM: West Conshohocken, PA, 2001. Lillard, R. S.; Willcutt, G. J.; Pile, D. L.; Butt, D. P. J. Nucl. Mater. 2000, 278, 277. Lillard, R. S.; Valot, C.; Hanrahan, R. J. Corrosion 2004, 60, 1134. Lillard, R. S.; Paciotti, M.; Tcharnotskaia, V. J. Nucl. Mater. 2004, 335, 487. Sikora, E.; Sikora, J.; Macdonald, D. D. Electrochim. Acta 1996, 41, 283. Lillard, R. S.; Vasquez, G. J. Electrochem. Soc. 2008, 155, C162. Macdonald, D. D. J. Electrochem. Soc. 1992, 139, 3434.
5.02
Water Chemistry Control in LWRs
C. J. Wood Electric Power Research Institute, Palo Alto, CA, USA
ß 2012 Elsevier Ltd. All rights reserved.
5.02.1
Introduction
18
5.02.2 5.02.2.1 5.02.2.2 5.02.2.3 5.02.2.4 5.02.2.5 5.02.3 5.02.3.1 5.02.3.2 5.02.3.3 5.02.3.4 5.02.4 5.02.4.1 5.02.4.2 5.02.4.3 5.02.4.4 5.02.5 5.02.6 References
BWR Chemistry Control Evolution of BWR Chemistry Strategies Mitigating Effects of Water Chemistry on Degradation of Reactor Materials Radiation Field Control Fuel Performance Issues Online Addition of Noble Metals PWR Primary Water Chemistry Control Evolution of PWR Primary Chemistry Strategies Materials Degradation PWR Radiation Field Control Fuel Performance PWR Secondary System Water Chemistry Experience Evolution of PWR Secondary Chemistry Strategies Chemistry Effects on Materials Degradation of SGs Control of Sludge Fouling of SGs Lead Chemistry Chemistry Control for FAC in BWRs and PWRs Water Chemistry Control Strategies
19 19 20 23 26 27 27 27 29 33 35 37 37 40 43 44 45 45 46
Abbreviations AO
AVT
BOP BRAC
BWR CGR CRUD DMA DZO EBA ECP
Axial offset, referring to localized flux depression in reactor core caused by buildup of boroncontaining deposits. Originally called AOA for axial offset anomaly. All-volatile treatment, suing ammonia for pH control in steam generators Balance of plant BWR radiation and control, referring to designated standard points in BWR reactors for radiation field measurements Boiling water reactor Crack growth rate Corrosion product deposits on fuel element surfaces Dimethylamine Depleted zinc oxide (BWRs) Enriched boric acid (PWRs) Electrochemical corrosion potential
ETA FAC GE
HWC HWC-L HWC-M IGA IGSCC LWR MOX MPA MRC NDE NMCA NWC OD IGA/SCC
Ethanolamine Flow-assisted corrosion General electric, the vendor for BWRs in the United States and some other countries Hydrogen water chemistry HWC (low) with 0.2–0.5 ppm hydrogen HWC (moderate) with 1.6–2.0 ppm hydrogen Intergranular attack Intergranular stress corrosion cracking Light water reactor Mixed oxide fuel 3-methoxypropylamine Molar ratio control (PWR secondary side) Nondestructive examination Noble metal chemical addition Normal water chemistry (BWRs) Outside diameter IGA/SCC in steam generator tubes
17
18
Water Chemistry Control in LWRs
OLNC OTSG PAA PbSCC PWR PWSCC SCC SG SHE
On-line noble chemistry Once through steam generator Poly acrylic acid Lead assisted stress corrosion cracking Pressurized water reactor Primary water stress corrosion cracking Stress corrosion cracking Steam generator Standard hydrogen electrode (for ECP measurements)
5.02.1 Introduction Other chapters of this comprehensive describe the various degradation processes affecting the structural materials used in the construction of nuclear power plants (see Chapter 5.04, Corrosion and Stress Corrosion Cracking of Ni-Base Alloys; Chapter 5.05, Corrosion and Stress Corrosion Cracking of Austenitic Stainless Steels; and Chapter 5.06, Corrosion and Environmentally-Assisted Cracking of Carbon and Low-Alloy Steels). This chapter describes the influence of water chemistry on corrosion of the most important materials in light water reactors (LWRs). In particular, alloys susceptible to intergranular attack (IGA) and stress corrosion cracking (SCC) are significantly impacted by water chemistry, most notably, sensitized 304 stainless steel in boiling water reactors (BWRs) and nickelbased alloys in pressurized water reactors (PWRs). Excellent water quality is essential if material degradation is to be controlled. In the early days of nuclear power plant operation, impurities in the coolant water were a major factor in causing excessive corrosion. Chlorides and sulfates are particularly aggressive in increasing intergranular stress corrosion cracking (IGSCC) and other corrosion processes. Transient increases of impurities in the coolant that occur during fault conditions (e.g., condenser leaks and ingress of oil or ion exchange resins) proved to be particularly damaging. Thus, water chemistry was traditionally regarded as a key cause of material degradation. Initial efforts to improve water quality brought about a slow but steady reduction in impurities through improved design and operation of purification systems. Not only were the average concentrations of impurities reduced over time, but the frequency and magnitude of impurity ‘spikes’ from transient fault conditions were also diminished.
However, excellent water chemistry alone was not sufficient to control corrosion. Hence, programs to modify water chemistry were introduced, including minimizing oxygen to reduce the electrochemical corrosion potential (ECP) in BWRs, and oxygen and pH control in PWRs. More recently, additives to further inhibit the corrosion process have been developed and are now in widespread use. As a result, water chemistry advances are now an important part of the overall operating strategy to control material degradation. Primary system water chemistry also affects fuel performance through the deposition of corrosion products on fuel pin surfaces, and influences radiation fields outside the core. Core uprating through increased fuel duty has reduced margins for tolerating corrosion products (CRUD) on BWR fuel pin surfaces. In PWRs, increasing fuel cycle duration has increased the challenge of controlling pH within the optimum range. At the same time, regulatory limits on worker radiation exposure are tending to be tightened worldwide, putting pressure on the operators to reduce radiation dose rates. Successful operation of PWR steam generators (SGs) and the remainder of the secondary system demand strict water chemistry control in secondary side systems if corrosion problems are to be avoided. Other operating parameters also influence the optimization process, for example, life extension (to 60 years) has emphasized the importance of controlling degradation of circuit materials. Therefore, although control of structural material degradation remains the highest priority, water chemistry must be optimized between the sometimes-conflicting requirements affecting other parts of the reactor. Advances in water chemistry have enabled plant operators to respond successfully to these technical challenges, and the overall performance has steadily improved in recent years.1 Plant-specific considerations sometimes influence or indeed limit the options for controlling water chemistry, so we see different chemistry specifications at different plants. This is especially true internationally and significant differences between countries are noted. The US industry started developing water chemistry guidelines 25–30 years ago, and these now provide the technical basis for guidelines in many other countries. The early editions of these guidelines presented impurity specifications and required action if limits were exceeded. When advanced water chemistries were developed and qualified, the guidelines evolved into a menu of options within an envelope of specifications that should not be
Water Chemistry Control in LWRs
exceeded. Guidance is now provided on how to select a plant-specific water chemistry strategy.2 The basis for water chemistry control was discussed in detail by Cohen.3 The remainder of this chapter describes more recent water chemistry developments for BWRs, PWR primary systems, and PWR secondary systems including SGs, with a short section on flow-assisted corrosion (FAC) in both BWRs and PWRs.
5.02.2 BWR Chemistry Control 5.02.2.1 Evolution of BWR Chemistry Strategies BWR water chemistry has to be optimized between the requirements to minimize material degradation, avoid fuel performance issues, and control radiation fields. These factors are depicted in Figure 1,4 which also includes the main chemistry changes involved in the optimization process. Plant-specific considerations sometimes influence or indeed limit the options for controlling water chemistry, so we see different chemistry specifications at different plants. This is especially true internationally and significant differences in chemistry strategies between countries are noted. Design features are an important reason for these different chemistry regimes, to which must be added the effects of different operational strategies in recent years. For example, a key issue facing BWRs in the United States concerns IGSCC of reactor internals, as discussed in other chapters. The occurrence of IGSCC resulted in the
Clad corrosion crud deposition: Limits on feedwater zinc
Impurity control: Monitoring/analysis required
implementation of hydrogen water chemistry, with or without noble metal chemical addition (NMCA), to ensure that extended plant lifetimes are achieved. German plants use 347 stainless steel, which is less susceptible to IGSCC than sensitized 304 stainless steel used originally in US-designed plants. Some Swedish and Japanese plants have replaced 304 stainless steel reactor internals with 316 nuclear grade material to minimize potential problems, as this material is less susceptible to IGSCC. As a result, many of these plants continue to use oxygenated normal water chemistry, whereas all US plants control IGSCC through the use of hydrogen water chemistry (HWC) with or without normal metal chemical addition to improve the efficiency of the hydrogen in reducing ECP. Second, BWRs in United States undoubtedly have greater cobalt sources than plants in most other countries, despite strong efforts to replace cobalt sources. This resulted in higher out-of-core radiation fields, leading all US plants to implement zinc injection to control fields, whereas only a small number of plants of other designs use zinc. Third, the move to longer fuel cycles and increased fuel duty at US plants, while having major economic benefits, has led to new constraints on chemistry specifications in order to avoid fuel performance issues. Figure 2 depicts the changing chemistry strategies over the past 30 years, showing the focus on improving water quality in the early 1980s and the move to educing chemistry to control IGSCC, which in turn resulted in increased radiation fields, subsequently controlled by zinc injection.
Materials degradation and mitigation
Water chemistry guidelines
Fuel performance
Chemistry control issues
Figure 1 Boiling water reactor chemistry interactions.
19
BWR internals IGSCC, IASCC: HWC or NMC required
Radiation exposure
Radiation fields crud bursts: Zinc required
20
Water Chemistry Control in LWRs
Increasing concerns about core internals cracking led to the need to increase hydrogen injection rates, which in turn resulted in the introduction of NMCA to reduce operating radiation fields from N-16. Figure 3 shows the rate of implementation of HWC, zinc and NMCA, and online noble metal addition (OLNC). The rationale and implications of these developments are discussed in greater detail in subsequent sections. The goal for BWRs is therefore to specify chemistry regimes that, together with the improved materials used in replacement components (e.g., 316 nuclear grade stainless steel), will ensure that the full extended life of the plants will be achieved without the need for further major replacements. At the same time, radiation dose rates, and hence worker radiation exposure, must be closely controlled, and fuel performance must not be adversely affected by chemistry changes.
The first requirement of plant chemistry is to maintain high-purity water in all coolant systems, including the need to avoid impurity transients, which are beyond the scope of this paper. The performance of all plants has improved steadily over the years, as shown by the trend for reactor water conductivity for GE-designed plants, given in Figure 4. This figure shows that conductivity now approaches the theoretical minimum for pure water. In fact, deliberately added chemicals, such as zinc (discussed in the following section), account for much of the difference between measured values and the theoretical minimum. The conductivity data are consistent with the reactor water concentrations for sulfate and chloride. In fact, sulfate is the most aggressive impurity from the viewpoint of IGSCC, and much effort has gone into reducing it. 5.02.2.2 Mitigating Effects of Water Chemistry on Degradation of Reactor Materials
1977: Neutral, oxygenated water
Corrosion, radiation buildup issues
1980s: Purer is better
IGSCC was first observed in small bore piping using sensitized 304 stainless steel fairly soon after BWRs started operation. Laboratory studies showed that impurities increased IGSCC rates, and in fact water quality in BWRs gradually improved in the early 1980s. However, the same studies found IGSCC in high-purity oxygenated water typical of good BWR operations. The key parameter affecting IGSCC was found to be ECP, as shown in Figure 5. In this laboratory test, the change from oxidizing conditions typical of normal water chemistry (NWC) operation
Chemistry guidelines
Late 1980s–1990s: HWC, zinc
Controlling IGSCC, radiation buildup
2000s: Noble metal chemical addition
Core internals cracking control with lower fields
Promising new option
2006–2008: Online Noblechem
Figure 2 Evolution of Boiling water reactor chemistry options from 1977 to 2008.
40
Number of BWRs
35
Zn injection
NMCA
HWC (no NMCA)
OLNC
30 25 20 15 10 5 0 1983
1988
1993
1998
2003
2008
Figure 3 Implementation of zinc injection, hydrogen water chemistry, noble metals chemical addition, and online noble metal at US boiling water reactors.
21
Water Chemistry Control in LWRs
0.40 0.35 EPRI action level 1
Conductivity ( µS cm–1)
0.30 0.25 0.20 0.15 0.10 0.05 Theoretical conductivity limit, 25 ºC
0.00 1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
Figure 4 Boiling water reactor mean reactor water conductivity at US boiling water reactor.
250
0.4950
0.4945 200
2.7 ⫻ 10−8 mm s–1 1 ⫻ 10−6 mm s–1
150
0.4935
0.4930
0.4925
100 CT2 #7-304SS 4 dpa Constant load, 19 ksi√in.
Dissolved O2
Outlet cond: 0.30 μS cm–1
50
Inlet cond: 0.27 μS cm–1 Na2SO4
0.4920
0.4915 1488
Dissolved oxygen (ppb)
Crack length (in.)
0.4940
1508
1528
1548
1568
1588
0 1608
Test time (h) Figure 5 Laboratory results showing the effect of reducing oxygen concentration on crack growth of 304 stainless steel.
to reducing conditions greatly reduced the rate of crack growth. Furthermore, hydrogen injection was effective at reducing the ECP in BWRs, as shown in Figure 6. In this figure, it can be seen that crack growth rates (CGR) for Alloy 182 were low in hydrogen water
chemistry (HWC), but increased greatly when the plant reverted to normal water chemistry (NWC). These results indicated that continuous hydrogen injection was required to fully mitigate cracking. Examination of extensive inspection data from several plants indicated that no IGSCC was observed with an
22
Water Chemistry Control in LWRs
901.00 900.00
HWC ECP = −510 mV (SHE)
NWC ECP = +110 mV (SHE)
Crack length
174 miles year-1
HWC
< 5 miles year-1
899.00 898.00 897.00 < 5 miles year -1
Alloy 182
896.00 895.00 800
900
1000
1100
1200
1300 Time (h)
1400
1500
1600
1700
Figure 6 Effect of hydrogen water chemistry on crack growth of Alloy 182.
ECP of 230 mV or lower, using a standard hydrogen electrode (SHE). This is the basis for the 230 mV requirement used by US plants for IGSCC control. In BWRs, the radiation field in the core decomposes water to hydrogen and oxygen species, most of which immediately recombine back to water. But some remain as oxygen or hydrogen peroxide, because some hydrogen is stripped into the steam phase before it can recombine. These same radiolysis reactions cause hydrogen to react with oxygen or peroxide to reduce ECP. These reactions occur mainly in the downcomer, and relatively low hydrogen concentrations are effective at lowering ECP in out-of-core regions of the system. More than half the BWRs in the United States adopted low hydrogen injection rates of 0.2–0.5 ppm (called HWC-L), which, coupled with the replacement of recirculation piping using 316 stainless steel, mitigated IGSCC of recirculation piping. In the 1990s, concerns about the cracking of core internals increased, but the low concentrations of hydrogen used to protect out-of-core regions were not sufficient to reduce ECP enough to mitigate IGSCC of in-core materials, because of the radiolysis of water occurring in the core. As a result, it was necessary to increase hydrogen concentrations to 1.6–2.0 ppm to lower the in-core ECP sufficiently to provide protection in the reactor vessel (termed HWC-M for moderate concentrations of hydrogen). Although this approach was effective in protecting core internals, it also increased radiation fields in the steam side of the circuit, including the turbines, as a result of carryover of nitrogen-16 under reducing chemistry. (Under the oxidizing conditions of NWC, most of the N-16 remains in the water as soluble
species such as nitrate, and only a small percent is transported with the steam.) In some plants, local shielding of turbine components has reduced the impact of the gamma radiation to acceptable levels, but the projected 4–6-fold increase did in fact curtail plans for increased hydrogen injection rates at many plants. Note that these N-16 radiation fields are a problem only when the plant is at power, as rapid decay occurs at shutdown because of the short halflife of N-16. (By contrast, out-of-core radiation fields from Cobalt-60 persist after shutdown and impact on maintenance work during outages.) NMCA was developed to increase the efficiency of hydrogen in BWR cores, to avoid high N-16 fields. In this process, a nanolayer of platinum þ rhodium is deposited on the wetted surfaces of the reactor. These treated surfaces catalyze the hydrogen redox reaction, converting oxygen back to water. When the addition of hydrogen to the feedwater raises the molar ratio of H2 to O2 to 2 or higher, the ECP of the treated surfaces drops to the hydrogen/oxygen redox potential, which is about 450 mV (SHE). This can be achieved with hydrogen concentrations of only about 0.2 ppm, and under these conditions, the main steam radiation level is not increased to an unacceptable level. The first plant used NMCA successfully in 1997, and over 25 plants have already followed, with excellent results. Field measurements show that NMCA has been effective in providing mitigation against IGSCC by lowering the ECP below the 230 mV (SHE) threshold with relatively low hydrogen injection rates. The NMCA process is typically applied at refueling outage, before new fuel is inserted into the core,
Water Chemistry Control in LWRs
additional benefit with NMCA on the upper, outer shroud regions, as indicated by the additional shading in the left-hand side of the figure5. It is estimated that noble metals protect slightly more of the outer core region than does moderate HWC (HWC-M), but the difference is not significant. Figure 8 shows the dramatic benefit of noble metals in reducing the rate of stub tube cracking at Nine Mile Point 1 since the application in 2000. Before 2000, several stub tubes had to be repaired or replaced at each outage, but since the application, only one tube leaked, and this was believed to have already cracked before NMCA. Recently, attention has been focused on the online application of noble metals, with the first application at the KKM plant in Switzerland. By April 2008, there were four applications in the United States. This is discussed in a later section.
HWC protected regions
NMCA protected regions
and is effective for about three fuel cycles, before reapplication is necessary. The regions of the reactor vessel internals that are protected by HWC-M or NMCA are shown in Figure 7. While both techniques offer significant areas of mitigation, there is an
5.02.2.3
Radiation Field Control
Corrosion products deposited on the fuel become activated, are released back into the coolant, and may be deposited on out-of-core surfaces. Both soluble and insoluble species may be involved, the latter tending to deposit in stagnate areas (‘crud traps’). The chemistry changes to control IGSCC resulted in increased out-of-core radiation fields, and the implementation by most plants of depleted zinc injection to
Figure 7 Mitigated regions of the boiling water reactor core.
Number of stub tubes identified with IGSCC throughwall cracking based on leakage
12
10
8
Noble metal applied mid cycle may 2000
6
4
2
0 1984–1985 1986–1987 1988–1990
23
1991
1993
1995
1997
1999
2001
2003
2005
2007
RFO-11
RFO-12
RFO-13
RFO-14
RFO-15
RFO-16
RFO-17
RFO-18
RFO-19
Year Figure 8 Mitigation of stub tube cracking at Nine Mile Point Unit 1.
24
Water Chemistry Control in LWRs
control dose rates, as discussed later in this section. During shutdowns, the major radiation source for personnel exposure is activated corrosion products, deposited on primary system surfaces. Exposures are generally accumulated at high-radiation field locations where maintenance work is frequently needed. Although improvement of maintenance equipment and procedures, reduction of maintenance requirements, increased hot-spot shielding, and control of contamination dispersion have significantly reduced total exposure, further reduction of radiation fields is a major goal in programs for minimizing occupational radiation exposure. The primary source of radiation field buildup on out-of-core surfaces in BWRs is 60Co, which in mature plants usually accounts for 80–90% of the total dose. 60Co has a relatively long half-life of 5.27 years. The higher the soluble 60Co concentration in the coolant, the more 60Co is incorporated and deposited on out-of-core systems and components, resulting in higher dose rates on recirculation piping, the reactor water cleanup system, dead legs, and other crud traps in the system. Other activated transition metals such as 54Mn, 58Co, 59Fe, and 65Zn contribute the remainder of the dose. 51Cr also contributes significantly to the piping dose in some NMCA plants. The radiation fields commonly measured in a BWR at the straight vertical section of recirculation pipes are considered to be more representative for the purposes of radiation buildup trending and comparison with other plants. These measurements are done in a prescribed manner developed under the EPRI BWR Radiation and Control program and are called BRAC point measurements. These measurements represent primarily the incorporation of soluble 60Co into the corrosion film on the piping surfaces and tend to be a fairly good predictor of drywell dose rates. The deposition of particulate oxides that contain 60Co and other activated species can also contribute significantly to outof-core radiation levels in BWRs, especially in hot spots. The particulate oxides, which vary in size, originate primarily from corrosion of the steam/condensate system and are introduced via the feedwater. The sole precursor of the gamma-emitting 60Co isotope is 59Co. 59Co is present as an impurity in the nickel in structural alloys used in BWRs (e.g., Type 304 stainless steel) and is the main constituent of wear-resistant alloys (e.g., Stellite), used as hard facing in valves and other applications requiring outstanding wear resistance. Corrosion and wear lead to release of 59Co into the coolant from these sources,
which is transported to the core and incorporated into the crud that deposits on the fuel rods. The 59 Co is activated to 60Co by neutron activation, released back into the coolant, and incorporated as a minor constituent into the passive films that form on components that are inspected, repaired, and replaced by maintenance personnel. Components in the neutron flux (e.g., the control blades) directly release 60Co. Cobalt source removal is clearly important if radiation fields are to be minimized. Another gamma-emitting isotope, 58Co, is formed by the activation of nickel from stainless steel and nickel-based alloys. 58Co has a shorter half-life and is not as major a contributor to radiation fields as 60Co in BWRs, but is much more significant in PWRs. Shutdown drywell dose rates increase when coolant chemistry is changed for the first time from oxidizing (NWC) to reducing (HWC) conditions. This results from a partial restructuring of the oxides formed under the oxidizing conditions of NWC (Fe2O3 type) to a more reducing spinel type oxide compound (Fe3O4 type). The oxides affected are the fuel deposits, the corrosion films on stainless steel piping, and out-of core deposits. This results in an increase in the chemical cobalt (and 60Co) concentration in the oxide because of the higher solid-state solubility of transition metals in the spinel structure. The presence of a higher soluble reactor 60Co concentration released from fuel crud while this conversion is occurring only aggravates the situation. The processes are depicted in Figure 9. The net result at most plants is a temporary increase in reactor water 60 Co, both soluble and insoluble forms, which leads to significantly increased shutdown dose rates because of both the increased reactor water concentrations and the increased capacity for transition metal uptake by the spinel phases.6
Oxide stable under normal water chemistry Fe2O3 (containing 60Co, 58Co, 54Mn, etc.)
• Corrosion films • Vessel crud • Fuel crud
Restructuring under HWC conditions
Fe3O4 form of oxide
Small insoluble particles containing 60Co, 54Mn, etc. Soluble 60Co, etc. released during restructure
Figure 9 Boiling water reactor oxide behavior under reducing conditions.
Water Chemistry Control in LWRs
25
0.8 Before Zn addition After Zn addition
RxW 60Co (Ci kg −1)
0.6
0.4
0.2
0 Brunswick-1
Brunswick-2
Dresden-2
Figure 10 Hydrogen water chemistry plant RxW
60
Duane Arnold
Fitz patrick
Monticello
Pilgrim
Co response to zinc addition.
As mentioned earlier, zinc addition reduces radiation field buildup. The mechanism of the zinc ion effect is complex, as release of 60Co from fuel crud is reduced, and deposition out-core is also reduced. Overall, reactor water 60Co is decreased significantly after zinc addition, as shown by plant data in Figure 10. Aqueous zinc ion promotes the formation of a more protective spinel-structured corrosion film on stainless steel, especially when reducing conditions are present. Second, both cobalt and zinc favor tetrahedral sites in the spinel structure, but the site preference energy favors zinc incorporation. Thus, the available sites have a higher probability of being filled with a zinc ion than a cobalt ion (or 60Co ion), and hence the uptake of 60Co into the film will be significantly less if zinc ion is present in the water. The 60 Co remains longer in the water and is eventually removed by the cleanup system. The zinc was originally added to the feedwater as ZnO, but it was quickly found that the 65Zn that was created by activation of the naturally occurring 64Zn isotope in natural zinc created problems. With the use of zinc oxide depleted in the 64Zn isotope, called depleted zinc oxide (DZO), this drawback was eliminated. Because of the high cost of DZO, feedwater zinc injection was not implemented widely until HWC shutdown dose issues emerged. For the case of plants treated with NMCA and injecting hydrogen, the oxidant concentration on the surface of the stainless steel is zero (due to the Pt
and Rh catalyzing the reaction of any oxidant with the surplus hydrogen). The net result is that the ECP is at or very near the hydrogen redox potential, typically about –490 mV (SHE) for neutral BWR water. This low potential causes a much more thorough restructuring of the oxides to the spinel state than observed under moderate hydrogen water chemistry (HWC-M). Feedwater iron ingress has a significant influence on the effectiveness of zinc injection. As discussed in the next section, deposits on fuel cladding surfaces (called ‘CRUD’) are mainly composed of iron oxides, with other constituents. Therefore, reducing iron ingress from the feedwater has the benefit of minimizing crud buildup, which is important for fuel reliability (next section). For these reasons, extensive efforts have been made to reduce iron ingress, with significant success. Furthermore, fuel crud has a large capacity for incorporating zinc and is in fact where most of the zinc ends up. The lower the amount of crud on the fuel, the greater the proportion of zinc that remains in solution and can subsequently be incorporated in out-of-core surfaces. Therefore, at plants with low feedwater iron, less zinc is captured by the crud on the fuel, so a relatively greater amount remains in solution and is available to control out-of core radiation fields. This is very important, as zinc injection rates are limited by fuel performance concerns, and hence lowering feedwater iron is essential for maintaining lower radiation fields.
26
Water Chemistry Control in LWRs
5.02.2.4
Fuel Performance Issues
Fuel durability has improved over the years, and failures have declined, helped by improvements in water purity. In operation, zircaloy fuel cladding develops a thin oxide layer (ZrO2), which typically does not adversely affect performance. However, an increase of deposition of corrosion product deposits (‘crud’) on this oxide film is undesirable because it can reduce heat transfer and increase fuel pin temperatures, with resultant increased corrosion of the fuel cladding, ultimately increasing the risk of fuel failure. Moreover, the addition of additives to control corrosion may increase the risk of crud buildup on the fuel. For example, zinc and noble metals in BWRs tend to increase the adherence of crud deposits on the fuel, which can result in undesirable oxide spalling in higher-rated cores. In fact, corrosion-related fuel failures occurred at four plants in the United States between 1999 and 2003. Although the precise root cause of fuel failures is often difficult to determine, it is clear that excessive crud buildup played a role in these failures. With progressive uprating of fuel duty in both PWRs (and BWRs), the margin to tolerate crud has been reduced and additional care has to be taken in specifying the water chemistry to avoid undesirable fuel performance issues. Despite these more demanding conditions, fuel failures have decreased in recent years. Concern about the possibility of adverse effects of NMCA on fuel has prompted imposition of a strict limit on the amount of noble metal that can end up on the fuel and guidance on the injection of zinc. Plant data indicate that spalling of the corrosion layer from
fuel cladding, which is often regarded as a precursor to cladding failure, is prevented if the cycle average feedwater zinc is maintained below 0.4 ppb in NMCA plants (0.6 ppb for non-NMCA plants). More recent data indicate that quarterly averages may be as high as 0.5 ppb for NMCA plants, without occurrence of spalling.5 These feedwater zinc data are the basis for limits in the water chemistry guidelines. The 2008 chemistry guidelines7 retain the cycle average feedwater zinc limit of 0.4 ppb (0.6 ppb for non-NMCA plants) but enable a slight increase in the quarterly average to 0.5 ppb, which may allow flexibility in controlling radiation buildup in parts of the cycle. The tighter control of water chemistry in recent years has been successful in controlling crud formation on fuel cladding, and Figure 118 shows failures from pellet–clad interaction causing SCC, fabrication defects, debris, and crud/corrosion. Note that there have been zero crud/cladding related fuel failures in US BWRs since 2004 (although assessment of 2007 failures is not yet complete, crud/corrosion is not believed to be a factor here). Analysis of recent plant data confirms that control of feedwater iron ingress has the positive benefit of reducing the amount of crud on the fuel. Control of copper, which generally originates from admiralty brass alloys, is also beneficial; not only can copper have detrimental effects on the fuel, but it also limits the ability of hydrogen to reduce the ECP, and it also leads to higher radiation fields. As a result, most US plants have replaced condensers containing brass tubing.
Number of failed assemblies
30 25 20
PCI-SCC Unknown Fabrication Debris Crud/corrosion
15 10 5 0 2000
2001
2002
2003 2004 EOC year
2005
2006
Figure 11 US boiling water reactor fuel failures by mechanism for each end-of-cycle (EOC) year.
2007
Water Chemistry Control in LWRs
5.02.2.5
Online Addition of Noble Metals
As discussed earlier, the classic NMCA process is generally applied during refueling outages before the new fuel is loaded into the core. Reapplication after about three cycles of operation takes approximately 2 days, while the plant maintains 107–154 C as it enters the refueling outage. To reduce this outage time, GE-Hitachi developed OLNC, first demonstrated at KKM (a GE design of plant in Switzerland) in 2005, with several more additions subsequently. Preliminary results indicate that there have been no unexpected chemistry effects during the first OLNC applications, and shutdown radiation fields actually decreased at KKM after OLNC.5 Subsequently, CGR of susceptible welds decreased significantly, as shown by the decrease in slopes in Figure 12 after OLNC initiation for two welds that have been monitored for several years. The effects of OLNC on fuel have been extensively studied in fuel removed from KKM, and no adverse effects have been observed. The jury is still out on this concern, but the general assessment is that OLNC will have no more impact than the classic application, and may well prove to be of less concern. More IGSCC and fuel measurements are planned, but with no issues emerging to date, it appears that OLNC applications about every 12 months would be effective and economical, avoiding the critical path time necessitated for the classic NMCA application during refueling outages. Initial OLNC applications have been carried out at plants that had previously applied noble metals in the classic off-power manner.
However, the first OLNC application at a plant that has not used noble meals previously occurred in late 2008, but no results are available.
5.02.3 PWR Primary Water Chemistry Control 5.02.3.1 Evolution of PWR Primary Chemistry Strategies In the very early days of PWR operation, heavy crud buildup on fuel cladding surfaces was caused by the transport of corrosion products from the SGs into the reactor core. As a result, activated corrosion products caused high-radiation fields on out-of-core surfaces (Figure 13), fuel performance was compromised, and even coolant flow issues were observed in some plants. These problems were initially mitigated by imposing a hydrogen overpressure on the primary system, to reduce the ECP, and raising the primary chemistry pH. Materials degradation in primary systems was then not a major concern, with most of the emphasis focused on secondary side corrosion issues in the SGs. Commercial PWR power plants use a steadily decreasing concentration of boric acid as a chemical shim (for reactor control) throughout the fuel cycle, which results in the use of lithium hydroxide to control pH. Some 30 years ago, the concept of ‘coordinated boron and lithium’ was developed, whereby the concentration of LiOH was gradually reduced in line with the boric acid reduction to maintain a constant pH.
300 NC appl.
HWC
Indication length (mm)
250 Ind 9
Ind10
Not inspected in 00, 01,04
200 150 100 50
27
OLNC 37 g OLNC 98 g OLNC 198 g OLNC 199 g
Indications 9,10 may be seeing mitigation by OLNC
0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Year Figure 12 Ultrasonic inspection results after online noble metal chemical addition.
28
Water Chemistry Control in LWRs
Corrosion products deposit out of core Pressurizer Steam generator Corrosion products activated in reactor core
Corrosion products released from SG tubing Coolant pump Reactor Primary loop
Figure 13 Transport and activation of corrosion products in pressurized water reactor primary systems.
Corrosion products released from the steam generator tubes are transported, dissolved, or deposited by the coolant on the basis of solubility differences. The solubility of nickel and iron depend on pH, temperature, and redox potential, all of which vary with location around the nonisothermal system. Originally, a constant at-temperature pH of 6.9 was recommended, based on the minimum temperature coefficient of solubility of magnetite. In fact, it was determined that heavy fuel crud buildup was avoided if a constant pH of at least 6.9 was maintained. This was possible with 12-month fuel cycles, but fuel cladding corrosion concerns limited the maximum LiOH concentration to 2.2 ppm. Consequently, plants often started the fuel cycle with pH below 6.9, which resulted in deposition of corrosion products on the fuel, activation of cobalt and nickel, and subsequent transport to out-of-core surfaces, resulting in radiation fields remaining relatively high. Even though detailed studies of fuel crud showed that the prime constituent of the crud was nickel ferrite (for which the optimum pH is 7.4), this coordinated chemistry had remained the standard for many years, until higher pHs became the norm in the 1990s. Although research and plant demonstrations showed that the 2.2 ppm limit was excessively conservative, the move to higher Li concentrations has been slow. However, detailed fuel examinations from a recent plant demonstration (that will be discussed later) have indicated that Li can be raised to as high as 6 ppm.
About 25 years ago, primary water stress corrosion cracking (PWSCC) of Alloy 600 SG tubes was observed in a few plants, leading to studies on mitigating this effect. Following successful demonstration of zinc injection in BWRs, initial field tests at PWRs showed that radiation fields were reduced, and laboratory studies indicating that PWSCC was reduced were eventually confirmed. As a result, zinc injection is being implemented at an increasing rate, although concerns about fuel performance at highduty plants have not been completely resolved. Most recently, buildup of boron-containing crud in areas of subcooled nucleate boiling leading to localized flux depression has encouraged the use of higher Li concentrations to minimize corrosion product transport. Concerns about the potential adverse effects of zinc deposited in high-crud regions have resulted in several highly rated plants applying in situ ultrasonic fuel cleaning before implementing zinc injection. Although zinc injection was developed for radiation field control, laboratory studies showed that it also inhibited SCC under PWR conditions. The identification of PWSCC in reactor vessel penetrations in the last 15–20 years has encouraged the use of zinc injection, but has also focused attention on the effects of dissolved hydrogen, for which the recommended range has remained 25–50 ml kg1 for 30 years. It now appears that raising hydrogen will reduce PWSCC rates, while lowering it may delay initiation of PWSCC. The interactions of materials, radiation fields, and fuels in PWR primary
Water Chemistry Control in LWRs
PWSCC: pH (Li, B) minimal effect Zn beneficial dissolved H2 effect
Plant operations
29
Dissolved H2 control range
Materials degradation
PWR chemistry control
Fuel performance
Plant dose rates Radiation fields: pH (Li, B), Zn beneficial
Crud deposition: Zn concern for highly rated cores
Figure 14 Pressurized water reactor primary chemistry optimization. Reproduced from Fruzzetti, K.; Perkins, D. PWR chemistry: EPRI perspective on technical issues and industry research. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008.
Dissolved (H2) range changes
Elevated constant pH (7.3/7.4) Ultrasonic fuel cleaning Elevated constant pH (7.1/7.2) Zinc injection Modified elevated lithium program EPRI water chemistry guidelines Elevated lithium program Constant pH 6.9 1975
1980
1985
1990
1995
2000
2006 2008
Figure 15 Pressurized water reactor primary chemistry changes at US plants.
chemistry and optimization issues covered in the water chemistry guidelines, which are discussed later, are depicted in Figure 14. The evolution of water chemistry control in PWR primary systems in the United States over the last 30 years is shown in Figure 15. The following sections address the three main factors – pH control, zinc injection, and dissolved hydrogen control – that have dominated PWR primary chemistry strategies in the past and continue to do so today.9 Each of these factors is considered from the viewpoint of materials degradation, radiation field control, and fuel performance concerns.
5.02.3.2
Materials Degradation
Materials degradation has been covered in detail in Chapter 5.04, Corrosion and Stress Corrosion Cracking of Ni-Base Alloys and Chapter 5.05, Corrosion and Stress Corrosion Cracking of Austenitic Stainless Steels, and here only the specific effects of water chemistry variables on materials in PWR primary systems will be reviewed, particularly those that may affect the chemistry of optimization process. Recent papers by Andresen et al.10,11 provide detailed results of a comprehensive study of the effects of PWR primary water chemistry on PWSCC of nickel-based alloys. Extensive studies have been carried out to determine the effect of lithium, boron, and pH on PWSCC, and the generally held conclusion is that any effects are minimal, especially compared to material susceptibility, stress state and temperature, and other operational issues. Crack initiation tests using the most reliable types of reverse U-bend specimens indicate that pH has a relatively small effect on crack initiation (generally less than a factor of 2). Although the most rapid crack initiation occurred at pH310 C 7.25, with slower rates at higher or lower pHs, CGR tests generally confirm that pH has minimal effect. The effect of lithium is even smaller than the pH effect, and the influence of boron is minor or nonexistent. Andresen et al. concluded that the effects of relevant variations in PWR primary water chemistry (B, Li,
30
Water Chemistry Control in LWRs
and pH) have little effect on the SCC growth rate in Alloy 600, and thus provide little opportunity for mitigation of PWSCC. Plant data have found no adverse effects from increasing lithium and pH in primary systems. As a result, it is considered that adjusting pH, lithium, or boron to minimize crack initiation may be of minimal value. The 2007 edition of the PWR Primary Water Chemistry Guidelines12 reviewed the most recent data and concluded that pH strategy changes based on PWSCC considerations are not warranted. This means that plants have the flexibility to pursue B/Li/pHt chemistry adjustments to minimize crud transport and radiation buildup without concern for negative effects on PWSCC susceptibility of nickelbased alloys, although of course chloride and sulfate impurities should continue to be minimized. Following good experience in BWRs, zinc injection has been implemented in the primary systems of PWRs, both to reduce primary side cracking of nickelbased alloys and to control dose rates. The qualification work for BWRs showed that zinc inhibited SCC, but the benefit was not sufficient to avoid the need for hydrogen water chemistry to mitigate IGSCC. Thus, the motivation for BWR zinc injection was exclusively radiation field control. The situation in PWRs is different, as laboratory work13 showed that initiation of PWSCC was significantly delayed by zinc injection, and hence the motivation for the initial applications of zinc in most US PWRs at the 10–30 ppb level was to control PWSCC of SG tubing. Additionally, German-designed PWRs and a few US plants used 5 ppb depleted zinc for radiation control.
Figure 16 shows the rate of introduction of zinc injection at PWRs worldwide. Zinc injection produces thinner, more protective oxides on stainless steel and Alloy 600, with zinc displacing Co2þ, Ni2þ, and Fe2þ from normal spinels to give ZnCr2O4, which is very stable. The benefits of PWR zinc injection have been clearly demonstrated in reducing PWSCC degradation (especially growth rate) of Alloy 600 SG tubes, and in controlling radiation fields. Evaluation of currently available laboratory data2 indicates that PWSCC initiation will be reduced, and PWSCC CGR may be reduced in thicker cross-section components, depending upon other factors such as the stress intensity factor of the specimen. Andresen et al.11 concluded that crack growth mitigation by adding Zn requires further study, although two of four tests show a decrease in growth rate of >3. Molander et al.14 also found that the effect of zinc on CGR was minor. Hence, more work is needed before making definitive conclusions from laboratory studies regarding the benefit of zinc in mitigating CGR. SG tube nondestructive examination (NDE) data from eight plants injecting zinc indicated reduction in the incidence of PWSCC by a factor of 2–10.9 An example from a 2-unit PWR showing the effect of zinc on SG tubing over successive cycles is given in Figure 17. The largest effect of zinc appears to be on initiation of cracking, with a smaller effect on CGR, with the data indicating a factor of 2–10 reduction for initiation and about a factor of 1.5 reduction in CGR, consistent with the extensive laboratory work,11
Application of zinc in world PWRs
Number of plants and percentage of PWR injecting zinc
50 45
Percent of PWRs injecting
Number of units injecting
40 35 30 25 20 15 10 5 0 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Year
Figure 16 Application of zinc injection in pressurized water reactors worldwide.
Water Chemistry Control in LWRs
31
140 X indicates last refueling outage before start of zinc injection
Number of new tubes affected
120 100 80
Unit 1 Unit 2 60 40 20 0
X−3
X−2
X−1
X+1 X Refueling outage
X+2
X+3
X+4
Figure 17 Effect of zinc on steam generator tube degradation at a US pressurized water reactor.
70 60 Number of plants
indicating that zinc inhibits mainly by delaying the initiation of PWSCC. However, the SG NDE data also showed that zinc reduced the rate of crack propagation (depth) by 17–60%. These results are consistent with initial laboratory test data indicating that zinc reduced crack propagation by a factor of approximately 3 at low stress intensities, but had no effect at higher stress intensities. In addition, the lack of cracking in the Farley PWR pressure vessel head penetrations (exposed to zinc for over 12 years), compared to PWSCC indications in similar pressure vessel heads in other plants, suggests that zinc addition is beneficial for Alloy 600 (and possibly Alloy 82/182) thick-section components under PWR primary service conditions. Recent work has studied the influence of dissolved hydrogen on PWSCC. In the early days of PWR operation, the lower limit on hydrogen was set at 25 ml kg1, to provide adequate margin against radiolysis and heavy crud formation. Plant tests in France showed that this limit was excessively conservative and that less than 10 ml kg1 would be satisfactory, provided good control of oxygen was maintained in makeup water. Several workers have found that the maximum in PWSCC CGR occurs close to the ECP corresponding to the Ni/NiO thermodynamic equilibrium condition.15 Although this potential is unaffected by lithium/boron/pH (consistent with the fact that these do not greatly influence PWSCC over the range of practical relevance), the equilibrium potential is significantly affected by the dissolved hydrogen
50 40 30 20 10 0 25–30 30–35 35–40 40–45 45–50 Cycle average hydrogen concentration (cm3 kg−1)
Figure 18 US plant data for dissolved hydrogen.
concentration. Andresen et al. found that the peak in SCC growth rate versus H2 fugacity was temperature dependent, but generally fell within the hydrogen concentration range used in PWRs. This provides an opportunity for mitigation, by perhaps a factor of 2 in Alloy 600 and a factor of 5 in Alloys 182, 82, and X750, as the median value of the dissolved hydrogen concentration for US plants is approximately 35 ml kg1. US PWRs currently operate within dissolved hydrogen within the recommended 25–50 ml kg1 range, with the majority in the 30–40 ml kg1 range, but none with more than 44 ml kg1 (Figure 18). The lower limit is set conservatively to provide an operating margin over the level of hydrogen required
32
Water Chemistry Control in LWRs
to suppress water radiolysis in the reactor core. Somewhat lower concentrations are used in other countries. The dissolved hydrogen concentrations corresponding to the peak CGR for a typical range of PWR primary operating temperatures are 4.3 ml kg1 at 290 C, 10.4 ml kg1 at 325 C, and 16.5 ml kg1 at 343 C.15 Andresen et al.11 published Figure 19, which indicates the proposed factors of improvement on changing from an initial hydrogen concentration of 25 ml kg1. It can be seen that raising the hydrogen provides benefit, but lowering it is detrimental below 330 C.
In response to the data showing the benefit of increasing hydrogen in reducing CGR, the US industry program in progress focuses on the extent to which dissolved hydrogen can be increased without adverse consequences to other parts of the system. Other countries, including Japan, are also investigating lowering hydrogen, because laboratory data suggest that the initiation of cracking is delayed at lower hydrogen concentrations. This is depicted in Figure 20, as discussed by Molander.14 The lower line in this figure shows the time to initiate cracking, based on laboratory tests using
Factor of improvement from H2
4.0 Based on Alloy 182, a current H2 level of 25 cm3 kg−1
3.5
70 cm3 kg-1 H2
25
3.0
45 cm3 kg-1 H2
25
2.5 2.0 1.5
Good 1.0 25 0.5 0.0 270
280
290
1 cm3 kg-1 H2
300 310 320 Temperature (⬚C)
4 cm3 kg-1 H2
25
330
340
Bad
350
Figure 19 Effect of dissolved H2 on primary water stress corrosion cracking crack growth rate at different temperatures.
ml H2/kg H2O (330 ˚C) 10
15
20
25
30
Crack initiation time (h)
Jenssen data on Alloy 600
35
1E−07
Growth 8E−08
20 000
6E−08
15 000
4E−08
10 000 Initiation
2E−08
5000
0 0
5
10 Hydrogen activity (kPa)
15
Crack growth rate (mm s–1)
5 25 000
0E+00 20
Figure 20 Dependencies between the dissolved hydrogen content in pressurized water reactor primary coolant on the crack initiation time observed on initially smooth surfaces and on the crack propagation rate.
Water Chemistry Control in LWRs
reverse U-bend specimens, whereas the upper line shows crack growth data over a similar concentration range. Thus, the lowering of hydrogen appears feasible. However, the relative importance of crack initiation and crack propagation is very dependent on material and plant conditions. In the United States, concern about increased crack propagation at low hydrogen and low temperatures, as shown in Figure 19, has resulted in moving to higher hydrogen being preferred to the alternative of reducing hydrogen. Several factors combine to make higher H2 the preferred way to mitigate SCC, including the importance of bottom-head penetrations (which are exposed to 290 C water) and the recent observation that the CGR in coldworked Alloy 600 is not mitigated at low H2.11 The preferred strategy in the United States is to gradually increase hydrogen to the upper end of the existing range, with the potential to move higher (say to 60 ml kg1) when the ongoing qualification work is completed. This will include evaluation of the effects of dissolved hydrogen on radiation fields and fuel performance, although any such effects are expected to be minimal.16 5.02.3.3
PWR Radiation Field Control
Corrosion products released from out-of-core materials (primarily SG tubing) deposit on the fuel and become activated, are released back into the coolant, and may be deposited on out-of-core surfaces. Both soluble and insoluble species may be involved, with the latter tending to deposit in stagnate areas (‘crud traps’). In addition to the chemistry items discussed later in this section, it must be stressed that other factors are important to the goal of reducing radiation fields. In particular, the success of the later German-designed plants in eliminating cobalt sources in hardfacing alloys, thereby achieving very low radiation fields, demonstrates the benefits of cobalt source reduction. With many plants replacing SGs, a correlation between recontamination rates and surface finish of the new SG tubing has been noted by Hussey et al.17 Typical PWR fuel cycles start with a relatively high boric acid concentration, which gradually reduces to zero at the end of the cycle. Lithium hydroxide is added to maintain an approximately constant pH. As the duration of fuel cycles increased, more boric acid was required at the start of cycle, which in turn necessitated increased LiOH to maintain the desired pH (Figure 21). As mentioned earlier, radiation field buildup can be controlled by minimizing corrosion product
33
transport and activation. Initially, coordination of lithium hydroxide with boron to maintain a constant at-temperature pH of 6.9 was recommended, based on the minimum solubility of magnetite. In fact, the prime constituent of the crud turned out to be nickel ferrite, requiring a pH of 7.4 for minimum solubility. Fruzzetti et al.15 have recently reviewed the data on elevated pH, which provides a number of benefits including decreased general corrosion (and thus reduced corrosion product transport to the core). Field-tests of pHs greater than 6.9 confirmed that radiation fields were lower. Although no adverse effects were observed on the fuel, many plants were slow to abandon a 2.2 ppm limit, established to avoid excessive zircaloy corrosion. However, there were indications of heavier crud formation after long periods operating below pH 6.9, and as fuel concerns relaxed, a gradual move toward a maximum of 3 ppm lithium resulted. Moreover, pHs in the range 7.1–7.2 became more popular in the late 1990s, with 7.3–7.4 eventually gaining favor. Figure 22 shows the maximum lithium concentrations reported by US PWRs in recent years. It can be seen that 95% are now using greater than 3 ppm at full power: a significant change from earlier in the decade. A demonstration of elevated Lithium/pH is in progress at Comanche Peak PWR.18 The goal was to reduce radiation fields and reduce susceptibility to the Axial Offset Anomaly (AOA) by reducing crud buildup. This test involved increasing the primary system pH from 7.1/7.2 to 7.3 and then two cycles at 7.4. No significant adverse trends have been noted, either in the area of chemistry or core performance. Radiation fields measured have shown a modest but continued improvement. On the basis of the positive trends and absence of any negative effects, Comanche Peak has established elevated constant pHTave 7.4 as the primary chemistry regime for both units. Without the increases in pH/lithium that have taken place, radiation fields would have been expected to increase significantly for longer fuel cycles. The increase in boiling in localized regions of the core (called subcooled nucleate boiling) in PWRs resulting from power uprating has resulted in higher crud buildup on the upper fuel surfaces, and there is growing evidence from US PWRs that radiation fields are indeed higher for the highest rated cores. Enriched boric acid (EBA), that is boric acid enriched with B-10, enables a given pH to be achieved with less lithium hydroxide, as the required concentration of B-10 can be obtained with less total
34
Water Chemistry Control in LWRs
Constant pH 7.2 6
Lithium ‘Li high limit’ ‘Li low limit’
5 Li target = 6.0 E−7 B2 + 0.0023B + 0.4413
Lithium (ppm)
4 3.5 ppm limit 3 2.2 ppm limit 2
Start of 18-month cycle
1
Start of 12-month cycle
0
20
80
140
200
260
320
380
440
500
560
620
680
740
800
860
920
980
1040
1100
1160
1220
1280
1340
1400
1460
1520
1580
1640
1700
0
Boron (ppm) Figure 21 Lithium concentrations required to maintain pH 7.2 for different fuel cycle lengths.
boric acid. EBA is used at several plants in Europe, typically to increase shutdown margin when using mixed oxide fuel (MOX), but has not been applied to date in the United States. However, consideration is being given to using EBA at some plants that will use MOX fuel in the future. Despite the transition to the use of EBA in operating plants, designing for it in new plants is recommended.19 As discussed earlier, the motivation for the initial applications of zinc in most US PWRs was to control PWSCC of SG tubing. However, German-designed PWRs and a few US plants used 5 ppb depleted zinc for radiation control, mostly with depleted zinc to avoid zinc-65 formation. A recent paper ‘‘Understanding the zinc behavior in PWR primary coolant: a comparison between French and German experience’’ by Tigeras et al.20 provides a European perspective on this topic. This paper concludes that ‘zinc injection seems to present the most positive and clearest results: in all the units injecting zinc, a dose rate reduction has been detected after a certain period of exposure without leading to any negative impact on plant
systems, components, and operation.’ Thus ‘zinc injection should be considered as a strategy with benefits in short, medium, and long term. Its application as soon as possible in the life of nuclear power plants and especially before SG replacement and fuel cycles modifications seems to be an excellent decision to contribute to ensuring the passivation process of new components, the fuel performance, the full power operation of the units, and the long life of materials and components.’ Figure 23 shows the effect of zinc in reducing radiation dose rates at several plants. It can be seen that the reduction factor approximately correlates with the cumulative zinc exposure in ppb months (the product of the average zinc concentration and the duration of zinc addition). As little as 5 ppb zinc has been shown to reduce radiation fields by 35–50% at operating plants, based on zinc exposures of 700 ppb months. There is relatively little difference between plants with Alloy 600/690 SG tubing and those with Alloy 800 tubing, but plants using depleted zinc show greater benefit than those using natural zinc, as shown in the figure.
Water Chemistry Control in LWRs
35
Percentage of units within range
70
3.5 ppm
60 50 40 30 20 10 0
2000
2001
2002
2003
2004
2005
2006
2007
EOC year Figure 22 Maximum reported coolant lithium (full power) at US pressurized water reactors.
Cumulative dose rate reduction fraction
1.2 Alloy 800 w/depleted zinc Alloy 600 and 690 w/depleted zinc Alloy 600 and 690 w/natural zinc Log Alloy 800 plants Log Alloy 600 and 690 w/depleted zinc Log Alloy 600 and 690 w/natural zinc
1
0.8
0.6
0.4
0.2
0 0
200
400
600 800 1000 1200 1400 1600 Cumulative zinc exposure (ppb months)
1800
2000
Figure 23 Effect of zinc injection on radiation dose rates.
5.02.3.4
Fuel Performance
With progressive uprating of fuel duty, the margin to tolerate crud has been reduced and additional care has to be taken in specifying the water chemistry to avoid undesirable fuel performance issues. Figure 24 shows the root causes of PWR fuel failures since 2000, including failures from pellet–clad interaction causing SCC, fabrication defects, debris, grid fretting, and crud/corrosion. In contrast to the BWR
situation, shown in Figure 11, very few failures in recent years have been attributed to crud/corrosion (the exceptions to this comment are discussed in a following section). A phenomenon called axial offset (AO) has caused concern over the past 10 years.21 AO is a measure of the relative power produced in the upper and lower parts of the core and is normally expressed as a percent, with a positive percent indicating that
36
Water Chemistry Control in LWRs
Number of failed assemblies
120 Unknown Debris Crud/corrosion
100
Fabrication PCI-SCC Grid fretting
80 60 40 20 0 2000
2001
2002
2003
2004
2005
2006
2007
EOC year Figure 24 US pressurized water reactor fuel failures by mechanism.
more power is produced in the upper part of the core. AOA occurs when boron concentrates in corrosion product deposits (crud) on the upper spans of fuel assemblies undergoing subcooled nucleate boiling, causing a reduction in neutron flux. AOA has affected at least 20 PWRs in the United States, as well as several in other countries. Clearly, fuel crud is involved in the AO phenomenon, and water chemistry effects must be considered in controlling AO. Besides their axial asymmetry, the composition of fuel deposits in boiling cores is different from nonboiling fuel. The nickel-rich deposits on boiling cores tend to be removed much less effectively by conventional chemistry shutdown evolutions than the nickel-ferrite deposits on nonboiling cores. Alternative methods are therefore required for removing corrosion product deposits from reload fuel from highduty cores, including ultrasonic fuel cleaning. An important difference exists between plants with Alloy 600 or 690 SG tubing and those (such as German-designed plants) with Alloy 800 tubing. The latter have a much lower proportion of nickel in fuel crud and have not experienced the AO phenomenon.22 Early work showed that lithium increased zircaloy oxidation rates, although the adverse effects were reduced in the presence of boric acid. As a result, a limit of 2.2 ppm lithium was generally imposed to reduce zircaloy corrosion, although excessive crud formation at low pHs was likely to be more detrimental to the cladding than higher lithium concentrations, especially as the resistance
to corrosion of zircaloy improved. This was confirmed by one of the few failures in recent years that was uniquely attributed to crud buildup. In this example, a move to a longer fuel cycle necessitated increasing the boron concentration at start of cycle; however, the 2.2 ppm lithium limit was retained, resulting in the pH being well below 6.9 for the initial period of the cycle. This in turn caused heavy crud formation, to which subsequent fuel failures were attributed. The move in the past ten years to greater fuel duty, with operation of fuel at higher temperatures (with localized subcooled nucleate boiling), has caused crud-related problems to reappear, particularly the localized flux depression as a result of buildup of boron-containing crud, which were discussed earlier. This in turn has renewed interest in elevated pH/ lithium to minimize corrosion product transport, the use of EBA and the more immediate mitigation that can be obtained from fuel cleaning. Fuel performance is always a concern with changes in water chemistry, such as zinc injection. On the basis of current experience, the impact appears to be minimal for the majority of plants, but insufficient data exist for plants with the highest fuel duties to allow application without postexposure fuel inspections. Data from US plants suggest little or no fuel concerns for coolant zinc levels up to 40 ppb for plants with less-highly rated cores. Extended experience at these plants, over at least 10 years of operation, indicates no adverse effects on fuel at zinc concentrations from 15 to 25 ppb. However, there have been no data
Water Chemistry Control in LWRs
available until recently for higher zinc concentrations in higher duty cores where significant subcooled nucleate boiling occurs on the fuel clad surface.23 Perkins et al.24 comment that fuel performance must be considered prior to injecting zinc and additional monitoring and fuel surveillances to understand and evaluate the impact and the role of zinc may be required in some circumstances.
5.02.4 PWR Secondary System Water Chemistry Experience 5.02.4.1 Evolution of PWR Secondary Chemistry Strategies The objectives of PWR secondary water chemistry control are to maximize secondary system integrity and reliability by minimizing impurity ingress and transport, minimizing SG fouling, and minimizing corrosion damage of SG tubes. Since secondary side corrosion damage of SG tubes is primarily caused by impurities in boiling regions, where high concentrations of impurities occur in occluded regions of the SG formed by corrosion product deposits, new approaches are continually sought to control corrosion product transport to and fouling within the SGs.25 PWRs have experienced IGA on both the primary and secondary sides of the Alloy 600 SG tubing, which has been a major contributing cause of the replacement of most of the SGs with mill-annealed tubing, not only in the United States but internationally. Figure 25 illustrates the various corrosion processes found in different locations in a recirculating SG.26 PWR secondary system water chemistry has evolved through many changes over the years, largely in response to emerging technical issues associated with this degradation of structural materials in SGs. In the early days of PWR operation, wastage became a problem in the secondary side of PWR SGs, resulting in a switch from the use of sodium phosphate inhibitor to all-volatile treatment (AVT) using ammonia, which in turn brought about the denting phenomenon. Tighter control of impurities, oxidizing potential, and pH were necessary to mitigate the denting problem. Despite continued chemistry improvements, many plants have had to replace SGs of earlier designs (e.g., those tubed with Alloy 600MA), as shown in Figure 26. Newer generation SGs are performing well, although there remain concerns about the adverse effects of lead impurity, causing Pb-assisted stress corrosion cracking (PbSCC), which is discussed later.
37
Lead has been observed in various flow streams (final feedwater, heater drains, etc.) in the secondary systems of PWRs. Lead is detected at some concentration in nearly all deposit analyses (SG and other locations). Lead is present in trace concentrations in secondary system materials of construction, as well as in chemical additives such as hydrazine.15 Figure 27 shows the worldwide causes of SG repairs through 2004. It can be seen that IGA is currently the most prevalent form of degradation. Figure 28 compares the behavior of three types of SG tubing, Alloy 600MA (mill-annealed material used in early plants, Alloy 600TT (thermally treated material used in later plants), and Alloy 690TT (an improved alloy used in most replacement SGs). This diagram is taken from the 2008 PWR Secondary Water Chemistry Guidelines,27 which contains a much more detailed account of corrosion processes. 600TT has reduced susceptibility under mildly oxidizing highalkaline conditions, that is, SCC is not observed until higher pH than for 600MA, and 600TT has approximately the same susceptibility as 600MA under acidic conditions. 690TT is indicated as having a still smaller region of susceptibility in the high-alkaline region and as having no susceptibility in the acid region except under highly oxidizing conditions that are unlikely to occur in plants. However, other work indicates that SCC can occur in 690TT at an acidic pH, especially if lead is present. Also, SCC occurs in both 600MA and 600TT in the mid pH region if lead is present. In the 1990s, improved pH control using amines became a regular practice, and fine-tuning, including using mixtures of different amines to control pH throughout the circuit and coordination with resin utilization, continues today. Hydrazine is used to remove oxygen from the system. Hydrazine levels have continually been reviewed and ‘optimized,’ with due regard to any impact on FAC in secondary systems, as FAC rates increase at very low oxygen concentrations. Molar ratio control (MRC) describes a control strategy that adjusts the bulk water chemistry, generally sodium and chloride, such that the solution that is developed in the flow-occluded region is targeted to be near neutral. MRC can involve the addition of chloride ions to ‘balance’ the cations that cannot be reduced via source term reduction programs. MRC was widely practiced to minimize SCC concerns, but has not been actively employed at plants replacing to SGs tubed with Alloy 690TT. With more plants replacing their SGs, less plants are adopting the MRC program. Only ten plants were doing MRC in 2007,28 and they are all with original SGs with
38
Water Chemistry Control in LWRs
U-bend cracks (PWSCC)
Fatigue
Free span ODSCC IGA
ODSCC
PWSCC
Expansion transition
PWSCC
PWSCC or ODSCC
ODSCC Denting
Fretting, wear, corrosion, thinning
Tubesheet Pitting IGA
Expansion transition
Tubesheet
ODSCC Sludge
Tubesheet
PWSCC tube-end cracking
Tubesheet
Figure 25 Corrosion processes in recirculating steam generators, showing primary water stress corrosion cracking and outside diameter stress corrosion cracking on the secondary side.
Water Chemistry Control in LWRs
39
140 Operating plants Plants w/replacement SGs 120
134
84 88
81
134 134
134 72
64 67
59
52
45 51
37
22 29
17
10
7
7
7
7
3 5 7
2
78
134
134
132
131 132
134 131 131
132
132 133
134
12
132
134
12
133
133
130
121
63 68
55
46
1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
1
22
20
31
41
40
0
11
93 99
60
108 115
127
80
79
Number of plants
100
Year Figure 26 Steam generator replacement status worldwide.
100 90 80
Percent
70 60 50 40 30 20 10 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
0 Year IGA Impingement Pitting Other Wear Thinning
Fatigue Unknown SCC Preventive
Figure 27 Worldwide causes of steam generator tube repair.
600MA and 600TT tubing. Currently, no plants with replaced SGs are believed to be using MRC. Titanium-based inhibitors to minimize corrosion are also employed at some plants. Boric acid
treatment (BAT) involves the addition of boric acid to feedwater. Such approaches are worthy of consideration, on the basis of plant-specific degradation mechanisms, operational considerations, and
40
Water Chemistry Control in LWRs
1.0
TT690
Potential (V vs. Ec )
0.8 Some tests indicate that 690 TT may be susceptible in the low pH region, especially if lead is 690 TT U-bend cracked present in near neutral AVT with lead and oxidizing sludge TT690 TT600 TT600 MA600
0.6
0.4
MA600 0.2 600 MA and 600 TT can be susceptible in mid pH range if lead or reduced sulfur is present.
0
2
3
4
5
6 7 8 pH 300 °C (572 °F)
9
10
11
12
Figure 28 Corrosion mode diagram for Alloys 600MA, 600TT and 690TT (based on Constant Extension Rate Tensile Tests at 300 C), showing regions where materials are susceptible to attack.
interactions. The most recent developments are aimed at reducing deposit buildup in crevices, including the use of dispersants, such as polyacrylic acid (PAA), that is discussed in more detail later. The historical trends in PWR secondary chemistry are shown in Figure 29. 5.02.4.2 Chemistry Effects on Materials Degradation of SGs Corrosion of SG tubes has been the major issue affecting selection of secondary water chemistry parameters. However, corrosion and FAC of SG internals and other secondary system components are also important concerns. Corrosion of SG tube materials is mainly affected by the following water chemistry related factors, in addition to nonwater chemistry factors such as material susceptibility, temperature, and stress: pH – Corrosion of several different types, including IGA/SCC and pitting, are strongly affected by the local pH. High pH (caustic conditions) and low pH (acidic conditions) accelerate the rates of IGA/SCC. ECP – The ECP is a measure of the strength of the oxidizing or reducing conditions present at the
metal surface. The rate of corrosion processes are strongly affected by the ECP. Secondary side SCC in tube alloys tends to be accelerated by increases in ECP, that is, by the presence of oxidizing conditions. Specific species – Some impurity species accelerate corrosion of tubing alloys as a result of their effects on pH and ECP. In addition, lead and reduced sulfur species (e.g., sulfides) appear to interfere with formation of protective oxide films on the tube metal surfaces, and thereby increase risks of IGA/SCC, independent of influences on pH or potential. Similarly, chlorides tend to increase the probability of pitting. These factors have been most thoroughly explored for mill-annealed Alloy 600 (600MA). As discussed in Chapter 5.04, Corrosion and Stress Corrosion Cracking of Ni-Base Alloys, tests indicate that the other tubing alloys, that is, stress-relieved Alloy 600 (600SR), thermally treated Alloy 600 (600TT), nuclear grade Alloy 800 (800NG), and thermally treated Alloy 690 (690TT), exhibit similar tendencies, but have increased resistance to corrosive attack, in the order listed, with 690TT having the highest resistance. Laboratory tests and plant experience indicate that 690TT has very high resistance to IGA/SCC on the outside diameter on tubing (OD IGA/SCC) in
Water Chemistry Control in LWRs
41
Pb remediation Dispersants Titanium Molar ratio control MPA, DMA ETA chemistry Morpholine chemistry Boric acid addition EPRI water chemistry guidelines Ammonia chemistry Phosphate 1975
1980
1985
1990
1995
2000
2005
Figure 29 Evolution of water chemistry for pressurized water reactor secondary systems.
normally expected crevice conditions, but OD IGA/ SCC could possibly occur as a result of upsets or as a result of long-term fouling and accumulation of aggressive species in deposit-formed crevices. Alloy 800NG also has high resistance to OD IGA/SCC, but laboratory tests indicate that it is about twice as susceptible as Alloy 690TT, and it has experienced limited amounts of IGA/SCC in plants, while no operation-related corrosion of 690TT has been reported. Laboratory tests and some plant experience indicate that 600TT is significantly more resistant than 600MA but less resistant than 800NG and 690TT. Water chemistry selected to protect SG tubes appears to be satisfactory for most balance-of-plant (BOP) components such as turbines. The main corrosion concerns in the BOP that affect secondary system water chemistry are FAC of carbon steel piping, tubing, and heat exchanger internals and shells, and ‘ammonia’ attack of copper and copper alloy tubes. In addition, FAC has also affected some recirculating SG internal components (e.g., feedrings, swirl vanes). FAC is mainly influenced by the at-temperature pH and oxygen content around the secondary system. ‘Ammonia’ attack of copper alloys is mainly influenced by the concentrations of ammonia and oxygen at the copper alloy locations, but is also accelerated by increases in concentrations and pH associated with other amines, although not as strongly as by increases in ammonia. Once-through steam generators (OTSGs) have different thermal hydraulics and (in original SGs)
tube materials than recirculating steam generators (RSGs). These differences have led to OTSGs having somewhat different tube corrosion experience than RSGs of the same vintage. For the most part, OTSGs have experienced somewhat lower rates of tube degradation. However, significant IGA/IGSCC has been detected in the upper bundle free spans of several units, especially at scratches, and SG replacement has been performed or is planned at all units. The locations in SGs that are most affected by IGA/IGSCC are those where free circulation of secondary water is impeded by the local geometry, for example, in crevices formed by tube support plates or by sludge piles that can accumulate on the tube sheet. Impurities in the secondary water can concentrate in these locations by boiling and evaporation in a process called ‘hideout.’ The key issue influencing water chemistry regimes in PWR secondary system is to minimize SG degradation by controlling sludge buildup, reducing (and balancing, e.g., MRC) the concentration of impurities (i.e., sodium, chloride and sulfate) in deposits at the tube-tubesheet and tube-tube support plate interfaces. The use of advanced amines to control pH has increased significantly in the past few years, as discussed in a following section. Figure 30 shows the main approaches used in typical chemistry control strategies. Impurities are removed from SGs by blowdown of the coolant. Over the past 20 years or so, average
42
Water Chemistry Control in LWRs
Key issue: Mitigating IGA/IGSCC in concentrating regions
Approach: Control local chemistry
Molar ratio control
Reduce Na increase Cl
Reduce iron
Redox potential
Amines dispersants
Reduce Cu increase N2H4
Inhibitors
Boric acid TiO2
Figure 30 Pressurized water reactor secondary chemistry control strategies.
blowdown impurity concentrations in US SGs have been reduced from several ppb to the sub-ppb range. Many PWRs today have SG blowdown concentrations near or below the analytical detection limit. Minimization of impurities is recommended but has been insufficient to prevent or completely mitigate IGA/IGSCC at most plants with susceptible tube material and design, as it can result in sodium-rich feedwater. Cations such as sodium can be more effectively retained by boiling in a crevice than chloride. Hence, excess cations over anions or anions over cations result in specific corrosion issues because of concentration processes in local environments. The original all-volatile treatment used ammonia to control pH, but a less-volatile chemical than ammonia would improve pH control throughout the circuit. Early work employed morpholine, but now several other amines are used. Since the initial application of advanced amine chemistry about 15 years ago, there has been tremendous success in reducing the transport of corrosion products to the SGs by improving the attemperature pH around the BOP, especially in the two-phase regions. This has resulted in mitigation of FAC and thus reduced generation of corrosion products that ultimately get transported to the SGs. Ethanolamine (ETA) remains the most used amine at US plants, with 75% of the US plants using ETA or ETA with other amines, such as dimethylamine (DMA) or 3-methoxypropylamine (MPA), to control secondary cycle pH, as shown in Figure 31.17 Several plants now use a mixture of amines to achieve the optimum pH throughout the secondary system, with 25% of the US plants using MPA or MPA with other amines while 12% of the plants use morpholine or morpholine with other amines.
0%
4% 5%
4%
16% 56%
2% 7% 6%
ETA MPA
ETA/DMA ETA/MPA MPA/DMA MPA/Morph Morph/DMA
ETA/Morph Morph
Figure 31 Amines used in the secondary systems of US pressurized water reactors.
The proper control of oxygen in pressurized water reactor (PWR) secondary feedwater, using an oxygen scavenger such as hydrazine and/or carbohydrazide, has been an enduring issue. The requirements for oxygen concentration necessitate that some optimization take place. Maintaining reducing conditions – that is, low electrochemical potential – in the SG is essential to minimize SCC. On the other hand, some oxygen in the feedwater counteracts corrosion of
Water Chemistry Control in LWRs
1985
< 1992
1993
1994
1995
1996
1998
43
2005
% of US PWRs using advanced water chemistries 120 99
100 80 60
76 62
73
40 58
40 68
44
44
20 23
0
3
Using advanced amines
57
30
23
30 30
50
41
31 33
28
0
3
On molar ratio control
Using >100 ppb FW N2H4
35
19 37 41
17
Using boric acid treatment
Figure 32 Pressurized water reactor secondary chemistry trends.
carbon steel surfaces and the transport of corrosion products to the SG. Recent work has investigated the effect of hydrazine and oxygen on the ECP of SG tubing materials (Alloys 600 and 690) as well as stainless steel (304 and 316) and carbon steel during PWR startup conditions. These laboratory studies have shown that changes in the concentration of hydrazine, used to ensure a reducing potential in the SG, within the typical range allowed and employed (e.g., 20–150 ppb) have no discernable effect on FAC at feedwater temperatures (e.g., 180–235 C). Figure 32 shows the trends in using advanced water chemistry regimes in the secondary systems at US plants. 5.02.4.3
Control of Sludge Fouling of SGs
Corrosion products in the secondary side of PWR SGs primarily deposit on the SG tubes. These deposits can inhibit heat transfer, lead to thermal–hydraulic instabilities through blockage of tube supports, and create occluded regions where corrosive species can concentrate along tubes and within tube-to-tube support plate crevices. The performance of the SGs can be compromised not only through formation of an insulating scale, but also through the removal of tubes from service due to corrosion. Although the application of various amines to control the at-temperature pH (pHT) in specific locations of PWR secondary systems has been
successful in reducing the corrosion of BOP metals and thus reducing corrosion product transport to SGs, a complementary strategy now exists for significantly reducing SG fouling through online application of dispersant, which inhibits deposition. By inhibiting deposition of the corrosion products, the dispersant facilitates more effective removal from the SGs via blowdown. This strategy has been employed at fossil boilers for many decades. However, due to the use of inorganic polymerization initiators (containing sulfur and other impurities), polymeric dispersants had not been utilized in the nuclear industry. Only recently has a PAA dispersant been available that meets the criteria for nuclear application, and progress has been made in reducing SG fouling by application of an online dispersant to substantially improve the efficiency of blowdown iron removal. Dispersant application is proving to be a highly promising technology for markedly decreasing SG fouling, delaying (or possibly eliminating) the need for expensive chemical cleaning and effectively reducing the frequency of sludge lancing for SG maintenance. Online application of PAA to the feedwater system has been successfully demonstrated to greatly increase the efficiency of the blowdown system in eliminating feedwater corrosion products from fouling the SGs. The first application occurred at Arkansas Nuclear One Unit 2 for a three-month trial in early 2000, which demonstrated a significant improvement in the blowdown iron removal efficiency from 2% to 60% with 4–6 ppb PAA in
44
Water Chemistry Control in LWRs
With dispersant
Iron removal efficiency
100
10 Without dispersant
1 0
0.5 1 1.5 2 Dispersant concentration (ppb)
Figure 33 Iron removal efficiency during dispersant test at McGuire pressurized water reactor.
the feedwater. The second application, in 2005–2006, at McGuire Unit 2 for a 6–9 month trial in their replacement SGs tubed with Alloy 690TT showed a similar significant improvement., as shown in Figure 33.15 The following conclusions are evident from the McGuire 2 demonstration described in the above reference: PAA is an effective dispersant. A feedwater PAA concentration approximately equal to the feedwater iron concentration (2 ppb in this case) appears to effectively remove approximately 50% of the influent feedwater iron under steady-state operating conditions. Although blowdown copper spikes with initial PAA application (albeit to a much lesser extent than iron), it quickly returns to normal levels and remains there. Filter element consumption is manageable. Blowdown cation conductivity and ammonia behavior changed during the trial, but these changes are believed to be mainly due to changes in plant configuration and not PAA. The SG thermal performance level has improved slightly with dispersant application, most likely due to slight beneficial changes in the tube deposit thermal properties. The long-term trial at McGuire 2 demonstrated the significant improvement in blowdown iron removal efficiency with application of PAA dispersant (a follow up to the successful short-term trial at ANO-2 in 2000). Based on the success of the McGuire 2 long-term trial, evaluations are in progress with SG vendors looking toward technical concurrence for long-term use in their fleet of recirculating SGs.
5.02.4.4
Lead Chemistry
PbSCC is a serious concern that can affect all SG tubing materials currently employed. A better understanding of lead behavior is needed at SG and feedwater temperatures before possible mitigation techniques can be successfully developed. It is well known that soluble lead at very low concentrations can contribute to SCC of nickel alloys. Likewise, it is well known that some locations on the secondary side of PWR SGs will accumulate lead in the solid state (i.e., deposit) with local concentrations considerably in excess of those observed to accelerate cracking in laboratory testing. Recent investigations using analytical transmission electron microscopy15 have identified lead in the cracks of many tubes pulled from PWRs. However, the absence of extensive operating SG tube failures at rates comparable to what might be predicted based on laboratory studies of PbSCC indicates that some mitigating phenomenon could be present.15 EPRI has published a sourcebook on lead29 that summarizes the state-of-knowledge regarding PbSCC and its effects on PWRs. It incorporates PbSCC laboratory testing, the current understanding of lead transport and other physical chemistry aspects of lead, and the accumulated industry experience regarding PbSCC and its mitigation. Three clearly understood and accepted facts regarding lead in PWR secondary water systems became clear as this sourcebook was being put together: In laboratory testing, the presence of lead accelerates SCC of mill-annealed 600MA, stress-relieved 600SR, and thermally treated 600TT stainless steels as well as thermally treated Alloy 690 (690TT). In operating PWRs, lead is present in the secondary system. In two cases, a large ingress of lead to the secondary system has occurred as a result of lead blankets having been left behind in SGs; the tubes in the affected SGs cracked sooner and faster than in the other SGs at the same units. Set against these known facts are the following four points: The mechanisms by which lead is transported from its ultimate source to the SG tube and into a crack are not well understood, and a comprehensive evaluation of possible mechanisms has not been performed.
Water Chemistry Control in LWRs
The threshold concentration at which lead will accelerate SCC in SGs is not well defined. No definitive indicator of PbSCC is available. There is no well-characterized mechanism by which lead accelerates SCC. Recent work has shown that adsorption/desorption of Pb on corrosion products and SG tubing surfaces could potentially be a major sink/source, respectively, for Pb microscopy.15 There is no direct evidence of adsorption in SGs; however, there is sufficient potential for this mechanism that direct high-temperature measurements under SG conditions have been performed. As a result of ongoing laboratory studies, microscopy15 speculates that formation of a lead layer slows repassivation, after a passive film at the crack tip is disrupted, potentially to an extent to which a crack can initiate and propagate.
5.02.5 Chemistry Control for FAC in BWRs and PWRs FAC causes wall thinning of carbon steel piping, vessels, and components, as discussed in Chapter 5.06, Corrosion and Environmentally-Assisted Cracking of Carbon and Low-Alloy Steels. The wall thinning is caused by an increased rate of dissolution of the normally protective oxide layer, for example, magnetite, that forms on the surface of carbon and low-alloy steels when exposed to highvelocity water or wet steam. The oxide layer reforms and the process continues. If the thinning is not detected in time, the reduced wall cannot withstand the internal pressure and other applied loads. The result can be either a leak or a complete rupture. The rate of wall loss (wear rate) of a given component is affected by temperature, fluid bulk velocity, the effect of component geometry on local hydrodynamics, the at-temperature pH, the liquid phase dissolved oxygen concentration, and the alloy composition. The addition of chromium to steels decreases the rate of FAC. Materials used to replace piping damaged by FAC include low-alloy steels containing chromium and molybdenum (P11, 1.25% Cr–0.5% Mo and P22, 2.25% Cr–1% Mo) and carbon or lowalloy steels clad with stainless steel. Corrosion models are used to estimate wall thinning and determine where monitoring is required. An example of the approach commonly used in the United States is described by Chexal and Horowitz.30
45
The main chemistry factors that affect the rate of FAC are pH and dissolved oxygen concentration. FAC is not an issue for PWR primary systems. As indicated earlier, laboratory studies have shown that changes in the concentration of hydrazine in the PWR secondary system feedwater, used to ensure a reducing potential in the SG, have no discernable effect on FAC at feedwater temperatures, within the typical range allowed and employed. The chemistry parameter that a BWR plant has some degree of control over is dissolved oxygen. Oxygen affects the form and solubility of the oxide layer, the dissolution of which is inherent in FAC. Several plants inject oxygen into the system, as the rate of FAC increases dramatically if the oxygen concentration is less than about 25 ppb. Plant data are shown in Figure 34. Use of HWC in a BWR can significantly reduce the amount of oxygen in the main steam, extraction steam, and heater drain systems, thus potentially increasing the FAC rates in these areas of the plant. The effect of NMCA on the corrosion behavior of carbon steel in 550 F (288 C) water containing various amounts of oxygen and hydrogen has been studied and the data confirm that there is no adverse effect of NMCA on FAC.7 The carbon steel segments of the BWR vessel bottom-head drain line have been identified as being FAC susceptible because of the flow conditions and the potential for low dissolved oxygen concentrations. However, a significant number of inspections have been performed recently at US plants and little thinning has been observed. The 2008 edition of the BWR Water Chemistry Guidelines7 recommends that feedwater oxygen should be maintained above 30 ppb to minimize FAC of carbon and low-alloy steels.
5.02.6 Water Chemistry Control Strategies Sometimes, step changes in chemistry strategy are unavoidable, as with the move to reducing chemistry in BWRs. In these cases, the operators must be prepared to deal with adverse effects. Some BWRs adopting reducing conditions experienced a large jump in out-of-core radiation fields, which may be avoided with prior zinc injection. Addition of new chemicals requires extensive qualification. For example, the successful demonstrations of BWR online noble
Water Chemistry Control in LWRs
Relative FAC wear rate (expressed as percentage of the average wear rate of components at 10 pbb oxygen)
46
160 140 Plant A 120
Plant A
100
Plant B Plant C
80
Plant D 60
Average
40 20 0 0
10
20
30
40
50
60
70
80
Dissolved oxygen (ppb) Figure 34 Plant data showing the relationship between flow-assisted corrosion and dissolved oxygen. (Oxygen values are localized, calculated by the CHECWORKS codes from measured values at condensate or feedwater locations.)
chemistry and PAA dispersants in PWR SGs resulted from detailed monitoring and evaluation during the first injections. If possible, changes in chemistry should be made in baby steps, with monitoring at each step, before further changes are implemented. Examples of this strategy are the gradual increases in lithium/pH and dissolved hydrogen in PWR primary systems. These incremental changes minimize adverse side-effects and allow a planned approach to the optimum plant-specific chemistry control program. The US nuclear power industry produces guidance documents to assist plant personnel in determining a plant-specific chemistry control strategy. The early versions of these documents, developed in the 1980s, listed water chemistry specifications and actions to be taken if the limits were exceeded. As more chemistry options became available, the guidelines evolved into providing guidance on selecting the most appropriate chemistry for a specific plant. Thus, the 2008 BWR Water Chemistry Guidelines7 offers recommendations on controlling ECP, zinc injection, and feedwater iron control. Likewise, the 2007 PWR Primary Water Chemistry Guidelines12 provides guidance on pH control and zinc injection, and the 2008 PWR Secondary Water Chemistry Guidelines27 discusses impurity control, amines, and dispersants. Theses documents are used by all US nuclear power plants and provide the technical basis for similar guidelines used in many other countries. Development of a strategic water chemistry plan, as discussed in these documents, is seen as crucial to controlling material degradation in the future.
References 1.
2.
3. 4. 5. 6.
7. 8. 9.
10.
11.
12.
Swan, T.; Wood, C. J. In Developments in Nuclear Power Plant Water Chemistry, VIIIth International Conference on Water Chemistry of Nuclear Reactor Systems, Oct 23–26, 2000; BNES: Bournemouth, UK, 2000. Fruzzetti, K.; Wood, C. J. In Developments in Nuclear Power Plant Water Chemistry. International Conference on Water Chemistry of Nuclear Reactor System, Jeju Island, Korea, Oct 23–26, 2006. Cohen, P. Water Coolant Technology of Power Reactors; Gordon and Breach: New York, 1969. Jones, R. L. In International Water Chemistry Conference, San Francisco, Oct 11–15, 2004; EPRI: Palo Alto, CA, 2004. Garcia, S.; Wood, C. Recent advances in BWR water chemistry. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. Cowan, R.; Hussey, D. Radiation field trends as related to chemistry in United States BWRs. In 2006 International Conference on Water Chemistry of Nuclear Reactor Systems, Jeju Island, Korea, Oct 23–26, 2006. EPRI. Boiling Water Reactor Water Chemistry Guidelines – 2008 Revision; EPRI: Palo Alto, CA, 2008. Edsinger, K. In Nuclear News; Tompkins, B., Ed.; 2008; pp 34–36. Fruzzetti, K.; Perkins, D. PWR chemistry: EPRI perspective on technical issues and industry research. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. Andresen, P.; Ahluwalia, A.; Hickling, J.; Wilson, J. Effects of PWR primary water chemistry on PWSCC of Ni alloys. In 13th International Conference on Environmental Degradation of Materials in Nuclear Power Systems, Whistler, Canada, Aug 19–23, 2007. Andresen, P.; Ahluwalia, A.; Wilson, J.; Hickling, J. Effects of dissolved H2 and Zn on PWSCC of Ni alloys. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. EPRI. Pressurized Water Reactor Primary Water Chemistry Guidelines: Revision 6; EPRI: Palo Alto, CA, 2007.
Water Chemistry Control in LWRs 13. Pathania, R.; Yagnik, S.; Gold, R. E.; Dove, M.; Kolstad, E. Evaluation of zinc addition to PWR primary coolant. In 7th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems, Breckenridge, CO, NACE: Houston, TX, 1995; pp 163–176. 14. Molander, A.; Jenssen, A.; Norring, K.; Ko¨nig, M.; Andersson, P.-O. Comparison of PWSCC initiation and crack growth data for Alloy 600. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. 15. Fruzzetti, K.; Rochester, D.; Wilson, L.; Kreider, M.; Miller, A. Dispersant application for mitigation of steam generator fouling: Final results from the McGuire 2 long-term trial and an industry update and EPRI perspective for long-term use. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. 16. Haas, C.; Ahluwalia, A.; Kucuk, A.; Perkins, D. PWR operation with elevated hydrogen. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. 17. Hussey, D.; Perkins, D.; Choi, S. Benchmarking radioactivity transport and deposition in PWRs. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. 18. Stevens, J.; Bosma, J. Elevated RCS pH program at Comanche peak. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. 19. Nordmann, F. Worldwide chemistry objectives and solutions for NPP. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008.
20.
21.
22. 23. 24.
25. 26. 27. 28. 29. 30.
47
Tigeras, A.; Stellwag, B.; Engler, N.; Bretelle, J.; Rocher, A. Understanding the zinc behavior in PWR primary coolant: A comparison between French and German experience. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. Frattini, P. L.; Blok, J.; Chauffriat, S.; Sawicki, J.; Riddle, J. In VIIIth International Conference on Water Chemistry of Nuclear Reactor Systems, Oct 23–26, 2000; BNES: Bournemouth, UK, 2000. Riess, R. Personal communication, 2008. Byers, W.; Wang, G.; Deshon, J. Limits of zinc addition in high duty PWRs. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. Perkins, D.; Ahluwalia, A.; Deshon, J.; Haas, C. An EPRI perspective and overview of PWR zinc injection. In VGB NPC’08 Water Chemistry Conference, Berlin, Sept 14–18, 2008. Millett, P. J; Hundley, F. Nucl. Energ. 1997; 36, pp 251–258. EPRI. Personal communication from K. Fruzzetti, 2009. EPRI. Pressurized Water Reactor Secondary Water Chemistry Guidelines – Revision 6; EPRI: Palo Alto, CA, 2008. Choi, S. Personal communication, 2009. EPRI. Pressurized Water Reactor Lead Sourcebook; EPRI: Palo Alto, CA, 2006. Chexal, V.; Horowitz, J. Chexal–Horowitz flowaccelerated corrosion model – Parameters and influences. In ASME PVP-Vol B, Current Perspectives of International Pressure Vessels and Piping Codes and Standards, Book No. H0976B, 1995.
5.03
Corrosion of Zirconium Alloys
T. R. Allen University of Wisconsin, Madison, WI, USA
R. J. M. Konings European Commission, Joint Research Centre, Institute for Transuranium Elements, Karlsruhe, Germany
A. T. Motta The Pennsylvania State University, University Park, PA, USA
ß 2012 Elsevier Ltd. All rights reserved.
5.03.1 5.03.2 5.03.2.1 5.03.2.2 5.03.2.3 5.03.3 5.03.3.1 5.03.3.2 5.03.3.3 5.03.3.4 5.03.3.4.1 5.03.3.4.2 5.03.3.4.3 5.03.4 5.03.5 5.03.5.1 5.03.5.2 5.03.5.3 5.03.5.4 5.03.5.5 5.03.6 5.03.7 References
Introduction General Considerations Oxidation Hydrogen Uptake Controlling Factors for Corrosion Uniform Oxidation Mechanism Temperature and Heat Flux Coolant Chemistry Irradiation Effects Radiolysis Irradiation effects in the oxide layer Changes in the metallurgical state of the metal Nodular Oxidation Hydrogen Embrittlement Hydrogen Production During Aqueous Corrosion of Zirconium-Base Materials Hydrogen Absorption Hydride Formation Hydride Formation Rates Formation of Hydride Rim Delayed Hydride Cracking Summary and Outlook
Abbreviations BWR CANDU CRUD DHC IAEA M5TM PWR tHM VVER ZIRLOTM Zry
Boiling water reactor Canadian Deuterium Uranium Chalk River unidentified deposits Delayed hydride cracking International Atomic Energy Agency Zirconium alloy material with niobium (AREVA) Pressurized water reactor Ton heavy metal Voda Voda Energy Reactor Zirconium alloy material with niobium, tin, and iron (Westinghouse) Zircaloy
49 50 50 51 52 53 53 57 57 59 59 60 60 61 61 62 62 62 63 64 65 66 66
5.03.1 Introduction Zirconium alloys are widely used for fuel cladding and in pressure tubes, fuel channels (boxes), and fuel spacer grids in almost all water-cooled reactors: light water reactors such as the pressurized water reactor (PWR) and the boiling water reactor (BWR) as well as the Canadian designed Canadian Deuterium Uranium (CANDU) heavy water reactor. Since its employment in the first commercial nuclear power plant (Shippingport) in the 1960s, Zircaloy, a zirconium–tin alloy, has shown satisfactory behavior during many decades. However, degradation due to waterside corrosion can limit the in-reactor design life of the nuclear fuel. The critical phenomenon is the
49
50
Corrosion of Zirconium Alloys
hydrogen ingress into the cladding during corrosion, which can cause cladding embrittlement. As utilities are striving to achieve higher fuel burnups, the nuclear industry has made several efforts to understand the mechanisms of corrosion and to mitigate its effects. In striving for increased burnup of the nuclear fuel from 33 000 to 50 000 MWd/tHM and beyond in PWRs, associated studies have shown that the corrosion of the Zircaloy-4 cladding accelerates under these higher burnup conditions. Although alloys that are more modern have not yet shown evidence of this high-burnup acceleration, this is a potential concern. Also, the efforts to increase the thermalcycle efficiency in PWRs by operating at higher temperatures (power uprates), combined with the more aggressive chemistry (introduction of B and Li for example) related to the use of high-burnup fuel, have resulted in increased fuel duty,1 and in increased corrosion rates. This has led to the introduction of cladding tubes of new zirconium alloys such as zirconium–niobium, which are much more corrosion resistant.2,3 With the introduction of these materials, the nuclear industry aims at zero tolerance for fuel failure in the future.4 Many reviews on the corrosion of zirconium alloys both out- and in-reactor, have been published.5–11 The extensive reviews made by an international expert group of the International Atomic Energy Agency (IAEA) and published as IAEA-TECDOCs 684 and 99612,13 are major references in this respect. As mentioned by Cox,6,7 ‘‘the number of publications on this topic is so enormous that it is impossible for a short review to be comprehensive.’’ This also applies to the current chapter, which therefore focuses on the main issues, naturally relying on the above-mentioned existing reviews and updating the information where possible with new results and insights.
protective, thus limiting the access of oxidizing species to the bare metal. Much evidence exists to indicate that Zr oxidation occurs by inward migration of oxygen ions through the oxide layer, either through grain boundaries or through the bulk.5,12,13 Zr þ O2 ¼ ZrO2 As shown in Figure 1, the growth of the oxide layer on the metal surface depends on the kinetics of the oxygen diffusion through this layer. Because the corrosion kinetics slow down as the oxide thickness increases, it has been argued that the rate controlling step in the oxidation process is the transport of atomic species in the protective oxide, by either oxygen diffusion through the oxide film14,15 or diffusion of electrons through the oxide film. These processes are necessarily coupled to maintain electroneutrality. Electron transport is, however, difficult in zirconium dioxide, as it is an electrical insulator when undoped. Although this is not positively confirmed, it is likely that the role of doping elements in the determination of corrosion kinetics is done through their influence on the electron or oxygen transport in the oxide layer. Several types of corrosion morphologies have been observed in nuclear reactors and in autoclave experiments, of which the most important are 1. Uniform: The formation of a thin uniform layer of zirconium dioxide on the surface of a zirconium alloy component (see Figure 2). 2. Nodular : The formation of local, small, circular zirconium oxide blisters (see Figure 3). 3. Shadow: The formation of local corrosion regions that mirror the shape (suggestive of a shadow) of other nearby noble reactor core components (Figure 4).
H2O ® O2− + 2H+
Coolant
5.03.2 General Considerations 5.03.2.1
Oxidation
Corrosion of zirconium alloys in an aqueous environment is principally related to the oxidation of the zirconium by the oxygen in the coolant, dissolved or produced by radiolysis of water. A small amount of oxygen can be dissolved in the metal, but once the thermodynamic solubility limit is exceeded, ZrO2 is formed on the metal. (All zirconium components normally have a thin oxide film (2–5 nm) on their surface in their as-fabricated state.) The oxide formed is
H+ O2−
H+
Oxide H+ + e
®
Zr + 2O2− ® ZrO2 + 4e−
H0
Metal
Figure 1 Schematic presentation of the corrosion of the zirconium alloys. Corrosion of zirconium alloys in nuclear power plants; TECDOC-684; International Atomic Energy Agency, Vienna, Austria, Jan 1993.
Corrosion of Zirconium Alloys
51
Zr + H2O = ZrO2 + H2
ZrH2−x ZrO2
Figure 2 Uniform oxide layer formation and hydride precipitation in Zircaloy cladding. © European Atomic Energy Commission.
The occurrence of these morphologies is strongly dependent on the reactor operating conditions and chemical environment (particularly the concentration of oxygen in the coolant), which are distinctly different in PWRs, BWRs, and CANDU (Table 1). In both BWRs and PWRs, a uniform oxide layer is observed, although its thickness is normally greater in PWR than in BWR, primarily because of the higher operating temperature. Nodular corrosion occurs occasionally in BWRs because a much higher oxygen concentration occurs in the coolant because of water radiolysis and boiling. Shadow corrosion is also occasionally observed in BWRs and is a form of galvanic corrosion.
1 mm
5.03.2.2
100 μm
Figure 3 General appearance of nodules formed on zirconium alloy following a 500 C steam test at 10.3 MPa. In the bottom, a cross-section view of a nodule is shown, exhibiting circumferential and vertical cracks. Photo courtesy of R. Ploc and NFIR (Nuclear Fuel Industry Research Group). Reproduced from Lemaignan, C.; Motta, A. T. Zirconium Alloys in Nuclear Applications, Materials Science and Technology, Nuclear Materials Pt. 2; VCH Verlagsgesellschaft mbH, Weinheim, Germany, 1994.
Hydrogen Uptake
The formation of an oxide layer would not bring severe consequences to cladding behavior were it not for the fact that in parallel with the corrosion process, a fraction of the hydrogen, primarily produced by the oxidation reaction as well as by radiolysis of water, diffuses through the oxide layer into the metal. Zirconium has a very low solubility for hydrogen (about 80 wt ppm at 300 C and 200 wt ppm at 400 C) and once the solubility limit is exceeded, the hydrogen precipitates as a zirconium hydride phase (Figure 2): ZrðH; slnÞ þ H2 ¼ ZrH1:6 or ZrH2 As a result, the following effects have been reported (although not all confirmed) to occur in the cladding: hydrogen embrittlement due to excess hydrogen or its localization into a blister or rim,16,17 loss of
52
Corrosion of Zirconium Alloys
fracture toughness, delayed hydride cracking (DHC), and acceleration of corrosion and of irradiation growth. Hydrogen embrittlement impacts the mechanical resistance of the Zircaloy cladding to
failure and it is thus of key importance to understand its underlying mechanisms. The ductility reduction due to hydrogen embrittlement is dependent on the volume fraction of hydride present, the orientation of the hydride precipitates in the cladding, and their degree of agglomeration.18,19
Oxide
5.03.2.3
Oxide
(b)
(a) Figure 4 Zirconium oxides near (b) and away from (a) a stainless steel control blade bundle, showing the effect of shadow corrosion. Reproduced from Adamson, R. B.; Lutz, D. R.; Davies, J. H. Hot cell observations of shadow corrosion phenomena. In Proceedings Fachtagung der KTG-Fachgruppe, Brennelemente und Kernbautelle, Forschungszentrum Karlsruhe, Feb 29–Mar 1, 2000. Table 1
Controlling Factors for Corrosion
The oxidation and hydrogen uptake of Zircaloy is of course determined by many factors. First of all, the chemical and physical state of the material: composition, metallurgical condition, and surface condition. These conditions are often specific to the material and sometimes batch-specific and also related to the fabrication process, as discussed in detail in Chapter 2.07, Zirconium Alloys: Properties and Characteristics. This is evident from the different behavior of Zircaloy and Zr–Nb alloys, as shown in Figure 5 for two different zirconium alloys employed in the French PWRs, Zircaloy and Zr1% Nb (M5). The peak oxide layer thickness of Zircaloy-4 (oxide thickness at the hottest fuel grid span) increases significantly with burnup (i.e., residence time in the reactor), whereas that of Zr1%Nb shows a moderate increase. In addition, a number of environmental factors affecting the corrosion of zirconium alloys must be considered: 1. Coolant Chemistry: It is obvious that the dissolved oxygen and hydrogen play a major role in the corrosion process, but other dissolved species must also be taken into account. To control the pH of the coolant at slightly alkaline conditions,
Typical reactor environments to which the zirconium alloys are exposed
Coolant Inlet temperature ( C) Outlet temperature ( C) Pressure (MPa) Neutron fluxa (n cm2 s1) Coolant chemistry [O2] (ppb) [H2] (ppm) pH B (as H3BO3) (ppm) Li (as LiOH) (ppm) Na (as NaOH) (ppm) K (as KOH) (ppm) NH3 (ppm)
BWR
PWR
VVER
CANDU
H2O 272–278 280–300 7 4–7 1013
H2O 280–295 310–330 15 6–9 1013
H2O 290 320 15 5–7 1013
D2O 255 300 10 2 1012
200 0.03 7 – – – – –
100 mm (700 wt ppm total hydrogen) exhibited brittle behavior, while those with a thickness