Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems

This two-volume set represents a collection of papers presented at the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors. The purpose of this conference series is to foster an exchange of ideas about problems and their remedies in water-cooled nuclear power plants of today and the future. Contributions cover problems facing nickel-based alloys, stainless steels, pressure vessel and piping steels, zirconium alloys, and other alloys in water environments of relevance. Components covered include pressure boundary components, reactor vessels and internals, steam generators, fuel cladding, irradiated components, fuel storage containers, and balance of plant components and systems.

132 downloads 5K Views 108MB Size

Report

Download pdf

Recommend Stories

Empty story

Idea Transcript

PROCEEDINGS OF THE

August 13–17, 2017 Marriott Portland Downtown Waterfront Portland, Oregon, USA

EDITED BY: John H. Jackson · Denise Paraventi · Michael Wright

The Minerals, Metals & Materials Series

John H. Jackson Denise Paraventi Michael Wright •

Editors

Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors

123

Editors John H. Jackson Idaho National Laboratory Idaho Falls, ID, USA

Michael Wright Chalk River Laboratories Canadian Nuclear Laboratories Chalk River, ON, Canada

Denise Paraventi Naval Nuclear Laboratory Pittsburgh, PA, USA

ISSN 2367-1181 ISSN 2367-1696 (electronic) The Minerals, Metals & Materials Series ISBN 978-3-030-04638-5 ISBN 978-3-030-04639-2 (eBook) https://doi.org/10.1007/978-3-030-04639-2 Library of Congress Control Number: 2018962127 © The Minerals, Metals & Materials Society 2019 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, speciﬁcally the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microﬁlms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a speciﬁc statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

The production of electricity from nuclear power reactors occupies a central role in meeting worldwide energy requirements. As of January 2017, there are 450 power reactors operating in 30 countries. The industry continues to grow, despite setbacks. There are 60 new construction projects underway in 15 different countries. The great majority of the new reactors are being built in countries with emerging economies. For the USA, western Europe, and Japan, although there are some new construction projects, the emphasis is on maintaining capacity and extending the operating lives of existing reactors. Of the worldwide fleet of 450 power reactors, 273 of them have been in service for over 30 years and 67 have been in service for over 40 years. If nuclear power is going to continue making a contribution to worldwide energy demands, power plants must be operated in a safe and cost-effective manner. Central to safety and reliability goals for these plants is effective materials degradation management. Effective materials degradation management must include the maintenance of operating plants, prolongation of the lifetime of plants, and choice and use of materials in new plants. Environmentally induced materials degradation represents a signiﬁcant fraction of materials-related problems in today’s nuclear power plant operation. Under extended lifetime to 60 years or potentially longer, understanding materials degradation will be even more important as these issues are of concern for both economic and safety reasons. Understanding today’s materials problems is also critical for the future in the advanced reactor sector. The purpose of the “Environmental Degradation of Materials in Nuclear Power Systems” conference series is to foster an exchange of ideas about such problems and their remedies. The conference series has been running since 1983. The meetings cover materials problems facing nickel base alloys, stainless steels, ferritic steels, zirconium alloys, and other alloys in water environments of relevance. Components covered include heat transport system pressure boundary components, reactor vessels and internals, steam generators, fuel cladding, irradiated components, fuel storage containers, and balance of plant components and systems. The scope and emphasis of the individual conferences have varied over the past 30 years v

vi

Preface

as different challenges arose and were resolved. Reflecting the growing concerns over materials aging in general, this, the 18th conference in the series, includes a session dedicated to cable insulation aging and concrete aging. This conference featured the presentation of 164 papers, spread over 25 sessions. The conference was attended by over 200 scientists and engineers from 17 countries, representing power plant operators, reactor vendors, regulators, national laboratories, research institutes, universities, and other organizations afﬁliated with the nuclear and nuclear service industries. However, this year’s meeting was sadly missing one of its longest serving and most influential contributors. January 2017 saw the passing of Dr. Roger Washburne Staehle (1934–2017). Roger was an international giant in the ﬁeld of metallurgy and corrosion. Roger made enormous scientiﬁc and engineering contributions to these ﬁelds, and he contributed to every one of the Environmental Degradation meetings and personally attended the majority. Roger studied metallurgical engineering at The Ohio State University (OSU), graduating in 1957. After college, he worked for Admiral Hyman Rickover, the father of the US Nuclear Navy, something Roger was immensely proud of. A successful and influential academic career followed, with a major focus on materials degradation in commercial nuclear power. In his later career, as an international consultant, mentor, and instigator, Roger was an aggressive advocate for the best experimental and analytical techniques. Roger was also a champion of international information exchange. Over his last 20 years, Roger focused considerable efforts to build collaborative interactions in China. Returning to the current meeting, the organizers recognize the help of a dedicated group of individuals who made this 18th edition of the conference a success and these proceedings possible. The Program Committee and Session Chairs listed in this volume’s front matter played vital roles in organizing the sessions and conducting paper reviews. These activities required a signiﬁcant commitment of their time, and they are greatly appreciated. The organizers are thankful to their respective employers for making them available for participating in planning, organizing, and serving at the conference. Special thanks go to the Technical Program Chair, John Jackson of Idaho National Laboratory, and the Assistant Technical Program Chair, Denise Paraventi of the (Bettis) Naval Nuclear Laboratory for their tireless efforts in ensuring the success of this conference. Finally, we thank The Minerals, Metals & Materials Society (TMS) for their assistance with conference planning, paper review, and publication of these proceedings. Michael Wright General Conference Chair

In Memoriam

Roger Washburne Staehle (February 4, 1934–January 16, 2017)

It is with great sadness that we mark the passing of Dr. Roger W. Staehle, an international giant in the ﬁeld of metallurgy and corrosion. Roger made enormous scientiﬁc and engineering contributions to these ﬁelds, but was best known for his friendship, generosity, and drive to support and recognize the work of others—no one “out-gave” Roger. Roger became ill after a fall during his morning walk in the snow and ice. He died several weeks later with members of his family at his bedside. He is survived by his brother, George, and his children, Elizabeth, Eric, Sara, Erin, and William. Roger Staehle studied metallurgical engineering at The Ohio State University (OSU), graduating in 1957 as the president of his senior class. After college, Roger fulﬁlled his Navy ROTC obligation on the staff of Admiral Hyman Rickover, who was the father of the US Nuclear Navy. Following this extensive education in nuclear technology, he returned to get his Ph.D. at OSU under the distinguished guidance of Prof. Mars Fontana. He graduated in 1965 and immediately joined the

vii

viii

In Memoriam

metallurgical engineering faculty. Roger founded the Fontana Corrosion Center (FCC) and built it into one of the largest and most influential academic corrosion laboratories in the world. By the end of the 70s, he had a group of over 40 people in the FCC, with a major focus on materials degradation in commercial nuclear power. He left OSU in 1979 to become Dean of the Institute of Technology at the University of Minnesota (College of Science and Engineering), a position he held until 1983. Roger remained very active to the end as a consultant specializing in the degradation of materials in nuclear reactors. Roger received many awards for his accomplishments. He was a fellow of NACE International and The Electrochemical Society and received the W.R. Whitney Award from NACE. He was elected to the US National Academy of Engineering in 1978, when he was 44 years old, one of the youngest to be elected to NAE at that time. Roger served as the editor for CORROSION journal from 1973 to 1981. Roger devoted enormous energy to coordinating and supporting research throughout the world. He made hundreds of trips to many countries, including Russia, Japan, China, Korea, Europe, South Africa, South America, the Middle East, investing consistently and heavily in promoting international collaborations. He was sufﬁciently proﬁcient in Russian at one time to be able to lecture in that language. Among the early examples of his efforts to develop international collaborations was his organization of major international conferences. Roger also promoted international collaboration by inviting and hosting innumerable prominent experts, as well as early-career scientists, to OSU, and it had a dramatic impact on the thinking and progress throughout the world (including the USA), in addition to fostering the careers of many dozens of international scientists. Roger was heavily engaged in the International Cooperative Group on Environmentally Assisted Cracking, whose charter is to promote international exchange of information, emerging data, and the best experimental techniques for addressing cracking in high-temperature water environments. In the last 20 years, Roger focused considerable efforts to build collaborative interactions in China, aimed primarily but not exclusively on issues related to materials degradation in commercial nuclear power reactors. He initiated and organized four major symposia in China, the ﬁrst of which was “Materials Problems in Light Water Nuclear Power Plants: Status, Mitigation, Future Problems” in 2005. Roger was single-handedly responsible for attracting the best of the world’s experts in each aspect of each symposium, and each symposium attracted about 200 engineers and scientists from throughout China. In 2008, Roger conceived and organized a conference on stress corrosion crack (SCC) initiation in Beaune, France, involving about 130 scientists. The week-long conference was designed to exchange ideas, deﬁne critical experiments, and discuss experimental technique. Many similar efforts were instigated by Roger in areas such as lead and sulfur effects in steam generators, alloys with improved SCC resistance. In 2010, Roger undertook a massive effort to bring together world experts in diverse ﬁelds who could speak to the issues of fundamental understanding and modeling of SCC. The series of four annual meetings brought together about 40 international

In Memoriam

ix

experts for a week-long workshop. Roger realized the importance of such a meeting, deﬁned the content, identiﬁed and invited the experts, and solicited support from about a dozen agencies and companies. These are but a few examples of Roger’s selfless, energetic efforts to promote international collaboration, which are in turn but a fraction of his efforts in consulting, educating, lecturing, mentoring, writing award nominations and references, etc. It is difﬁcult to imagine a world without Roger’s presence, friendship, vision, and energy. Peter L. Andresen, Ron Latanision, and Gerald S. Frankel First Published by NACE in CORROSION—Vol. 73, No. 3

Contents

Part I

PWR Nickel SCC—SCC

Scoring Process for Evaluating Laboratory PWSCC Crack Growth Rate Data of Thick-Wall Alloy 690 Wrought Material and Alloy 52, 152, and Variant Weld Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Amanda R. Jenks, Glenn A. White and Paul Crooker

3

Applicability of Alloy 690/52/152 Crack Growth Testing Conditions to Plant Components. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Warren Bamford, Steve Fyﬁtch, Raj Pathania and Paul Crooker

15

SCC of Alloy 152/52 Welds Defects, Repairs and Dilution Zones in PWR Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Peter L. Andresen, Martin M. Morra and Kawaljit Ahluwalia

27

NRC Perspectives on Primary Water Stress Corrosion Cracking of High-Chromium, Nickel-Based Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . Greg Oberson, Margaret Audrain, Jay Collins and Eric Reichelt

55

Stress Corrosion Cracking of Alloy 52/152 Weldments Near Dissimilar Metal Weld Interfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Alexandreanu, Y. Chen, W.-Y. Chen and K. Natesan

67

Stress Corrosion Crack Growth Rate Testing of Composite Material Specimens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David S. Morton, John V. Mullen, Eric Plesko, John Sutliff, Robert Morris and Nathan Lewis

87

Investigation of Hydrogen Behavior in Relation to the PWSCC Mechanism in Alloy TT690 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Takumi Terachi, Takuyo Yamada, Nobuo Totsuka and Koji Arioka

xi

xii

Part II

Contents

PWR Nickel SCC—Initiation

Crack Initiation of Alloy 600 in PWR Water . . . . . . . . . . . . . . . . . . . . . . 121 Peter Andresen and Peter Chou SCC Initiation Behavior of Alloy 182 in PWR Primary Water . . . . . . . . 137 Mychailo Toloczko, Ziqing Zhai and Stephen Bruemmer Multiple Cracks Interactions in Stress Corrosion Cracking: In Situ Observation by Digital Image Correlation and Phase Field Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 J. Bolivar, T.T. Nguyen, Y. Shi, M. Fregonese, J. Réthoré, J. Adrien, A. King, J.Y. Bufﬁere and N. Huin Stress Corrosion Cracking Initiation of Alloy 82 in Hydrogenated Steam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 E. Chaumun, J. Crépin, C. Duhamel, C. Guerre, E. Héripré, M. Sennour and I. de Curières Application of Ultra-High Pressure Cavitation Peening on Reactor Vessel Head Penetration, BMN and Primary Nozzles . . . . . . . . . . . . . . . 191 Daniel Brimbal, Gary Poling, Darren Wood, Antoine Marion, Nicolas Huin and Olivier Calonne The Effect of Surface Condition on Primary Water Stress Corrosion Cracking Initiation of Alloy 600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 S.R. Pemberton, M.A. Chatterton, A.S. Grifﬁths, S.L. Medway, D.R. Tice and K.J. Mottershead Microstructural Effects on SCC Initiation in PWR Primary Water for Cold-Worked Alloy 600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 Ziqing Zhai, Mychailo Toloczko and Stephen Bruemmer Part III

PWR Nickel SCC—Aging Effects

A Kinetic Study of Order-Disorder Transition in Ni–Cr Based Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 B. Stephan, D. Jacob, F. Delabrouille and L. Legras The Role of Stoichiometry on Ordering Phase Transformations in Ni–Cr Alloys for Nuclear Applications . . . . . . . . . . . . . . . . . . . . . . . . . 251 Fei Teng, Li-Jen Yu, Octav Ciuca, Emmanuelle Marquis, Grace Burke and Julie D. Tucker The Effect of Hardening via Long Range Order on the SCC and LTCP Susceptibility of a Nickel-30Chromium Binary Alloy . . . . . . . . . . . . . . . . 261 Tyler E. Moss, Catherine M. Brown and George A. Young

Contents

xiii

PWSCC Initiation of Alloy 600: Effect of Long-Term Thermal Aging and Triaxial Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 Seung Chang Yoo, Kyoung Joon Choi, Seunghyun Kim, Ji-Soo Kim, Byoung Ho Choi, Yun-Jae Kim, Jong-Sung Kim and Ji Hyun Kim Stress Corrosion Cracking Behavior of Alloy 718 Subjected to Various Thermal Mechanical Treatments in Primary Water . . . . . . . . . . . . . . . . 293 Mi Wang, Miao Song, Gary S. Was and L. Nelson Time- and Fluence-to-fracture Studies of Alloy 718 in Reactor . . . . . . . . 307 C. Joseph Long Development of Short-Range Order and Intergranular Carbide Precipitation in Alloy 690 TT upon Thermal Ageing . . . . . . . . . . . . . . . . 321 Roman Mouginot, Teemu Sarikka, Mikko Heikkilä, Mykola Ivanchenko, Unto Tapper, Ulla Ehrnstén and Hannu Hänninen Part IV

PWR Nickel SCC—Alloy 600 Mechanistic

Diffusion Processes as Possible Mechanisms for Cr Depletion at SCC Crack Tip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337 Josiane Nguejio, Jérôme Crepin, Cécilie Duhamel, Fabrice Gaslain, Catherine Guerre, François Jomard and Marc Maisonneuve Role of Grain Boundary Cr5B3 Precipitates on Intergranular Attack in Alloy 600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359 Daniel K. Schreiber, Matthew J. Olszta, Karen Kruska and Stephen M. Bruemmer Advanced Characterization of Oxidation Processes and Grain Boundary Migration in Ni Alloys Exposed to 480 °C Hydrogenated Steam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375 S.Y. Persaud, B. Langelier, A. Eskandari, H. Zhu, G.A. Botton and R.C. Newman Exploring Nanoscale Precursor Reactions in Alloy 600 in H2/N2–H2O Vapor Using In Situ Analytical Transmission Electron Microscopy . . . . 399 M.G. Burke, G. Bertali, F. Scenini, S.J. Haigh and E. Prestat Electrochemical and Microstructural Characterization of Alloy 600 in Low Pressure H2-Steam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409 L. Volpe, G. Bertali, M. Curioni, M.G. Burke and F. Scenini Effect of Dissolved Hydrogen on the Crack Growth Rate and Oxide Film Formation at the Crack Tip of Alloy 600 Exposed to Simulated PWR Primary Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423 Johan Stjärnsäter, Jiaxin Chen, Fredrik Lindberg, Peter Ekström and Pål Efsing

xiv

Contents

A Mechanistic Study of the Effect of Temperature on Crack Propagation in Alloy 600 Under PWR Primary Water Conditions . . . . . 439 Zhao Shen and Sergio Lozano-Perez Part V

PWR Nickel SCC—Alloy 690 Mechanistic

Grain Boundary Damage Evolution and SCC Initiation of Cold-Worked Alloy 690 in Simulated PWR Primary Water . . . . . . . . 457 Ziqing Zhai, Mychailo Toloczko, Karen Kruska, Daniel Schreiber and Stephen Bruemmer PWSCC Susceptibility of Alloy 690, 52 and 152 . . . . . . . . . . . . . . . . . . . . 485 Takaharu Maeguchi, Kimihisa Sakima, Kenji Sato, Koji Fujimoto, Yasuto Nagoshi and Kazuya Tsutsumi Relationship Among Dislocation Density, Local Strain Distribution, and PWSCC Susceptibility of Alloy 690 . . . . . . . . . . . . . . . . . . . . . . . . . . 501 Tae-Young Ahn, Sung-Woo Kim, Seong Sik Hwang and Hong-Pyo Kim Morphology Evolution of Grain Boundary Carbides Precipitated Near Triple Junctions in Highly Twinned Alloy 690 . . . . . . . . . . . . . . . . . . . . . 509 Hui Li, Xirong Liu, Kai Zhang, Wenqing Liu and Shuang Xia A Mechanistic Study of Stress Corrosion Crack Propagation in Heavily Cold Worked TT Alloy 690 Exposed to Simulated PWR Primary Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519 Toshio Yonezawa, Masashi Watanabe and Atsushi Hashimoto Microstructural Study on the Stress Corrosion Cracking of Alloy 690 in Simulated Pressurized Water Reactor Primary Environment . . . . . . . 535 Wenjun Kuang, Miao Song, Chad M. Parish and Gary S. Was Part VI

Irradiation Damage—Stainless Steel

Effect of Strain Rate and High Temperature Water on Deformation Structure of VVER Neutron Irradiated Core Internals Steel . . . . . . . . . 549 Anna Hojna, Jan Duchon, Patricie Halodova and Hygreeva Kiran Namburi Radiation-Induced Precipitates in a Self-ion Irradiated Cold-Worked 316 Austenitic Stainless Steel Used for PWR Bafﬂe-Bolts . . . . . . . . . . . . 565 Jan Michalička, Zhijie Jiao and Gary Was In Situ and Ex Situ Observations of the Inﬂuence of Twin Boundaries on Heavy Ion Irradiation Damage Effects in 316L Austenitic Stainless Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 581 G. Meric de Bellefon, J.C. van Duysen and K. Sridharan

Contents

xv

In Situ Microtensile Testing for Ion Beam Irradiated Materials . . . . . . . 593 H.T. Vo, A. Reinhardt, D. Frazer, N. Bailey, P. Chou and P. Hosemann Development of High Irradiation Resistant and Corrosion Resistant Oxide Dispersion Strengthened Austenitic Stainless Steels . . . . . . . . . . . . 605 Takahiro Ishizaki, Yusaku Maruno, Kiyohiro Yabuuchi, Sosuke Kondo and Akihiko Kimura Spherical Nanoindentation Stress-Strain Analysis of Ion-Irradiated Tungsten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 617 Siddhartha Pathak, Jordan S. Weaver, Cheng Sun, Yongqiang Wang, Surya R. Kalidindi and Nathan A. Mara Part VII

Irradiation Damage—Swelling

Formation of He Bubbles by Repair-Welding in Neutron-Irradiated Stainless Steels Containing Surface Cold-Worked Layer . . . . . . . . . . . . . 639 Masato Koshiishi and Naoto Shigenaka Predictions and Measurements of Helium and Hydrogen in PWR Structural Components Following Neutron Irradiation and Subsequent Charged Particle Bombardment. . . . . . . . . . . . . . . . . . . . . . . 651 F.A. Garner, L. Shao and C. Topbasi Emulating Neutron-Induced Void Swelling in Stainless Steels Using Ion Irradiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 669 C. Sun, L. Malerba, M.J. Konstantinovic, F.A. Garner and S.A. Maloy Carbon Contamination, Its Consequences and Its Mitigation in IonSimulation of Neutron-Induced Swelling of Structural Metals . . . . . . . . 681 Lin Shao, Jonathan Gigax, Hyosim Kim, Frank A. Garner, Jing Wang and Mychailo B. Toloczko Void Swelling Screening Criteria for Stainless Steels in PWR Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 695 Sarah Davidsaver, Steve Fyﬁtch, Daniel Brimbal, Joshua McKinley and Kyle Amberge Theoretical Study of Swelling of Structural Materials in Light Water Reactors at High Fluencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 709 S.I. Golubov and A.V. Barashev Part VIII

Irradiation Damage—Nickel Based and Low Alloy

High Resolution Transmission Electron Microscopy of Irradiation Damage in Inconel X-750 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 727 C.D. Judge, H. Rajakumar, A. Korinek, G. Botton, J. Cole, J.W. Madden, J.H. Jackson, P.D. Freyer, L.A. Giannuzzi and M. Grifﬁths

xvi

Contents

In Situ SEM Push-to-Pull Micro-tensile Testing of Ex-service Inconel X-750 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 743 C. Howard, C.D. Judge, H.T. Vo, M. Grifﬁths and P. Hosemann Microstructural Characterization of Proton-Irradiated 316 Stainless Steels by Transmission Electron Microscopy and Atom Probe Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 759 Yun Soo Lim, Dong Jin Kim and Seong Sik Hwang Part IX

PWR Stainless Steel SCC and Fatigue—SCC

Microstructural Effects on Stress Corrosion Initiation in Austenitic Stainless Steel in PWR Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775 D.R. Tice, V. Addepalli, K.J. Mottershead, M.G. Burke, F. Scenini, S. Lozano-Perez and G. Pimentel Oxidation and SCC Initiation Studies of Type 304L SS in PWR Primary Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 793 F. Scenini, J. Lindsay, Litao Chang, Y.L. Wang, M.G. Burke, S. Lozano-Perez, G. Pimentel, D. Tice, K. Mottershead and V. Addepalli SCC Initiation in the Machined Austenitic Stainless Steel 316L in Simulated PWR Primary Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 811 Litao Chang, Jonathan Duff, M. Grace Burke and Fabio Scenini High-Resolution Characterisation of Austenitic Stainless Steel in PWR Environments: Effect of Strain and Surface Finish on Crack Initiation and Propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 829 G. Pimentel, D.R. Tice, V. Addepalli, K.J. Mottershead, M.G. Burke, F. Scenini, J. Lindsay, Y.L. Wang and S. Lozano-Perez SCC of Austenitic Stainless Steels Under Off-Normal Water Chemistry and Surface Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849 Nicolas Huin, Olivier Calonne, Matthias Herbst and Renate Kilian SCC of Austenitic Stainless Steels Under Off-Normal Water Chemistry and Surface Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 865 Matthias Herbst, Renate Kilian, Nicolas Huin and Olivier Calonne The Effect of Microchemistry on the Crack Response of Lightly Cold Worked Dual Certiﬁed Type 304/304L Stainless Steel After Sensitizing Heat Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 881 K.B. Fisher, B.D. Miller, E.C. Johns, R. Hermer, C. Brown and E.A. Marquis

Contents

Part X

xvii

PWR Stainless Steel SCC and Fatigue—Fatigue

The Effect of Load Ratio on the Fatigue Crack Growth Rate of Type 304 Stainless Steels in Air and High Temperature Deaerated Water at 482 °F . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 895 D.J. Paraventi, C.M. Brown, L.B. O’Brien and B.A. McGraw Electrical Potential Drop Observations of Fatigue Crack Closure. . . . . . 913 E.A. West The Effect of Environment and Material Chemistry on Single-Effects Creep Testing of Austenitic Stainless Steels . . . . . . . . . . . . . . . . . . . . . . . 927 L.B. O’Brien and B.D. Miller Corrosion Fatigue Behavior of Austenitic Stainless Steel in a Pure D2O Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 943 L. Yu, R.G. Ballinger, X. Huang, M.M. Morra, L.B. O’Brien, D.J. Paraventi, V.S. Smentkowski and P.W. Stahle Mechanistic Understanding of Environmentally Assisted Fatigue Crack Growth of Austenitic Stainless Steels in PWR Environments . . . . 957 S.L. Medway, D.R. Tice, N. Platts, A. Grifﬁths, G. Ilevbare and R. Pathania Study on Hold-Time Effects in Environmental Fatigue Lifetime of Low-Alloy Steel and Austenitic Stainless Steel in Air and Under Simulated PWR Primary Water Conditions . . . . . . . . . . . . . . . . . . . . . . . 987 M. Herbst, A. Roth and J. Rudolph Part XI

Special Topics I—Materials

Evaluation of Additively Manufactured Materials for Nuclear Plant Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1009 R.M. Horn, M. Connor, D. Webber, J. Jackson and F. Bolger Hot Cell Tensile Testing of Neutron Irradiated Additively Manufactured Type 316L Stainless Steel. . . . . . . . . . . . . . . . . . . . . . . . . . 1021 Paula D. Freyer, William T. Cleary, Elaine M. Ruminski, C. Joseph Long and Peng Xu Computational and Experimental Studies on Novel Materials for Fission Gas Capture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039 Shenli Zhang, Haoyan Sha, Erick Yu, Maria Perez Page, Ricardo Castro, Pieter Stroeve, Joseph Tringe and Roland Faller Hydrogen Assisted Cracking Studies of a 12% Chromium Martensitic Stainless Steel—Inﬂuence of Hardness, Stress and Environment . . . . . . 1051 D.A. Horner, M. Lowden, P. Nevitt and G. Quirk

xviii

Contents

Investigation of Flow Accelerated Corrosion Models to Predict the Corrosion Behavior of Coated Carbon Steels in Secondary Piping Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1067 Seunghyun Kim and Ji Hyun Kim Effect of High-Temperature Water Environment on the Fracture Behaviour of Low-Alloy RPV Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1077 Z. Que, H.P. Seifert, P. Spätig, G.S. Rao and S. Ritter Corrosion Fatigue Testing of Low Alloy Steel in Water Environment with Low Levels of Oxygen and Varied Load Dwell Times . . . . . . . . . . . 1101 Cybele Gabris Feasibility Study of the Internal Zr/ZrO2 Reference Electrodes in Supercritical Water Environments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1109 Yu-Hsuan Li, Yu-Ming Tung, Tsung-Kuang Yeh and Mei-Ya Wang Part XII

Special Topics II—Processes

Investigation of Pitting Corrosion in Sensitized Modiﬁed HighNitrogen 316LN Steel After Neutron Irradiation . . . . . . . . . . . . . . . . . . . 1125 D.A. Merezhko, M.S. Merezhko, M.N. Gussev, J.T. Busby, O.P. Maksimkin, M.P. Short and F.A. Garner Quantifying Erosion-Corrosion Impacts on Light Water Reactor Piping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1141 C.E. Guzmán-Leong, J.W. Cluever and S.R. Gosselin Effect of Molybdate Anion Addition on Repassivation of Corroding Crevice in Austenitic Stainless Steel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1157 Shun Watanabe, Tomohiro Sekiguchi, Hiroshi Abe and Yutaka Watanabe Effect of pH on Hydrogen Pick-Up and Corrosion in Zircaloy-4 . . . . . . 1169 James Sayers, Susan Ortner, Kexue Li and Sergio Lozano-Perez Oxidation Kinetics of Austenitic Stainless Steels as SCWR Fuel Cladding Candidate Materials in Supercritical Water . . . . . . . . . . . . . . . 1181 Hiroshi Abe, Ryuichi Suzuki and Yutaka Watanabe A Recent Look at CANDU Feeder Cracking: High Resolution Transmission Electron Microscopy and Electron Energy Loss Near Edge Structure (ELNES) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195 C.D. Judge, S.Y. Persaud, A. Korinek and M.D. Wright

Contents

Part XIII

xix

Cables and Concrete Aging and Degradation–Cables

Simultaneous Thermal and Gamma Radiation Aging of Electrical Cable Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1219 Leonard S. Fiﬁeld Principal Component Analysis (PCA) as a Statistical Tool for Identifying Key Indicators of Nuclear Power Plant Cable Insulation Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1227 Chamila C. De Silva, Scott P. Beckman, Shuaishuai Liu and Nicola Bowler How Can Material Characterization Support Cable Aging Management? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1241 David Rouison, Marzieh Riahinezhad and Anand Anandakumaran Aqueous Degradation in Harvested Medium Voltage Cables in Nuclear Power Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1257 R.C. Duckworth, A. Ellis, B. Hinderliter, E. Hill and M. Maurer-Jones Frequency Domain Reﬂectometry Modeling and Measurement for Nondestructive Evaluation of Nuclear Power Plant Cables . . . . . . . . . 1267 S.W. Glass, L.S. Fiﬁeld, A.M. Jones and T.S. Hartman Aging Mechanisms and Nondestructive Aging Indicator of Filled Cross-Linked Polyethylene (XLPE) Exposed to Simultaneous Thermal and Gamma Radiation . . . . . . . . . . . . . . . . . 1281 Shuaishuai Liu, Leonard S. Fiﬁeld and Nicola Bowler Successful Detection of Insulation Degradation in Cables by Frequency Domain Reﬂectometry . . . . . . . . . . . . . . . . . . . . . . . . . . 1293 Yoshimichi Ohki and Naoshi Hirai Capacitive Nondestructive Evaluation of Aged Cross-Linked Polyethylene (XLPE) Cable Insulation Material . . . . . . . . . . . . . . . . . . 1303 Z.H. Shao and N. Bowler Tracking of Nuclear Cable Insulation Polymer Structural Changes Using the Gel Fraction and Uptake Factor Method . . . . . . . . . . . . . . . 1315 Miguel Correa, Qian Huang and Leonard S. Fiﬁeld Degradation of Silicone Rubber Analyzed by Instrumental Analyses and Dielectric Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1323 Yoshimichi Ohki, Naoshi Hirai, Daomin Min, Liuqing Yang and Shengtao Li

xx

Part XIV

Contents

Cables and Concrete Aging and Degradation–Concrete

Automated Detection of Alkali-Silica Reaction in Concrete Using Linear Array Ultrasound Data . . . . . . . . . . . . . . . . . . . . . . . . . . 1335 Dwight A. Clayton, Hector Santos-Villalobos, N. Dianne Bull Ezell, Joseph Clayton and Justin Baba Coupled Physics Simulation of Expansive Reactions in Concrete with the Grizzly Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1347 Benjamin W. Spencer and Hai Huang Overview of EPRI Long Term Operations Work on Nuclear Power Plant Concrete Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1359 Joe Wall and Sam Johnson The Effects of Neutron Irradiation on the Mechanical Properties of Mineral Analogues of Concrete Aggregates . . . . . . . . . . . . . . . . . . . 1367 Thomas M. Rosseel, Maxim N. Gussev and Luis F. Mora Part XV

Accident Tolerant Fuel Cladding

Accident Tolerant FeCrAl Fuel Cladding: Current Status Towards Commercialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1381 Kevin G. Field, Yukinori Yamamoto, Bruce A. Pint, Maxim N. Gussev and Kurt A. Terrani Interdiffusion Behavior of FeCrAl with U3Si2 . . . . . . . . . . . . . . . . . . . 1391 Rita E. Hoggan, Lingfeng He and Jason M. Harp Mechanical Behavior of FeCrAl and Other Alloys Following Exposure to LOCA Conditions Plus Quenching . . . . . . . . . . . . . . . . . . 1401 Evan J. Dolley, Michael Schuster, Cole Crawford and Raul B. Rebak Mechanical Behavior and Structure of Advanced Fe-Cr-Al Alloy Weldments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1417 M.N. Gussev, K.G. Field, E. Cakmak and Y. Yamamoto Investigating Potential Accident Tolerant Fuel Cladding Materials and Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1431 K. Daub, S.Y. Persaud, R.B. Rebak, R. Van Nieuwenhove, S. Ramamurthy and H. Nordin Steam Oxidation Behavior of FeCrAl Cladding . . . . . . . . . . . . . . . . . . 1451 B.A. Pint, K.A. Terrani and R.B. Rebak In-Situ Proton Irradiation-Corrosion Study of ATF Candidate Alloys in Simulated PWR Primary Water . . . . . . . . . . . . . . . . . . . . . . 1461 Peng Wang and Gary S. Was

Contents

xxi

Hydrothermal Corrosion of SiC Materials for Accident Tolerant Fuel Cladding with and Without Mitigation Coatings . . . . . . . . . . . . . 1475 Stephen S. Raiman, Caen Ang, Peter Doyle and Kurt A. Terrani Characterization of the Hydrothermal Corrosion Behavior of Ceramics for Accident Tolerant Fuel Cladding . . . . . . . . . . . . . . . . 1485 Peter J. Doyle, Stephen S. Raiman, R. Rebak and Kurt A. Terrani Corrosion of Multilayer Ceramic-Coated ZIRLO Exposed to High Temperature Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1497 Kiran K. Mandapaka and Gary S. Was Part XVI

General SCC and SCC Modeling

Calibration of the Local IGSCC Engineering Model for Alloy 600 . . . 1511 Thierry Couvant, Jacqueline Caballero, Cécilie Duhamel, Jérôme Crépin and Takaharu Maeguchi Prediction of IGSCC as a Finite Element Modeling Post-analysis . . . . 1535 Thierry Couvant Monte Carlo Simulation Based on SCC Test Results in Hydrogenated Steam Environment for Alloy 600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1551 Yohei Sakakibara, Ippei Shinozaki, Gen Nakayama, Takashi Nan-Nichi, Tomoyuki Fujii, Yoshinobu Shimamura and Keiichiro Tohgo Protection of the Steel Used for Dry Cask Storage System from Atmospheric Corrosion by Tio2 Coating . . . . . . . . . . . . . . . . . . . . . . . . 1563 Jing-Ru Yang, Mei-Ya Wang, Tsung-Kuang Yeh and Peter Chen Predictive Modeling of Bafﬂe-Former Bolt Failures in Pressurized Water Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1573 Gregory A. Banyay, Matthew H. Kelley, Joshua K. McKinley, Matthew J. Palamara, Scott E. Sidener and Clarence L. Worrell Technical Basis and SCC Growth Rate Data to Develop an SCC Disposition Curve for Alloy 82 in BWR Environments . . . . . . . . . . . . 1589 Katsuhiko Kumagai, Yusuke Sakai and Takayuki Kaminaga Part XVII

BWR SCC and Water Chemistry

SCC and Fracture Toughness of XM-19 . . . . . . . . . . . . . . . . . . . . . . . 1607 Peter Andresen, Martin Morra and Robert Carter On the Effect of Preoxidation of Nickel Alloy X-750 . . . . . . . . . . . . . . 1623 Silvia Tuzi, Kenneth Göransson, Fang Liu, Mattias Thuvander and Krystyna Stiller

xxii

Contents

Microstructures of Oxide Films Formed in Alloy 182 BWR Core Shroud Support Leg Cracks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1633 Jiaxin Chen, Daniel Jädernäs, Fredrik Lindberg, Henrik Pettersson, Martin Bjurman, Kwadwo Kese, Anders Jenssen, Massimo Cocco and Hanna Johansson Effect of Chloride Transients on Crack Growth Rates in Low Alloy Steels in BWR Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1649 Xiaoyuan Lou and Raj Pathania Electrochemical Behavior of Platinum Treated Type 304 Stainless Steels in Simulated BWR Environments Under Startup Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1663 Chu-Yung Yuan, Tsung-Kuang Yeh and Mei-Ya Wang Investigations of the Dual Beneﬁts of Zinc Injection on Cobalt-60 Uptake and Oxide Film Formation Under Boiling Water Reactor Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1673 Samuel Holdsworth, Fabio Scenini, M. Grace Burke, Tsuyoshi Ito, Yoichi Wada, Hideyuki Hosokawa, Nobuyuki Ota and Makoto Nagase SCC Mitigation in Boiling Water Reactors: Platinum Deposition and Durability on Structural Materials . . . . . . . . . . . . . . . . . . . . . . . . 1685 Pascal V. Grundler, Stefan Ritter and Lyubomira Veleva Conﬁrmation of On-Line NobleChem™ (OLNC) Mitigation Effectiveness in Operating Boiling Water Reactors (BWRs) . . . . . . . . . 1701 Joe Kopcash, Juan Varela, Hubert Huie and G. Depta Development of the Fundamental Multiphysics Analysis Model for Crevice Corrosion Using a Finite Element Method . . . . . . . . . . . . . . . 1713 Masahiko Tachibana, Yoichi Wada, Takayuki Arakawa, Yoshiharu Kikuchi and Takehiro Seto In Situ Electrochemical Study on Crevice Environment of Stainless Steel in High Temperature Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1725 Y. Soma, C. Kato and F. Ueno Part XVIII

Zirconium and Fuel Cladding

Corrosion Fatigue Crack Initiation in Zr-2.5Nb . . . . . . . . . . . . . . . . . . 1741 H.M. Nordin, A.J. Phillion, T.M. Karlsen and S. Persaud Cluster Dynamics Model for the Hydride Precipitation Kinetics in Zirconium Cladding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1759 Donghua Xu and Hang Xiao

Contents

xxiii

Modeling Corrosion Kinetics of Zirconium Alloys in Loss-of-Coolant Accident (LOCA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1769 Léo Borrel and Adrien Couet Progressing Zirconium-Alloy Corrosion Models Using Synchrotron XANES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1781 Michael Moorehead, Adrien Couet, Jing Hu and Zhonghou Cai Advanced Characterization of Hydrides in Zirconium Alloys . . . . . . . 1793 S.M. Hanlon, S.Y. Persaud, F. Long and M.R. Daymond Inﬂuence of Alloying Elements and Effect of Stress on Anisotropic Hydrogen Diffusion in Zr-Based Alloys Predicted by Accelerated Kinetic Monte Carlo Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1815 Jianguo Yu, Chao Jiang and Yongfeng Zhang Part XIX

Stainless Steel Aging and CASS

Inﬂuence of d-Ferrite Content on Thermal Aging Induced Mechanical Property Degradation in Cast Stainless Steels . . . . . . . . . . . . . . . . . . . 1829 Thak Sang Byun, Timothy G. Lach, Ying Yang and Changheui Jang Microstructure and Deformation Behavior of Thermally Aged Cast Austenitic Stainless Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1841 Y. Chen, C. Xu, X. Zhang, W.-Y. Chen, J.-S. Park, J. Almer, M. Li, Z. Li, Y. Yang, A.S. Rao, B. Alexandreanu and K. Natesan Microstructural Evolution of Cast Austenitic Stainless Steels Under Accelerated Thermal Aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1859 Timothy G. Lach and Thak Sang Byun Electrochemical Characteristics of Delta Ferrite in Thermally Aged Austenitic Stainless Steel Weld . . . . . . . . . . . . . . . . . . . . . . . . . . 1869 Gokul Obulan Subramanian, Sunghoon Hong, Ho Jung Lee, Byeong Seo Kong, Kyoung-Soo Lee, Thak Sang Byun and Changheui Jang Effect of Long-Term Thermal Aging on SCC Initiation Susceptibility in Low Carbon Austenitic Stainless Steels . . . . . . . . . . . . . . . . . . . . . . 1879 So Aoki, Keietsu Kondo, Yoshiyuki Kaji and Masahiro Yamamoto Crack Growth Rate and Fracture Toughness of CF3 Cast Stainless Steels at ~3 DPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1889 Y. Chen, W.-Y. Chen, B. Alexandreanu, K. Natesan and A.S. Rao Effects of Thermal Aging and Low Dose Neutron Irradiation on the Ferrite Phase in a 308L Weld . . . . . . . . . . . . . . . . . . . . . . . . . . 1905 Z. Li, Y. Chen, A.S. Rao and Y. Yang

xxiv

Contents

Microstructural Evolution of Welded Stainless Steels on Integrated Effect of Thermal Aging and Low Flux Irradiation . . . . . . . . . . . . . . . 1919 Martin Bjurman, Kristina Lindgren, Mattias Thuvander, Peter Ekström and Pål Efsing Part XX

Welds, Weld Metals, and Weld Assessments

The Use of Tapered Specimens to Evaluate the SCC Initiation Susceptibility in Alloy 182 in BWR and PWR Environments . . . . . . . . 1929 Juxing Bai, Stefan Ritter, Hans-Peter Seifert, Marc Vankeerberghen and Rik-Wouter Bosch Effect of Thermal Aging on Fracture Mechanical Properties and Crack Propagation Behavior of Alloy 52 Narrow-Gap Dissimilar Metal Weld . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1949 Matias Ahonen, Sebastian Lindqvist, Teemu Sarikka, Jari Lydman, Roman Mouginot, Ulla Ehrnstén, Pekka Nevasmaa and Hannu Hänninen Distribution and Characteristics of Oxide Films Formed on Stainless Steel Cladding on Low Alloy Steel in Simulated PWR Primary Water Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965 Qi Xiong, Hongjuan Li, Zhanpeng Lu, Junjie Chen, Qian Xiao, Jiarong Ma, Xiangkun Ru and Xue Liang Microstructural Characterization of Alloy 52 Narrow-Gap Dissimilar Metal Weld After Aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1979 Teemu Sarikka, Roman Mouginot, Matias Ahonen, Sebastian Lindqvist, Ulla Ehrnstén, Pekka Nevasmaa and Hannu Hänninen A Statistical Analysis on Modeling Uncertainty Through Crack Initiation Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1995 Jae Phil Park, Chanseok Park and Chi Bum Bahn Part XXI

Plant Operating Experience

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2015 James Hyres, Rocky Thompson and Jim Batton Root Cause Analysis of Cracking in Alloy 182 BWR Core Shroud Support Leg Cracks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2035 Martin Bjurman, Daniel Jädernäs, Kwadwo Kese, Anders Jenssen, Jiaxin Chen, Massimo Cocco and Hanna Johansson

Contents

xxv

Microbially Induced Corrosion in Fireﬁghting Systems—Experience and Remedies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2047 Ulla Ehrnstén, Leena Carpén and Kimmo Tompuri Managing the Ageing Degradation of Concealed Safety Relevant Cooling Water Piping in European S/KWU LWRs . . . . . . . . . . . . . . . 2057 Martin Widera, Gerd Ahlers, Bernd Gruhne and Thomas Wermelinger Identiﬁcation of PWR Stainless Steel Piping Safety Signiﬁcant Locations Susceptible to Stress Corrosion Cracking . . . . . . . . . . . . . . . 2069 R. Hosler, A. Kulp, P. Stevenson and S. Petro Part XXII

IASCC Testing—Characterization

On the Use of Density-Based Algorithms for the Analysis of Solute Clustering in Atom Probe Tomography Data . . . . . . . . . . . . . . . . . . . . 2097 Emmanuelle A. Marquis, Vicente Araullo-Peters, Yan Dong, Auriane Etienne, Svetlana Fedotova, Katsuhiko Fujii, Koji Fukuya, Evgenia Kuleshova, Anabelle Lopez, Andrew London, Sergio Lozano-Perez, Yasuyoshi Nagai, Kenji Nishida, Bertrand Radiguet, Daniel Schreiber, Naoki Soneda, Mattias Thuvander, Takeshi Toyama, Faiza Sefta and Peter Chou Comparative Study on Short Time Oxidation of Un-Irradiated and Protons Pre-Irradiated 316L Stainless Steel in Simulated PWR Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2115 M. Boisson, L. Legras, F. Carrette, O. Wendling, T. Sauvage, A. Bellamy, P. Desgardin, L. Laffont and E. Andrieu Hydrogen Trapping by Irradiation-Induced Defects in 316L Stainless Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2135 Anne-Cécile Bach, Frantz Martin, Cécilie Duhamel, Stéphane Perrin, François Jomard and Jérôme Crépin Grain Boundary Oxidation of Neutron Irradiated Stainless Steels in Simulated PWR Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2153 Takuya Fukumura, Koji Fukuya, Katsuhiko Fujii, Terumitsu Miura and Yuji Kitsunai Irradiation Assisted Stress Corrosion Cracking (IASCC) of Nickel-Base Alloys in Light Water Reactors Environments—Part I: Microstructure Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2165 M. Song, M. Wang, G.S. Was, L. Nelson and R. Pathania Irradiation Assisted Stress Corrosion Cracking (IASCC) of Nickel-Base Alloys in Light Water Reactors Environments Part II: Stress Corrosion Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . 2177 Mi Wang, Miao Song, Gary S. Was, L. Nelson and R. Pathania

xxvi

Contents

Solute Clustering in As-irradiated and Post-irradiation-Annealed 304 Stainless Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2189 Yimeng Chen, Yan Dong, Emmanuelle Marquis, Zhijie Jiao, Justin Hesterberg, Gary Was and Peter Chou Part XXIII

IASCC Testing—Initiation and Growth

Irradiation-Assisted Stress Corrosion Cracking Initiation Screening Criteria for Stainless Steels in PWR Systems . . . . . . . . . . . . . . . . . . . . 2211 Steve Fyﬁtch, Sarah Davidsaver and Kyle Amberge Novel Technique for Quantitative Measurement of Localized Stresses Near Dislocation Channel—Grain Boundary Interaction Sites in Irradiated Stainless Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2221 D.C. Johnson and G.S. Was IASCC Susceptibility of 304L Stainless Steel Irradiated in a BWR and Subjected to Post Irradiation Annealing . . . . . . . . . . . . . . . . . . . . 2231 Justin R. Hesterberg, Zhijie Jiao and Gary S. Was Irradiation Assisted Stress Corrosion Cracking Susceptibility of Alloy X-750 Exposed to BWR Environments . . . . . . . . . . . . . . . . . . 2243 S. Teysseyre, J.H. Jackson, P.L. Andresen, P. Chou and B. Carter Evaluation of Crack Growth Rates and Microstructures Near the Crack Tip of Neutron-Irradiated Austenitic Stainless Steels in Simulated BWR Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2255 Yasuhiro Chimi, Shigeki Kasahara, Hitoshi Seto, Yuji Kitsunai, Masato Koshiishi and Yutaka Nishiyama Effect of Specimen Size on the Crack Growth Rate Behavior of Irradiated Type 304 Stainless Steel . . . . . . . . . . . . . . . . . . . . . . . . . 2271 A. Jenssen, P. Chou and C. Tobpasi Plastic Deformation Processes Accompanying Stress Corrosion Crack Propagation in Irradiated Austenitic Steels . . . . . . . . . . . . . . . . . . . . . 2289 M.N. Gussev, G.S. Was, J.T. Busby and K.J. Leonard Part XXIV

PWR Oxides and Deposits

Effect of Grain Orientation on Irradiation Assisted Corrosion of 316L Stainless Steel in Simulated PWR Primary Water . . . . . . . . . 2303 Rigel D. Hanbury and Gary S. Was Finite Element Modelling to Investigate the Mechanisms of CRUD Deposition in PWR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2313 Jiejie Wu, Nicholas Stevens, Fabio Scenini, Brian Connolly, Andy Banks, Andrew Powell and Lara-Jane Pegg

Contents

xxvii

Properties of Oxide Films on Ni–Cr–xFe Alloys in a Simulated PWR Water Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2327 Xiangkun Ru, Zhanpeng Lu, Junjie Chen, Guangdong Han, Jinlong Zhang, Pengfei Hu, Xue Liang and Wenqing Liu Part XXV

PWR Secondary Side

Effect of Applied Potential and Inhibitors on PbSCC of Alloy 690TT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2345 Brent Capell, Jesse Lumsden, Michael Calabrese and Rick Eaker Corrosion of SG Tube Alloys in Typical Secondary Side Local Chemistries Derived from Operating Experience . . . . . . . . . . . . . . . . . 2361 Ian de Curieres Investigation on the Effect of Lead (Pb) on the Degradation Behavior of Passive Films on Alloy 800 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2381 J. Ulaganathan and H. Ha Inﬂuence of Alloying on a-aʹ Phase Separation in Duplex Stainless Steels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2399 David A. Garﬁnkel, Jonathan D. Poplawsky, Wei Guo, George A. Young and Julie D. Tucker Stress Corrosion Cracking of Alloy 800 in Secondary Side Crevice Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2409 Maria-Lynn Komar and Guylaine Goszczynski Using Modern Microscopy to “Fingerprint” Secondary Side SCC in Ni–Fe Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2421 S.Y. Persaud, J.M. Smith, C.D. Judge, M. Bryk, R.C. Newman, M.G. Burke, I. de Curieres, B.M. Capell and M.D. Wright Appendix: 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors: Questions and Answers . . . . . . . . 2453 Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2513 Subject Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2519

About the Editors/Organizers

Technical Program Chair: John Jackson, Idaho National Laboratory Dr. John H. Jackson is a Distinguished Staff Scientist/Engineer at the Idaho National Laboratory (INL) in Idaho Falls, Idaho. He currently has dual responsibility as the Gateway for Accelerated Innovation in Nuclear (GAIN) Technical Interface and as the Industry Program Lead for the Nuclear Science User Facilities (NSUF). In these capacities, John works closely with the DOE Ofﬁce of Nuclear Energy and the nuclear industry to ensure that DOE facilities are used effectively to maintain the current reactor fleet and to enable innovation. John has nearly twenty years of experience in the areas of mechanical testing and fracture mechanics. He also has over three years of experience in extreme environment materials characterization and drilling mechanics at the ExxonMobil Upstream Research Company in Houston, Texas. John holds Ph.D. (2001) and M.S. (1998) degrees in Mechanical Engineering from the University of Washington, Seattle, WA, and a B.S. in Mechanical Engineering Technology (1995) from Central Washington University in Ellensburg, WA.

xxix

xxx

About the Editors/Organizers

Assistant Technical Program Chair: Denise Paraventi, Naval Nuclear Laboratory Dr. Denise Paraventi is an Advisor Engineer at the Bettis site of the Naval Nuclear Laboratory in Pittsburgh, PA, and has worked at Bettis since 2000. Her career has been dedicated to testing to understand environmental degradation of materials in nuclear power systems, primarily focused on stress corrosion cracking of nickel-based alloys and corrosion fatigue of austenitic stainless steels. She has extensive experience with mechanical and high-temperature autoclave testing of these materials. She has also been involved with Electric Power Research Institute (EPRI) efforts to evaluate Alloy 690 stress corrosion crack growth rates for use in commercial industry disposition curves. Denise holds a Ph.D. in Materials Science and Engineering from the University of Michigan in Ann Arbor and a B.S. degree in Metallurgical Engineering from Michigan Technological University in Houghton, Michigan. Conference General Chair: Michael Wright, Canadian Nuclear Laboratories Dr. Michael Wright has worked at the Canadian Nuclear Laboratories' (CNL) Chalk River site since 1993. His career at Chalk River has exposed him to a wide range of degradation issues and materials: nickel alloys, ferritic steels, stainless steels, aluminum, and zirconium. Mike has applied his expertise largely to corrosion and cracking issues (fatigue, creep-cracking, hydride cracking, stress corrosion cracking, and corrosion fatigue), but he has also worked on assessments of irradiation embrittlement of aluminum- and nickel-based alloys. Much of his work has been in direct support of utility operations largely related to steam generators or primary heat transport system degradation issues. He has also provided support for nuclear operations at the Chalk River site (e.g., high-level-waste storage-tank corrosion, researchreactor vessel repair). Mike is currently working in a business development role at CNL. Before moving to Canada, and the nuclear industry, Mike worked for four years at The Welding Institute (TWI) in the UK

About the Editors/Organizers

xxxi

where he led welding-related commercial research projects and failure investigations. Work at TWI started Mike’s career-long interest in environmentally assisted cracking. Mike holds a Ph.D. from Birmingham University (UK) for research on fatigue of aerospace alloys.

Organizing Committee for Environmental Degradation Conferences

Todd Allen, Third Way and University of Wisconsin Peter Andresen, Andresen Consulting Steve Bruemmer, Paciﬁc Northwest National Laboratory Jeremy Busby, Oak Ridge National Laboratory Thierry Couvant, Electricite de France Ian de Curieres, IRSN Pal Efsing, Vattenfall, Sweden Ulla Ehrnsten, VTT Technical Research Centre of Finland Lionel Fournier, AREVA, France Steve Fyﬁtch, AREVA, USA Barry Gordon, Structural Integrity Associates Inc. Catherine Guerre, CEA En-Hou Han, Institute of Metal Research, China Ron Horn, GE-Hitachi, retired Il Soon Hwang, Seoul National University, Korea Gabriel Ilevbare, Idaho National Laboratory John Jackson, Idaho National Laboratory Anders Jenssen, Studsvik, Sweden Renate Kilian, AREVA, GmbH Hong Pyo Kim, KAERI, Korea Peter King, PJ King Consulting, Canada Stuart Medway, AMEC Foster Wheeler Dave Morton, Naval Nuclear Laboratory Larry Nelson, JLN Consulting Greg Oberson, US Nuclear Regulatory Commission Denise Paraventi, Naval Nuclear Laboratory Hans-Peter Seifert, PSI, Switzerland Robert Tapping, Canadian Nuclear Laboratories

xxxiii

xxxiv

Organizing Committee for Environmental Degradation Conferences

Gary Was, University of Michigan Yutaka Watanabe, Tohoku University Mike Wright, CNL, Canada TK Yeh, Nat Tsing Hua University, Taiwan Toshio Yonezawa, Tohoku University, Japan

Session Chairs and Co-chairs

Note The primary session chair is italicized. Zirconium and Fuel Cladding Jacki Stevens, AREVA Inc. Evan Dolley, GE Global Research George Jiao, University of Michigan Accident Tolerant Fuel Cladding Gary Was, University of Michigan Bruce Pint, Oak Ridge National Laboratory Cem Topbasi, Electric Power Research Institute Cables and Concrete Aging and Degradation–Cables Leo Fiﬁeld, Paciﬁc Northwest National Laboratory Robert Duckworth, Oak Ridge National Laboratory David Rouison, Kinectrics Cables and Concrete Aging and Degradation–Concrete Thomas Rosseel, Oak Ridge National Laboratory Joe Wall, Electric Power Research Institute BWR SCC and Water Chemistry Bob Carter, Electric Power Research Institute Earl Johns, Naval Nuclear Laboratory Susan Garcia, Electric Power Research Institute Plant Operating Experience Maria-Lynn Komar, Kinectrics Inc. Peter King, PJKing Consulting Inc. Irradiation Damage–Nickel Based and Low Alloy Mychailo Toloczko, Paciﬁc Northwest National Laboratory Maxim Gussev, Oak Ridge National Laboratory Myles Connor, General Electric–Hitachi

xxxv

xxxvi

Irradiation Damage–Stainless Steel Larry Nelson, JLN Consulting Sarah Davidsaver, AREVA GmbH Anna Hojna, Centrum vyzkumu Rez Irradiation Damage–Swelling Frank Garner, Radiation Effects Consulting Cheng Sun, Idaho National Laboratory Sebastien Teysseyre, Idaho National Laboratory PWR Ni Alloy SCC–Alloy 600 Mechanistic Steve Bruemmer, Paciﬁc Northwest National Laboratory Thierry Couvant, EDF Tony Horner, Rolls Royce plc PWR Ni Alloy SCC–Alloy 690 Mechanistic Stuart Medway, AMEC Foster Wheeler Matt Olszta, Paciﬁc Northwest National Laboratory Hannu Hänninen, Aalto University PWR Ni Alloy SCC–SCC Bogdan Alexandreanu, Argonne National Laboratory Sonya Pemberton, AMEC Foster Wheeler Grace Burke, The University of Manchester PWR Ni Alloy SCC–Initiation Dave Morton, Naval Nuclear Laboratory Ziqing Zhai, Paciﬁc Northwest National Laboratory Meg Audrain, US Nuclear Regulatory Commission PWR Ni Alloy SCC–Aging Effects Tyler Moss, Naval Nuclear Laboratory Peter Chou, Electric Power Research Institute Dan Schreiber, Paciﬁc Northwest National Laboratory IASCC Testing–Initiation and Growth Peter Andresen, Andresen Consulting Yiren Chen, Argonne National Laboratory Colin Judge, Canadian Nuclear Laboratory IASCC Testing–Characterization Anders Jenssen, Studsvik Nuclear AB Masato Koshiishi, Nippon Nuclear Fuel Development Mike McMurtrey, Idaho National Laboratory PWR Stainless Steel SCC and Fatigue–Fatigue Denise Paraventi, Naval Nuclear Laboratory Barry Gordon, Structural Integrity Associates Inc. Renate Killian, AREVA GmbH

Session Chairs and Co-chairs

Session Chairs and Co-chairs

PWR Stainless Steel SCC and Fatigue–SCC Gabriel Ilevbare, Idaho National Laboratory Keith Leonard, Oak Ridge National Laboratory Elaine West, Knolls Atomic Power Laboratory PWR Secondary Side Ian De Curieres, IRSN Jared Smith, Canadian Nuclear Laboratory Brent Capell, Electric Power Research Institute Special Topics II: Processes Ulla Ehrnstén, VTT Technical Research Centre of Finland Ltd. TK-Yeh, National Tsing Hua University George Young, Dominion Engineering Welds, Weld Metals, and Weld Assessments Catherine Guerre, CEA Bryan Miller, Naval Nuclear Laboratory Hans Peter Seifert, Paul Scherrer Institute PWR Oxides and Deposits Cecilie Duhamel, MINES ParisTech Fabio Scenini, The University of Manchester General SCC and SCC Modeling Raj Pathania, Electric Power Research Institute David Tice, AMEC Foster Wheeler Jean Smith, Electric Power Research Institute Stainless Steel aging and CASS Steve Fyﬁtch, AREVA Inc. Jeremy Busby, Oak Ridge National Laboratory TS Byun, Paciﬁc Northwest National Laboratory Special Topics I: Materials Pål Efsing, Ringhals AB Rory Kennedy, Idaho National Laboratory Peter Hosemann, University of California, Berkeley

xxxvii

Part I

PWR Nickel SCC—SCC

Scoring Process for Evaluating Laboratory PWSCC Crack Growth Rate Data of Thick-Wall Alloy 690 Wrought Material and Alloy 52, 152, and Variant Weld Material Amanda R. Jenks, Glenn A. White and Paul Crooker

Abstract Due to the widespread use of thick-wall Alloy 690 and its corresponding weld metals Alloys 52 and 152 in various replacement, repair, mitigation, and new plant pressurized water reactor (PWR) applications, there is an industry need for an equation or methodology to predict crack growth rates (CGRs) for primary water stress corrosion cracking (PWSCC) of these materials. An international PWSCC CGR Expert Panel was organized by EPRI, with the participation of national laboratories sponsored by the US NRC, to support the development of such PWSCC CGR equations. A database of over 500 Alloy 690 CGR data points and over 130 Alloy 52/152 CGR data points from seven research laboratories was compiled, evaluated and scored for data quality, and assessed to determine the effects of numerous parameters such as temperature, crack-tip stress intensity factor, yield strength, and crack orientation. The process by which these data were evaluated and scored is presented in this paper. Keywords PWSCC

Crack growth rate Expert panel Alloy 690

The participation of NRC is authorized by a legal Memorandum of Understanding between NRC and EPRI. The conclusions of this paper do not reflect technical or regulatory positions of NRC. A.R. Jenks (&) G.A. White Dominion Engineering, Inc., 12100 Sunrise Valley Drive, Suite 220, Reston, VA 20191, USA e-mail: [email protected] P. Crooker Electric Power Research Institute, 3420 Hillview Ave., Palo Alto, CA 94304, USA © The Minerals, Metals & Materials Society 2019 J.H. Jackson et al. (eds.), Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, The Minerals, Metals & Materials Series, https://doi.org/10.1007/978-3-030-04639-2_1

3

4

A.R. Jenks et al.

Introduction Alloys 690 and 52/152 were selected as replacement materials for Alloys 600 and 82/182/132, respectively, because of their much greater resistance to primary water stress corrosion cracking (PWSCC) due, apparently, to their higher chromium contents. Almost 30 years of plant experience to date has supported this decision in that no PWSCC indications have been observed in the replacement materials. However, meticulous laboratory testing has shown that these high-chromium materials are not immune to PWSCC growth and, in fact, may yield relatively fast crack growth rates (CGRs) under certain material and environmental conditions. Therefore, given the possibility of PWSCC initiation or PWSCC crack growth from a manufacturing flaw conservatively considered analogous to an initiated crack, the rate at which cracks grow in Alloy 690 and Alloy 52/152 due to PWSCC is of signiﬁcant importance to the industry. Much of the work to develop deterministic models to describe PWSCC CGRs in Alloys 690 and 52/152 is rooted in work for Alloy 600 that produced MRP–55 in 2002 [1, 2] and for Alloy 82/182/132 that produced MRP–115 in 2004 [3, 4]. Those reports detailed the efforts to develop CGR disposition equations for the lower-chromium materials, including compiling CGR databases, screening the data, and performing statistical analyses investigating multiple factors to describe the CGR behavior. One of the signiﬁcant differences between the MRP–55 and MRP–115 efforts and the current one is the data screening process. The more sophisticated laboratory techniques necessary for measurement of the lower CGRs in Alloys 690 and 52/152 require that the data undergo more detailed evaluation than was used for Alloys 600 and 82/182/132 to ensure the quality of the data used for CGR model development. This paper describes the data evaluation and scoring process used by the Data Evaluation Group of the PWSCC CGR Expert Panel. This scoring process, as well as the full statistical analysis of the CGR data and the resulting models, is documented more extensively in MRP–386 [5].

Data Compilation CGR data for Alloys 690 and 52/152 were obtained from seven laboratories located in four countries: Amec Foster Wheeler, ANL, Bettis/KAPL (now Naval Nuclear Laboratory), CIEMAT, GE–GRC, PNNL, and Tohoku University. Only data without cyclic loading (i.e., data typically under constant load or constant stress intensity factor (CL/CK)) were included in the database, as the effects of periodic partial unloading on Alloys 690 and 52/152 SCC CGRs are often large, inconsistent, and generally not well characterized at this time. Alloy 690 HAZ data were included in the Alloy 690 database, and variants of Alloys 52 and 152 (speciﬁcally, 52i, 52M, 52MSS, and 152M) were included in the Alloy 52/152 database (and

Scoring Process for Evaluating Laboratory PWSCC …

5

throughout this paper are implicitly included when the “52/152” terminology is used). Weld interface data, including crack growth data near the fusion line and within the dilution zone, are not included in the assessments presented in this paper because of the limited amount of available data in these conditions. A single test specimen may have several distinct test conditions, i.e., “test segments,” each of which was considered on an individual basis. Over 40 attributes were recorded for each CL/CK test segment, including: • Material heat/weld wire • Product form and post-supplier processing, including heat treatment and applied cold work (thickness reduction percentage and method) • Yield strength and hardness values • Crack growth orientation versus product form and versus applied cold work • Stress intensity factor • Temperature • Dissolved hydrogen concentration in the water • Test segment duration • Crack length increment • Intergranular morphology percentage • Crack growth rate Although only CL/CK test segments were used for CGR model development, information on all segments throughout the test, including pre-cracking and transitioning test segments, were generally available. Data were provided in tabular form and/or in crack length vs. time plots. Macro- and micrographs of the fracture surface were also often available. The Alloy 690 database prior to scoring totaled 532 test segments. The Alloy 52/152 database prior to scoring totaled 137 test segments.

Scoring Process As with numerous similar efforts, the compiled laboratory data were evaluated prior to use in the CGR model development to enhance the quality of the ﬁnal models. The evaluation, or scoring, procedure was geared to be data-centered and to focus on the experimental techniques, overall data quality, and testing credibility, rather than on expected dependencies or relevance of the material or testing condition to plant applications. (These latter issues were addressed by the Applications Group of the Expert Panel during the statistical analysis process.) The data evaluation method used for the MRP–55 [1] and MRP–115 [3] efforts was a screening process, in which ﬁxed criteria were selected and applied to all data, such that any test segments that did not pass all requirements were screened out of the ﬁnal database used for model development. Expert input was key to developing appropriate screening criteria for Alloys 600 and 82/182/132, but the

6

A.R. Jenks et al.

experts were not required to do a detailed assessment of each data point. In contrast, this effort for Alloys 690 and 52/152 involved an expert scoring process, in which the Data Evaluation Group members evaluated each data point in detail and assigned scores based on guidelines or general criteria identiﬁed by the group. (A similar process was used for the IASCC database and modeling effort [6].) The categories of criteria are similar to those in MRP–115, but because testing CGRs of Alloys 690 and 52/152 requires greater sophistication, evaluating the results requires a much more nuanced approach than was used for the more susceptible low-Cr alloys. Therefore, it was decided that it was more appropriate to use expert judgment to evaluate the data using a scoring process than to apply hard threshold criteria in a screening process.

Scoring Approach The Data Evaluation Group of the Expert Panel held extensive discussions to decide on the approach to evaluate the data and to determine the general criteria to be used. Because of the complexities involved in obtaining CGR data for Alloys 690 and 52/152 and the fact that every test has its own set of challenges, the interpretation and weight given to each criterion varied among experts and among the different data points; this is expected and is the essence of expert judgment. One expert disagreed with this scoring approach and opted to write a dissent, although he still participated in the data scoring process.1 Nevertheless, there was generally good agreement across the board in the data scoring among all experts. One of the unique characteristics of Alloys 690 and 52/152 CGR data is the presence of low and very low growth rates, i.e., less than 1 10−11 m/s. CGRs of this order of magnitude are common for Alloys 690 and 52/152 in the as-received or lightly cold worked conditions. However, these rates are nearly impossible to validate fractographically, which can raise issues when trying to determine whether crack advance occurred during cycling or at constant load/K. This also means that the minimum crack growth increment used in MRP–115 of 0.5 mm is impractical for these PWSCC-resistant alloys, as it would take over a year and a half for the crack to grow to such an extent, based on a growth rate of 1 10−11 m/s. Because low growth rates inherently have more uncertainty associated with their exact value due, for example, to the lack of fractographic support and the necessary corrections for DCPD resistivity drift, it was acknowledged that this uncertainty could negatively impact scores of low growth rate test segments. This is undesirable for a number of reasons, primarily because it unfairly penalizes the material for being so resistant to PWSCC and biases the database to higher CGRs. Therefore, a two-column scoring approach was used. The ﬁrst, and primary, score was the

1

This dissent, as well as a response by the other members of the Expert Panel, can be found in Appendix C of the upcoming report MRP–386 [5].

Scoring Process for Evaluating Laboratory PWSCC …

7

Segment Credibility score, to rate the test segment for data quality and assess how credible the growth rate observation is. The scoring criteria detailed in the next section were used to assign this score on a scale of 1 (high quality test) to 5 (low quality test). The second score was for CGR Uncertainty, to approximately quantify how much conﬁdence one has in the exact value of the reported crack growth rate on a scale of 1 (high conﬁdence) to 5 (low conﬁdence). For example, a test that was run in an exemplary fashion but had a CL/CK CGR of 5 10−13 m/s could get a very good Segment Credibility score but a poor CGR Uncertainty score due to the lack of conﬁdence of whether the actual CGR was exactly 5 10−13 m/s or if it could have been 1 10−13 m/s or 1 10−12 m/s. This score was implemented ensure that low growth rates could be viewed as very credible observations even if there was great uncertainty in the precise CGR, rather than having the test segment be scored low simply because of the low CGR. The Segment Credibility score was then used to determine which data were included in the statistical modeling process, using an average score threshold; any data with an average score better than this threshold would be included in the statistical modeling database, while the others would be removed for the initial modeling but later compared to the CGR model. In all cases, it was important that the data were allowed to speak for themselves, rather than using expectation-based scoring using, for example, prior experience with similar specimens or material conditions or using a speciﬁc value calculated from cyclic data. Discussion was often a key element in reconciling differences in scoring. Individual scores that deviated substantially from the average score of all experts were highlighted to provide that person the opportunity to reassess their score, articulate the basis for their score, and understand the reasons for other scores. Often this produced a change in scoring, but not always, as is appropriate.

Scoring Criteria Numerous criteria were considered when assessing the Alloy 690 and Alloy 52/152 CGR data. The following sections describe several of these topics.

Testing Environment The testing environment should be relevant to PWR primary water conditions, although some variation was allowed. In particular, the following conditions should be met: • Temperature should be 280–360°C. • Impurity levels should be controlled and kept to a minimum. • Dissolved hydrogen, oxygen, pH, boron, and lithium concentrations are known and controlled within appropriate ranges.

8

A.R. Jenks et al.

Loading Conditions Loading conditions should be stable and accurate; i.e., unintentional load fluctuations should not occur during the test, particularly during CL/CK test segments.

Test Segment Duration and Growth Increment Test segment duration and crack growth increment within a given CL/CK test segment should be sufﬁcient for the CGR to develop and stabilize, and there should be a high signal-to-noise ratio. Ideally, the crack increment would be large enough to represent a reasonable amount of the microstructure and be validated fractographically. However, it was recognized that this is often not possible due to the low CGRs of as-received and low-CW material. Test segment duration was ideally >1000 h, although 15 MPa m) and frequency (often >0.5 Hz). The fatigue precrack should be fully nucleated along the machined notch and of uniform depth. Transitioning is then implemented to extend the crack to (1) achieve 100% engagement (growth) from the fatigue precrack, (2) extend the crack so that the fatigue-hardened plastic zone becomes a monotonic plastic zone, and (3) provide a good opportunity for the crack to ﬁnd the most susceptible path, which in austenitic materials is often intergranular (IG). The crack may not become fully IG, however, if the SCC susceptibility of the material is sufﬁciently low because the grain boundaries have very low susceptibility. Thus, an important aspect of performing the transitioning steps is to provide opportunity (e.g., through sufﬁciently large crack increments and increasingly gentle cycling) for the crack to behave as if it were always an SCC crack, regardless of whether the crack morphology is IG or not. Actively loaded specimens that exhibited no crack growth (or little or no IG morphology) during the test segment were scored primarily based on the evidence of good transitioning/opportunity.

Post-Test Correction Errors in DCPD measurements can occur during SCC tests for many reasons. One cause is the presence of an uneven crack front, where the error increases as the difference between the average and minimum crack lengths within a specimen increases. If the error is large (e.g., >10%), post-test correction of crack lengths, crack growth rates, and K values should be performed. The standard method of correction is to perform a linear correction of crack length over the entire test (post-fatigue precracking), particularly when TG growth from cyclic test segments and (mostly) IG growth from CL/CK test segments cannot be clearly identiﬁed on the fracture surface. This correction factor is then applied to the CGR and K values.

10

A.R. Jenks et al.

Additional Criteria Additional criteria that were considered during the scoring process include the following: • Key details were documented, including crack length versus time plots and fractography. • K was appropriate for the specimen size and was sufﬁciently constant during CL/CK test segments. • Comments were provided by the investigator regarding unusual test conditions.

Scoring Results After scoring, the average score for each test segment was calculated and a threshold average value for the Test Segment Credibility score was selected. Any data with an average score of this threshold value or better was included in the statistical modeling process, while any data with a worse average score was excluded. A threshold value of 3 was selected. This value has historical precedent— during the IASCC effort [6], it was observed that the data scatter relative to the ﬁnal model predictions increased signiﬁcantly for data that had received scores worse than 3. The scoring range in the current effort was 1 (best) to 5 (worst), so data with scores of 1–3 (inclusive) were kept in the database, while scores of 3 (exclusive) to 5 were removed. The scoring process for Alloy 690 resulted in approximately 6% of the original database being removed due to low average scores, resulting in a ﬁnal database consisting of 499 test segments. Tables 1 and 2 show some of the key attributes for the scored-in Alloy 690 data, and Fig. 1 plots the as-reported data (i.e., adjusted for post-test correction but unadjusted for any parameters, such as temperature, dissolved hydrogen, cold work, etc., thought to affect CGRs). Figure 2 plots the average Test Segment Credibility score as a function of as-reported (unadjusted) CGR for the Alloy 690 data. This plot shows that higher CGRs generally received excellent scores, but that below *1 10−10 m/s, there was limited correlation

Table 1 Distribution of scored-in Alloy 690 CGR data points, specimens, and heats by testing laboratory AMEC

ANL

Bettis

CIEMAT

GE-GRC

PNNL

Tohoku

Totala

No. Points 16 20 95 27 135 191 15 499 No. Specimens 11 9 95 19 58 36 15 243 No. Heats 4 3 2 2 15 12 7 30 a Total numbers of points and specimens are a sum of the respective values for each of the testing laboratories. Some heats were tested by multiple laboratories, however, so the total number of heats is less than the sum of those tested at each laboratory

Scoring Process for Evaluating Laboratory PWSCC …

11

Table 2 Range of several key parameters for scored-in Alloy 690 CGR data Parameter

Minimum value

Maximum value

K T [H2] DECPNi/NiO CW t

11.2 290 1 −88.3 0 40

78 360 81 45.1 41 14,680

Units p MPa m °C cc/kg mV % hr

1E-08

Crack Growth Rate (m/s)

1E-09

MRP-55 curve (75th percentile)

1E-10

1E-11

1E-12 Unadjusted Alloy 690 data. Data at 1E-13 m/s were reported as "no growth."

1E-13 0

10

20

30

40

50

60

70

80

90

Stress Intensity Factor (MPa√m)

Fig. 1 Scored-in Alloy 690 database: all data, unadjusted for any parameters (e.g., temperature, dissolved hydrogen, cold work level, etc.); shown with the MRP–55 75th percentile curve for as-received Alloy 600

between CGRs and scores, as was desirable. That is, the use of expert judgement was intended to ensure that low susceptibility materials did not selectively receive poor scores. In both Figs. 1 and 2, the “no-growth” data were assigned a very low CGR of 1 10−13 m/s. For Alloy 52/152, approximately 4% of the original database was scored out, leaving 132 test segments available for model development. Tables 3 and 4 show some of the key attributes for the scored-in Alloy 52/152 data, and Fig. 3 plots the as-reported data (i.e., adjusted for post-test correction but unadjusted for any parameters, such as temperature or dissolved hydrogen, thought to affect CGRs). (No weld metal was cold worked.) Figure 4 plots the average Segment Credibility score as a function of as-reported (unadjusted) CGR for the Alloy 52/152 data.

12

A.R. Jenks et al. 5.0 Scored-In Data Scored-Out Data

Average Test Segment Credibility Score

4.5

4.0

3.5

3.0

2.5

2.0

1.5

1.0 1.E-13

1.E-12

1.E-11 1.E-10 Reported Crack Growth Rate (m/s)

1.E-09

1.E-08

Fig. 2 Comparison between average test segment credibility score and reported (unadjusted) CGR for Alloy 690

Table 3 Distribution of Alloy 52/152 CGR data points, specimens, and welds by testing laboratory ANL

CIEMAT

GE-GRC

PNNL

Totala

No. Points 42 9 38 43 132 No. Specimens 9 8 16 15 48 No. Welds 4 2 12 7 18 a Total numbers of points and specimens are a sum of the respective values for each of the testing laboratories. Some welds were tested by multiple laboratories, however, so the total number of welds is less than the sum of those tested at each laboratory

Table 4 Range of several key parameters for Alloy 52/152 CGR data Parameter

Minimum value

Maximum value

K T [H2] DECPNi/NiO t

25 300 13 −0.5 51

51.7 360 81 31.6 4481

Units p MPa m °C cc/kg mV hr

Similar to the Alloy 690 scored-out data, the Alloy 52/152 scored-out data showed no correlation between CGR and score. In both Figs. 3 and 4, the “no-growth” data were assigned a very low CGR of 1 10−13 m/s.

Scoring Process for Evaluating Laboratory PWSCC …

13

1E-08

Crack Growth Rate (m/s)

1E-09

MRP-115 curve (75th percentile)

1E-10

1E-11

1E-12 Unadjusted Alloy 52/152 data. Data at 1E-13 m/s were reported as "no growth."

1E-13 0

10

20

30

40

50

60

70

80

90

Stress Intensity Factor (MPa√m)

Fig. 3 Scored-in Alloy 52/152 database: all data, unadjusted for any parameters (e.g., temperature, dissolved hydrogen, etc.); shown with the MRP–115 75th percentile curve for Alloy 182

Conclusions The scoring of the Alloy 690 and 52/152 CGR data was a long and complex process that involved extensive discussions to ensure that the data were evaluated based on all relevant considerations. Numerous criteria were identiﬁed by the experts for scoring the data, including experimental techniques, overall data quality, and testing credibility. Despite the use of expert judgment for these evaluations, there was good agreement among almost all of the experts’ scores. The fact that few data were scored out of the modeling databases for both Alloy 690 and Alloy 52/152 suggests that testing practices have improved over time (e.g., compared to the amount of data screened out during the MRP–55 and MRP–115 efforts), and it reflects the overall quality of the data, as well as some opportunities for improvements. The data that passed the scoring threshold will be used to develop the CGR models and disposition curves. Several factors will be included in the models, including those for temperature, dissolved hydrogen, yield strength, and crack growth orientation, the dependencies for which will be derived from the scored data. Acknowledgements The authors would like to acknowledge the members of the Data Evaluation Subgroup of the Expert Panel, which developed and applied the scoring procedure discussed in this paper, in addition to providing most of the CGR data. In particular, thanks goes out to Peter Andresen (GE-GRC, now retired), Bogdan Alexandreanu (ANL), Stuart Medway (Amec Foster Wheeler), F.J. Perosanz and Lola Gomez-Briceño (CIEMAT), Denise Paraventi (Bettis, now NNL), and Mychailo Toloczko (PNNL).

14

A.R. Jenks et al. 5.0 Scored-In Data Scored-Out Data

Average Test Segment Credibility Score

4.5

4.0

3.5

3.0

2.5

2.0

1.5

1.0 1.E-13

1.E-12

1.E-11 1.E-10 Reported Crack Growth Rate (m/s)

1.E-09

1.E-08

Fig. 4 Comparison between average test segment credibility score and reported (unadjusted) CGR for Alloy 52/152

References 1. Materials Reliability Program (MRP) Crack Growth Rates for Evaluating Primary Water Stress Corrosion Cracking (PWSCC) of Thick-Wall Alloy 600 Materials (MRP-55) Revision 1, EPRI, Palo Alto, CA: 2002. 1006695 2. G.A. White, J. Hickling, and L.K. Mathews, “Crack Growth Rates for Evaluating PWSCC of Thick-Wall Alloy 600 Material,” in Proceedings of the 11th International Conference of Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, 2003, pp. 166–179 3. Materials Reliability Program Crack Growth Rates for Evaluating Primary Water Stress Corrosion Cracking (PWSCC) of Alloy 82, 182, and 132 Welds (MRP-115), EPRI, Palo Alto, CA: 2004. 1006696 4. G.A. White, N.S. Nordmann, J. Hickling, and C.D. Harrington, “Development of Crack Growth Rate Disposition Curves for Primary Water Stress Corrosion Cracking (PWSCC) of Alloy 82, 182, and 132 Weldments,” in Proceedings of the 12th International Conference on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, 2005, pp. 511–531 5. Materials Reliability Program: Crack Growth Rates for PWSCC of Alloy 690 and Alloy 52, 152, and Variants Welds (MRP-386), EPRI, Palo Alto, CA: 2017. 3002010756 6. Models of Irradiation-Assisted Stress Corrosion Cracking of Austenitic Stainless Steels in Light Water Reactor Environments: Volume 1: Disposition Curves Development; Volume 2: Disposition Curves Application, EPRI, Palo Alto, CA: 2014. 3002003103

Applicability of Alloy 690/52/152 Crack Growth Testing Conditions to Plant Components Warren Bamford, Steve Fyﬁtch, Raj Pathania and Paul Crooker

Abstract A great deal of PWSCC testing has been conducted on a range of A690 materials. An expert panel organized by EPRI is working to collect all the available laboratory data for its applicability to actual plant components. This paper will review the considerations that are being used to perform this evaluation, and the results of that evaluation. One of the key variables is cold work, and detailed studies have been conducted to measure the residual strains, so as to determine the amount of cold work that can exist in a heat-affected-zone (HAZ) region of a typical weld. In addition to these measurements on weldments, limitations on bulk cold work on base metal imposed by all vendors will be reviewed. The effect of the orientation of the specimens tested will be compared with the orientations of typical flaws in operating plants, so as to determine the relevance of the data to those plants.

Keywords Alloy 690 Alloy 152 Alloy 52 PWSCC Cold work Orientation

Introduction The effects of cold work are signiﬁcant on the stress corrosion cracking (SCC) of austenitic alloys such as stainless steels and nickel-base alloys. The primary water stress corrosion cracking (PWSCC) behavior of Alloy 690 and its associated welds is therefore expected to follow the same pattern. A good example of these effects was summarized by Bruemmer [1] in 2015, as shown in Fig. 1. There is a relatively small effect until a cold work level of about 15%, and then the effect takes off, producing very striking effects at cold work levels approaching 40%. W. Bamford (&) Westinghouse, Pittsburgh, PA, USA e-mail: [email protected] S. Fyﬁtch AREVA Inc., Lynchburg, VA, USA R. Pathania P. Crooker EPRI, Palo Alto, CA, USA © The Minerals, Metals & Materials Society 2019 J.H. Jackson et al. (eds.), Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, The Minerals, Metals & Materials Series, https://doi.org/10.1007/978-3-030-04639-2_2

15

16

W. Bamford et al.

The purpose of this paper is to discuss the basis for these effects, and determine the levels of cold work that exist in typical plant components. In this manner, the range of cold work that should be included in the ﬁnal recommended model and disposition curves for PWSCC of these materials can be determined. The effects have been separated into base metal effects, including the HAZ, and weld metal effects. In addressing this issue, it is essential to separate the effects of bulk cold work from those of surface cold work. There are several examples of surface cold work in operating plants: • CRDM and other nozzles with near surface cold worked layers from gun drilling or other forms of aggressive machining that are then deformed into an oval shape during welding to the upper head • Steam Generator divider plates with compressive layers induced by planing and/or grinding that are pulled into tension during the initial hydrotest • Dissimilar metal welds of Alloy 52/152 that are typically ground to prepare the surface of the weld for non-destructive examination.

Fig. 1 Cold Work effects on measured SCC growth rates at K = 30–32 MPa-rt-m, for Alloy 690 Materials in the as-received MA or TT condition [1]

Applicability of Alloy 690/52/152 Crack Growth …

17

Note that these examples do not involve cold working of the entire cross section, which is either prohibited or tightly controlled to limit these deformations, as will be discussed below. The laboratory PWSCC data being collected on specimens with bulk cold work are not particularly relevant to the above applications, since the surface layer present in plant applications is very thin, and the effect of cold work there is very limited. Since these increased strains are highly localized at the surface, any crack that initiates will not propagate far unless there is serious susceptibility to SCC of the substrate material. This is the case for Alloy 600 in PWR primary water but not for materials like Alloy 690.

Base Metal Effects All nuclear equipment vendors have processes in place to limit the cold work present in operating plant components. Such limitations are standard practice, and the ASME Boiler and Pressure Vessel Code also contains requirements to limit the amount of cold work in components. Some of these limitations will be reviewed below. One of the key examples of limitations on cold work is for CRDM head adapters. There have been historical issues with maintaining perpendicularity after J-welding the CRDM adapter into the vessel heads. The paragraph below identiﬁes the applicable ASME code sections for maintaining cold work of Alloy 690 CRDM components to a maximum of 5% strain. These controls have been in place since the initial usage of Alloy 690 CRDM components (about 1993). ASME Code Requirements for Cold Straightening. ASME Section III Division I Subsection NH [2] provides the following guidance regarding cold forming of reactor components subsequent to solution anneal and thermal treatment: NH-4212 Effects of Forming and Bending Processes The rules of this paragraph shall supplement those of NB-4212 and NB-4213. Any process may be used to form or bend pressure retaining materials, including weld metal, provided that the requirements of the subparagraphs below are met. (a) Post fabrication heat treatment [in accordance with (b) below] of materials which have been formed during fabrication, shall be required unless one of the following conditions are met. (1) Maximum fabrication induced local strains1 do not exceed 5%2, regardless of the service temperature.

Strain is deﬁned as the maximum local ﬁber elongation or contraction per unit length; and where more than one strain increment occurs (e.g., biaxiality or reversed bending), it shall be the sum of the absolute values of all the strain increments. 2 Strain resulting from ﬁnal straightening operations performed on materials furnished in the solution annealed or heat treated condition need not be included in the computation of strain. 1

18

W. Bamford et al.

(2) Written technical justiﬁcation shall be provided in the Design Report for not performing heat treatment, subsequent to straining, or for the use of an alternate heat treatment procedure, to that speciﬁed in 2. below for fabrication induced strains greater than 5%. The commentary in footnote 2 above regarding straightening would apply to the as-received Alloy 690 tube, and not to the ﬁnal as-welded component. As such, manufacturers would need to limit maximum bending strain to 5%. The strain requirement should be calculated based on the actual post-weld distortion values, but is not expected to approach this 5% limit. In addition, material speciﬁcations that govern the use of Alloy 690 materials generally refer to ASME B&PV Code Section II material speciﬁcations (e.g., SB-166 for rod and bar or SB-167 for seamless pipe and tube) with the requirement that material is to be furnished in the solution-annealed condition. These speciﬁcations also typically require supplemental thermal treatments for Alloy 690 materials that are well-known to produce what has generally been considered to be an optimal microstructure (i.e., controlled grain size and carbide precipitation) for Alloy 600 PWSCC resistance [3], and vendors typically require suppliers to provide micrographs so that the grain structure can be evaluated. Procedures to limit residual weld stresses at the surface of attachment welds (e.g., J-groove welds for CRDM nozzles) and to limit residual stresses resulting from mechanical loads associated with the ﬁt-up of the component item have also been developed. Cold straightening of Alloy 690 materials is generally prohibited. Cold bending operations performed on nickel-based materials are required to be followed by a stress relieving heat treatment. Repairs performed on nickel-based materials are generally followed by surface conditioning to remove residual surface tensile stresses resulting from the repair. MHI material speciﬁcations for applications in the United States are generally in accordance with the ASME B&PV Code Section II (e.g., SB-166 for rod and bar or SB-167 for seamless pipe and tube). With these material speciﬁcations, MHI controls the chemical composition and the ﬁnal annealing and thermal treatment temperatures for obtaining the required mechanical properties and optimum microstructure for PWSCC resistance. Following the thermal treatment, no further heat treatments (at temperatures greater than the thermal treatment temperature) are permitted. No speciﬁc bulk cold work limitations are speciﬁed; however, a maximum hardness (220 HV) is speciﬁed, which is in accordance with EPRI “Alloy 690 Procurement Speciﬁcation 1026723” [4]. The Swedish requirements for Alloy 690 component items are collected into a document called TBM (Technical Regulations for Mechanical Equipment). In this document it is stipulated that cold work levels exceeding 3% are not acceptable, and that, as a general condition, iron- and nickel-based austenitic materials shall be delivered in the solution annealed condition. The basis for these strict requirements is from operating experience with stress corrosion cracking in stainless steel tubing and bends during the initial years of operation of the BWR fleet. These limitations

Applicability of Alloy 690/52/152 Crack Growth …

19

have been transferred to the PWR fleet. During manufacturing of replacement reactor vessel heads, these requirements have been stipulated for all relevant components even if the manufacturer’s original scope already required adherence to the restrictions associated with ASME Code Article NH 4212 [2].

Crack Growth Orientation Versus Product Form Figure 2 shows the nomenclature for CT specimen orientations with respect to a cylindrical forged product form, such as a CRDM nozzle. The orientation of an in-service crack propagating through-wall can be represented by the C-R or C-L orientations for an axial crack and by the L-R or L-C orientations for a circumferential crack. The R-L and R-C orientations are parallel to the pressurized surface, often termed the laminar direction; flaws in these orientations have never been observed in service. This is because there is little or no stress in the radial direction of a pressure vessel. Radial stresses must be in equilibrium with the stress at the surface, and that stress is either zero or compressive and equal to the internal pressure, for a pressurized cylinder. Tests conducted with flaws in the R-L or R-C orientation are therefore not applicable for service conditions. Figure 3 shows the nomenclature for CT specimen orientations with respect to a plate or weld product form. The orientation of an in-service crack propagating through-wall can be represented by the T-S or T-L orientations for an axial crack and by L-S or L-T for a transverse crack. The S-L and S-T orientations correspond to the laminar direction, in which no cracks have been observed in service. Again, tests conducted with flaws in the S-L or S-T orientation are not useful for applications. For welds, the product form orientation is typically more straightforward due to the formation of similarly oriented dendrites. Weld metal forms by solidiﬁcation from a molten state, which leads to the formation of dendrites growing in the direction of the heat flow, i.e., perpendicular to the solid material on which the weld is deposited. Most welds are made with multiple passes. The grain structure of dendrites in subsequent passes is normally related to that of previous passes as a result of epitaxy, i.e., by the tendency of a crystal forming on a substrate to have the same structural orientation as the substrate. This results in the dendrites persisting through some or many weld passes. The dendrites tend to be perpendicular to the base material at the weld-base material interface, and tend to become vertical (root to crown direction) as the weld thickness increases. The dendrites are mainly vertical in the central region of the weld, or, in other words, parallel to the faces being joined, and perpendicular to the pressure boundary. The orientation of the crack in the weld, i.e., relative to the weld’s columnar microstructure, has a strong influence on the CGR. Thus, it is necessary to include

20

Fig. 2 Specimen orientations for a cylindrical product form

Fig. 3 Specimen orientations for a plate or weld product form

W. Bamford et al.

Applicability of Alloy 690/52/152 Crack Growth …

21

the relative crack orientation (i.e., parallel or perpendicular to the direction of the weld dendrites) in the development of a CGR model. Cracks grow fastest along high energy grain boundaries in the direction of grain growth, and next fastest along high energy grain boundaries perpendicular to the direction of grain growth, i.e., parallel to the welding direction. Cracks that grow perpendicular to the high energy grain boundaries, i.e., perpendicular to the columnar dendrites, grow signiﬁcantly slower. Cracks in both orientations have been seen in service, but it is conservative to use crack growth results obtained with flaws oriented along the dendrites, and this orientation has been almost exclusively used for testing. Using the terminology above, typical weld crack orientations have been in the T-S or T-L orientations.

Measurement of Welding Residual Strains in Prototypic Welds GE Global Research has completed a very large number of measurements of the residual plastic strain associated with structural weldments in general, and Alloy 690 weldments, in particular, and these results will be summarized below, taken from the description provided in [5]. The weld HAZ was examined using a recently developed electron back scattered pattern (EBSP) technique that quantiﬁes strain by measuring the amount of intra-grain misorientation [4]. The samples for EBSP examination were slices perpendicular to the fusion line that were ground and given a ﬁnal polish using a vibratory polisher with 0.05 micron colloidal silica to produce a surface with minimal preparation-induced surface strain. To calibrate the EBSP strain measurements, as-received Alloy 690 material was solution annealed at 1050 °C/0.5 h, slow cooled, machined into tensile specimens and then stress relieved at 500 ° C/100 h and slow cooled. These tensile specimens were strained to 1, 2, 5, 10, 15 and 20% strain at room temperature. These specimens were also ground and polished using colloidal silica. All samples were examined using a CamScan CS44 Scanning Electron Microscope (SEM). Back-scattered Electron Images (BEI) were obtained at normal beam incidence with 30 kV electrons at a 20 degree angle of incidence to the sample surface. Patterns were obtained in 1 micron steps from the fusion line out to 26 mm. Ten scans were obtained, each 220 microns apart. The high angle boundaries were identiﬁed (>10° misorientation) and each location within a grain was indexed and compared to all the other locations within the grain to obtain the average misorientation for that grain. Results of Residual Plastic Strain Measurements in Alloy 690/52 and Alloy 690/152 Weld Mockups. EPRI-MRP sponsored an extensive study of residual plastic strain measurements by electron back scattered diffraction (EBSD) technique in Alloy 690/52 and Alloy 690/152 weld mock-ups by Martin Morra and

22

W. Bamford et al.

co-workers at GE Global Research [4]. The mock-ups were prepared from commercial Alloy 690 plate material welded with Alloy 52 or 152 ﬁller metal by various organizations [EPRI, Knolls Atomic Power Laboratory (KAPL), ENSA (Spain) and Mitsubishi Heavy Industries (MHI)]. In addition, a CRDM J-groove weld produced by Vallinox was also examined. The results of this work have been reviewed at the EPRI PWSCC Research Collaboration Meetings in Tampa from 2011–15. The objectives of this work were to characterize the composition, microstructure, and plastic strains in the weld metal, heat-affected-zone (HAZ) and base metal. Many Alloy 690 crack growth specimens in the Alloy 690 database were prepared from cold worked plate material. The Applications Group of the Expert Panel has used the measured plastic strains in the base metal and HAZ of weld mockups to select the appropriate amount of cold work that is applicable to plant components. Figure 4 (Fig. 16 from [4]) shows an example of the plastic strain proﬁle in Alloy 52 weld metal, weld dilution zone, partially melted and un-melted heat affected zone and base metal. Results from more than a dozen weld mockups show that the measured plastic strains in the mid-thickness locations are in the range of: • 1–5% in Alloy 690 base metal • 4–14% in Alloy 690 HAZ • 8–36% in the 52 and 152 weld metals

Fig. 4 Illustration of Morra’s results for residual strains [5]

Applicability of Alloy 690/52/152 Crack Growth …

23

Fig. 5 Cumulative distribution of HAZ strains in weld mockups

Figure 5 shows the cumulative distributions of plastic strains in the HAZ at three different locations and the overall cumulative distribution plot. The strains decrease through the thickness from ID (weld root) to OD (weld crown). Statistical treatment of these results shows that the 50th, 75th, and 95th percentile values of overall plastic strains in the HAZ are 8.4, 10.6, and 17.2% respectively. These results suggest that in order to simulate the amount of plastic strain expected in the HAZ of welds, the amount of additional cold work applied to the Alloy 690 base metal CGR specimens should be limited to about 10%, although a speciﬁc number has not yet been ﬁnalized.

Weld Metal Effects and HAZ During production of various types of welds during component and structure manufacturing, it is a common practice in some areas to grind the surfaces for reasons such as inspectability and minimization of local flow disturbances. The grinding procedure will likely induce local increases of tensile stresses and strains on the surfaces of the areas (i.e. cold work). This grinding process, and the resulting local cold work, only has an impact on the initiation process, due to the limited thickness of the cold worked layer. Crack propagation is a function of the SCC resistance of the underlying material, which is not cold worked in this situation. However, the welding process itself induces residual strains by nature through the entire volume of the material in the vicinity of 10–20% residual plastic strain in

24

W. Bamford et al.

the axial direction of the weld. The amount depends on the manufacturing technology and process as well as direction in the weld material. The additional plastic strain on the surface may affect the initiation behavior of a flaw. However the effect on the continuous process of PWSCC growth does not need to be separately considered, because the majority of the tests are performed on as-welded material.

Cold Work Versus Warm Work The mechanisms of damage are different for the bulk cold work imposed on the test specimens used for SCC testing compared to the warm working that results from the welding process. Cold work is produced by deformation that occurs at temperatures below 0.5 TM, while warm work can be deﬁned as a process that occurs at temperatures above *0.5 TM. [Note that TM is the homologous melting temperature.] The resulting hardened material from cold work contains randomly arrayed tangles of dislocations that resist further dislocation motion, while the dislocation structure from warm work is actually ‘recovered’, meaning it has a reduced number of dislocations in thermally rearranged arrays. Both cold and warm work are produced by transgranular shear, but cold work entails no grain boundary sliding, as opposed to warm work, for which such deformation has other potential creep-like effects, including grain boundary sliding and dislocation climb. These can relieve elastic mismatch across grain boundaries, and result in reduced increments of grain boundary stress. Cold work has no grain boundary motion and creep, so it results in a higher increment of local grain boundary stresses. Because of the different structures produced, warm work is not as effective at changing some properties as cold work—e.g. a greater level of warm work is needed to induce the same level of microstructural damage (dislocation entanglement) as cold work. Therefore some properties may be very much more affected by cold work than by warm work. In other words, for the same change in the property, a much lower level of cold work would be needed than for warm work. This leads to the conclusion that the effect of welding strains on crack growth will likely be lower than the equivalent effects of strain induced by cold work on such behavior. Also, it is important to mention that much of the effect associated with residual strains that exist in a weldment are captured within a specimen, even after it is machined from the weldment; therefore no additional effects of cold work need to be addressed for these situations.

Applicability of Alloy 690/52/152 Crack Growth …

25

Summary and Conclusions In this paper, the available experimental crack growth results have been summarized as a function of cold work, using for example the data of Bruemmer [1]. These results show that cold work can signiﬁcantly increase PWSCC rates in Alloy 690, for bulk cold work levels above about 15%. The levels of cold work that exist in nuclear components are limited by material speciﬁcations as well as ASME Boiler and Pressure Vessel Code requirements, to a bulk cold work level of about ﬁve percent. The question of the level of work that exists in Alloy 52 and 152 weldments was addressed by compiling an extensive series of measurements by Morra [6], which leads to a statistical prediction that the maximum level of plastic strain (a measure of cold work) is about 10% in the HAZ adjacent to these welds. For the base metal, the use of crack growth rate data from as-received materials is recommended, with added bulk cold work not to exceed 10%. In addition, crack growth rate data from tests of HAZ specimens are directly applicable to plant components, and are suitable for development of disposition curves with no corrections for cold work effects. Crack growth disposition curves for Alloys 52, and 152welds should be based on CGR tests conducted on weld specimens with no additional cold work because these specimens retain the plastic strain produced during welding. Acknowledgements Valuable contributions to this work were obtained from Martin Morra and Peter Andresen of GE Global Research, Mike Burke and Dave Love of Westinghouse, and Peter Scott. The authors also wish to acknowledge the contributions of other members of the Alloy 690 Applications Expert Panel: Amanda Jenks, Pal Efsing, Dave Morton, and Toshio Yonezawa. This work was sponsored by the Electric Power Research Institute.

References 1. S.M. Bruemmer et al, Cold-work effects on stress corrosion crack growth in alloy 690 tubing and plate materials, in Proceedings, 17th Conference on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, August 2015 2. ASME Boiler and Pressure Vessel Code, Section III, Rules for Construction of Nuclear Power Plants, ASME International, New York, NY 3. T.M. Angeliu et al, Intergranular stress corrosion cracking of unsensitized stainless steels in BWR environments, in Proceedings, 9th Conference on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, August 1999 4. Alloy 690 procurement speciﬁcation, Rev. 1, Electric Power Research Institute, Palo Alto, CA, report 1026723, 2012 5. P.L. Andresen, M.M. Morra, K. Ahluwalia, SCC of alloy 690 and its weld metals, Corrosion 2012, Paper C2012-0001187, NACE International, 2012 6. Martin M. Mora, M. Othon, Weld Slice DDC-52-LAS-2: 52 M Repair/52 M Weld Interfaces, Alloy 690/52/152 PWSCC Research Collaboration Meeting (Tampa, FL, 2016)

SCC of Alloy 152/52 Welds Defects, Repairs and Dilution Zones in PWR Water Peter L. Andresen, Martin M. Morra and Kawaljit Ahluwalia

Abstract Extensive SCC growth rate measurements have been performed on Alloy 690 and its weld metals in the past, and this paper focuses on SCC growth rate evaluation of Alloy 52/152 welds with a variety of defects and/or weld repairs, and in the dilution zone. Ductility dip cracking dominated the weld defects, and weld repair mockups were fabricated by EPRI Charlotte to be 20% or 50% excavation and repair, as well as welds with a refuse pass every layer. Only low and very low SCC growth rates were observed in all cases. Studies on weld dilution zone effects of varying Cr content were evaluated using welds created with variable ratios of dual-ﬁller-wire feel, which permits deﬁnitive SCC growth rate measurements in a homogenous weld without the ambiguity of having the crack front in undeﬁned composition of an actual weld dilution zone.

Keywords Alloy 52 Alloy 152 Alloy 52i Weld metal Weld defects Weld repairs Stress corrosion cracking Crack growth rate High temperature water

Introduction Stress corrosion cracking (SCC) has been observed in many light water reactors structural components [1–10], with the rate of SCC initiation and crack growth vary widely. Newer materials have been adopted in the last 2+ decades, primarily Alloy 690 (UNS N06690) and its weld metals, Alloy 52 (UNS W86052) and Alloy 152 (UNS W86152). Before careful testing was performed, these materials were widely viewed as immune to SCC, but it is now recognized that these materials exhibit at P.L. Andresen (&) Andresen Cons., 12204 Wildwood Park Pl, Bakersﬁeld, CA 93311, USA e-mail: [email protected] M.M. Morra GE Research, One Research Circle CE 2545, Schenectady, NY 12309, USA K. Ahluwalia EPRI, 3420 Hillview Ave, Palo Alto, CA 94304, USA © The Minerals, Metals & Materials Society 2019 J.H. Jackson et al. (eds.), Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, The Minerals, Metals & Materials Series, https://doi.org/10.1007/978-3-030-04639-2_3

27

28

P.L. Andresen et al.

least some limited susceptibility to SCC [1, 2, 8–30]. Because historical testing was often relatively simple and insensitive, combinations of materials, environments, and stressing conditions—such as unsensitized stainless steel and Alloy 690 in both pressurized water reactor (PWR) and boiling water reactor (BWR) water—once generally viewed as immune to SCC have now been shown to be susceptible [1, 2, 5–7]. The higher Cr content in Alloy 690 and its weld metals create greater challenges in melting and processing, and in welds, solidiﬁcation causes partitioning of some elements to dendrite boundaries, and there is a greater propensity for various weld defects to occur. About 14 years ago, a crack growth rate program [8–18] was initiated to determine if Alloy 690 could exhibit any susceptibility to SCC in representative PWR primary water under constant stress intensity factor (K) conditions. Key elements to this effort were excellent crack length resolution (*1 µm on a 1T CT specimen) and careful transitioning from the fatigue precrack so that the crack behaved as if it had always grown as an SCC crack [11–15, 18–22]. The data accumulated since that time by many labs [8–30] has shown that growth rates in some instances (primarily cold worked materials tested in the S-L orientation relative to the plane of cold work) can be very high (e.g., >1 10−6 mm/s) in PWR primary water and at moderate stress intensity factors (Fig. 1). Component lifetime is often viewed as being controlled by SCC initiation resistance, but there are dozens of materials and thousands of cases where the expected resistance to initiation was not realized. There is intrinsic optimism in the ASME Pressure Vessel and Boiler Codes [31], which simply mandate that SCC immunity exists, but typical 1–2 year evaluations in simple laboratory tests is not a

Fig. 1 Range of crack growth rate data in Alloy 690 in PWR primary water [19, 21]. The MRP-55 reference curve for Alloy 600 is shown for 325 °C

SCC of Alloy 152/52 Welds Defects, Repairs …

29

strong basis for conﬁdence in a 60+ year lifetime. The impetus behind the Codes was >50,000 deaths in one year from steam boiler explosions in the U.S. alone [32] (most caused by SCC), but 175 years later, the Code only addresses fracture and fatigue. To provide greater conﬁdence in a 60–100 year lifetime, both SCC initiation and growth should be optimized, and indeed both are being pursued for Alloy 690 and it is weld metals. To help deﬁne an adequately low growth rate, crack advance as a function of the residual stress proﬁle in a butt weld was calculated for materials of different susceptibility. The resulting crack trajectories are shown in Fig. 2, and the trajectory that p provides *80–100 year life was chosen, then the growth rate at a *22 MPa m m was identiﬁed for that case, which is *4 10−9 mm/s. A low growth rate has been deﬁned as below *5 10−9 mm/s, a medium growth rate as up to 5 10−8 mm/s, high as up to 3 10−7 mm/s, and very high above that value. Weld residual stresses and strains are the origins of most of the SCC in LWR structural materials. In stainless steels [33–35], more of the weld residual strain develops in the heat affected zone (HAZ) than in Alloy 690 (Fig. 3) [36–41]. These strains are accurately quantiﬁed using electron back-scattered diffraction (EBSD), which is a very high resolution and precise technique for evaluating the misorientation that characterizes dislocation density from residual deformation [37–40]. Remarkably, the weld residual strains in Alloy 52/152 weld metal can be very high in isolated dendrites (Fig. 3), exceeding 40% and even 50% in isolated dendrites.

Fig. 2 Example of a crack growth calculation in which a crack is grown from a 50 lm defect in a butt weld. Calculations are performed for materials of differentpSCC susceptibility. The growth rate is noted when the stress intensity factor achieves 22 MPa m for the curve that represents a *100 year life

30

P.L. Andresen et al.

Weld metal

690 composition

Fig. 3 (Top) Characterization of a steam generator divider plate mock-up weld of Alloy 52M showing that the base metal had *8% residual strain (from incomplete annealing), which increased to 12–14% in the base metal microstructure and 17–22% in the 690 composition (but different microstructure) of the partially melted zone [36]. (Bottom left) EBSD strain map showing highly localized strains in isolated dendrites. (Bottom right) Residual strains in V-groove Alloy 52 welds, heat NX4467JK [36]. Unlike stainless steel welds [37–40], which have higher strain in the HAZ (often 20–30%) than in the weld metal, the 152/52 weld metal strains increase to much higher levels in local areas

Into the partially melted zone and HAZ, the weld residual strains drop below *20%. However, there has been little evaluation of the effect of welding defects on SCC. This paper reports the progress in an on-going research program on the SCC growth rate of Alloy 52, 152 and 52i weld metals.

Experimental Procedures The experimental systems and DC potential drop measurement of crack growth used in these studies is described elsewhere [6, 8–10, 42, 43]. 0.5T compact type specimens were used, with 5% hemispherical side grooves on each side. Servo loading systems under computer control allowed testing under constant Kmax (load shedding) conditions. The water chemistry was refreshed at an average rate of *3–4 volume exchanges per hour, with stable, measured ionic and dissolved gas chemistry. Low impurity levels were achieved by recirculating the water through a full-flow mixed-bed demineralizer equilibrated to 600 ppm B as boric acid and 1 ppm Li as LiOH. Apart from a few tests to study the effect of dissolved hydrogen,

SCC of Alloy 152/52 Welds Defects, Repairs …

31

it was maintained at 26 cc/kg by bubbling pure hydrogen at 6.9 psig. This corresponds to the Ni/NiO phase boundary in 360 °C water, which is where a peak in crack growth rate is observed in nickel alloys [44–46]. The corrosion potentials of the CT specimen and a Pt coupon were measured using a zirconia membrane reference electrode [47]. The weld wires were mostly commercial heats, although some were custom heats of 25 and 27% Cr materials. Some welds were archive/qualiﬁcation welds from PWR head replacements, and most welds were fabricated by nuclear-qualiﬁed vendors. Table 1 summarizes the weld metals evaluated for SCC response. Alloy 52 welds are generally produced by automated gas tungsten arc welding (GTAW, or tungsten inert gas, TIG) and gas metal arc welding (GMAW, or metal inert gas, MIG). Alloy 152 weld are produced by manual shielded metal arc welding (SMAW). DC potential drop provides a very high resolution in situ monitoring of crack length versus time, but poses a number of usage and interpretation challenges. The resistivity of high nickel alloys changes with time in the vicinity of *300–360 °C, and without compensation this appears to be crack advance. Cold work extends this temperature regime of concern downward by perhaps 30 °C, and the extent of the concern is proportional to the crack growth rate being measured. That is, for materials that exhibit a growth rate above 10−7 mm/s, only large shifts in resistivity are of signiﬁcant concern. But to measure growth rates below 10−8 mm/s, and especially below 10−9 mm/s, even minor changes in resistivity become important. The resistivity shift approaches saturation versus time much more rapidly at higher temperature, and the initial slope is much higher at high temperature, but the duration of the effect is much longer at lower temperature. Compensation for resistivity drift in these weld metals is crucial to accurately determine their crack growth rate, and it should be recognized that the inherent inhomogeneity in weld metals may lead to differences in local resistivity shift that translate to small over- or under- compensation. At 10−9 mm/s, 2.6 lm of growth requires one month, and this is difﬁcult to conﬁrm fractographically, doubly so in weld metal with large dendrites aligned in the direction of crack advance (the T-S orientation). Crack growth rates below about 5 10−9 mm/s must be viewed with less conﬁdence than higher growth rates, and even if the growth rate were exactly correct, the conﬁdence in the overall behavior is lower because very little microstructure is sampled. The large grain size is also a challenge to testing, as the intent is to transition from transgranular (TG) fatigue to SCC so that the crack behaves as if it had developed as an SCC crack (from a cyclically hardened to a monotonic plastic zone), but under loading conditions that will advance the crack gently enough so that it can follow an IG path if it is preferred. This may require weeks or months in large grained materials, and there may essentially no preference for an IG path in highly resistant materials. Thus, many variations in transitioning technique have been attempted, including lower load ratio (R) values, longer hold times, asymmetric waveforms (rapid drop, slow rise in load), etc. There are no recipes that guarantee successful and complete SCC transitioning, and the underlying philosophy is to transition as gently as possible under conditions where the crack is advancing and able to encounter grain

32

P.L. Andresen et al.

Fig. 4 Example of the extensive transitioning sequences used to promote the shift from the TG fatigue precrack to an IG SCC crack. IG Transitioning requires opportunity, which is related to sufﬁcient growth under sufﬁciently gentle loading conditions so that grain boundaries can be encountered and a transition made. Note that even for once-per-day periodic partial unloading, the growth rate can be biased upward by 3 or more in SCC resistant materials

boundaries (Fig. 4). Faster rise times and higher DK promote TG cracking, and a cyclically hardened plastic zone. Active, in situ crack monitoring is crucial. Precracking and transitioning at the same Kmax is helpful, and there is some beneﬁt of using asymmetric waveforms since the period of decreasing load doesn’t promote SCC. In some tests, little IG was achieved, and in most or all cases, some to extensive TG crack growth morphology was observed (Fig. 8), indicating that when the grain boundary path is sufﬁciently resistant, the IG crack path will not be consequentially favored, and extensive TG SCC will be observed. Engagement from the fatigue precrack and %IG are important considerations in evaluating crack growth rate, especially in weld metals. However, there are challenges in using and interpreting %IG because if TG growth does occur, one cannot simply take the growth (or growth rate) indicated by DC potential drop and divide it by the %IG based on the assumption TG growth is not occurring, or that the entire crack front has the same SCC susceptibility a few isolated IG areas. Corrections for engagement from the fatigue precrack were more critical for Alloy 182/82 weld metals, where very large differences in crack advance occurred because parts of the crack remained pinned at the fatigue precrack. This primarily occurred in tests where transitioning was not attempted or was done poorly, which generally produced a crack front where the maximum crack advance might be >2 mm, but much of the crack front, no crack advance had occurred from the fatigue precrack. DC potential drop is very strongly biased by uneven crack fronts (as is compliance), so that the indicated crack advance would appear to be small. In such cases, corrections for the trajectory of cracking throughout the test is a signiﬁcant problem, because the evolution of crack front unevenness versus time cannot be determine from post-test fractography. However, for Alloys 152/52, the extent of crack front unevenness is often small, and under these conditions DC potential drop represents something closer to an area average. Finally, some testing has been performed under a single, ﬁxed condition, and growth rates were determined from the post-test fractography of the engaged areas

SCC of Alloy 152/52 Welds Defects, Repairs …

33

and the total test time [48]. From a correction perspective, this is superior, especially if DC potential drop is used to provide evidence of early crack nucleation and stable growth (so the use of total test time to determine growth rate is appropriate). However, if the “single condition” involves cycling under periodic partial unloading conditions (as it does in [48]), there is very often a signiﬁcantly bias in the SCC growth rate data in Alloy 152/52. At high crack growth rates (such as with Alloy 182 weld metal), periodic partial unloading at R = 0.7, 0.001 Hz once every 8 or 24 h has a small effect (less than *1.5) on the measured growth rates. However, in very low growth rate materials, it usually has a large effect (*10). While estimates for the effect of periodic partial unloading can be made, they are not reliable or consistently accurate, and it is much better to use constant K (no cycling) data to determine SCC (constant K, no cycling) growth rates in resistant materials. There have been formulations for over 30 years for the effect of slow cycling and periodic partial unloading, but even in air they do a poor job at high load ratio (R), especially at low frequency and long hold time, where the data itself is variable. When hot water is added into the scenario, there are large variations in behavior that are observed, esp. at high R. Indeed, at R 0.9, analyses have shown that the crack growth data are on average lower than constant K data [49]. Thus, using a simple calculational approach to adjust cyclic crack growth data to determine the crack growth rate at constant load is unreliable. Additionally, extensive efforts to evaluate SCC have shown that the growth rate in some materials drops signiﬁcantly at constant load compared to slow cyclic loading or periodic partial unloading.

Results on Alloy 152/52/52i Weld Metals In all of the data generated in this program, very extensive efforts were made to transition from the TG fatigue precrack to an SCC crack, and complete engagement from the fatigue precrack was obtained in all cases. 100% IG morphology was never achieved—nor has it ever been in Alloy 152/52 in PWR primary water (e.g., Fig. 8), despite a broad variety of techniques and >315,000 of hours of effort. This must be viewed as a measure of the resistance of the grain boundaries to (IG or TG) crack advance in these weld metals. Additionally, in all cases two or three attempts were made to evaluate the SCC response in different microstructures. This was accomplished by advancing the crack in fatigue (typically by *1 mm) and re-transitioning to SCC. No great differences in response were observed, and the data presented reflect the highest growth rates observed for a given specimen. Graphs where the data have been correctly based on post-test fractography are marked “Corrected” at the upper left, although the use of anticipatory correction in our DC potential drop software made the post-test corrections a relatively small issue in most tests. An example of a typical IG transitioning sequence is shown in Fig. 4, where slow cycling at 0.001 Hz and R = 0.5–0.7 permitted a moderate amount of crack

34

P.L. Andresen et al.

advance. It was followed by the introduction of a 1000–3000 s hold time at Kmax, again for extended periods of time and moderate amounts of crack advance. Then changes in hold time were made to *9000 s hold; then to *27,300 s hold, then to *85,900 s hold, then to constant K (no cycling). Many variations were evaluated, including the use of asymmetric loading (e.g., 5% of the time for unloading, 95% for re-loading). The overall objective is to alter the crack tip conditions—from cyclically hardened to monotonic plastic zone—so that it behaves as if it has always grown as an SCC crack. It is designed to maximize the opportunity for the crack to become intergranular (IG) if indeed there is a consequentially greater susceptibility of the grain boundary over the transgranular (TG) path—as grain boundaries become increasingly resistant, the preference for the IG path becomes very small. The high Cr weld metals evaluated in this program at our laboratory (and most others) exhibited low growth rates (e.g., Figs. 5, 6, 7 and 8), as summarized in Fig. 9. The main exception is an artiﬁcial case where 20% cold work was applied after welding, and the observed SCC growth rates were high (Fig. 9). These data encompass a variety of weld metals and welding techniques, including 25–31% Cr weld metals, manual and automated welds, V-groove and narrow gap welds, etc. Figures 5 and 6 show the SCC response and fractography of specimens c499 (Alloy 52i) and c500 (Alloy 152i), welds provided by KAPL. The SCC growth rates were observed, *2 10−9 mm/s for c499 and *4 10−9 mm/s for c500,

Fig. 5 Crack length versus time and fractography for specimen c499 of as-welded Alloy 52i weld metal, heat 187775 in 360 °C PWR primary water

SCC of Alloy 152/52 Welds Defects, Repairs …

35

Fig. 6 Crack length versus time and fractography for specimen c500 of as-welded Alloy 152i weld metal, heat 187775 in 360 °C PWR primary water

Fig. 7 Crack length versus time and fractography for specimen c501 of as-welded Alloy 152 weld metal, heat 747136 in 360 °C PWR primary water

and many different attempts were made over the 6111 h test duration. There was extensive evidence of IG cracking in both cases, but this did not translate into higher growth rates than other specimens, some of which exhibited little IG morphology. The SCC response and fractography of specimen c541 (Alloy 152) are shown in Figs. 7 and 8. Only low SCC growth rates were observed, *8 10−10 mm/s

36

P.L. Andresen et al.

Fig. 8 Crack length versus time and fractography for specimen c541 of as-welded Alloy 152 weld metal, heat WC04F6 in 360 °C PWR primary water. A detailed evaluation of the fractography by PNNL is shown at the bottom [50]

Fig. 9 (Left) Summary of SCC growth rates obtained in this program on Alloy 52, 152 and 52i weld metals in 360 °C PWR primary water. (Right) Summary of SCC growth rates for Alloy 52, 152 and 52i weld metals in aggressive 288 °C water with 2 ppm O2 and 30–50 ppb sulfate or chloride

SCC of Alloy 152/52 Welds Defects, Repairs …

37

based on multiple attempts over the 12,264 h duration of the test. There was strong evidence of IG cracking, and the fractographic images were examined by Paciﬁc Northwest National Lab, with IG regions highlighted in yellow. The %IG was similar in tests on the same weld at GE, ANL and PNNL [50].

Weld Repairs Two weld repair mock-ups, both performed at EPRI Charlotte under the supervision of Steve McCracken, were evaluated for SCC response. Specimen c637 was a 50% excavation and repair weld using Alloy 152 M, heat 098934. The CT specimen was oriented to advance the crack near the edge of the repair weld and into the underlying original weld. Two attempts were made, each involving extensive transitioning (in time and crack increment) and both resulted in very low growth rates of *2 10−10 mm/s (Fig. 10), a rate that results in a crack increment over a 1300 h constant K period of only *1 µm. The second weld repair evaluated was a 20% excavation and repair, specimen c768, Alloy 52 M, heat NX8956TK, tested in as-welded condition. Figure 11 shows the transitioning sequence, which has maintained crack advance over very long times, long hold times and moderate crack increments (for this material). Even at one cycle per day, the growth rates are only *2 10−9 mm/s, which would clearly be lower under constant K conditions.

Weld With Refuse Passes A weld was fabricated at EPRI Charlotte under the supervision of Steve McCracken that incorporated refuse passes every layer. A refuse pass is performed with little or no ﬁller metal, and is designed to make the layer of weld beads more planar to improve inspectability of the weld as it is made. Specimen c636 was machined from such as weld, fabricated with Alloy 52 M, heat NX7859TK. The crack was positioned to grow in the vicinity of the refuse passes. Two attempts were made over the nearly 12,000 h test (Fig. 12), each involving extensive transitioning (in time and crack increment) and both resulted in very low growth rates of *2 10−10 mm/s.

Welds With Ductility Dip Cracking (DDC) Two welds were evaluated to determine the effect of DDC on SCC growth rates. One weld was not fabricated to enhance the likelihood of DDC, but extensive DDC occurred (Fig. 13). Another weld was fabricated by EPRI Charlotte using techniques designed to enhance the density of DDC, but the amount of DDC was

38

P.L. Andresen et al.

Fig. 10 Crack length versus time for specimen c637 of as-welded Alloy 152 M weld metal, heat 098934, where the original weld was 50% excavated and repair welded

limited compared to the other weld evaluated. In both cases, the crack plane was positioned in regions where characterization showed extensive DDC. Specimen c669 was fabricated from Alloy 52i, heat 187775, with extensive DDC, and two

SCC of Alloy 152/52 Welds Defects, Repairs …

39

Fig. 11 Crack length versus time for specimen c768 of as-welded Alloy 52 M weld metal, heat NX8956TK, where the original weld was 20% excavated and repair welded

Fig. 12 Crack length versus time for specimen c636 of as-welded Alloy 52 M weld metal fabricated with refuse passes every layer of weld metal

attempts over the 13,666 h test failed to show consequential growth rates at constant K—in one case showing *0 mm/s after transitioning, and in the second case showing *6 10−10 mm/s (Fig. 14). Specimen c770 was fabricated from Alloy 52, heat NX6523JK in an area where DDC was observed. Figure 15 shows the data obtained to date, which even during transitioning at one cycle per day, exhibits a low growth rate of 1.4 10−9 mm/s. Among the many debates that experimentalists have when evaluating these highly SCC resistant materials is whether the constant K rate is representative of SCC. That is, should it be assumed that resistant materials should automatically

40

P.L. Andresen et al.

Fig. 13 Micrographs showing examples of DDC in high Cr weld metals [13]

Fig. 14 Crack length versus time for specimen c669 of as-welded Alloy 52 M weld metal that exhibited extensive ductility dip cracking (DDC)

exhibit 100% IG morphology if the test is transitioned successfully, or are there more meaningful measures of good transitioning related to the opportunity provided to grow as the crack wishes? It is crucial to recognize that in the weld metals, highly IG morphology is rare, and even the higher growth rates measured at one lab [50] show no difference in %IG relative to other labs that have found only *10 lower growth rates (Fig. 8). Indeed, some TG morphology is observed in plant

SCC of Alloy 152/52 Welds Defects, Repairs …

41

Fig. 15 Crack length versus time for specimen c770 of as-welded Alloy 52 weld metal that some ductility dip cracking (DDC)

components and most laboratory specimens of all materials, and as the grain boundary path becomes increasingly resistant, there should be an increasing tendency for the crack morphology to become TG.

Welds Dilution Zones Whenever a high Cr weld is made—whether to low alloy steel, stainless steel, Alloy 600, Alloy 182/82, or, to a much lesser extent, Alloy 690—there will be dilution in the ﬁrst 1–3 welding passes (layers). For welds on Alloy 690, the change in Cr is minimal, but there is a variation in other elements, such as Nb, Ti and Mo in the weld metal. The largest dilution is when welding to carbon or low alloy steel, and the areas of largest dilution create mixtures that are iron-base rather than nickel-base. To evaluate the effects of Cr dilution, which has a pronounced effect on SCC growth rates, welds were made by the U.S. Naval Nuclear Labs using dual-wire feed of different relative rates. For example, a 16% Cr Alloy 600 wire fed at the same rate as a 30% Cr “Alloy 52” wire would have a Cr content of 23%. Each weld is a sizeable, multi-pass, multi-layer weld so that specimens of relatively homogeneous composition can be cut from material that is a well deﬁned yet simulates what develops in the ﬁrst pass or two. While some may argue that looking an actual dilution zones makes more sense, there is wide variability in dilution and

42

P.L. Andresen et al.

its depth and uniformity near the weld fusion line. Across the width and growth increment of an SCC crack, it is very challenging to determine the exact composition, and it certainly will be non-uniform. Complementing such measurements with homogenous weld metals of well-deﬁned composition permits a much more deﬁnitive evaluation of the effects of dilution n SCC growth rate. The actual two-wire-feed weld compositions and associated specimens number are listed below. All but the c736 (14% Cr weld) represent dilutions between *16% Cr and *28% Cr weld wires. C736 represents a weld created by a *50:50 dilution of low alloy steel and *28% Cr weld wires. • • • • • • •

C736—14% Cr, NNL, JT21008-G Weld—1:1 dilution with low alloy steel wire C730—16.6% Cr, NNL, JT21008-D Weld C731—20.3% Cr, NNL, JT21008-B Weld C732—21.7% Cr, NNL, JT21008-C Weld C733—23.5% Cr, NNL, JT21008-A Weld C737—24% Cr, NNL, JT21008-H Weld C738—28% Cr, NNL, JT21008-F Weld

The response of the 14% Cr two-wire-feed weld (specimen c736) is shown in Fig. 16, and its low growth rate and inability to sustain crack advance (in two long-term attempts) may seem surprising given the other specimens and compositions. However, this weld is a 50:50 mix of low alloy steel and a *28% Cr nickel alloy, so the iron content is dramatically higher than in other welds—that is, its Cr variation is small compared to Fe. Figures 15, 16, 17, 18, 19, 20, 21 and 22, detail the SCC response (typically there were two long-term evaluations separated by *1 mm crack advance in fatigue and transitioning to SCC) of the 16.6, 20.3, 21.7, 23.5, 24 and 28% Cr welds, and show a progressively lower growth rate with increasing Cr content. Note that Cr content alone does not control SCC response, but perhaps ten primary and hundreds of secondary factors are important. Nonetheless, for these welds made using nominally identical weld parameters, it’s clear the Cr content is a major factor. These data are summarized in Fig. 24, which shows a broadly parallel response versus Cr content to that observed by Morton and Young [45] on other welds under other test conditions.

Discussion Extraordinary efforts have been made in this program to obtain realistic growth rates in a wide variety of high Cr nickel alloy weld metals. All welds to date have been tested in the as-welded condition (no post-weld heat treatment). SCC susceptibility waspenhanced by testing at moderate to high stress intensity factor (*30–45 MPa m), at high temperature (360 °C), and at the Ni/NiO phase boundary (26 cc/kg H2 at 360 °C). While the effect of these parameters was not

SCC of Alloy 152/52 Welds Defects, Repairs …

43

Fig. 16 Crack length versus time for specimen c736 of a dual-wire-feed weld involving low alloy steel and *28% Cr weld wires with 14% Cr tested in as-welded condition. Note that the 14% Cr point (in red) was a 50:50 mix of low alloy steel and a *28% Cr nickel alloy, so the iron content is dramatically higher than in other welds

Fig. 17 Crack length versus time for specimen c730 of a weld involving only a 16.6% Cr weld wire tested in as-welded condition

44

P.L. Andresen et al.

Fig. 18 Crack length versus time for specimen c731 of a dual-wire-feed weld involving a *16% Cr and a *28% Cr weld wires to yield a 20.3% Cr composition tested in as-welded condition

Fig. 19 Crack length versus time for specimen c732 of a dual-wire-feed weld involving a *16% Cr and a *28% Cr weld wires to yield a 21.7% Cr composition tested in as-welded condition

SCC of Alloy 152/52 Welds Defects, Repairs …

45

Fig. 20 Crack length versus time for specimen c733 of a dual-wire-feed weld involving a *16% Cr and a *28% Cr weld wires to yield a 23.5% Cr composition tested in as-welded condition

Fig. 21 Crack length versus time for specimen c737 of a dual-wire-feed weld involving a *16% Cr and a *28% Cr weld wires to yield a 24% Cr composition tested in as-welded condition

46

P.L. Andresen et al.

Fig. 22 Crack length versus time for specimen c737 of a dual-wire-feed weld involving a *16% Cr and a *28% Cr weld wires to yield a 24% Cr composition tested in as-welded condition

Fig. 23 Crack length versus time for specimen c738 of a dual-wire-feed weld involving only a *28% Cr weld wires to yield a 28% Cr composition tested in as-welded condition

SCC of Alloy 152/52 Welds Defects, Repairs …

47

directly evaluated on the high Cr weld metals, they have been evaluated on Alloy 690 [8–30] and in more extensive data on Alloy 600 and its weld metals [44–46, 51, 52]. So it’s reasonable to assume that these conditions promote SCC compared to lower K or temperature, or other dissolved H2 concentrations. Additionally, very extensive efforts—probably more than any historical effort— have been made to transition the crack from the TG fatigue precrack to SCC (e.g., Fig. 4). These included very long times in cycle + hold (at Kmax) so that extensive crack advance occurred, which provided an opportunity for the crack to encounter and follow an IG path. Transitioning is more challenging in weld metals, because the diameter and length of the dendritic grains are much greater than the grain size in most wrought materials. In many cases more than 2000 h per specimen per transitioning attempt were used, and in some cases much more. Variations were considered, such as the use of a range of load ratio (R) values from 0.5 to 0.7, asymmetric waveforms (SCC is considered to be unaffected by fall time, so waveforms comprised of 90–95% rise time were used in many cases), many hold time steps (e.g., 1000, 3000, 9000, 27,300 and 85,900 s) before changing to constant K (no cycling). While increasing the care and time of transitioning did appear to help to sustain subsequent SCC, it did not lead to growth rates about *4 10−9 mm/s, and in most cases the growth rates fell below 1–2 10−9 mm/s. Nor did the many additional hold time steps or longer transitioning times (and associated crack advance) produce signiﬁcantly greater %IG. Note that all specimens were engaged from the fatigue precrack—the lack of any subsequent crack advance after fatigue precracking was found to be a big issue in distinguishing resistance to SCC from a crack that was simply pinned at the end of the high DK fatigue precrack [52]. It is important to obtain data at constant K (no cycling) rather than claiming that occasional partial unloading has little effect on the data and therefore can be considered constant-K data. The effect of a given load cycle has to be viewed in the context of the SCC growth rate. For susceptible materials and water chemistries, such as for the Alloy 182/82 weld metals considered in MRP-115 [52] (or in poor quality water with oxygen), growth rates in the vicinity of 10−6 mm/s can occur, and the effect of a periodic partial unloading cycle (e.g., R = 0.7, 500 s rise time and 9000 s hold at Kmax) is very small, generally 10. In our data, even a once-a-day cycle (85,900 s hold) produced on average about a 10 increase in growth rate above completely constant K (no cycling) data. Other laboratories [25, 26] propose that a good estimate can be made of the constant K (no cycling) response from the periodic partial unloading (PPU) response, but data from ﬁve laboratories has been shown to be unreliable in that some resistant materials grow much slower at constant K than expected from the PPU data—indeed, some show crack arrest. In general, the use of a formula to estimate the air (inert) fatigue growth rate is not consistently reliable, and indeed the air rates have rarely or never been measured under such conditions (300+ °C, high

48

P.L. Andresen et al.

R, 7000 h, such as in Figs. 20 and 22—where no such upward trend is observed. The resistance to SCC of these high Cr weld metals is also exempliﬁed in growth rate data [53] obtained in aggressive BWR water using 2 ppm dissolved O2, 30– 50 ppb sulfate or chloride, and high stress intensity factors (Fig. 9). Despite these harsh conditions, the three weld metals tested (multiple times/microstructures per specimen) exhibited only low growth rates in the same vicinity as the testing done in this program in PWR primary water. Much of the focus of this study has been on weld defects, weld repairs and weld dilution zones, and only low growth rates have been observed (Figs. 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 and 24). While some challenges remain in fabricating optimum high Cr welds, these data indicate that large defects can be tolerated because the SCC growth rates remain low.

Fig. 24 Summary of data on two-wire-feed welds that simulate weld dilution effects. Note that the 14% Cr point (in red) was a 50:50 mix of low alloy steel and a *28% Cr nickel alloy, so the iron content is dramatically higher than in other welds

50

P.L. Andresen et al.

Conclusions For the *30% Cr nickel alloys, there can be no question that some inherent susceptibility to SCC exists, and the SCC growth rate response of various heats, microstructures and types of cold work of Alloy 690 were evaluated Alloy 690 in a related program [8–18]. Clearly the concept of SCC immunity should be replaced with the concept of adequately low crack growth rate. The SCC resistance of 33 specimens of many Alloy 52/152/52i weld metals was evaluated over >315,000 h, including a number with weld defects and weld repairs, and their SCC growth rates were almost uniformly low. These consistently low growth rates were observed in spite of many (reasonable) p accelerants: moderate to high stress intensity factor (*30–45 MPa m), high temperature (360 °C) and dissolved H2 values (26 cc/kg at 360 °C) that are at the growth rate peak (at the Ni/NiO phase boundary). Additionally, very extensive and diverse approaches to IG transitioning were used over very long periods of time (over 171,000 h of total testing time), and multiple microstructures were examined in each weld (by advancing the crack by *1 mm in fatigue and re-transitioning to condition that favor IG SCC). All of the specimens tested at GE were fully engaged from the TG fatigue precrack. The extent of %IG ranged from small to medium (*50%), and did not appear to depend on the speciﬁc transitioning technique used. Higher %IG did not obviously translate to higher growth rates (within the context of the low growth rates at %IG observed), in part because TG crack advance was observed in most cases at constant K (no cycling). It should not be surprising that as alloys become highly resistant, the IG path becomes only slightly favored over the TG path. Good agreement on the crack growth rates of the high Cr weld metals—including on speciﬁc welds tested by multiple laboratories—was obtained. The only outliers related to Alloy 152 welds fabricated by ANL, and only when tested at ANL—other laboratories were unable to reproduce the higher growth rates despite extensive attempts (*50,000 h) on multiple specimens and multiple areas of each weld microstructure. There is at time no explanation for the discrepancy, but in the broad spectrum of the international results, they are reasonably considered outliers at this time. The excellent resistance—although not immunity—to SCC of these weld metals extends to aggressive BWR water chemistry conditions [53], although it’s clear that the SCC resistance can be compromised by cold working the welds, which shouldn’t occur in plant components.

References 1. P.L. Andresen, Perspective and direction of stress corrosion cracking in hot water, in Proceedings of 10th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems, NACE, 2001

SCC of Alloy 152/52 Welds Defects, Repairs …

51

2. P.L. Andresen, T.M. Angeliu, L.M. Young, Immunity, Thresholds, and Other SCC Fiction (TMS, Proc. Staehle Symp. on Chemistry and Electrochemistry of Corrosion and SCC, 2001) 3. P.L. Andresen, L.M. Young, P.W. Emigh, R.M. Horn, Stress corrosion crack growth rate behavior of Ni alloys 182 and 600 in high temperature water, corrosion/02, paper 02510, NACE, 2002 4. P.L. Andresen, Conceptual similarities and common predictive approaches for SCC in high temperature water systems, paper 96258, corrosion/96, NACE, 1996 5. F.P. Ford, P.L. Andresen, Corrosion in nuclear systems: environmentally assisted cracking in light water reactors, in Corrosion Mechanisms, ed. By P. Marcus and J. Ouder, Marcel Dekker, pp. 501–546, 1994 6. P.L. Andresen, C.L. Briant, Environmentally Assisted Cracking of Types 304L/316L/316NG Stainless Steel in 288 °C Water. Corrosion 45, 448–463 (1989) 7. P.L. Andresen, Environmentally assisted growth rate response of nonsensitized AISI 316 grade stainless steels in high temperature water. Corrosion. 44 (7), 450 (1988) 8. P.L. Andresen, J. Hickling, PWSCC growth rates of cold worked alloy 690, Conference on Alloy 600, Santa Ana Pueblo, NM, March 2005, EPRI, Palo Alto 9. P.L. Andresen, Development of advanced testing techniques to quantify the improved PWSCC resistance of alloy 690 and its weld metals (MRP-123), EPRI, Palo Alto, CA, 2004. Technical report 1010269 10. P.L. Andresen, M.M. Morra, Materials reliability program: Resistance of alloy 690/52/152 to SCC crack growth in simulated PWR Primary Water (MRP-196), EPRI, Palo Alto, CA, 2006, Technical report 1013516 11. P.L. Andresen, M.M. Morra, K. Ahluwalia, Effect of deformation temperature, orientation and carbides on SCC of alloy 690, in Proceedings of 16th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, NACE, Aug 2013 12. P.L. Andresen, M.M. Morra, K. Ahluwalia, SCC of alloy 152/52/52i weld metals in PWR primary water, in Proceedings of 16th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, NACE, Aug 2013 13. P.L. Andresen, Overview and Perspective on Stress Corrosion Cracking of Alloy 690 and Alloy 152/52 Weld Metals, Report 3002009966 (EPRI, Palo Alto, 2017) 14. P.L. Andresen, M.M. Morra, K. Ahluwalia, SCC of alloy 690 and its weld metals, in Proceedings of 15th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, The Metallurgical Society, Aug 2011 15. P.L. Andresen, M.M. Morra, K.S. Ahluwalia, J. Wilson, Effect of deformation and orientation on SCC of Alloy 690, in Proceedings of 14th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, American Nuclear Society, Aug 2009 16. P.L. Andresen, M. M. Morra, J. Hickling, K.S. Ahluwalia, J.A. Wilson, PWSCC of alloys 690, 52 and 152, in Proceedings of 13th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, Canadian Nuclear Society, Aug 2007 17. P.L. Andresen, The Present and Future of SCC Testing of Alloy 690 and Its Weld Metals, Alloy 690 Experts Meeting (EPRI, Tampa, 2009) 18. P.L. Andresen, SCC of Alloy 690 Base Metal and HAZ in PWR Primary Water and of Alloy 52 & 152 Weld Metals in PWR Primary Water, Alloy 690 Experts Mtg (EPRI, Tampa, 2010) 19. M.B. Toloczko, S.M. Bruemmer, Crack Growth Response of alloy 690 in simulated PWR Primary Water, in Proceedings of 14th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, American Nuclear Society, Aug 2009 20. M.B. Toloczko, M.J. Olszta, N.J. Overman, S.M. Bruemmer, observations and implications of intergranular stress corrosion crack growth of alloy 152 weld metals in simulated PWR primary water, in Proceedings of 16th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, NACE, Aug 2013

52

P.L. Andresen et al.

21. S.M. Bruemmer, M. Olszta, M.B. Toloczko, Microstructural effects on stress corrosion crack growth in cold-worked alloy 690 tubing and plate materials, in Proceedings 16th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, NACE, Aug 2013 22. M.B. Toloczko, M.J. Olszta, S.M. Bruemmer, 1D cold rolling effects on SCC growth in alloy 690 tubing and plate materials and SCC crack growth testing of alloy 52 M in simulated PWR Primary Water, in Proceedings of 15th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, The Metallurgical Society, Aug 2011 23. D.J. Paraventi, W.C. Moshier, Alloy 690 SCC growth rate testing, in Workshop on Cold Work in Iron- and Nickel-Base Alloys, ed. by R.W. Staehle, J. Gorman, June 2007, EPRI, Palo Alto 24. D.J. Paraventi, W.C. Moshier, Alloy 690 SCC growth rate testing, in Proceedings of EPRI Alloy 690 Workshop, Atlanta, 31 Oct 2007 25. B. Alexandreanu, O. K. Chopra, W.J. Shack, The stress corrosion cracking behavior of alloys 690 and 152 weld in a PWR environment, PVP2008-61137, in Proceedings of ASME PVP, 2008 26. B. Alexandreanu, The stress corrosion cracking behavior of alloys 690 and 152 weld in a PWR environment, in Proceesings of 14th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, American Nuclear Society, Aug 2009 27. D. Tice, Personal communication on crack growth rate of alloy 690, AMEC (Serco), April 2010 28. D. Tice, S. Medway, N. Platts, J. Stairmand, I. Armson, Crack growth testing of cold worked alloy 690 in primary water environment, in Proceedings of 15th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, The Metallurgical Society, Aug 2011 29. D. Gomez-Briceno et al., “CGR testing of alloy 690 and weld metals 52 and 152”, alloy 690 experts meeting (EPRI, Tampa, 2010) 30. A.R. Jenks, G.A. White, P. Crooker, Assessment of laboratory PWSCC crack growth rate data for Nickel-based alloys, in International Light Water Reactor Materials Reliability Conference and Exhibition 2016, 1–4 Aug 2016, EPRI, Palo Alto, 2016. (and ﬁnal EPRI report, in press) 31. ASME boiler and pressure vessel code, Sections III and XI, ASME, NY 32. M.J. Esmacher, SCC of industrial utilities: Boiler and cooling water systems, in Stress Corrosion Cracking: Mechanisms, Materials and Application to Industrial Problems, eds. by V.S. Raja, T. Shoji, Woodhead Publishing, 2010 33. P.L. Andresen, P.W. Emigh, M.M. Morra, R.M. Horn, Effects of yield strength, corrosion potential, stress intensity factor, silicon and grain boundary character on the SCC of stainless steels, in Proceedings of 11th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, ANS, 2003 34. P.L. Andresen, M.M. Morra, SCC of stainless steels and Ni alloys in high temperature water. Corrosion 64, 15–29 (2008) 35. P.L. Andresen, G.S. Was, SCC of unirradiated stainless steels and Nickel alloys in hot water, in 17th International Corrosion Congress, Las Vegas, NACE, Houston, 2008 36. M.M. Morra, M. Othon, E. Willis, S. McCracken, “Characterization of Structures and Strains in 52-type and 152 Welds”, Alloy 690/52/152 PWSCC Research Collaboration Meeting (Tampa, Florida, 2011) 37. T.M. Angeliu, P.L. Andresen, J.A. Sutliff and R.M. Horn, intergranular stress corrosion cracking of unsensitized stainless steels in BWR Environments, in Proceedings of 9th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, AIME, 1999 38. T.M. Angeliu, P.L. Andresen, E. Hall, J.A. Sutliff, S. Sitzman, Strain and microstructure characterization of austenitic SS weld HAZs”, Corrosion/2000, Paper 00186, NACE, 2000

SCC of Alloy 152/52 Welds Defects, Repairs …

53

39. J.A. Sutliff, An investigation of plastic strain in copper by automated-EBSP, Microsc. Microanal. 5(Supp. 2), 236 (Proc: Microscopy & Microanalysis ‘99) (1999) 40. M.A. Othon, M.M. Morra, EBSD characterization of the deformation behavior of alloy 182 weld metal, Microscopy and Microanalysis, 11(Suppl 2), 522–523CD Cambridge University Press, (2005). doi:10.1017/S1431927605506949 41. G.A. Young, M.J. Hackett, J.D. Tucker, T.E. Capobianco, Welds for Nuclear Systems, Comprehensive Treatise on Materials for Nuclear Systems, ed. by R. Konings, Editor in Chief (Elsevier Science, to be published in 2011) 42. P.L. Andresen, C.L. Briant, Role of S, P and N segregation on intergranular environmental cracking of stainless steels in high temperature water, in Proceedings of 3rd International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, AIME, pp. 371–382, 1988 43. P.L. Andresen, The Effects of Aqueous Impurities on Intergranular Stress Corrosion Cracking of Sensitized Type 304 Stainless Steel, Final Report NP3384 Contract T115-3 (EPRI, 1983) 44. P.L. Andresen, J. Hickling, K.S. Ahluwalia, J.A. Wilson, Effects of hydrogen on SCC growth rate of Ni alloys in high temperature water. Corrosion 64(9), 707 (2008) 45. P.L. Andresen, J. Hickling, K.S. Ahluwalia, J.A. Wilson, effects of PWR Primary Water Chemistry on PWSCC of Ni alloys, in Proceedings of 13th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, Canadian Nuclear Society, Aug 2007 46. S.A. Attanasio, D.S. Morton, Measurement of the Ni/NiO transition in Ni–Cr–Fe alloys and updated data and correlation to quantify the effect of aqueous hydrogen on primary water SCC, in Proceedings 11th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems, ANS, 2003 47. L.W. Niedrach, A new membrane type pH sensor for use in high temperature high pressure water. J. Electrochem. Soc. 127, 2122 (1980) 48. Elaine West, D.S. Morton, John Mullen, Bob Etien and Heather Mohr, SCC behavior of EN52/EN52i in deaerated water, in Proceedings of 16th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, NACE, Aug 2013 49. BWR vessel and internals project, evaluation of crack growth in BWR stainless steel RPV internals (BWRVIP-14-A), EPRI Report TR-105873, Mar 1996 50. S.M. Bruemmer, Unpublished data (Paciﬁc Northwest Nat, Lab, 2016) 51. G. White, Crack Growth Rates for Evaluating Primary Water Stress Corrosion Cracking (PWSCC) of thick-wall alloy 600 materials, MRP-55 Revision 1, Final Report 1006695, EPRI (Palo Alto, CA, 2002) 52. Crack Growth Rates for Evaluating Primary Water Stress Corrosion Cracking (PWSCC) of Alloy 82, 182, and 132 Welds (MRP-115), EPRI Final Report 1006696, Nov 2004 53. P.L. Andresen, SCC of high Cr alloys in BWR environments, in Proceedings of 15th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, The Metallurgical Society, Aug 2011

NRC Perspectives on Primary Water Stress Corrosion Cracking of High-Chromium, Nickel-Based Alloys Greg Oberson, Margaret Audrain, Jay Collins and Eric Reichelt

Abstract High-chromium, nickel-based alloys, including alloy 690 base material and alloys 52 and 152 weld ﬁller metals, are used in the primary system of new pressurized water reactors (PWRs), as well as for replacement and mitigation of components in existing reactors. These materials are thought to be highly resistant to primary water stress corrosion cracking (PWSCC), which is observed in plant service for components fabricated from the low-chromium alloys 600, 82, and 182. For over 10 years, the U.S. Nuclear Regulatory Commission has sponsored a laboratory testing program to measure the PWSCC growth rates of alloys 690, 52, and 152 in environmental conditions representative of PWRs, with the intent to support technical bases for the determination of appropriate in-service inspection requirements. In many tests, the low crack growth rates are conﬁrmed. For certain cases, however, such as in highly cold worked alloy 690 and at dilution zones between high-chromium weld metals and low-chromium base metals, PWSCC growth rates are reported to be similar to those observed in alloys 600, 82, and 182. Challenges arise in the use of these data for predictive models given the relatively few numbers of tests performed for some material conditions and uncertainties about the correlation between conditions of test materials and those found in the ﬁeld. This paper will present perspectives on factors that may be considered for the application of these data to the analysis of plant components.

G. Oberson (&) M. Audrain U.S. Nuclear Regulatory Commission, Ofﬁce of Nuclear Regulatory Research, Washington DC 20555, USA e-mail: [email protected] J. Collins U.S. Nuclear Regulatory Commission, Ofﬁce of Nuclear Reactor Regulation, Washington DC 20555, USA E. Reichelt U.S. Nuclear Regulatory Commission, Ofﬁce of New Reactors, Washington DC 20555, USA © The Minerals, Metals & Materials Society 2019 J.H. Jackson et al. (eds.), Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, The Minerals, Metals & Materials Series, https://doi.org/10.1007/978-3-030-04639-2_4

55

56

G. Oberson et al.

Introduction Primary water stress corrosion cracking (PWSCC) in nickel-base alloy primary pressure boundary components is a degradation mechanism that can affect the operational safety of pressurized water reactors (PWRs). PWSCC preferentially occurs in components that operate at high temperatures and pressures, including steam generator tubes, reactor vessel head penetrations, nozzles, and dissimilar metal piping welds. At the time of plant construction, the primary nickel-base alloys used in PWRs included alloy 600 base metal and alloy 82 or 182 weld ﬁller metal. The composition of these alloys is about 14 to 22 percent chromium. PWSCC was ﬁrst observed in thin-walled steam generator tubes in the 1970s and 1980s, but did not become apparent in thick-section nozzles and welds until the 1990s and early 2000s. Examples of operational experience include circumferential cracking of a hot leg dissimilar metal piping weld at V.C. Summer in 2000 [1], cracking of reactor pressure vessel head penetrations at North Anna [2] and Arkansas Nuclear One in 2001 [3] and circumferential cracking of a pressurizer surge nozzle at Wolf Creek in 2006 [4]. Most notably, PWSCC of a control rod drive mechanism (CRDM) nozzle at Davis-Besse and subsequent boric acid leakage led to signiﬁcant wastage of the reactor pressure vessel head in 2002 [5]. In response to the operational experience for PWSCC in alloys 600, 82, and 182, the industry began to replace or repair nickel-base components and welds with alloys 690, 52, and 152, which are thought to be more corrosion resistant because of a higher chromium content in the range of 28–31%. Actions included replacing reactor pressure vessel heads and overlaying welds with more resistant material. The U.S. Nuclear Regulatory Commission’s (NRC’s) regulatory approach for ensuring that structural integrity is maintained in plant components that could potentially be affected by PWSCC involves, in part, the periodic inspection of these components by non-destructive methods, such as by visual, surface, or volumetric examination. In particular, Title 10 of the Code of Federal Regulations, Part 50.55a incorporates by reference the in-service inspection requirements set forth in Section XI of the ASME Boiler and Pressure Vessel Code and approved Code Cases. Moreover, NRC identiﬁed PWSCC as an aging effect that should be managed for both initial and subsequent license renewal of U.S. power reactors, following aging management programs described in NUREG-1801 [6] and NUREG-2191 [7], respectively. Although NRC believes that this regulatory framework provides reasonable assurance that PWSCC will not compromise reactor safety, an active research program is maintained to conﬁrm the validity of technical positions taken by licensees in their safety analyses and to proactively identify issues that may eventually affect plant components. For approximately the past 10 years, the main focus of the NRC research program for PWSCC issues has involved laboratory crack growth rate testing of high-chromium alloys 690, 52, and 152. To date, the operating experience of alloys 690, 52 and 152 in power reactors worldwide is largely positive, with no indications yet for the occurrence of PWSCC in thin-walled steam generator tubes or

NRC Perspectives on Primary Water Stress Corrosion Cracking …

57

thick-section components and welds. This observation is consistent with the ﬁndings of laboratory testing which generally show substantially lower crack growth rates for alloys 690, 52, and 152 compared to the low-chromium alloys when exposed to the same environmental conditions. In light of the operational and laboratory testing experience for alloys 690, 52, and 152, the U.S. nuclear power industry has proposed that the current in-service inspection requirements could safely be reduced. A representative technical analysis to support such a proposal is given in the Electric Power Research Institute (EPRI) report MRP-375 [8]. In the course of reviewing the body of laboratory-generated crack growth rate data, it became apparent to NRC staff that important questions remain with respect to its interpretation and its applicability to plant components. NRC and industry, represented by EPRI, agreed that the appropriate path forward would be to rigorously evaluate the quality of the data for alloys 690, 52, and 152, such that analytical models for crack growth in plant components could then be developed. This conceptual approach is similar to that followed for alloys 600, 82, and 182, as documented in the EPRI reports MRP-55 [9] and -115 [10]. The purpose of this paper is to synopsize NRC perspectives on some of the unique technical issues that have emerged in the data evaluation. These include the interpretation of periodic partial unload (PPU) test data, the effects of cold work on PWSCC susceptibility, cracking along weld interfaces, and heat-to-heat testing variations. No regulatory positions will be presented here. Rather the intent is to briefly identify issues that may warrant further consideration and discussion in the model development process.

Interpretation of PPU Data PWSCC growth rate tests for alloys 690, 52, and 152 typically involve the initiation of a transgranular fatigue precrack in a compact tension specimen, followed by a loading sequence intended to transition the crack to an intergranular morphology for a constant load or constant stress intensity factor (constant K) growth rate measurement (hereafter collectively referred to as constant load). The loading sequence can progress stepwise from higher to lower frequency fatigue loading cycles, then through steps where the load is held constant for certain durations between load changes. Depending on the load waveform, the latter may be referred to as cycle-plus-hold or PPU steps (hereafter collectively referred to as PPU). Finally, constant load may be established. The process of attempting to transition the cracks to intergranular morphology in this manner is well established, though there is no prescriptive testing standard that governs such parameters as applied stress levels, fatigue frequencies, or loading durations. These are generally left to the discretion of the investigator. The interpretation of crack growth rate data acquired by PPU has been a topic of debate among the technical community. Conceptually the PPU growth rate can be

58

G. Oberson et al.

understood as the time-weighted summation of the fatigue crack growth rate and the constant load crack growth rate that collectively encompass the PPU loading step. As such, a constant load crack growth rate could be calculated from the measured PPU growth rate by subtracting out the fatigue contribution, which would be known from pure fatigue testing. Such an approach, however, necessarily assumes that the periodic load change has no signiﬁcant mechanistic effect to enhance the crack growth rate during the constant load portion of the step. An analysis for alloys 600 and 182 was presented in MRP-115, which compared constant load crack growth rates to those measured in PPU for hold periods (time interval between load changes) between 600s and 1 day. The PPU growth rates were generally no more than a factor of 2 higher than those measured in constant load. PPU growth rates were thus considered to be of sufﬁciently representative of the constant load condition to use in analytical models. Comparisons between constant load and PPU growth rates for alloys 690, 52, and 152 have proven, in many cases, to show greater divergence. For instance, Paciﬁc Northwest National Laboratory (PNNL) [11], General Electric Global Research [12], Naval Nuclear Laboratories (NNL) [13], and others have reported PPU growth rates may be over 10 times faster than constant load for alloys 690, 52, and 152. However, the enhancement is not consistent for different test specimens, or even within different regions of a single specimen under similar loading conditions. Argonne National Laboratory (ANL) also did a systematic comparison of PPU and constant load growth rates and found that the relative enhancement of the PPU rate was greater at low levels of intergranular crack engagement, generally corresponding to lower growth rates, but would tend to converge at higher growth rates [14]. PNNL indicated that the effects of periodic load cycling are particularly pronounced for materials that tend to maintain contact ligaments behind the most advanced part of the crack front, where the dynamic strain appears to fracture the ligaments. Because the effects of PPU on alloys 690, 52, and 152 are not sufﬁciently understood to explain why the crack growth enhancement relative to constant load is not consistent from test-to-test, it is in uncertain whether the data should be used in analytical models, as was the case for the low-chromium materials. Various approaches could be taken. Including of all PPU data in the models could be considered as the most conservative, though potentially unrepresentative of constant load if there is a clearly demonstrated bias towards higher growth rates. Excluding all of the PPU data from the models has the advantage of simplifying the analysis, but at the possible expense of increased model uncertainty. A third approach would be to include PPU data in the model development, but only if they meet certain criteria. Such criteria could involve speciﬁcations for load hold time, test step duration, or others. NRC has not drawn any conclusions with respect to the acceptability of any of these approaches, but would recommend a clear explanation for the treatment of the PPU data in describing the model technical basis.

NRC Perspectives on Primary Water Stress Corrosion Cracking …

59

Treatment of Cold Worked Material A substantial amount of effort has been applied by the technical community to the testing of cold worked alloy 690, and to a lesser extent, alloys 52 and 152. The reasoning for this is twofold. First, it is thought that some components in plant service will experience cold work. Second, cold work appears to increase the susceptibility of the materials to cracking, thereby allowing for the study of crack growth phenomenology in shorter timeframes compared to non-cold worked material. The focus on the cold worked condition for the high-chromium alloys stands in general contrast to alloys 600, 82, and 182, for which data were generally reported in the non-cold worked condition in MRP-55 and -115, though there has been more testing done since those reports were published. One issue to be addressed for cold worked material is how to account for this condition in crack growth rate models. It is not likely that simply specifying cold work as a percentage reduction in dimension is, in itself, adequate. This does not provide potentially relevant information about the mode of deformation (e.g., forging, rolling, tensile straining), nor does it recognize possible variations in the properties of the starting material. The most comprehensive study of this topic to date was undertaken by PNNL under sponsorship by NRC, and focused on characterizing the material properties and crack growth rates of alloy 690 [15]. PNNL recognized that the application of cold work will have an effect on at least two related but independently measurable properties: the material hardness, and internal misorientation strain. If material hardness is assumed to be linearly related to yield strength, this effectively introduces a third parameter, but essentially interchangeable with hardness. PNNL showed that exponential and power law-type relationships can reasonably, but with some degree of scatter, correlate the hardness/yield stress or misorientation strain to the crack growth rate over certain ranges for these parameters. The scatter appears at least partly attributable to microstructural features, such as differences in dislocation densities near grain boundaries, even in materials with similar hardness and misorientation strain measurement. The extent to which such differences affect any proposed correlations in the crack growth rate models should be identiﬁed. Another issue that will clearly need to be addressed is the amount of cold work anticipated in plant components. Cold work could arise from material shaping, bending, surface grinding, or other types of preparation. Stresses induced by these processes may not be fully relieved by thermomechanical treatments. Cold work from shaping or bending is likely to be of greater concern for alloy 690 components than the welds, the most typical example being the straightening of CRDM nozzles in reactor pressure vessel upper heads after welding. It is recognized that high stresses could arise from surface grinding of alloy 690 or weld metals, but these are expected to be very localized, and not of great concern for component-scale structural integrity [16]. The other location for alloy 690 that is treated within the scope of cold work evaluations is the weld heat affected zone (HAZ). Internal strains can arise in the HAZ during welding and the technical community has

60

G. Oberson et al.

assessed whether this could be a preferential location for crack growth compared to unaffected base material. While attempts have been made to orient test specimens to grow cracks in the HAZ, this is challenging because of its small size and wavy boundaries. An alternative approach is to quantify the strain or hardness in the HAZ and to relate that to bulk cold worked material with equivalent strain or hardness, but which could be more readily tested. The amount of cold work or strain in alloy 690 plant components is not likely to be directly measured in-service, but rather could be inferred from knowledge of component fabrication and installation processes as well as measurements on what are thought to be representative mockup specimens. With respect to the former, if the model assumes that the amount of shaping or bending-induced cold work is effectively limited by fabrication and installation processes, it should be demonstrated that these processes are governed by rigorous ASME Code or vendor-speciﬁc requirements, and not simply assumed based on good-practices. Further, it should be recognized that even if the fabrication and installation requirements are based on a percentage of allowable cold work, there can be a substantial variation in hardness or internal strain for materials at the same cold work level depending on the material microstructure and thermomechanical treatment [15]. This would in turn affect that cracking susceptibility. Establishing the amount of cold work in mockups would likely involve measurements of hardness and/or internal strain, as by electron backscatter diffraction (EBSD), and correlating these to the same measurements from material of known cold work level. A complication that arises in such an approach, particularly as it relates to the HAZ, is that the strain develops at elevated welding temperatures, so it may be more appropriately be considered as warm work, rather than cold work. Some test data indicate that for warm work and cold work to the same reduction, the warm worked condition is less susceptible to cracking [17]. If this is the case, treating warm work as cold work in crack growth modeling would be a conservative interpretation.

Weld Interface and Dilution Zone (DZ) Testing The higher chromium content in alloys 52 and 152 compared to alloys 82 and 182 contributes to their greater PWSCC resistance. For some welds, alloy 52 or 152 will be joined to metal with lower chromium content. Examples include welds between carbon steel nozzles and stainless steel safe ends, J-groove welds to stainless steel clad, low-alloy steel reactor pressure vessel heads, and weld overlays, inlays, or onlays to mitigate alloy 82 and 182 welds. Near the weld interfaces, there will be a transition region where compositional mixing of the respective alloys dilutes the chromium content from the amount found in the starting material. The potential for enhanced PWSCC susceptibility near weld interfaces has been recognized and studied by the technical community. The PWSCC susceptibility in the DZ may depend, in part, on the type of interface. For instance, the amount of chromium dilution in the DZ would likely be

NRC Perspectives on Primary Water Stress Corrosion Cracking …

61

different for alloy 52 or 152 joined to stainless steel which contains signiﬁcant chromium itself, compared to joining with low alloy steel which has very little chromium. It is possible for the former case that, even if there is some dilution, the chromium content remains high enough to maintain PWSCC resistance. In laboratory testing, higher crack growth rates have been reported for interfaces with low alloy steel compared to stainless steel [18]. Further, the selection of weld parameters like arc power and travel speed can affect the amount of base metal melting, and thus the size of the DZ. This may provide some insight into the relative susceptibility of welds fabricated by different methods. In absence of the ability to test all potential variations of weld ﬁller metal, base metal, and welding process, a reasonable approach may be to focus testing an analysis on what are thought to be the bounding conditions for the extent of dilution Another complication for assessing the PWSCC susceptibility of the DZ is that the mechanisms of enhanced susceptibility are not well understood. Though it is anticipated that reducing the chromium content will, in itself, increase the susceptibility near the interface, it appears that microstructural features could also play an important role. For an ANL weld of alloy 152 to low alloy steel, tests near the interface at three laboratories have given constant load crack growth rates as low as the order of 10−9 mm/s to as high as the order of 10−7 mm/s. The difference in crack growth rates does not seem to be explicable on the basis of chromium content difference alone. Microstructural examination of the ANL weld and other DZ specimens identiﬁed boundaries between dendrite packets that can be aligned along the cracking direction over distances of hundreds of microns or more [19]. There is some evidence that the could form a preferential crack path that correlates with high crack growth measurements in localized regions of the material, but this postulation is still under investigation. If this is conﬁrmed to be a contributing factor to DZ SCC susceptibility, it will be important to know whether the spatial distribution of the boundary features in typical plant welds is such that cracks could propagate along them for signiﬁcant distances.

Stress Intensity Factor and Temperature Dependencies The models for alloys 600 and 82/182 reported in MRP-55 and -115, respectively, include terms to account for the dependency between crack growth rate with crack tip stress intensity factor and temperature, among other parameters. The temperature dependency is addressed by an activation energy, whereas there is a power-law type relationship for the crack tip stress intensity factor. The activation energy was calculated based on tests conducted between about 290 and 360 °C, while the stress intensity varied between approximately 15 and 60 MPa√m. The change in crack growth rate over the parameter ranges was generally greater than an order of magnitude. For alloys 690, 52, and 152, however, the very low growth rates in some conditions approach the limits of measurement resolution and clearly discernable dependencies are confounded by the scatter.

62

G. Oberson et al.

In alloy 690, for instance, growth rates of sufﬁcient magnitude to calculate activation energy, generally above 10−9 mm/s, are only reported in moderately to highly cold worked material in the range of about 17–30% reduction. If the crack growth rate model applies this activation energy to material with no or low cold work, it would be for a condition at which it has not been directly measured. The adequacy of technical basis for such an assumption, would depend on the conﬁdence that the rate controlling deformation process is the same at the lower and higher cold worked conditions, or at least that there is not a stronger temperature dependence for the lower cold worked condition. Moreover, even for the higher cold worked condition, there is a degree of scatter in the measure of the activation energy itself. For various heats of alloy 690, PNNL calculated activation energies from less than 100 kJ/mol to over 160 kJ/mol [15], and there is perhaps even greater variance when considering the measurements of other laboratories. Fewer attempts have been made to calculate the activation energy for alloys 52 and 152 compared to alloy 690. Most of those are for as-welded material, not having the accelerating effect of added cold work, making the measurements even more uncertain. The challenge with calculating the dependence between crack growth rate and crack tip stress intensity factor are similar to those for the activation energy. The models presented in MRP-55 and -115 show about 10x increase in crack growth rate as K increases between 15 and 50 MPa√m. For alloys 690, 52, and 152, there appears to be a signiﬁcant variation in the K-dependency from heat-to-heat of material. PNNL reported very strong K dependencies, similar to those in MRP-55 and -115 for some heats of alloy 690, such as 31% CF heats of Valinox RE243 and Sumitomo E67074C. On the other hand, little to no increase in crack growth rate (1. For the current study two of the selected tube sections of the 100 dpa tube are chosen that were either just on or beyond the shoulder region of the dpa proﬁle and it is therefore important to assess whether the spectral environment of the two shoulder specimens are very different from that of the flat spectral region specimen. Fukuya has provided an excellent example of a radial trace through a PWR core and the surrounding internals [16], showing in Fig. 10 that while fast neutrons and gamma rays pass through material discontinuities with relatively modest inflections, the thermal neutron fluxes and therefore the T/F ratio experience abrupt variations at every metal/water discontinuity.

Fig. 9 Schematic representation of the proﬁles of fast (E > 0.1 MeV) and thermal fluxes of a typical PWR [1, 10]

660

F.A. Garner et al.

Fig. 10 Schematic representation of the proﬁles of gamma rays, fast neutrons and gamma rays along a radial trace though a PWR core and surrounding components [16]

However, on a more local scale, even in the core flat-spectral region, the neutron physics codes used to calculate the neutron-spectra are not based on a sufﬁciently ﬁne enough scale to allow a clear deﬁnition of the local time-averaged T/F ratio for a thin flux-thimble tube with an abrupt material discontinuity, with water on the outside and gas on the inside of the tube, surrounded by other discontinuities associated with the fuel/cladding/water. Intuitively, the accumulated thermal neutron fluence cannot be conﬁdently speciﬁed to within *50–75%. Therefore there is comparable uncertainty in the gas generation rates of the same order, even though we know the full details of the various reaction cross-sections. Several studies have shown that the T/F ratio can change even in the absence of a discontinuity, such as along the axis of a baffle bolt as the influence of fuel absorption of thermal neutrons decreases when moving away from the core [17, 18]. However, knowing that the helium content is dominated by the accumulated thermal neutron fluence and the nickel content, we can use the measured helium as a “retrospective dosimeter” if we have a comparable flux thimble tube, especially from the same reactor with the displacement damage calculated with the same degree of uncertainty. While hydrogen easily moves in steel at elevated temperatures and can escape the metal, rendering it unsuitable as a dosimeter, helium is essentially immobile because it is very quickly trapped by vacancies and vacancy clusters, and can be easily measured. Helium can therefore serve as an excellent thermal neutron fluence dosimeter. Luckily for this project, there are published data on three previously determined helium levels for Ringhals thimble tubes. In an earlier study two tube sections at *33 and *70 dpa positions of a peak 75 dpa tube was removed from the Ringhals reactor and examined by electron microscopy [19]. As shown in Fig. 11 the microstructure of the steel at *70 dpa was dominated by a very high density of gas-ﬁlled nano-cavities. A lesser but still signiﬁcant concentration of nano-cavities was observed in the 33 dpa section [19].

Predictions and Measurements of Helium and Hydrogen …

661

Fig. 11 High density of nano-cavities observed in a Ringhals flux thimble tube at 70 dpa. The cavities are thought to contain both helium and hydrogen [19]

Measurement of the gas content in another section of the same tube at 65 dpa revealed that *625 appm of helium and *2500 appm of hydrogen were in the steel [19]. The hydrogen was several times larger than expected from transmutation alone and it was deduced from this and other studies [9–12] that, in addition to transmutant hydrogen, there was a large amount of environmentally-generated hydrogen also being stored in the helium-nucleated cavities. There are other helium determinations that were performed by researchers on thimble tubes extracted from various PWRs [20–22]. In one of the Japanese-led PWR studies seven data points were derived from different positions along a single tube, thereby providing data going through the over-the-shoulder region of flux-spectra [20]. In another study conducted by the International IASCC Advisory Committee tubes from three different reactors were examined (Ringhals, H.B. Robinson Unit #2, Beaver Valley Unit #2) [22]. This study has two Ringhals data points at 35 and 76 dpa from a single tube to add to the single data point from the study by Edwards [19], allowing a comfortable extrapolation to the 100 dpa level with no need to extrapolate to the 75 dpa level of the current study. One data point in this study was generated beyond the shoulder region, allowing another check on our concern about over-the-shoulder spectral variations. With the exception of the Fukuya–Fujii specimens at 12.65% nickel, all the various steels had nickel levels of 13.3%. Note that the 76 dpa Ringhals tube in the IASCC study was measured to contain 1055 appm of helium in the flat portion of the flux-spectra, so the extrapolation to 100 dpa using a linear dependence on thermal neutron fluence is expected to produce a very high helium level on the order of *2000 appm and perhaps more, as shown in Fig. 12 in the right-side graph. Such a high level has not previously been observed in PWR components and prudence requires that this extrapolated estimate be experimentally conﬁrmed.

662

F.A. Garner et al.

Fig. 12 Compilation of helium measurements from tubes removed from various PWR reactors. The Fukuya–Fujii data in this plot have not been corrected for their lower nickel level, 12.6 versus 13.3%

Direct measurements of both gases can be made in neutron-irradiated material using mass dilution isotopic spectroscopy [6–8], but this technique cannot be used following additional accumulation of ion-induced dpa in “preconditioning” experiments due to the short range of injected ions. Figure 12 also shows that the helium generation rate per dpa in this data ensemble bottoms out a little below *10 appm per dpa, a rate similar to that shown for the baffle bolt shown in Fig. 8. This data ensemble is not single-variable in that both the dpa level (determined by fast neutrons) and the helium per dpa rate (determined by the T/F ratio) are varying with position on the tube, a situation that partially obscures the dose dependence. The sharp increase of the He/dpa values at low doses, especially below *20 dpa, conﬁrms our earlier-expressed concern that the increasing T/F neutron ratio outside of the core might not allow a straight-forward comparison of flat and over-the-shoulder sections of the Ringhals tube. However, looking at the relative invariance of the helium/dpa ratio in the 30–65 dpa region, it appears to be fortuitous that the over-the-shoulder concern will not signiﬁcantly impact the proposed conduct of this experiment using the 50 and 75 dpa sections of the 100 dpa tube. It appears that there may possibly be an increase in the He/dpa ratio above *60 dpa, an observation that would influence the extrapolation to 100 dpa, but this tendency is not clearly evident in this data ensemble. We can test this possibility, however, by using the measured helium content as a retrospective dosimeter to determine the total neutron fluence, the current concentration of 59Ni and current rate of helium generation. Since the overwhelming majority of the helium is produced by thermal neutrons the relationship is well established. Figure 13 shows the predicted helium production via the 59Ni reaction versus thermal fluence in both pure Ni [3] and in 316 stainless steel calculated with 13.3% Ni. Eventually, all 58Ni will be consumed and all 59Ni will be consumed sometime thereafter, leading to a saturation of helium at very high neutron exposures, as shown in the left-hand graph. However, we are more interested in neutron

Predictions and Measurements of Helium and Hydrogen …

663

Fig. 13 Production of helium via the 59Ni (n, a) reaction in pure nickel [3] and in 316 stainless steel with 13.3% nickel. The accumulated helium concentration depends only on the thermal neutron fluence

exposures typical of thimble tubes in the 10–20 year residence period, involving thermal neutron exposures of 0.1 MeV) equals *7 dpa in PWR spectra [1, 2], a dose of 76 dpa is generated by 1.1 1023 fast

664

F.A. Garner et al.

neutrons (E > 0.1 MeV), indicating that the life-averaged T/F ratio of this section is *3/11 = 0.27. Such a moderately high T/F value is consistent with the spectral environment expected for an unfueled tube surrounded by water.

Recommendations on Gas Co-injection Rates for Ion Irradiations Having established the *2000 appm predicted level for the 100 dpa specimen, what helium coinjection rate should be speciﬁed for the dose range between 100 and 160 dpa, representing a 60% extension in dose? Note that Fig. 7 indicates that the 59Ni concentration will fall 1043 887 16.6 78 dpa, 3 at.% Center Unknown 1091 N/A 1136 1136 2.3 He, (speculated Tirr = 180 °C to be TB) HAGB high angle grain boundary, TB twin boundary, CRSS critical resolved shear stress

Discussion The bulk yield strength of non-irradiated Inconel X-750 material is noted to be 1070 MPa in [18]. The new push-to-pull, in situ, micro-tensile testing method presented here measures the yield strength of two non-irradiated specimens to be 1001 MPa and 1043 MPa, respectively. It should be noted that this new SSMT technique estimates the bulk yield strength of the material well, to within 70 MPa, based on initial test cases. This is due to the fact that the strength determining features in non-irradiated Inconel X-750 are the second-phase ordered c′ nanoprecipitates. Because these precipitates average 15 nm in size and are dispersed throughout the matrix, specimen size effects can be considered negligible since micro-tensile specimens are *1–1.5 lm in length and width and *2.5 lm in gauge length and thus contain many of the c′ nanoprecipitates. Using a volume fraction for c′ of 4.5%, which is the lower bound for the Inconel X-750 material studied here as reported in [24], and the average volume of a micro-tensile specimen, it can be calculated that >600 ordered nanoprecipitates are contained in the non-irradiated tensile specimens on average. Thus, the specimens should behave similar to bulk specimens because c′ remains the most important strength determining feature and is abundant. Slight deviations from the bulk value can be attributed to the heterogeneous, non-uniform distribution of c′ throughout the

754

C. Howard et al.

matrix as mentioned in [19], as well as contributions from the differences in orientation of the weakly oriented single grains which deform with respect to the loading direction compared to deformation, which is averaged over many different grains and grain orientations in a bulk sample. The good correspondence of micro-scale push to pull tensile tests with the bulk yield strength of Inconel X-750 in the case of non-irradiated specimens is very encouraging. However, in cases where observable plastic deformation in the form of slip steps is observed, grain orientation should be taken into account. A critical resolved shear stress (CRSS) parameter which accounts for the Schmid factor of the deformed grain will provide for a more accurate comparison between non-irradiated and irradiated specimens with plasticity. Preliminary analysis on the two non-irradiated specimens with signiﬁcant plasticity produce CRSS parameters in good agreement, 454 ± 2 and 461 MPa respectively. It has been well documented that cold working prior to precipitation hardening increases the fraction of cellular carbides, and this structure is associated with low ductility and a greater degree of chromium depletion in the surrounding matrix area [25, 26]. In the case of our Inconel X-750 material, the components underwent a cold coiling process after the aging treatment, but micro-tensile testing on non-irradiated outer edge specimens suggests similar low ductility. It is possible that the grain boundary failure can be attributed to stress concentrations from the presence of a large number of grain boundary carbides and η phase Ni3Ti precipitates which replaces the c′ strengthening nanoprecipitates at the outer edge. Rather abrupt grain boundary decohesion with associated flaking in the case of the high angle grain boundary outer edge specimen was observed. Another possibility to account for the localized failure at the grain boundaries is the zone denuded of c′ precipitates, a precipitate free region (PFZ), that is typically found in these alloys. The combination of carbide accumulation and η phase accumulation on the boundaries and a c′ PFZ would mean that easiest pathway for dislocation motion would be within the limited space of the precipitate free region (PFZ) directly adjacent to the grain boundary. The localized deformation in the PFZ and the grain boundary precipitate structure [27] could account for the intergranular failure mode observed, i.e. flaking along the boundary surface and multiple dimples within the fracture surface seen in Fig. 4a. Mills also recognized that plasticity around the grain-boundary carbides leads to a stress concentration at the carbide-matrix interface, which results in a decohesion between the carbide particles and the nickel matrix. The M23C6 carbides serve as initiation points for micro-voids, which grow and coalesce within the grain boundary denuded zone leading to intergranular failure via a dimple rupture mechanism at room temperature [28]. Unlike high angle grain boundaries, twin boundaries arise as a result of mirrored stacking fault errors and do not serve as nucleation sites for M23C6 carbides. Thus, as seen in Fig. 4b, the twin boundary does not fracture, although losses in ductility due to cold-working are still evidenced.

In Situ SEM Push-to-Pull Micro-tensile Testing …

755

While signiﬁcant losses in ductility are evidenced in the heavily cold-worked outer edge of non-irradiated components, the grain boundary strength of the specimen tested from this region is much higher than the yield strength of the same component in its center. Thus, non-irradiated components will yield in the center long before fracturing at their outer edge. It is thought that after high dose and lengthy service in CANDU reactors, the grain boundary strength (fracture strength) of the material approaches the original non-irradiated yield strength of the component, 1136 MPa, with ductility almost completely erased. It is possible that, because the irradiated specimen that was tested had a high concentration of He bubbles segregated at all boundaries and interfaces as seen in [16] a different grain boundary failure mechanism exists such that the grain boundary ruptures along the bubble-matrix interface. Although there are complications caused by surface ﬁnishing for material tested at the edge of the specimens, both the bulk component crush tests and the micro-scale testing (Table 2) indicates that the ductility of the irradiated material is lower than the non-irradiated material. However, whereas the bulk component crush tests show a consistently lower failure load for the irradiated material, to date only one successful micro-scale irradiated test has been performed, which appears to lie at the lower bound of the non-irradiated failure strength. This may reflect the fact that small volumes of material are sampled in the latter case, and additional tests must be performed. Failure in the bulk tests is dictated by the weakest part of the material. Because bulk tests sample many different parts simultaneously, there is a better chance of sampling the weakest parts thus establishing a clearer difference between the strength of the non-irradiated compared with the irradiated material. Nevertheless, by identifying and testing at particular features, such as high angle grain boundaries, the small scale testing showed a reduction in failure strength (boundary strength) of 456 MPa after irradiation that is somewhat consistent with the crush test data. Above all, the small scale testing provides a direct measure of the mechanical properties that can only be inferred from ﬁnite element analyses in the case of bulk component crush test material. Continuing work using this micro-tensile testing method on more irradiation doses and temperatures is ongoing in order to better understand and quantify the reductions in grain boundary strength and ductility. In addition, single-grained specimens will be fabricated and tested in order to quantify changes in the matrix’s CRSS as a function of dose and irradiation temperature. Transmission Electron Microscopy (TEM) will be used to investigate the relationships between helium bubble grain boundary coverage and grain boundary strength. Attempts to observe fracture surfaces from broken micro-tensile tests in the TEM will also be made. This will help to better understand matrix-bubble and bubble-grain boundary interactions under load to failure and provide input for improving models such as those discussed in [29].

756

C. Howard et al.

Conclusions A ﬁrst-of-its-kind, in situ, micro-tensile testing method was successfully developed for use in a Scanning Electron Microscope (SEM) for real time quantitative mechanical data acquisition and deformation imaging on ex-service Inconel X-750 components. This new SSMT technique characterized the yield strength, failure strength, and failure mechanisms of non-irradiated and high dose (78 dpa, 3 at.% He, Tirr = 180 °C) material by fabricating specimens containing single grain boundaries in order to assess the material’s grain boundary strength as a function of irradiation condition after it had been found to fail in an intergranular manner after spending signiﬁcant time in a CANDU reactor. Specimen dimensions on the order of 1 lm 1 lm 2.5 lm (thickness width gauge length) provided yield strength values within 70 MPa of the bulk value of 1070 MPa for non-irradiated material and quantiﬁed relative differences due to pre-existing cold working in the control material in terms of failure strengths and ductility losses. Boundary strength is estimated to be reduced by *456 MPa after irradiation to 78 dpa at an average irradiation temperature of 180 °C. The ductility in the center of the components decreases from 16.6% total elongation to 2.3% after the same irradiation conditions. Acknowledgements The authors of this manuscript would like to acknowledge Canadian Nuclear Laboratories and the CANDU Owners Group (COG) for their donation of sample material and the Nuclear Science User Facility (NSUF) sample library at Idaho National Lab (INL) operated through the U.S. Department of Energy (DOE), Nuclear Energy University Program (NEUP), Rapid Turnaround Post-Irradiation Experiment (RT-PIE) for instrument time, sample management and preparation. Grant Bickel and Don Metzger are acknowledged for useful discussion, and Marc Paulseth for temperature estimates. The authors would like to thank COG for ﬁnancial support for some of this work and permission to use the data. In addition, the authors would like to thank the Biomolecular Nanotechnology Center (BNC) at the University of California, Berkeley (UCB) for the use of the FEI Quanta 3D FEG. We would also like to recognize INL microscopist James W. Madden for sample preparation and Leandro Von Werra at the University of Bern for use of his digital image correlation software.

References 1. W.J. Mills, B. Mastel, Deformation and fracture characteristics for irradiated Inconel X-750. Nucl. Technol. 73(1), 102–108 (1986) 2. M. Grifﬁths. The effect of irradiation on Ni-containing components in CANDU® reactor cores: a review. AECL Nucl. Rev. 2(1), (2013) 3. L.C. Walters, W.E. Ruther, In-reactor stress relaxation of Inconel X-750 springs. J. Nucl. Mater. 68, 324–333 (1977) 4. J.J. Olivera et al., Failure of Inconel X-750 bolts of internals of the CHOOZ-A nuclear power plant. CEA Centre d’Etudes Nucleaires de Fontenay-aux-Roses, 92 (France). Dept. d’Analyse de Surete, 1989 5. L.J. Balog, D.E. Boyle, Four pin mounting system for nuclear reactor control rod guide tubes. U.S. Patent No. 4,937,039. 26 Jun. 1990 6. Y. Katayama et al., Jet pump beam and method for producing the same. U.S. Patent No. 8,879,683. 4 Nov. 2014

In Situ SEM Push-to-Pull Micro-tensile Testing …

757

7. A.W. Fanning, W.G. Jameson Jr., V.E. Hazel, Nuclear fuel assembly. U.S. Patent No. 4,314,884. 9 Feb. 1982 8. Y.J. Kim et al., Fuel rod assembly and method for mitigating the radiation-enhanced corrosion of a zirconium-based component. U.S. Patent No. 8,792,607. 29 Jul. 2014 9. B.H. Sencer et al., Microstructural evolution of alloy 718 at high helium and hydrogen generation rates during irradiation with 600–800 MeV protons. J. Nucl. Mater. 283 (2000), 324–328 10. F. Carsughi et al., Investigations on Inconel 718 irradiated with 800 MeV protons. J. Nucl. Mater. 264(1), 78–88 11. N. Hashimoto et al., Microstructural analysis of ion-irradiation-induced hardening in Inconel 718. J. Nucl. Mater. 318, 300–306 (2003) 12. J.D. Hunn et al., Ion-irradiation-induced hardening in Inconel 718. J. Nucl. Mater. 296(1), 203–209 (2001) 13. Ji-Jung Kai, R.D. Lee, Effects of proton irradiation on the microstructural and microchemical evolution of Inconel 600 alloy. J. Nucl. Mater. 207, 286–294 (1993) 14. M. Grifﬁths, G.A. Bickel, S.A. Donohue, P. Feenstra, C.D. Judge, D. Poff, L. Walters, M.D. Wright, L.R. Greenwood, F.A. Garner, Degradation of Ni-alloy components in a CANDU reactor core. 16th International Symposium on Environmental Degradation in Materials, Asheville, NC (2013) 15. C.D. Judge, M. Grifﬁths, L. Walters, M. Wright, G.A. Bickel, O.T. Woo, M. Stewart, S.R. Douglas, F. Garner, Embrittlement of nickel alloys in a CANDU reactor environment, in Effects of Radiation on Nuclear Materials, vol. 25, ASTM International, ed. by T. Yamamoto (Anaheim, CA, 2012), pp. 161–175 16. C.D. Judge, N. Gauquelin, L. Walters, M. Wright, J.I. Cole, J. Madden, G.A. Botton, M. Grifﬁths, Intergranular fracture in irradiated Inconel X-750 containing very high concentrations of helium and hydrogen. J. Nucl. Mater. 457, 165–172 (2015) 17. H.K. Zhang, Z. Yao, G. Morin, M. Grifﬁths, TEM characterization of in-reactor neutron irradiated CANDU spacer material Inconel X-750. J. Nucl. Mater. 451, 88–96 (2014) 18. Special Metals Corporation Publication No. SMC-067, Sept 2004 19. O.T. Woo et al., The microstructure of unirradiated and neutron irradiated Inconel X750. Microsc. Microanal. 17(2), (2011) 20. L.R. Greenwood, A new calculation of thermal neutron damage and helium production in nickel. J. Nucl. Mater. 115, 137–142 (1983) 21. L.R. Greenwood, F.A. Garner, Hydrogen generation arising from the 59Ni (n, p) reaction and its impact on ﬁssion-fusion correlations. J. Nucl. Mater. 233-237(2), 1530–1534 (1996) 22. L.R. Greenwood, R.K. Smither, SPECTER: neutron damage calculations for materials irradiations. ANL/FPP/TM-197, Argonne National Laboratory (1985). https://www-nds.iaea. org/irdf2002/codes/index.htmlx 23. https://www.hysitron.com/media/1637/ptpou-r1f.pdf 24. J.W. Ha et al., Effect of cold drawing ratio on c′ precipitation in Inconel X-750. Mater. Charact. 96, 1–5 (2014) 25. J.L. Nelson, S. Floreen, Second International Symposium on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors (American Nuclear Society, LaGrange Park, 1986), p. 632 26. K. Hosoi, S. Hattori, Y. Urayama, I. Masaoka, R. Sasaki, International Symposium on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors (NACE, Houston, 1984), p. 34 27. T. Yonezawa, K. Onimura, O. Sakamoto, N. Sasaguri, H. Nakata, H. Susukida, International Symposium on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors (NACE, Houston, 1984), p. 345 28. W.J. Mills, Effect of temperature on the fracture toughness behavior of Inconel X-750. Fractography and Materials Science (ASTM International, Pennsylvania, 1981) 29. D.V. Yuryev, M.J. Demkowicz, Modeling growth, coalescence, and stability of helium precipitates on patterned interfaces. Modell. Simul. Mater. Sci. Eng. 25(1), 015003 (2016)

Microstructural Characterization of Proton-Irradiated 316 Stainless Steels by Transmission Electron Microscopy and Atom Probe Tomography Yun Soo Lim, Dong Jin Kim and Seong Sik Hwang

Abstract Proton irradiation is a well-known useful experimental technique to study neutron irradiation-induced phenomena in reactor core materials. Type 316 austenitic stainless steel was irradiated with 2 meV protons to doses up to 10 displacement per atom at 360 °C, and the various effects of the proton irradiation on the microstructural changes were characterized with transmission electron microscopy and atom probe tomography. Typical irradiation damage mainly consisted of small dislocation loops, cavities, tiny precipitates and network dislocations. Ni and Si were enriched, whereas Cr, Mn and Mo were depleted on the grain boundaries associated with irradiation-induced segregation. Ni–Si rich clusters were also found in the matrix. A new method to prepare TEM specimens of a proton-irradiated material is suggested, which was shown to be a relatively simple and effective method to chemically eliminate the inherent surface damage induced by a conventional high-energy focused ion beam and subsequent low-energy ion milling treatments.

Keywords Stainless steel Proton irradiation TEM APT Defects Radiation induced segregation

Introduction The proper management of irradiation-assisted stress corrosion cracking (IASCC) of internal components of a pressurized water reactor (PWR) is considered critical for safe long-term operation. Some cracking of internals made of stainless steel (SS), such as guide tube support pins, baffle former bolts and so on, has been

Y.S. Lim (&) D.J. Kim S.S. Hwang Nuclear Materials Research Division, Korea Atomic Energy Research Institute, 1045 Daedeok-Daero, Yuseong-Gu, Daejeon 305-353, South Korea e-mail: [email protected] © The Minerals, Metals & Materials Society 2019 J.H. Jackson et al. (eds.), Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, The Minerals, Metals & Materials Series, https://doi.org/10.1007/978-3-030-04639-2_49

759

760

Y.S. Lim et al.

identiﬁed [1, 2]. The IASCC mechanism is not fully understood; however, it appears to be closely related to microstructural defects caused by neutron irradiation and their resultant effects on cracking under an environment characterized by high fluence, stress, and temperature levels. The changes in the microstructure of austenitic stainless steels under neutron irradiation consist of the formation of point defects, interstitial dislocation loops (Frank loops), voids, cavities, precipitates and/or solute clusters (Ni–Si, c′ phase Ni3Si and carbides). Moreover, the chemical composition of a grain boundary is modiﬁed under irradiation. Cr depletion, and Ni and Si enrichment are observed. Point defects can enhance solute atom diffusion resulting in enhanced precipitation, or couple differentially with solute and solvent atoms resulting in radiation induced segregation (RIS). Therefore, studies on microstructural changes should be the ﬁrst step to understand how neutron irradiation defects affect the cracking behavior in these alloys. Proton irradiation is a useful experimental technique to study irradiation-induced phenomena in nuclear core materials instead of neutrons [3, 4]. The challenging issue is that proton irradiation-induced defects accumulate in the vicinity of the surface within a narrow range of about a few tens of micrometers. Recently, a focused ion beam (FIB) system has been shown to be a powerful tool for the preparation of transmission electron microscopy (TEM) specimens of ion irradiated materials. The most unique advantage of the FIB system for TEM specimen preparation is that speciﬁc regions and/or orientations of interest can be picked with precise accuracy for cross-sectional as well as plane-view conﬁgurations. However, one of the drawbacks of the FIB technique is that TEM specimens inevitably suffer from some amount of damage at the surface because their preparation involves the interaction with a highly energetic ion beam. A number of studies have reported on FIB-induced damages introduced during the TEM sample preparation [5, 6]. A subsequent treatment with low energy ion milling (IM) using Ar ions after the FIB process is the most popular way to resolve the problem. However, contamination, re-deposition and extensive destructive damage can also develop during the IM process even when using low energy Ar, which also has impeded TEM analyses. Therefore, a post treatment to eliminate completely the surface artifacts produced during the FIB process is critical to reveal the original defects by the irradiation itself for the TEM specimens prepared by the FIB and/or IM methods. The objectives of the present work were to investigate the irradiation defects and RIS using TEM and atom probe tomography (APT) in type 316 austenitic SS after proton irradiation and to discuss their influence on the mechanical properties and IASCC. A newly developed chemical polishing (CP) method is also presented, which was shown to be a relatively simple and effective method to chemically eliminate the surface defects induced by the FIB and/or IM methods during the TEM sample preparation of proton-irradiated materials.

Microstructural Characterization of Proton-Irradiated …

761

Experimental Materials and Proton Irradiation Type 316 austenitic SS, commonly used as forged bolts near the core region of a PWR, was used in this study. The chemical composition and mechanical properties of the alloy are given in Tables 1 and 2, respectively. The test alloy was solution annealed at 1100 °C and ﬁnally water quenched. Before the proton irradiation, the surfaces of the specimens were mechanically ground down with SiC sand papers No. 2400 and then electrochemically polished in a solution of 50 vol.% phosphoric acid + 25 vol.% sulfuric acid + 25 vol.% glycerol for 15–30 s at room temperature. The proton irradiation was performed with the General Ionex Tandetron accelerator at the Michigan Ion Beam Laboratory at the University of Michigan. The irradiation processes were conducted using 2.0 meV protons at a current range of 40 lA. Details of the irradiation procedure can be found in the literature [7]. The specimens were exposed to four levels of irradiation of 0.4, 1.6, 2.7, and 4.2 displacements per atom (dpa) at 360 °C and were subsequently cooled for 3–7 days at room temperature to permit the short-lived isotopes to decay. The radiation damage levels in dpa of the samples were calculated with the Stopping and Range of Ions in Matter (SRIM) Version 2008 program using a displacement energy of 40 eV [8] in the ‘quick calculation’ mode. It has recently been clearly shown that the Kinchin-Pease mode (quick calculation in SRIM) is the correct method for producing dpa values relevant to neutron dpa, while calculations in the ‘full cascade’ mode in SRIM over-estimate the dose by a factor of 2.4 for a comparison to neutron irradiation levels [9].

Sample Preparation of the Irradiated Materials The specimens for optical microscopy and SEM were prepared by chemical etching in a solution of 2 vol.% bromine + 98 vol.% methanol. TEM foils containing the irradiated area were prepared with conventional electropolishing (EP) and focused

Table 1 Chemical composition of the 316 austenitic SS under test (wt%) Cr

Ni

C

Mo

Mn

Si

Al

Cu

P

S

Fe

16.14

10.41

0.047

2.11

1.08

0.66

0.1

0.1

0.003

0.001

Bal.

Table 2 Mechanical properties of the 316 austenitic SS under test YS (MPa)

UTS (MPa)

Elongation (%)

301

574

67

762

Y.S. Lim et al.

ion beam (FIB) milling techniques. To prepare the TEM specimens with an EP method, thin foils were mechanically thinned until they were less than 40 lm thick and then electro-jet polished in a 7 vol.% perchloric acid + 93 vol.% methanol solution at—40 °C with a current of approximately 80 mA. For the FIB TEM sample preparation, a dual-beam Hitachi FIB-2100 system was used with a Ga+ incident beam energy of 30 kV and a current of 1–5 nA. Subsequent ion milling (IM) was done for the FIB TEM specimens with Ar ions with an incident beam energy level of 300 V at an incidence angle of 10° for 10 min. Chemical polishing (CP) of the FIB and/or IM TEM specimens to eliminate the damages induced by energetic Ga ions during the FIB and/or Ar ions during the IM was carried out in dilute bromine (Br) solutions. Before applying CP to a TEM specimen, the thickness reduction rate by the CP was estimated with a weight loss method [10] using non-irradiated 316 SS plates with dimensions of 10 10 1 mm3. It is known that general corrosion occurs with dilute Br solutions without any localized corrosion, such as pitting and intergranular attack, on Ni-based alloys and stainless steels.

Microstructural Analysis The proton-irradiated specimens were investigated using various types of microscopic equipment. The SEM imaging and orientation imaging microscopy (OIM) by electron back-scattered diffraction (EBSD) were done with the JEOL 5200 (operating voltage 25 kV) and JEOL 6300 (operating voltage 20 kV), respectively. TEM analysis was done with a JEOL JEM-2100F (operating voltage 200 kV) to determine the typical microstructural changes due to the proton irradiation. Speciﬁcally, of interest was the RIS at the grain boundaries and the formation of irradiation defects including black dot damage, Frank loops, voids, precipitates and clusters. The defects in irradiated and non-irradiated microstructures were observed with bright/dark ﬁeld imaging and week beam techniques in a two-beam condition. APT specimens were prepared by a standard approach based on FIB lift-out and annular milling. APT analyses were performed using the CAMECA 4000X HR local electrode atom probe (LEAP) microscope operated in the voltage mode with a specimen temperature of 45 K and an evaporation rate of *0.005 atom/pulse. Data reconstruction and data analysis were performed using the IVAS 3.6.4 software from CAMECA. RIS and Ni-Si clustering were analyzed using a customized approach based on the maximum separation distance.

Microstructural Characterization of Proton-Irradiated …

763

Results and Discussions Microstructure of the Non-irradiated Material Figure 1 shows the typical microstructure of the as-received alloy before proton irradiation. From Fig. 1a showing an OIM image obtained from the EBSD measurement, it is clear that the alloy has a homogeneous and isotropic microstructure with many twin boundaries. The relative population of the twin boundaries to the total for the boundaries was measured to be 36–43% depending on the measurement location. Type 316 austenitic SS has a low stacking fault energy [11]; therefore, twins as well as stacking faults can easily be generated. Because the alloy was solution annealed without an additional heat treatment, Cr carbides were rarely found on the grain boundaries and inside the grains as well shown in Fig. 1b, c. Images originating from stacking faults are frequently seen in Fig. 1c. The

Fig. 1 a OIM. b SEM. c TEM images of the as-received 316 SS before proton irradiation

764

Y.S. Lim et al.

dislocation density was very low in Fig. 1c where, the image was obtained in a two-beam condition with a [110] beam direction, as shown in the ﬁgure insert.

Microstructure of the Proton-Irradiated Materials The irradiation damages depending on the dpa and the depth of 316 SS were calculated with the Stopping and Range of Ions in Matter (SRIM) 2008 program using a displacement energy of 40 eV in the ‘quick calculation’ mode. These results are given in Fig. 2a. The damage proﬁles calculated using the full cascade mode exhibit plateaus roughly up to a depth of 15 lm, as well as subsequent damage peaks near the depth of 20 lm from the irradiated surfaces. Figure 2b is a SEM image showing a typical cross-sectional view of a specimen proton-irradiated to 2.7 dpa. A ditch-like band is visible at about 20 lm deep from the irradiated surface, and it was approximately 2 lm wide shown in Fig. 2b. This ditch-like band shows a severely etched area. The depth and the width of the ditch-like band are in good agreement with those of the irradiation damage peak calculated with the SRIM for the 316 SS in Fig. 2a. The material in the damage peak near the depth of 20 lm and the width of 2 lm from the surface suffers from severe damage by proton irradiation. Therefore, this area can experience a drastic change in the microstructure leading to a different etching behavior from the surrounding areas as shown in Fig. 2b. Irradiation of materials with particles that are sufﬁciently energetic to create atomic displacements can induce a signiﬁcant microstructural alteration, such as dislocation loops, network dislocations, dislocation channeling, radiation induced precipitation (RIP), voids, cavities, RIS and so on [12, 13]. To reveal the microstructural defects in the 316 SS generated by proton irradiation, TEM

Fig. 2 a Calculated dpa-depth proﬁles of the proton-irradiated 316 SS using the ‘quick calculation’ mode of the SRIM software. b SEM image showing the cross-sectional view of the specimen proton-irradiated to 2.7 dpa

Microstructural Characterization of Proton-Irradiated …

765

Fig. 3 TEM images showing a cavities, b Frank loops in bright-ﬁeld, and c Frank loops in dark-ﬁeld in the proton-irradiated 316 SS

specimens were prepared with a typical EP method. The locations of the TEM examination were roughly at a depth of 10–15 lm. Signiﬁcant irradiation-induced defects were clearly visible in the TEM images, and the results for various types of defects found in the proton-irradiated 316 SS are given in Fig. 3. Figure 3a is an image showing the micro-cavities found in the specimen irradiated to 0.4 dpa. The ﬁgure was obtained in the under-focus condition, from which the image of the cavities could be visualized clearly. Figure 3b shows a bright-ﬁeld image of Frank loops obtained in the specimen irradiated to 1.6 dpa. Finally, Fig. 3c shows a dark-ﬁeld image of Frank loops obtained using the rel-rod method from the specimen irradiated to 2.7 dpa. All the present ﬁndings are in good agreement with the previously reported results on neutron and proton-irradiated stainless steels [14–17]. The density of the network dislocations also increased considerably by proton irradiation, and the degree of increase seemed to depend on the dose of the proton irradiation. Figure 4 shows the changes of the dislocation morphology in the specimens that were in a non-irradiated condition (Fig. 4a), irradiated to 0.4 dpa (Fig. 4b), and irradiated to 1.6 dpa (Fig. 4c). All the images were obtained in a two beam condition with an electron beam direction of [110] zone axis. The as-received alloy was solution annealed at a high temperature of 1100 °C; therefore, the dislocation density was low before the proton irradiation. From the ﬁgures, it is obvious that proton irradiation induces many dislocations, especially in the form of

766

Y.S. Lim et al.

Fig. 4 Dislocation morphologies in a non-irradiated, b 0.4 dpa proton-irradiated, and c 1.6 dpa proton-irradiated 316 SS

a network dislocation structure. During the irradiation, the microstructural evolution typically involves the formation and growth of faulted dislocation loops, loop unfaulting to create perfect dislocation loops, and then loop interaction/ impingement to form network dislocation structures [12]. The network dislocations are typically randomly distributed and are often heavily jogged as opposed to the relatively straight dislocations found in non-irradiated metals, as shown in the ﬁgures.

APT Analysis of the Proton-Irradiated Materials The degree of grain boundary segregation can change depending on the grain boundary character [18]. Therefore, the APT specimens for the observation of the grain boundary segregation in the present study were taken from random high angle grain boundaries. From the APT analysis of a non-irradiated specimen, it was found that only Mo, C, B and P were lightly enriched at a grain boundary before the proton irradiation. This ﬁnding agrees well with the previous result obtained from the solution annealed, non-irradiated austenitic 316 SS [19]. Figure 5 shows

Microstructural Characterization of Proton-Irradiated …

767

Fig. 5 a, b Concentration proﬁles of solute atoms across a grain boundary in 316 SS proton-irradiated to 2.7 dpa obtained from APT

Fig. 6 a Atom maps of Ni and Si, and b concentration proﬁles of Ni and Si across clusters in 316 SS proton-irradiated to 2.7 dpa obtained from APT

one-dimensional element concentration proﬁles across a grain boundary of the specimen irradiated to 2.7 dpa. The concentration proﬁles were obtained along a cylinder 3 nm in diameter positioned normal to a grain boundary. The main element Fe did not change signiﬁcantly at a grain boundary; however, the concentrations of Cr, Mo and Mn decreased at the grain boundary, while Ni, Si, B and P were enriched. The width of the segregated zone for the elements was approximately 10 nm. The phenomenon of Cr depletion and Ni and Si enrichment at a grain boundary due to proton or neutron irradiation is well known [4, 7, 18, 20]. From a recent APT observation of irradiated stainless steels [20], it was also identiﬁed that Mo and Mn were depleted at a grain boundary. It is known that Ni and Si elements segregate at grain boundaries and also react with each other to form clusters in the matrix. Figure 6a shows the spatial

768

Y.S. Lim et al.

distribution of the identiﬁed Ni- and Si-rich clusters, and Fig. 6b is the concentration proﬁles of the Ni and Si across clusters in the specimen proton-irradiated to 2.7 dpa. The concentrations of Ni and Si periodically change from high to low values in roughly the same manner in Fig. 6b, which means that the Ni–Si clusters are distributed over the specimen. From Fig. 6b, it can be presumed that the average size of the Ni–Si clusters is below 10 nm. Chen et al. [21] studied solute clustering in 304 SS proton-irradiated to 10 dpa at 360 °C, and found that proton irradiation caused signiﬁcant changes in the microstructure with segregation of Ni, Si and P to dislocations and the formation of high number densities of Ni–Si clusters. Ni–Si clusters in the matrix have been regularly found in neutron- and proton-irradiated stainless steels, and they are identiﬁed as the c′ phase [22, 23], which has a Ni3Si composition and which is known consist of typical radiation-induced precipitates [17].

Effect of Irradiation on the Mechanical Properties and IASCC The characteristics of microscopic changes by irradiation and their role in changes of material behavior and IASCC have been extensively studied [24, 25]. Irradiation of a material causes hardening, RIS and localized deformation through microstructural and microchemical changes. Radiation-induced hardening basically originates from the introduction of a large amount of small Frank loops (Fig. 3) and network dislocations (Fig. 4) in the material, which increases the yield stress and lowers the fracture toughness. The increase in the yield strength and the decrease in the fracture stress are large and rapid. Uniform elongation is also reduced sharply when a material is irradiated. As such, the alloy becomes signiﬁcantly embrittled in the PWR operation condition, which in turn makes it much more susceptible to IASCC by irradiation. RIS leads to the changes in the grain boundary composition. Undersized atoms such as Ni are enriched at the grain boundary, while oversized atoms such as Cr and Mn are depleted. In the present study, it was found that Cr, Mo and Mn were depleted at the grain boundary, while Ni, Si, B and P were enriched (Fig. 5). It was also reported that Cr depletion at the grain boundaries was closely correlated with component failures [26], which is a phenomenon very similar to that Cr depletion at the grain boundaries in austenitic alloys due to thermal sensitization as the main cause of enhanced intergranular cracking [27]. Another factor playing an important role in IASCC is localized deformation. The formation of intense deformation channels that transmit dislocations to the grain boundaries can result in localized slip or sliding of the grain boundary and cause the initiation of intergranular cracks [28]. As a result, the combined effect of hardening, RIS and localized deformation by irradiation can signiﬁcantly reduce the resistance to IASCC of a structural material under irradiation.

Microstructural Characterization of Proton-Irradiated …

769

Chemical Polishing Effects on the FIM TEM Specimens Figure 7 shows the CP effect on the non-irradiated 316 SS in a dilute Br solution. The total thickness reduction, calculated from the weight loss method [10], depending on time in a solution of 0.05 vol.% Br + 99.95 methyl alcohol is shown in Fig. 7a. As expected, the total thickness of a sample in a ﬁxed solution decreased in a roughly linear manner as the time increased. A SEM image showing the surface morphology after CP is shown in Fig. 7b. The sample was corroded uniformly over the entire surface without any intergranular attack. Only minor small pits by chemical polishing are visible in the ﬁgure. TEM images of differently prepared specimens for the non-irradiated 316 SS are shown in Fig. 8. All the images were obtained in a two beam condition. The FIB specimen itself (Fig. 8a) has a complicated surface damage layer caused by the interaction with a highly energetic Ga ion beam [7, 8], for which defects such as dislocations are invisible even in a two beam condition. Figure 8b shows the effect of the subsequent IM of the FIB specimen, which was done with Ar ions with an incident beam energy level of 300 V at an incidence angle of 10° for 10 min. In Fig. 8b, it can be seen that almost all of the surface damage layer generated during the FIB process was removed by ion milling, and some dislocations are visible. However, the appearance of dislocations is not clear and some defects generated by the FIB process are still present. Figure 8c shows a FIB specimen that was immersed in a solution of 0.2 vol.% Br + 99.8 vol.% methyl alcohol for 15 s to chemically remove the surface damage layer. Clarity is greatly improved, and no other defects by the FIB process are visible. Therefore, it can be conﬁrmed that the CP method is a powerful way to chemically eliminate the surface defects induced by FIB and/or IM during TEM sample preparation. This method was also identiﬁed to be reproducible and controllable. Moreover, it was shown that it could also make thick TEM specimens thin enough for electron transparency in a TEM observation as well.

Fig. 7 a Variation of the thickness reduction of 316 SS depending on time and b SEM image after CP in a solution of 0.05 vol.% Br + 99.95 vol.% methyl alcohol for 25 s

770

Y.S. Lim et al.

Fig. 8 TEM images showing the surface morphologies of the a FIB, b FIB ! IM, and c FIB ! CP processed 316 SS specimens

Conclusions In the present study, type 316 austenitic SS was irradiated using 2 meV protons up to 4.2 dpa at 360 °C, and the various defects generated by proton irradiation were characterized with transmission electron microscopy and atom probe tomography. Additionally, the effects of chemical polishing on the FIB TEM specimens were investigated to eliminate the surface damage layer caused during the FIB and/or IM treatments. The test alloy before proton irradiation had a homogeneous and isotropic microstructure, and the dislocation density was very low without any intergranularand intragranular precipitates. Typical irradiation damage mainly consisted of Frank loops, cavities, tiny precipitates and network dislocations. As the dose of the proton irradiation increased, the density of the network dislocations increased. Ni, Si, B and P were enriched, whereas Cr, Mn and Mo were depleted on the grain boundaries associated with RIS. Ni–Si rich clusters, identiﬁed as c’ phase from the previous studies, were also found in the matrix. All the ﬁndings were in good agreement with the previous reports.

Microstructural Characterization of Proton-Irradiated …

771

All of the microstructural and microchemical changes induced by irradiation degrade the mechanical properties of a material. Radiation-induced hardening originating from the introduction of a large amount of small Frank loops and network dislocations in a material causes embrittlement of the material. Cr depletion at the grain boundaries induced by RIS enhances the intergranular cracking of an irradiated material. Therefore, irradiation can signiﬁcantly reduce the resistance to IASCC of structural components installed near the core region of a PWR. A new method to prepare TEM specimens of a proton-irradiated material is suggested, which was shown to be a relatively simple and effective method to chemically eliminate the inherent surface damage induced by a conventional high-energy FIB and subsequent low-energy IM treatments. This method can also be used to make any type of overly thick TEM specimens thin enough for electron transparency. Acknowledgements This work was ﬁnancially supported by the Korean Nuclear R&D Program organized by the National Research Foundation (NRF) in support of the Ministry of Science, ICT and Future Planning (2017M2A8A4015155), and by the R&D Program of Korea Atomic Energy Research Institute.

References 1. J. McKinley, R. Lott, B. Hall, K. Kalchik, Examination of baffle-former bolts from D.C. cook unit 2, in Proceedings of the 16th International Conference on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2013 2. J. Pakarinen, U. Ehrnstén, H. Keinänen and W. Karlsen, Microstructural characterization of irradiated baffle bolts removed from a ﬁnnish VVER and a French PWR, in Proceedings of the 16th International Conference on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2013 3. J. Gan, G.S. Was, Microstructure evolution in austenitic Fe-Cr-Ni alloys irradiated with rotons: comparison with neutron-irradiated microstructures. J. Nucl. Mater. 297, 161–175 (2001) 4. K.J. Stephenson, G.S. Was, Comparison of the microstructure, deformation and crack initiation behavior of austenitic stainless steel irradiated in-reactor or with protons. J. Nucl. Mater. 456, 85–98 (2015) 5. J. Yu, J. Liu, J. Zhang, J. Wu, TEM investigation of FIB induced damages in preparation of metal material TEM specimens by FIB. Mater. Lett. 60, 206–209 (2006) 6. A. Aitkaliyeva, J.W. Madden, B.D. Miller, J.I. Cole, J. Gan, Comparison of preparation techniques for nuclear materials for transmission eletron microscopy (TEM). J. Nucl. Mater. 459, 241–246 (2015) 7. G.S. Was, J.T. Busby, T. Allen, E.A. Kenik, A. Jenssen, S.M. Bruemmer, J. Gan, A.D. Edwards, P.M. Scott, P.L. Andresen, Emulation of neutron irradiation effects with protons: validation of principle. J. Nucl. Mater. 300, 198–216 (2002) 8. ASTM Standard E693-01, Standard practice for characterizing neutron exposure in iron and low alloy steels in terms of displacements per atom (dpa) (2001) 9. R.E. Stoller, M.B. Toloczko, G.S. Was, A.G. Certain, S. Dwaraknath, F.A. Garner, On the use of SRIM for computing radiation damage exposure. Nucl. Instrum. Methods Phys. Res. B 310, 75–80 (2013)

772

Y.S. Lim et al.

10. ASTM G1-03, Standard practice for preparing, cleaning, and evaluating corrosion test specimens. ASTM, West Conshohocken, PA (2011) 11. R.E. Schramm, R.P. Reed, Stacking fault energies of seven commercial austenitic stainless steels. Metall. Trans. A 6A, 1345–1351 (1975) 12. S.J. Zinkle, Radiation-Induced Effects on Microstructure. in Reference Module in Materials Science and Materials Engineering: Comprehensive Nuclear Materials (Elsevier, Amsterdam, 2012), pp. 65–98 13. A.J. Ardell, P. Bellon, Radiation-induced solute segregation in metallic alloys. Curr. Opin. Solid State Mater. Sci. 20, 115–139 (2016) 14. A.R. Laborne, P. Gavoille, J. Malaplate, C. Pokor, B. Tanguy, Correlation of radiation-induced changes in microstructure/microchemistry, density and thermo-electric power of type 304L and 316 stainless steels irradiated in the Phénix reactor. J. Nucl. Mater. 460, 72–81 (2015) 15. Z. Jiao, Y. Chen, J. Hesterberg, E.A. Marquis and G. Was, Post-irradiation annealing in mitigation of IASCC in proton-irradiated stainless steel, in Proceedings of the 16th International Conference on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactor, 2013 16. W.V. Renterghem, M.J. Konstantinovic, M. Vankeerberghen, Evolution of the radiation-induced defect structure in 316 type stainless steel after post-irradiation annealing. J. Nucl. Mater. 452, 158–165 (2014) 17. P.J. Maziasz, Overview of microstructural evolution in neutron-irradiated austenitic stainless steels. J. Nucl. Mater. 205, 118–145 (1993) 18. C.M. Barr, G.A. Vetterick, K.A. Unocic, K. Hattar, X.-M. Bai, M.L. Taheri, Anisotropic radiation-induced segregation in 316L austenitic stainless steel with grain boundary character. Acta Mater. 67, 145–155 (2014) 19. M. Tomozawa, Y. Miyahara, K. Kako, Solute segregation on R3 and random grain boundaries in type 316L stainless steel. Mater. Sci. Eng. A 578, 167–173 (2013) 20. K. Fujii, K. Fukuya, Irradiation-induced microchemical changes in highly irradiated 316 stainless steel. J. Nucl. Mater. 469, 82–88 (2016) 21. Y. Chen, P.H. Chou, E.A. Marquis, Quantitative atom probe tomography characterization of microstructures in a proton irradiated 304 stainless steel. J. Nucl. Mater. 451, 130–136 (2014) 22. T. Toyama, Y. Nozawa, W. Van Renterghem, Y. Matsukawa, M. Hatakeyama, Y. Nagai, A. Al Mazouzi, S. Van Dyck, Irradiation-induced precipitates in a neutron irradiated 304 stainless steel studied by three-dimensional atom probe. J. Nucl. Mater. 418, 62–68 (2011) 23. W. Van Renterghem, A. Al Mazouzi, S. Van Dyck, Influence of post irradiation annealing on the mechanical properties and defect structure of AISI 304 steel. J. Nucl. Mater. 413, 95–102 (2011) 24. G.S. Was, S.M. Bruemmer, Effects of irradiation on intergranular stress corrosion cracking. J. Nucl. Mater. 216, 326–347 (1994) 25. P.L. Andresen, G.S. Was, Irradiation assisted stress corrosion cracking, in Comprehensive Nuclear Materials, ed. by R.J.M. Konings, T.R. Allen, R.E. Stoller, S. Yamanaka (Elsevier Ltd., Amsterdam, 2012), pp. 177–295 26. A.J. Jacobs, G.P. Wozadlo, G.M. Gordon, Low-temperature annealing: a process to mitigate irradiation-assisted stress corrosion racking. Corrosion 51, 731–737 (1995) 27. S.M. Bruemmer, G.S. Was, Microstructural and microchemical mechanisms controlling intergranular stress corrosion cracking in light-water-reactor systems. J. Nucl. Mater. 216, 348–363 (1994) 28. M.D. McMurtrey, B. Cui, I. Robertson, D. Farkas, G.S. Was, Mechanism of dislocation channel-induced irradiation assisted stress corrosion crack initiation in austenitic stainless steel. Curr. Opin. Solid State Mater. Sci. 19, 305–314 (2015)

Part IX

PWR Stainless Steel SCC and Fatigue—SCC

Microstructural Effects on Stress Corrosion Initiation in Austenitic Stainless Steel in PWR Environments D.R. Tice, V. Addepalli, K.J. Mottershead, M.G. Burke, F. Scenini, S. Lozano-Perez and G. Pimentel

Abstract Although service experience of austenitic stainless steels exposed to PWR primary coolant has been good, stress corrosion crack propagation has been observed in laboratory tests in the presence of 15% cold work. Data on crack initiation are much more limited and this study therefore aims to improve the understanding of the conditions under which crack initiation and subsequent development of stress corrosion cracking might be possible. Testing was performed on two heats of Type 304/304L stainless steel under slow strain rate tensile loading. A range of analytical techniques were used to characterize the resultant precursor features and cracking, and digital image correlation before and after testing was also used to evaluate the influence of localized deformation on SCC. The results indicate that crack initiation can occur in austenitic stainless steels exposed to good quality primary coolant under dynamic straining conditions; additional testing underway under more plant-representative conditions will be reported later. Signiﬁcant influences of steel microstructure on crack initiation susceptibility were observed.

D.R. Tice (&) V. Addepalli K.J. Mottershead Amec Foster Wheeler, Warrington, UK e-mail: [email protected] V. Addepalli e-mail: [email protected] K.J. Mottershead e-mail: [email protected] M.G. Burke F. Scenini The University of Manchester, Manchester, UK e-mail: [email protected] F. Scenini e-mail: [email protected] S. Lozano-Perez G. Pimentel University of Oxford, Oxford, UK e-mail: [email protected] G. Pimentel e-mail: [email protected] © The Minerals, Metals & Materials Society 2019 J.H. Jackson et al. (eds.), Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, The Minerals, Metals & Materials Series, https://doi.org/10.1007/978-3-030-04639-2_50

775

776

D.R. Tice et al.

Keywords Stress corrosion cracking Initiation Crack growth

SCC Stainless steel PWR LWR

Introduction In the primary circuit of pressurized water reactor (PWR) plants, stress corrosion cracking (SCC) of nickel-based alloys (mainly Alloy 600 and weld metals of similar composition such as Alloys 182 and 82) has been the main cause of in-service component degradation [1]. In contrast, Type 300 series stainless steels have generally performed well, with only a relatively small number of failures due to SCC. Evaluation of ﬁeld experience of austenitic stainless steels in PWRs [2–4] conﬁrmed that the majority of incidences of SCC in austenitic stainless steels in high-temperature PWR primary water occurred in occluded regions in the plant, where water quality can differ signiﬁcantly from that of free flowing bulk chemistry, including the presence of trapped oxygen and, in some cases, anionic impurities. However, after long-term PWR plant operation, a few instances of SCC have been reported in locations exposed to free-flowing, nominally good quality primary water chemistry. All these instances appear to be associated with relatively high levels of surface and/or bulk cold work [5]. Stainless steels are generally used in the solution annealed and quenched condition and there are restrictions on the amount of deformation allowed during component manufacture and assembly. Nevertheless, deformation cannot always be totally avoided, and can be introduced by processes such as machining, local grinding, weld-induced shrinkage, and bending or straightening during installation. Laboratory studies on austenitic stainless steels in PWR primary coolant environments have shown that propagation of an existing crack by stress corrosion cracking can occur if the material has been subjected to cold work to a level of about 15% or greater [6–9]. Crack growth rates can be high under some conditions, with rates of several millimetres per year having been measured for 20% cold work and loading normal to the cold working direction [7]. Despite these observations from crack growth tests, the good plant experience of stainless steels exposed to primary coolant suggests that crack initiation is difﬁcult under plant loading conditions [10]. The work described in this paper is part of an ongoing collaborative study between Amec Foster Wheeler and the Universities of Manchester and Oxford which is aimed at improving understanding of the conditions under which crack initiation and subsequent development of stress corrosion cracking of stainless steels in PWR primary coolant might be possible. The current paper focuses on the influence of material microstructure on SCC initiation.

Microstructural Effects on Stress Corrosion Initiation …

777

Materials Testing has been performed on two heats of Type 304/304L stainless steel, of chemical compositions shown in Table 1. Heat A was Type 304 and therefore had a much higher carbon content than heat B (304L). The nickel content was substantially higher in heat A, whereas heat B contained much more molybdenum. The microstructures of the materials are illustrated in Fig. 1. Heat A has a rather high average grain size of 135 lm, nearly 4 times as large as heat B (35 lm). As a result of the lower nickel content and higher Mo content, heat B contained substantially greater d-ferrite, approximately 5%, compared to 50 40

814

L. Chang et al.

Fig. 1 Schematic of the location of the 316L samples and the orientation of the marks in the machined surface

using 600 grit grinding paper. The ﬁnal polishing using the OPS was used to create reference surfaces and compared with the machined surfaces. The dimension of the machined cold-rolled 316L sample (hereafter refers to CR316L) was the same as samples 316L-1 and 316L-2. The roughness of the surfaces was measured by a Wyko NT1100 White light interferometer, and residual stresses were measured by X-ray diffraction using the sin2w-d method. A Mn tube was used for X-ray generation and 11 angles ranging from −27° to 27° were selected for the measurements. The {311} planes were selected for strain determination, and the Bragg angle for this plane was 152.8°. Cross-sectional microstructures of the machined surfaces were characterized using a Zeiss Sigma Scanning electron microscopy. The samples were tested at 300 °C, in high purity water doped with 2 ppm of Li (added as LiOH), 3 ppm hydrogen, and oxygen content of the water was controlled to be less than 2 ppb. The samples were strained at 10−5/s until the load approached 80% of the samples’ yield strength. Thereafter, the strain rate was reduced to 2 10−8/s until the samples were plastically deformed by 5% (nominal strain). Samples 316L-1 and 316L-2 were not strained during the ﬁrst test because of a fault with the machine’s gear box, so the total time that they were exposed to the PWR water was longer than that of sample CW316L. After tests, both the machined and polished surfaces were ﬁrst examined in plan view and then longitudinally sectioned and examined in the cross section. Elongations for the three samples were 4.0% for samples 316L-1 and 316L-2 and 3.5% for sample CR316L after the tests.

Results and Discussion Pre-test Characterization Surface Roughness, Residual Stresses, and Machining Marks Orientation Table 2 lists the roughness and residual stress of the machined surfaces of the samples. Roughness (Ra) of all the machined surfaces was similar; the as-machined

SCC Initiation in the Machined Austenitic Stainless Steel 316L …

815

surface of sample 316L-1 was a little bit rougher than that of 316L-2 and CR316L. Residual stresses in the machined surfaces of 316L-1 and 316L-2 were tensile both in the longitudinal and transverse directions. Transverse stresses were always larger than longitudinal stresses. The residual stress in the machined surface of 316L-2 was larger than that of 316L-1, but they were all larger than the tensile strength of the as-annealed material. This was not the case for CR316L, where the residual stress in the longitudinal direction was compressive while it was tensile in the transverse direction. Residual stresses in both directions were relatively small compared to the strength of the baseline material. The roughness (Ra) of all the OPS-polished surfaces was zero, and residual stresses on these surfaces were all nearly zero. The machining marks of samples 316L-1 and CR316L were nearly perpendicular to sample longitudinal direction, the angle between the machining marks and the sample longitudinal direction of sample 316L-2 ranged from 36 to 51°.

Cross Sectional Microstructure of the Machined Surfaces Figure 2 shows cross-sectional microstructures of the machined surfaces. A machining-induced deformation layer formed on the top of all the cross-section samples, see Fig. 2a, b and d. The deformation layer had a gradated microstructure, which consisted of a top ultraﬁne grain layer and deformation bands underneath. Details of the ultraﬁne grain layer in the machined surface of 316L-1 and CR316L are shown in Fig. 2c, e. The thickness of the machined layer in sample 316L-1 and 316L-2 was about 50–90 lm, and it was a little bit thicker for 316L-1 than that of 316L-2. The average thickness of the ultraﬁne grain layer on 316L-1 was 3.8 ± 1.1 lm while it was 3.4 ± 1.4 lm for 316L-2. Underneath the deformation layer, was the substrate, which mainly consisted of annealed coarse grains and a small amount of delta ferrite. The machining-induced deformation layer on the cold-rolled substrate was much thinner than that on the annealed substrate. Table 2 Roughness, residual stresses, machining marks angle and thickness of the ultraﬁne grain layer of the machined and polished surfaces Surface

316L-1 machined 316L-2 machined CR316L machined OPS polished a Refers to the

Ra, lm

Longitudinal residual stress (MPa)

Transverse residual stress (MPa)

Machining mark anglea

Ultraﬁne grain layer thickness (lm)

2.97

598

826

73–83°

3.8 ± 1.1

2.25

837

1007

36–51°

3.4 ± 1.4

2

−170

262

70–80°

0.8

0

−20 to 20

−20 to 20

/

0

angle between the machining marks and the longitudinal direction of the sample

816

L. Chang et al.

Fig. 2 Backscattered electron (BSE) images of the machined surfaces. a Cross-section of sample 316L-1; b cross-section of sample 316L-2; d cross-section of sample CR316L; c, e outer ultraﬁne grain layer on samples 316L-1 and CR316L, respectively

Numerous deformation bands formed in the substrate during cold-rolling, and it was difﬁcult to distinguish the deformation bands that formed during cold-rolling from which formed during machining, as can be seen from Fig. 2d, e. The thickness of the ultraﬁne grain layer on CR316L was measured, and the average thickness was around 0.8 lm (Fig. 1e). The thicknesses of both the total machining-induced deformation layer and the ultraﬁne grain layer were not uniform throughout the sample, they varied from area to area, as can be seen in Fig. 2a, b.

Post-test Microstructural Characterization Plan-View Characterization of Specimens Figure 3 shows the plan-view examination results of the machined surfaces of samples 316L-1 (a, b) and 316L-2 (c) and the OPS polished surface of sample 316L-1 (d) after the test, the loading direction for the samples during the test was in the vertical direction of the pictures. The image for the OPS polished surface of sample 316L-2 is not shown here because it was effectively identical to that of 316L-1. Many cracks were found in the machined surface of sample 316L-1, these cracks were initiated at the machining grooves and propagated along them, as

SCC Initiation in the Machined Austenitic Stainless Steel 316L …

817

shown in Fig. 3a, b; Fig. 3a shows the long and wide cracks along the machining marks, (b) shows the micro-cracks at a higher magniﬁcation. Cracks crossing the machining marks was not observed on the surface. No cracks were found in the machined surface of sample 316L-2 and the OPS polished surfaces of both 316L-1 and 316L-2, as can be seen in Fig. 3c, d. These observations indicated that the orientation of the machining marks in respect to the loading direction is an important factor that influences cracks initiation, while the residual stress in the machined surface did not appear to play any role; in fact, 316L-2 which had higher residual stresses in both directions did not crack whilst sample 316L-1 which had lower residual stresses cracked. Figure 4 shows the plan view examination results of the machined and OPS polished surfaces of CR316L after the test. A number of cracks were observed on the machined surface, as shown in Fig. 4a. As the surface was relatively flat, correlation between the orientation of the machining marks and the crack initiation site was not found on this surface; typical cracks found on the surface are shown in Fig. 4a, b, they are relatively short and always perpendicular to the loading direction. Two types of cracks were found on the OPS polished surface, as shown in Fig. 4c, d. The ﬁrst type crack shown in Fig. 4c was judged to be intergranular based on their morphology. Some of them initiated in the grain boundaries and some initiated at the interface between the austenite grain and the delta ferrite. The

Fig. 3 Secondary electron (SE) images taken on the machined surfaces of the as-tested samples a, b 316L-1, and c 316L-2, d the OPS polished surface of samples 316L-1 (surface of OPS polished 316-2 was very similar)

818

L. Chang et al.

second type of crack is shown in Fig. 4d, and is narrower than the ﬁrst type of crack and they are always perpendicular or inclined to perpendicular to the loading direction.

Characterisation of Cross-Section Specimens Figure 5a shows the crack initiated on the machined surface of 316L-1, Fig. 5b, c show the oxide layer formed on both the machined surface (b) and OPS polished surface (c). During the cross-sectional examination, it was observed that a number of cracks initiated and penetrated into the ultraﬁne grain layer and deformation bands in the machined surface of sample 316L-1, but crack penetration into the ultraﬁne grain layer was not observed in the machined surface of 316L-2 and cracking was also not observed in the OPS polished surfaces of samples 316L-1 and 316L-2. The average penetration depth of the cracks in the machined surface of sample 316L-1 was 1.2 lm and the maximum crack penetration depth was 2.3 lm. Most of these cracks stopped in the ultraﬁne grain layer. Only a few of them penetrated into the deformation bands, see the crack in Fig. 5a. The double-layered structured oxide was also revealed in cross-section, Fig. 5b, c showing the oxide layer formed on the machined surface and the OPS polished surface of 316L-1 respectively. From these images, it can be seen that the oxide formed on the

Fig. 4 Secondary electron images taken on the machined (a, b) and OPS polished surfaces (c, d) of the as-tested CR316L

SCC Initiation in the Machined Austenitic Stainless Steel 316L …

819

machined surface was much thicker than that of the OPS polished surface; the thickness of the inner oxide layer on the machined surface was measured to be around 511 ± 100 nm while it was about 325 ± 52 nm for that of the OPS polished surface. Cracks were also observed in the inner oxide layer of the surfaces, but they rarely penetrated to the interface between the oxide and the metal on the OPS polished surface (Fig. 5c). Figure 6 shows cracks identiﬁed in the cross-sections of both the machined side (a, b) and OPS polished side (c, d) of CR316L. The cracks identiﬁed in the machined surface were all very shallow, most of them had similar morphology as the ones shown in Fig. 6a, which stopped at the interface between the ultraﬁne grain layer and the deformation bands; some of the cracks penetrated through to the deformation bands, the deepest one observed in the cross section is shown in Fig. 6b. Both intergranular and transgranular cracks were identiﬁed in the cross section of the OPS polished surface (Fig. 6c, d) and were mainly transgranular and very shallow, see Fig. 6c. A limited number of cracks were intergranular, and a typical one is shown in Fig. 6d. The penetration depth for these cracks was about several microns, and the deepest one was measured to be 20 lm. Figure 7 shows the cross section of the oxide layer formed on the machined surface (a) and OPS polished surface (b) of sample CR316L. It is shown that the thickness of the inner oxide layer on both the machined surface and OPS polished surface was not uniform. The thickness of the inner oxide layer on the machined surface was much thicker than that on the OPS polished surface, and it was measured to be 848 ± 110 nm for the machined surface and 447 ± 72 nm for the OPS polished surface. As the thickness of the inner oxide layer on the machined surface nearly approached the thickness of the original ultraﬁne grain layer, most of the ultraﬁne grain was consumed during the growth of the inner oxide layer during the test. From Fig. 7, it can also be seen that the inner oxide layer on the OPS-polished surface was more compact than that on the machined surface, as many very ﬁne voids were present in the inner oxide layer on the machined surface. A summary of the inner oxide layer thickness measurement and SCC behavior of all the samples is shown in Table 3.

Fig. 5 Back-scattered electron images taken on the cross section of the as-tested sample 316L-1: a SCC on the machined surface, b the cross-section of the oxide layer formed on the machined surface, c the oxide layer on OPS polished surface

820

L. Chang et al.

Fig. 6 Back-scattered electron images obtained from cross-section samples of CR-316L showing the cracks formed on: a, b the machined surface, and c, d OPS polished surface

Fig. 7 Back-scattered electron images obtained from cross-section samples of CR-316L showing the oxide layer formed on: a the machined surface, and b OPS-polished surface

Surface

Machined OPS polished CR316L Machined OPS a Based on the examination

316L-1

Sample

SCC/type of crack

511 ± 100 Yes 325 ± 52 No 848 ± 110 Yes/intergranular and transgranular 447 ± 72 Yes/intergranular and transgranular of a 12 mm long cross section

Inner oxide layer thickness (nm) 1.23 – 1.08 2.95

Average crack depth (lm)

Table 3 Summary of the oxide layer thickness measurement and cracking behavior of 316L-1 and CR316L

2.29 – 3.6 20.7

Maximum crack depth (lm)

19 – 11 43

Number of crack identiﬁeda

SCC Initiation in the Machined Austenitic Stainless Steel 316L … 821

822

L. Chang et al.

Discussion The results reported in the section above can be summarized as follows. Machining increased the SCC initiation susceptibility of the annealed material; however, the cracks did not propagate in the annealed substrate. Conversely, machining increased the SCC initiation resistance of the cold-worked stainless steel. Furthermore, the orientation of the machining marks influenced SCC initiation whilst evidence showed that the residual stress appeared to have no effect. In the followed sections, these points will be discussed.

Influence of the Machining-Induced Deformation Layer on SCC Initiation The structure of the machining-induced deformation layer present on the annealed substrate was very similar to the one present on the cold-worked substrate; consisting of an ultraﬁne grain layer on top of the material and slip bands underneath. However, the thickness of the deformation layer on the annealed substrate was greater than that on the cold-worked substrate and the residual stresses in the machined layer on the annealed substrate were greater than that in the machined layer on the cold-worked substrate. Jacobus et al. [27] proposed two types of interactions between the milling cutter and the material during machining: the mechanical interaction and the thermal interaction. Mechanical interaction induces deformation of the material, and the thermal interaction induces both residual stress and microstructural changes. Based on this, the ultraﬁne grain layer was formed due to both the mechanical effect and thermal effect. The thermal effect induced the recrystallization of the mechanically induced highly deformed layer on the top whilst the slip bands layer formed mainly due to the mechanical effect. The cold-worked material cannot be as easily deformed during machining as much as the annealed material due to its lower ductility and higher strength, and consequently it develops a thinner deformation layer than the annealed material. Machined surfaces of samples 316L-1 and CR316L both cracked during the SSRT test, this indicates that the machined surfaces are SCC susceptible in PWR primary water environment during the accelerated SSRT test, although the cracks in the machined surfaces are relatively shallow. Three reasons may account for the initiation of SCC in the machined surfaces: (1) formation of thick and brittle oxide layer; (2) depletion of Cr in the ultraﬁne grain layer underneath the inner oxide due to the formation of the thick inner oxide layer; (3) large stress in the ultraﬁne grain layer during the SSRT test. These possibilities will be discussed in detail in the followed paragraphs. One important feature of the ultraﬁne grain materials is the high diffusivity of the alloying elements in the material because of the high volume fraction of grain boundaries [33, 34]. Results by Scenini et al. [35] and Ostwald et al. [36] showed

SCC Initiation in the Machined Austenitic Stainless Steel 316L …

823

that surface deformation of nickel-based alloy and chromium steels enhanced the diffusion of Cr to the surface to promote a more continuous external oxide due to the higher Cr diffusivity. In the present study, the inner oxide layer on the machined surface was much thicker than that formed on the OPS polished surface. Ziemniak et al. [37] investigated the corrosion behavior of the machined and polished surfaces of stainless steel 304L in PWR water and found that the chemistry of the inner oxide layer formed on the machined surface and polished surface were almost identical. The inner oxide on the polished surface had a lower porosity than that formed on the machined surface (based on Ziemniak’s work and the present work), and this results in a lower mass transport rate and lower corrosion rate. The inner oxide formed on the stainless steel was rich in Cr with the Cr diffusing from the material underneath. Therefore, in this work, the thick inner oxide layer formed on the machined surface, would possibly lead to a shallow Cr depleted ultraﬁne grain layer underneath the oxide that has less Cr than the material underneath the thinner inner oxide layer of the polished surface. The oxide formed on the material surface has lower plasticity and a large Young’s modulus compared to the substrate. Therefore, during the SSRT, cracking of the oxide layer might be possible which re-exposes the material and results in further corrosion and stress corrosion cracking. Two types of OPS-polished surfaces were investigated in the present study, one was on the annealed stainless steel substrate and the other was on the cold-worked stainless steel substrate. However, these two OPS-polished surfaces are different because of the different processing histories of the baseline materials. For the material that was solution and annealed, the material experienced fully recrystallization and the internal stress formed during processing was released. The grain boundaries in this material are with relatively small amount of defects. For the cold-worked material, formation of the deformation bands during cold-working in different grains lead to the strain incompatibilities at the grain boundaries, these grain boundaries would act as fast diffusion paths during the test. As indicated in the experimental section, polishing using OPS produced nearly strain-free surfaces. When the samples with OPS-polished surfaces were tested under SSRT test condition, the grain boundaries were exposed to the water. The grain boundaries in the cold-worked material would be easier to be preferentially oxidized than the grain boundaries in the annealed material because of the strain incompatibilities along the grain boundaries, and this ﬁnally leads to the initiation of intergranular SCC [38, 39]. Nonetheless, when an ultraﬁne grain layer was present on the cold-worked material, it stopped the preferentially oxidation of the grain boundaries of the coarse grains, then the SCC initiation resistant of the cold-worked material was increased.

824

L. Chang et al.

The Influences of the Residual Stress and Orientation of the Machining Marks Tensile stresses are important for SCC initiation. For real components, shrinkage of the weld, bending or deformation during component fabrication are the main causes of tensile stress [40, 41]. These residual stresses combined with the susceptible environment, can lead to the initiation of SCC. For components under static tensile residual stress tested in accelerated condition, the magnitude of the residual stress may influence crack initiation and propagation. Zhang et al. [30] found that a critical stress exists for SCC initiation in annealed 316L with a machined surface during immersion testing in boiling MgCl2 solution and that the crack initiation rate increased with the increase of the residual stress. However, for the samples tested under slow strain rate tensile test condition in the present study, a similar phenomenon was not observed. The machined surface of sample 316L-1 with a low tensile residual stress and machined surface of CW316L with compressive stress cracked while the machined surface of 316L-2 did not crack. The reason for this is that all the samples were mostly in the plastic deformation region during the test. Therefore, the stress in the material during the test was not the original residual stress induced by machining as discussed below. The origin of the residual stress in the machining-induced deformation layer is the un-relaxed elastic strain within the layer, so the residual stress must be lower than the yield strength of the material where it exists (except where tri-axial stress is present). In static test conditions, such as the above mentioned boiling MgCl2 solution, samples with different residual stress may have different cracking behavior, but when plastically strained, the stress in the material during the test is its internal flow stress. As indicated in Section “Cross Sectional Microstructure of the Machined Surfaces”, an ultraﬁne grain layer formed on the top of all the machined surfaces, so the yield strength of the ultraﬁne grain layer for the samples investigated is expected to be similar. The strain hardening component for the ultraﬁne grain layer must be small because of its considerably higher strength and ﬁne grain size. The samples investigated in the present study were stretched to 3–4% plastic strain, so both the sample substrate and the deformation layer were mostly in the plastic deformation region during the test. Once the deformation layer plastically deforms, the original difference in residual stress owing to the difference in residual strain is minimized, and the original compressive stress will transition to a tensile stress. Based on these concepts, the stresses in the machined surfaces would be similar during the test. The stresses in the machined surfaces must be the largest followed by the stress in the OPS polished surface of the cold-worked 316L followed by the stress in the OPS polished surface of the annealed 316L. From the results showed in Section “Plan-View Characterization of Specimens”, the cracks prefer to initiate and propagate along the machining marks. This is because the machining marks can act as stress concentrators. For samples with marks perpendicular to the loading direction, the highest normal stress would be applied in the machined layer. While for samples with marks having a smaller angle

SCC Initiation in the Machined Austenitic Stainless Steel 316L …

825

with the loading direction, the normal stress was reduced, so this surface is less SCC initiation susceptible.

Conclusions Annealed and cold-worked stainless steel samples with machined surfaces have been tested in simulated PWR primary water environment under SSRT conditions to investigate the effect of machining on SCC initiation in austenitic stainless steel. Conclusions can be drawn as follows: (1) Machining introduced a deformation layer into both the annealed and cold worked stainless steels. This deformation layer can be sub-divided into the ultraﬁne grain layer on top and the deformation bands underneath. (2) The deformation layer formed on the annealed material was thicker than that formed on the cold-worked material and residual stresses were present in the deformed layer of both the annealed and cold-rolled materials. (3) Machining appears to increase the SCC initiation susceptibility of the annealed stainless steel 316L, whilst cracking did not occur in the OPS polished surface of the annealed 316L. (4) Orientation of the machining marks with respect to the loading direction appeared to have influence on crack initiation in the machined surface during the SSRT test. The sample with machining marks nearly perpendicular to the loading direction was easy to initiate SCC, although the cracks were shallow and most of them stopped in the ultraﬁne grain layer. (5) Machining appeared to increase the SCC initiation resistance of the cold-worked 316L by introducing a ultraﬁne grain layer as compared to the OPS polished surface. Both the density and penetration depth of the cracks initiated in the machined surface were less than the cracks initiated in the OPS polished surface. Acknowledgements The authors would like to acknowledge the ﬁnancial support of the New Nuclear Manufacturing (NNUMAN) program sponsored by EPSRC (grant EP/JO21172/1), Dr. Agostino Maurotto of the Nuclear Advanced Manufacturing Research Centre (University of Shefﬁeld) for providing the machined plates, and Dr. Kudzanai Mukahiwa for beneﬁcial discussions.

References 1. T. Yonezawa, M. Watanabe, H. Atsushi, in 17th International Conference on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, Canadian Nuclear Society, Ottawa, Ontario, Canada, 2015, pp. 496–513

826

L. Chang et al.

2. T. Couvant, L. Legras, C. Pokor, F. Vaillant, Y. Brechet, J. Boursier, P. Moulart, in 13th International conference on environmental degradation of materials in nuclear power systems, Whistler, BC (Canada), 2007, pp. 19–23 3. L. Tribouilloy, F. Vaillant, J. Olive, M. Puiggali, L. Legras, T. Couvant, J. Boursier, Y. Rouillon, C. Amzallag, Adv. Mater. Sci. 7, 11 (2007) 4. O. Raquet, E. Herms, F. Vaillant, T. Couvant, J.-M. Boursier, in 12th International Symposium on Environmental Degradation of Materials in Nuclear Power Reactors, ed. by T.R. Allen, P.J. King, L. Nelson (Salt Lake City, USA, 2005), pp. 1049–1058 5. P. Huguenin, J. Crepin, C. Duhamel, H. Proudhon, in 16th International Conference on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, NACE International, Asheville, North Carolina, USA, 2012, pp. 1414–1433 6. R. Staehle, in Workshop on Detection, Avoidance, Mechanisms, Modeling and Prediction of SCC Initiation in Water-Cooled Nuclear Reactor Plants, Beaune, France, 2008 7. T. Couvant, P. Moulart, L. Legras, P. Bordes, J. Capelle, Y. Rouillon, T. Balon, in 12th International Symposium on Environmental Degradation of Materials in Nuclear Power Reactors, ed. by T.R. Allen, P.J. King, L. Nelson (TMS, Salt Lake City, USA, 2005), pp. 1069–1079 8. S.W. Sharkawy, Z. Xia, Z. Szklarska-Smialowska, J. Nucl. Mater. 195, 184–190 (1992) 9. Z. Lu, T. Shoji, F. Meng, Y. Qiu, T. Dan, H. Xue, Corros. Sci. 53, 247–262 (2011) 10. J. Chen, Z. Lu, Q. Xiao, X. Ru, G. Han, Z. Chen, B. Zhou, T. Shoji, J. Nucl. Mater. 472, 1–12 (2016) 11. D. Du, K. Chen, L. Yu, H. lu, L. Zhang, X. Shi, X. Xu, J. Nucl. Mater. 456 (2015), 228–234 12. L. Zhang, J. Wang, Corros. Sci. 87, 278–287 (2014) 13. S. Yaguchi, T. Yonezawa, Corros. Sci. 86, 326–336 (2014) 14. T. Couvant, F. Vaillant, J. Boursier, D. Delafosse, Euro. Fed. Corros. Publ. 51, 87 (2007) 15. G. Han, Z. Lu, X. Ru, J. Chen, Q. Xiao, Y. Tian, J. Nucl. Mater 467, Part 1, 194–204 (2015) 16. Y. Han, J. Mei, Q. Peng, E.-H. Han, W. Ke, Corros. Sci. 98, 72–80 (2015) 17. Y. Han, J. Mei, Q. Peng, E.-H. Han, W. Ke, Corros. Sci. (2016) 18. F. Scenini, A. Sherry, Corrosion 68, 1094–1107 (2012) 19. M.J. Olszta, L.E. Thomas, K. Asano, S. Ooki, S.M. Bruemmer, in 14th International Conf. on Environmental Degradation of Materials in Nuclear Power Systems, American Nuclear Society, Virginia Beach, VA, 2009, pp. 549–561 20. Y. Miura, Y. Miyahara, K. Kako, M. Sato, in Corrosion, NACE International, 2013 21. Y. Miura, Y. Miyahara, M. Sato, K. Kako, J.I. Tani, in 15th International Conference on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, ed. by J. T. Busby, G. Ilevbare, P.L. Andersen, (Wiley Online Library, 2011), pp. 439–449 22. U. Ehrnsten, P. Aaltonen, P. Nenonen, H. Hänninen, C. Jansson, T. Angeliu, in 10th Symposium on Environmental Degradation of Materials in Nuclear Power Systems–Water Reactors. Nevada, USA, 2001, pp. 5-9 23. R. Horn, G. Gordon, F. Ford, R. Cowan, Nucl. Eng. Des. 174, 313–325 (1997) 24. T. Shoji, in 11th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems-Water Reactors, NACE Stevenson Washington, USA, 2003, pp. 588-598. 25. A. Turnbull, K. Mingard, J. Lord, B. Roebuck, D. Tice, K. Mottershead, N. Fairweather, A. Bradbury, Corros. Sci. 53, 3398–3415 (2011) 26. G. Hinds, L. Wickström, K. Mingard, A. Turnbull, Corros. Sci. 71, 43–52 (2013) 27. K. Jacobus, R. DeVor, S. Kapoor, J. Manuf. Sci. Eng. 122, 20–31 (2000) 28. D.Y. Jang, T.R. Watkins, K.J. Kozaczek, C.R. Hubbard, O.B. Cavin, Wear 194, 168–173 (1996) 29. R. M. Saoubi, J.C. Outeiro, B. Changeux, J.L. Lebrun, A. Morão Dias, J. Mater. Process. Technol. 96 225–233 (1999) 30. W. Zhang, K. Fang, Y. Hu, S. Wang, X. Wang, Corros. Sci. 108, 173–184 (2016) 31. K.N. Lyon, T.J. Marrow, S.B. Lyon, J. Mater. Process. Technol. 218, 32–37 (2015)

SCC Initiation in the Machined Austenitic Stainless Steel 316L …

827

32. D.R. Tice, S. Nouraei, K.J. Mottershead, J.W. Stairman, in 14th International Conference on Environmental Degradation of Materials in Nuclear Power Systems, American Nuclear Society, Virginia Beach, VA, 2009, pp. 158-171 33. Z.B. Wang, N.R. Tao, W.P. Tong, J. Lu, K. Lu, Acta Mater. 51, 4319–4329 (2003) 34. R.K. Gupta, N. Birbilis, Corros. Sci. 92, 1–15 (2015) 35. F. Scenini, R.C. Newman, R.A. Cottis, R.J. Jacko, Corrosion 64, 824–835 (2008) 36. C. Ostwald, H.J. Grabke, Corros. Sci. 46, 1113–1127 (2004) 37. S.E. Ziemniak, M. Hanson, P.C. Sander, Corros. Sci. 50, 2465–2477 (2008) 38. K. Kruska, S. Lozano-Perez, D.W. Saxey, T. Terachi, T. Yamada, G.D.W. Smith, Corros. Sci. 63, 225–233 (2012) 39. S. Lozano-Perez, T. Yamada, T. Terachi, M. Schröder, C.A. English, G.D.W. Smith, C.R.M. Grovenor, B.L. Eyre, Acta Mater. 57, 5361–5381 (2009) 40. P. Withers, H. Bhadeshia, Mater. Sci. Technol. 17, 366–375 (2001) 41. C.C. Hot, H. Seam, in Handbook of Residual Stress and Deformation of Steel, Residual Stress Formation Processes During Welding and Joining, 2002, pp. 391

High-Resolution Characterisation of Austenitic Stainless Steel in PWR Environments: Effect of Strain and Surface Finish on Crack Initiation and Propagation G. Pimentel, D.R. Tice, V. Addepalli, K.J. Mottershead, M.G. Burke, F. Scenini, J. Lindsay, Y.L. Wang and S. Lozano-Perez

Abstract Initiation and propagation of cracks under simulated primary water conditions and different slow strain rates have been studied for an austenitic 304-type stainless steel. Two surface ﬁnishes were used to better understand the conditions that trigger stress corrosion cracking (SCC). The main objective is to identify the mechanism(s) that govern the initiation and propagation of SCC and the influence of microstructure. Crack morphology, stress localisation and local chemical composition were characterized for all samples studied. The characterization methodology includes scanning electron microscopy (SEM), 3D focused ion beam (FIB), Transmission Kikuchi Diffraction (TKD), and analytical scanning transmission electron microscopy (STEM).

Keywords Stress corrosion cracking (SCC) Slow strain rate test (SSRT) Transmission kikuchi diffraction (TKD) Electron energy loss spectroscopy (EELS)

G. Pimentel (&) S. Lozano-Perez Department of Materials, University of Oxford, Parks Road, Oxford OX1 3PH, UK e-mail: [email protected] D.R. Tice V. Addepalli K.J. Mottershead AMEC Foster Wheeler, Walton House Birchwood Park, Birchwood Warrington, Cheshire WA3 6GA, UK M.G. Burke F. Scenini J. Lindsay Y.L. Wang University of Manchester, Material Performance Centre, Manchester, UK © The Minerals, Metals & Materials Society 2019 J.H. Jackson et al. (eds.), Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, The Minerals, Metals & Materials Series, https://doi.org/10.1007/978-3-030-04639-2_53

829

830

G. Pimentel et al.

Introduction Stainless steel alloys, like austenitic type 304, are often used for structural components in the primary circuit of ﬁssion reactors due to their excellent behaviour under high-temperatures and corrosive environments [1]. However, despite their corrosion resistance, the combination of stress in a corrosive environment, during long operation periods, can, under some conditions, trigger the appearance of microscopic cracks that penetrate into the material (known as stress corrosion cracking (SCC)) [2, 3]. Even though SCC has been widely studied over the past two decades [1, 4, 5], there is still limited understanding of the mechanisms that govern the initiation and propagation of these cracks and, although a range of models have been proposed [6], still there is no agreement on a common and general model. Since intergranular oxide penetrations were observed, many of these previous SCC studies were based on internal oxidation or oxygen diffusion ahead of the crack tip [7, 8] and rely on the fact that oxygen locally diffuses at the same rate or faster than the crack growth rate and that the oxides forming are more brittle than the metal matrix. Therefore, a variety of electron microscopy methods have been used in order to analyse oxide chemistry, crack morphology and local changes around the crack tip [9–11]. However, most studies do not take into account crystallographic aspects such as grain misorientation, local plastic deformation and strain localization in the area surrounding the crack tip. In this effort, transmission Kikuchi diffraction (TKD) has been proven to be very beneﬁcial in revealing the extent of the strain concentration around the crack tips of 316 stainless steel and its effect on SCC [12, 13]. Therefore, an in-depth characterization effort providing information at high resolution and nm length scales combining chemical analysis Electron Energy Loss Spectroscopy (EELS) and Energy-dispersive X-ray spectroscopy, (EDX) and local stress and strain measurements (TKD) could contribute to solve some of the remaining questions. This is the focus of the current work. Moreover, since SCC propagation depends on multiple parameters (such as test temperature, material composition and water chemistry) two specimens of 304 stainless steel under different slow strain rate conditions and one specimen tested under constant load will be compared in order to study the effect that effective strain and microstructure (separate from all other factors) have on SCC.

Materials and Methods The material used for this work is an austenitic 304 stainless steel. Its chemical composition is shown in Table 1. The material was then forged to a 20% reduction in the temperature window 170–360 °C in order to suppress any martensitic transformation induced by strain. Then, three flat tensile samples (with a gauge length of 24 mm and a cross section

Si

0.36 P 0.023

Al

S C gL ln gL exp gL þ gU ln gU exp gU b P P > > si si 4 5¼0 h i h i > þ ln cj b cj b > g g > < i¼1 j¼1 exp gL exp gU 2 cj b h cj b i cj b h cj b i3 > > S b C gL exp gL gU exp gU P P > b si > 4 b 5 ¼ 0: h b i h b i > þ > g g g cj cj : i¼1 j¼1 exp L exp U g

g

ð9Þ

A Statistical Analysis on Modeling Uncertainty Through Crack …

2001

It would be extremely difﬁcult to determine a general analytical solution for Eq. (9); therefore, we used a numerical approach. In this case, the MATLAB (R2015b, MathWorks, Natick, MA, USA, 2015) offers the numerical nonlinear simultaneous equation solver fsolve.

Monte Carlo Simulation Experimental Factors Experimental factors (e.g., number of specimens) can affect the uncertainties of Weibull estimators. In the present study, a Monte Carlo simulation was performed to investigate their corresponding quantitative effects. The experimental factors considered in the simulation study include (1) true Weibull parameters; (2) the number of specimens; (3) end cracking fractions (ECF); and (4) length of censoring interval (LCI).

True Weibull Parameters As previously mentioned, it could be reasonably assumed that the inherent cracking probability was Weibull distributed at a macroscopic scale. The study investigated as to whether the estimation uncertainties were affected by the parameters of the given Weibull distribution (i.e., inherent cracking probability behavior). These were termed as the true Weibull parameters ðbtrue ; gtrue Þ, which are generally unknown to experimenters. It should be noted that the scale parameter g was considered as a nuisance parameter in several applications [19]. For example, if the relative errors (RE) of estimators that were deﬁned as follows: ^ ¼ RE(bÞ

^b ^ g gtrue b true ; RE(^gÞ ¼ ; btrue gtrue

ð10Þ

were affected by the value of the true scale parameter ðgtrue Þ, then only changing the time unit (e.g., hours to seconds) could affect the relative estimation errors. This is contradictory. Therefore, the gtrue is just a scale factor, and relative estimation errors should not be affected by the value of gtrue [17]. Without loss of generality, the value of gtrue was ﬁxed at 100 in this simulation study. Nevertheless, the value of the true Weibull shape parameter ðbtrue Þ could be a factor affecting the relative estimation errors. Therefore, the shape parameter was a main parameter of the Weibull distribution [19]. Several values of btrue (2, 3 and 4) were selected as the simulation inputs. In previous studies, the value of the estimated Weibull shape parameter for an SCC initiation time ranged from two to four [6, 20–22].

2002

J.P. Park et al.

Number of Specimens ^ g ^) could be reliable It is expected that the estimated Weibull parameters (i.e., b; with a large number of specimens. However, the SCC initiation test for nuclear reactor materials requires a corrosive environment with high temperatures and pressures. Thus, it is difﬁcult to test a large number of specimens simultaneously. Hence, the simulation range of the specimen number was set from 5 to 50.

End Cracking Fraction During the performance of the SCC test, cracking does not necessarily occur for every specimen within the available testing time. Thus, the test duration was considered as a factor of estimation uncertainties in an earlier study [18]. However, the results indicated that there were deﬁciencies to using the test duration as a factor of estimation uncertainties. First, experimenters do not know their relative test duration (RTD), which is deﬁned as follows: RTD ¼

test duration ; gtrue

ð11Þ

this is because the experimenters do not know the exact value of gtrue . Additionally, it is possible to continue the simulation even after all of the specimens cracked when the test duration is used as a ﬁxed input for a simulation study [18]. Therefore, the end cracking fraction (ECF) is considered as an alternative factor of estimation uncertainties. The ECF is a proportion of cracked specimen at the end of a test, which used for a criterion of the test termination. For example, if the value of ECF corresponded to 0.6, then the test ended when more than or equal to 60% of the specimens cracked. The simulation range for ECF was set from 0.6 to 1.0.

Length of Censoring Interval It is expected that a shorter LCI is better to estimate reliable Weibull parameters. We considered a situation that the SCC test was carried out by inserting a large number of U-bend specimens into the autoclave. To censor the specimens, the experiment has to be stopped. Therefore, frequent censoring causes inconveniences for experimenters, and hence it is important to set a reasonable LCI for a cracking test. Although there are some advanced test methods (e.g., direct current potential drop) which can measure the crack initiation time without stopping the experiment, however, it is difﬁcult to test a lot of specimens simultaneously through that kind of advanced test methods. Therefore, one of the interesting motivation of the simulation study is to ﬁnd which of the following is more helpful to estimate the more reliable Weibull model – ‘a lot of inaccurate cracking time data’ or ‘a few accurate cracking time data’.

A Statistical Analysis on Modeling Uncertainty Through Crack …

2003

Thus, in order to investigate the general effect of LCI, the simulation range for starting LCI was set from 5 to 50% of gtrue , although the value of gtrue was unknown in the real testing case. It is assumed that length of censoring interval does not vary with time.

Simulation Procedure Table 1 shows the simulation range of the study. A total of 900 (= 1 3 10 3 10) experimental cases were considered, and 20,000 random iterations were performed for each experimental case. From the resulting simulation data, the ^ ^g) were calculated by the MLE method. Figure 2 shows Weibull estimators (i.e., b; the schematic procedure of the simulation study. The simulation procedure can be summarized as follows: (1) Based on the assumed true Weibull distribution (see black solid lines in Fig. 2) and given experimental condition (e.g., number of specimens, ECF), we obtain the result of virtual cracking time data through the Monte Carlo simulation (see black squared dots in Fig. 2). (2) We estimate the Weibull parameters by MLE method with the previous cracking time data (see red solid lines in Fig. 2). (3) We iterate (1) and (2) 20,000 times to obtain a distribution of Weibull estimates at the given experimental condition. (4) With changing to another experimental condition, we iterate the above procedure 900 times.

Table 1 Experimental factors considered in the Monte Carlo simulation True Weibull parameters gtrue btrue (Dim’less time)

Number of specimens

ECF

LCI (% of gtrue )

100 – – – – – – – – –

5 10 15 20 25 30 35 40 45 50

0.6 0.8 1.0 – – – – – – –

5 10 15 20 25 30 35 40 45 50

2 3 4 – – – – – – –

2004

J.P. Park et al. ECF = 1 1

(Iter.1)

0.9

Cracking Fraction

1

1

(Iter.2)

0.9

0.9

0.8

0.8

0.8

0.7

0.7

0.7

0.6

0.6

0.6

0.5

0.5

0.5

0.4

0.4

0.4

0.3

0.3

0.3

0.2

0.2

0.2

0.1 0

0.1

0

20 40 60 80 100 120 140 160 180 200

0

(Iter.3)

20,000 iterations

0.1

0

0

20 40 60 80 100 120 140 160 180 200 0

Time

Time

20 40 60 80 100 120 140 160 180 200

Time

900 experimental cases

Fig. 2 Schematic procedure of the Monte Carlo simulation (btrue : 3; LCI: 20% of gtrue ; ECF: 1.0)

Results and Discussion From the random simulation results, the 5th, 50th and 95th percentiles ^ ;b ^ ;b ^ ; ^g ; ^g ; ^g ) of 20,000 replicates of Weibull estimates could (b 5% 50% 95% 5% 50% 95% be derived for each experimental case. The selected median estimates (i.e., ^ ; ^g ) were converted to the relative error ðRE50% Þ to represent the bias of b 50% 50% estimators, which is deﬁned as follows: ^ ^ g gtrue ^ ¼ RE(b ^ Þ ¼ b50% btrue ; RE50% ð^gÞ ¼ RE(^ g50% Þ ¼ 50% RE50% ðbÞ 50% btrue gtrue ð12Þ In order to quantify the dispersion of estimators, a relative length of a 90% conﬁdence interval ðRLCI90% Þ was utilized, which is deﬁned as follows: ^ ¼ RE(b ^ Þ RE(b ^ Þ; RLCI90% ð^ RLCI90% ðbÞ gÞ ¼ RE(^ g95% Þ RE(^ g5% Þ: ð13Þ 95% 5% As an example, Fig. 3 shows the effect of the number of specimens on estimation uncertainties when the other experimental factors are ﬁxed at the certain values. It is well represented that as the number of specimens is large, estimators becomes reliable (i.e., little bias and short length of conﬁdence interval). For estimating the shape parameter b, it is likely that the shape parameters are over^ estimated when the number of specimens is relatively low (i.e.,RE50% ðb MLE Þ [ 0). For the scale parameter g estimation, the bias is barely noticeable even when the number of specimens is low [i.e., RE50% ð^gMLE Þ 0]. The relative length of 90% conﬁdence interval (i.e., RLCI90% ð^gMLE Þ) are much lower than that of the shape ^ parameter b (i.e., RLCI90% ðb MLE Þ).

A Statistical Analysis on Modeling Uncertainty Through Crack …

2005

^MLE Uncertainty of b ^ Figure 4 shows the contour plots of RE50% ðb MLE Þ, indicating a relative estimation ^ bias in the Weibull shape parameter. It was likely that b MLE showed a tendency to be overestimated irrespective of the value of ECF and btrue when the number of specimens was relatively small. It is shown that the value of LCI does not much ^ affect RE50% ðb MLE Þ especially when the ECF is relatively high. The unusual result that occurred in the long LCI region may not be reliable due to the low convergence ratio [15]. ^ Figure 5 shows the contour plots of RLCI90% ðb MLE Þ. As expected, the disper^ sion in b was large when the number of specimens was relatively small. In contrast, the effect of LCI was not noticeable. Additionally, some critical regions were ^ were produced [18]. The gradients of observed, in which very wide RLCI90% ðbÞ ^ were very high around the critical region. It is likely that this critical RLCI90% ðbÞ region was an inherent behavior of estimation uncertainty because another estimation method, which has a unity convergence ratio for the same experimental condition, also showed the critical region [15]. Experimenters should plan the crack initiation testing so that they can avoid this critical region. We suggest that it is very dangerous to estimate b with less than ten specimens regardless of the censoring interval and ECF.

Uncertainty of ^ gMLE Figures 6 shows the contour plots of RE50% ð^gMLE Þ. The simulation results indicated that the value of RE50% ð^gMLE Þ was almost zero in every experimental condition.

(a) 2.4

RE_5% RE_50% RE_95%

2.1 1.8

(b) 2.4 1.8

1.5

1.5

1.2

1.2

0.9

0.9 0.6

0.6 0.3

0.3

OveresƟmated

0.0

0.0

-0.3

-0.3

-0.6

RE_5% RE_50% RE_95%

2.1

0

5

10

15

20

25

30

35

40

Number of Specimen (ea.)

45

50

-0.6

0

5

10

15

20

25

30

35

40

45

50

Number of Specimen (ea.)

^ and (b) RE(^ Fig. 3 Effects of the number of specimens on (a) RE(bÞ gÞ (btrue : 3; LCI: 20% of gtrue ; ECF: 1.0)

2006

J.P. Park et al.

50

Number of Specimen (ea.)

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

5

10

15

20

25

30

35

40

45

50

5

10

15

20

25

30

35

40

45

50

5

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

10

15

20

25

30

35

40

45

50

50

5

10

15

20

25

30

35

40

45

50

5

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

10

15

20

25

30

35

LCI (% of ηtrue)

40

45

50

5

0.04000 0.08000 0.1200 0.1600 0.2000

ECF = 1.0 5

10

15

20

25

30

35

40

45

50 -0.2000

(f)

-0.1600 -0.1200 -0.08000 -0.04000 0.000 0.04000 0.08000 0.1200 0.1600 0.2000

ECF = 0.8 5

10

15

20

25

30

35

40

45

50 -0.2000

(i)

45

40

5

0.000

50

(h)

45

40

5

-0.04000

10

5

50

(g)

45

-0.08000

45

40

5

-0.1200

50

(e)

45

40

5

-0.1600

10

5

50

(d)

45

-0.2000

(c)

45

40

50

Number of Specimen (ea.)

(b)

45

40

5

Number of Specimen (ea.)

50

50

(a)

45

-0.1600 -0.1200 -0.08000 -0.04000 0.000 0.04000 0.08000 0.1200 0.1600 0.2000

10

5

10

15

20

25

30

35

LCI (% of ηtrue)

40

45

50

5

ECF = 0.6 5

10

15

20

25

30

35

40

45

50

LCI (% of ηtrue)

^ Fig. 4 Effects of the number of specimens and LCI on RE50% ðb MLE Þ when ECF = 1.0 and (a) btrue ¼ 2, (b) btrue ¼ 3, (c) btrue ¼ 4; when ECF = 0.8 and (d) btrue ¼ 2, (e) btrue ¼ 3, (f) btrue ¼ 4; when ECF = 0.6 and (g) btrue ¼ 2, (h) btrue ¼ 3, (i) btrue ¼ 4 [15]

This suggests that g^MLE is always unbiased irrespective of the combination of experimental conditions in the simulation range. gMLE Þ. When compared to the Figure 7 shows the contour plots of RLCI90% ð^ ^ gMLE Þ was quite small. This case of RLCI90% ðbMLE Þ, the overall value of RLCI90% ð^ indicates that the estimated scale parameter is more reliable than the shape parameter under the same experimental conditions [18]. The dispersion in ^ gMLE was large when: (1) the number of specimens was small; (2) the value of ECF was low; and (3) the value of btrue was small. LCI showed no noticeable effect on gMLE Þ. RLCI90% ð^gMLE Þ. The critical region did not appear in the case of RLCI90% ð^ The simulation result suggests that a relatively reliable ^ g could be obtained with only ﬁve specimens regardless of the other experimental conditions.

A Statistical Analysis on Modeling Uncertainty Through Crack …

50

Number of Specimen (ea.)

50

(a)

45

40

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

5

10

15

20

25

30

35

40

45

50

50

5

10

15

20

25

30

35

40

45

50

5

(e)

45 40

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

5

50

10

15

20

25

30

35

40

45

50

10

15

20

25

30

35

40

45

50

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

10

15

20

25

30

35

LCI (% of ηtrue)

40

45

50

5

1.200 1.400 1.600 1.800 2.000

ECF = 1.0 5

10

15

20

25

30

35

40

45

50 0.000

(f)

0.2000 0.4000 0.6000 0.8000 1.000 1.200 1.400 1.600 1.800 2.000

ECF = 0.8 5

10

15

20

25

30

35

40

45

50 0.000

(i)

45

40

5

1.000

50

(h)

45

5

40

5

0.8000

10

5

50

(g)

45

5

0.6000

45

40

5

0.4000

50

50

(d)

45

Number of Specimen (ea.)

5

0.2000

10

10

10

0.000

(c)

45

40

5

Number of Specimen (ea.)

50

(b)

45

2007

0.2000 0.4000 0.6000 0.8000 1.000 1.200 1.400 1.600 1.800 2.000

10

5

10

15

20

25

30

35

LCI (% of ηtrue)

40

45

50

5

ECF = 0.6 5

10

15

20

25

30

35

40

45

50

LCI (% of ηtrue)

^ Fig. 5 Effects of the number of specimens and LCI on RLCI90% ðb MLE Þ when ECF = 1.0 and (a) btrue ¼ 2, (b) btrue ¼ 3, (c) btrue ¼ 4; when ECF = 0.8 and (d) btrue ¼ 2, (e) btrue ¼ 3, (f) btrue ¼ 4; when ECF = 0.6 and (g) btrue ¼ 2, (h) btrue ¼ 3, (i) btrue ¼ 4 [15]

2008

J.P. Park et al.

50

50

(a)

Number of Specimen (ea.)

45

40

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

5

10

15

20

25

30

35

40

45

50

50

10

15

20

25

30

35

40

45

50

5

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

10

15

20

25

30

35

40

45

50

50

5

10

15

20

25

30

35

40

45

50

5

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

10

5

5

15

20

25

30

35

LCI (% of ηtrue)

40

45

50

0.04000 0.08000 0.1200 0.1600 0.2000

ECF = 1.0 5

10

15

20

25

30

35

40

45

50 -0.2000

(f)

-0.1600 -0.1200 -0.08000 -0.04000 0.000 0.04000 0.08000 0.1200 0.1600 0.2000

ECF = 0.8 5

10

15

20

25

30

35

40

45

50

5

10

-0.2000

(i)

45

40

10

0.000

50

(h)

45

40

5

-0.04000

10

5

50

(g)

45

-0.08000

45

40

5

-0.1200

50

(e)

45

40

5

-0.1600

10

5

50

(d)

45

Number of Specimen (ea.)

5

-0.2000

(c)

45

40

5

Number of Specimen (ea.)

50

(b)

45

15

20

25

30

35

LCI (% of ηtrue)

40

45

50

5

-0.1600 -0.1200 -0.08000 -0.04000 0.000 0.04000 0.08000 0.1200 0.1600 0.2000

ECF = 0.6 5

10

15

20

25

30

35

40

45

50

LCI (% of ηtrue)

Fig. 6 Effects of the number of specimens and LCI on RE50% ð^ gMLE Þ when ECF = 1.0 and (a) btrue ¼ 2, (b) btrue ¼ 3, (c) btrue ¼ 4; when ECF = 0.8 and (d) btrue ¼ 2, (e) btrue ¼ 3, (f) btrue ¼ 4; when ECF = 0.6 and (g) btrue ¼ 2, (h) btrue ¼ 3, (i) btrue ¼ 4 [15]

A Statistical Analysis on Modeling Uncertainty Through Crack …

50

50

(a)

Number of Specimen (ea.)

45

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

5

10

15

20

25

30

35

40

45

50

10

15

20

25

30

35

40

45

50

5

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

10

5

5

15

20

25

30

35

40

45

50

5

10

15

20

25

30

35

40

45

50

50

50

(g)

45

5

40

35

35

35

30

30

30

25

25

25

20

20

20

15

15

15

10

10

10

5

5

15

20

25

30

35

LCI (% of ηtrue)

40

45

50

1.000 1.200 1.400 1.600 1.800 2.000

ECF = 1.0 5

10

15

20

25

30

35

40

45

50 0.000

(f)

0.2000 0.4000 0.6000 0.8000 1.000 1.200 1.400 1.600 1.800 2.000

ECF = 0.8 5

10

15

20

25

30

35

40

45

50

5

10

0.000

(i)

45

40

10

0.8000

50

(h)

45

40

5

0.6000

45

40

10

0.4000

50

(e)

45

40

5

0.2000

10

5

50

(d)

45

Number of Specimen (ea.)

5

0.000

(c)

45

40

50

Number of Specimen (ea.)

50

(b)

45

40

5

2009

15

20

25

30

35

LCI (% of ηtrue)

40

45

50

5

0.2000 0.4000 0.6000 0.8000 1.000 1.200 1.400 1.600 1.800 2.000

ECF = 0.6 5

10

15

20

25

30

35

40

45

50

LCI (% of ηtrue)

Fig. 7 Effects of the number of specimens and LCI on RLCI90% ð^ gMLE Þ when ECF = 1.0 and (a) btrue ¼ 2, (b) btrue ¼ 3, (c) btrue ¼ 4; when ECF = 0.8 and (d) btrue ¼ 2, (e) btrue ¼ 3, (f) btrue ¼ 4; when ECF = 0.6 and (g) btrue ¼ 2, (h) btrue ¼ 3, (i) btrue ¼ 4 [15]

2010

J.P. Park et al.

Conclusions The main goal of this study is to provide quantitative estimation uncertainties for experimenters developing a Weibull distribution model via cracking tests. The MLE method was performed with respect to the Weibull estimation. Monte Carlo simulations were used in order to quantify uncertainties estimators in various experimental conditions by considering the effects of: (1) true Weibull parameters; (2) the number of specimens; (3) end cracking fractions; and (4) length of censoring interval. The following conclusions were drawn from the study: • The Weibull distribution was appropriate for the statistical model of crack initiation time at a macroscopic scale. ^ • In the simulation range, b MLE showed a tendency to be overestimated and dispersed when the number of specimens was small and the value of ECF was ^ low. The value of LCI does not much affect bias of b MLE especially when the ECF is relatively high. It was shown that there were critical regions, in which the dispersions were extremely large. Thus, experimenters should avoid this critical region when establishing an SCC test plan. We suggest that it is very dangerous to estimate b with less than ten specimens regardless of the censoring interval and ECF. • ^gMLE showed almost zero bias in all simulation ranges. In most cases, the LCI did not affect the estimation uncertainty of ^g. The overall bias and dispersion of ^ in the simulation study range. Therefore, the ^g were much lower than those of b estimated scale parameter would be more reliable than the estimated shape parameter from the cracking tests. Actually, the simulation result suggests that a relatively reliable ^g could be obtained with only ﬁve specimens regardless of the other experimental conditions. Therefore, It was shown that increasing the number of specimen would be more efﬁcient to reduce the uncertainty of estimators than the more frequent censoring. Acknowledgements This work was supported by the Nuclear Safety Research Program through the Korea Foundation of Nuclear Safety (KOFONS) granted ﬁnancial resource from the Nuclear Safety and Security Commission (NSSC), Republic of Korea (No. 1403006), and was supported by “Human Resources Program in Energy Technology” of the Korea Institute of Energy Technology Evaluation and Planning (KETEP), who granted the ﬁnancial resources from the Ministry of Trade, Industry & Energy, Korea. (No. 20164010201000).

References 1. W. Lunceford, T. DeWees, P. Scott, EPRI Materials Degradation Matrix, Rev. 3. EPRI, Palo Alto, CA, USA 3002000628, 2013 2. P. Scott, M.-C. Meunier, Materials Reliability Program: Review of Stress Corrosion Cracking of Alloys 182 and 82 in PWR Primary Water Service (MRP-220). EPRI, Palo Alto, CA, USA Rep. No. 1015427, 2007

A Statistical Analysis on Modeling Uncertainty Through Crack …

2011

3. K.J. Kim and E.S. Do, Technical Report: Inspection of Bottom Mounted Instrumentation Nozzle (in Korean). Korea Institute of Nuclear Safety (KINS), Daejeon, Korea KINS/RR-1360, 2015 4. C. Amzallag, S.L. Hong, C. Pages, A. Gelpi, Stress Corrosion Life Assessment of Alloy 600 PWR Components. in 9th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, 1999, pp. 243–250 5. Y.S. Garud, Stress Corrosion Cracking Initiation Model for Stainless Steel and Nickel Alloys. EPRI, Palo Alto, CA, USA 1019032, 2009 6. M. Erickson, F. Ammirato, B. Brust, D. Dedhia, E. Focht, M. Kirk, et al., Models and Inputs Selected for Use in the xLPR Pilot Study. EPRI, Palo Alto, CA, USA 1022528, 2011 7. G. Troyer, S. Fyﬁtch, K. Schmitt, G. White, C. Harrington, Dissimilar Metal Weld PWSCC Initiation Model Reﬁnement for XLPR Part I: A Survey of Alloy 82/182/132 Crack Initiation Literature. in 17th International Conference on Environmental Degradation of Materials in Nuclear Power Systems—Water Reactors, Ottawa, Ontario, Canada, August 9–13, 2015 8. W. Weibull, A statistical distribution function of wide applicability. J. Appl. Mech. 18, 293– 297 (1951) 9. E. Eason, Materials Reliability Program: Effects of Hydrogen, pH, Lithium and Boron on Primary Water Stress Corrosion Crack Initiation in Alloy 600 for Temperatures in the Range 320–330°C (MRP-147). EPRI, Palo Alto, CA, USA 1012145, 2005 10. I.S. Hwang, S.U. Kwon, J.H. Kim, S.G. Lee, An intraspecimen method for the statistical characterization of stress corrosion crack initiation behavior. Corrosion 57, 787–793 (2001) 11. J. McCool, Using the Weibull distribution: reliability, modeling, and inference (Wiley, Hoboken, N.J., USA, 2012) 12. S.M. Ross, Introduction to Probability and Statistics for Engineers and Scientists, 4th edn. (Elsevier Academic Press, USA, 2009) 13. R.A. Fisher L.H.C. Tippett, Limiting Forms of the Frequency Distribution of the Largest or Smallest Member of a Sample. in Mathematical Proceedings of the Cambridge Philosophical Society, 1928, pp. 180–190 14. R.L. Wolpert, Extremes, Available online: https://www2.stat.duke.edu/courses/Fall15/sta711/ lec/topics/extremes.pdf, 2014 15. J.P. Park, C. Park, J. Cho, C.B. Bahn, Effects of cracking test conditions on estimation uncertainty for Weibull parameters considering time-dependent censoring interval. Materials 10, 3 (2016) 16. D. McFadden,Modeling the Choice of Residential Location. Transportation Research Record 1978 17. U. Genschel, W.Q. Meeker, A comparison of maximum likelihood and median-rank regression for Weibull estimation. Q. Eng. 22, 236–255 (2010) 18. J.P. Park, C.B. Bahn, Uncertainty evaluation of Weibull estimators through Monte Carlo simulation: applications for crack initiation testing. Materials 9, 521 (2016) 19. F.R. Hampel, E.M. Ronchetti, P.J. Rousseeuw, W.A. Stahel, Robust statistics: the approach based on influence functions vol 114, Wiley, 2011 20. J.D. Hong, C. Jang, T.S. Kim, PFM application for the PWSCC integrity of Ni-Base alloy welds—development and application of PINEP-PWSCC. Nuclear Eng. Technol. 44, 961–970 (2012) 21. K. Dozaki, D. Akutagawa, N. Nagata, H. Takihuchi, K. Norring, Effects of dissolved hydrogen condtent in PWR primary water on PWSCC initiation property. E-J. Adv. Main. 2, 65–76 (2010) 22. Y.S. Garud, SCC initiation model and its implementation for probabilistic assessment. in ASME Pressure Vessels & Piping Division, July 18–22, 2010

Part XXI

Plant Operating Experience

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3 James Hyres, Rocky Thompson and Jim Batton

Abstract This paper covers the results of laboratory examinations performed on a leaking letdown cooler from Oconee Unit 3. The laboratory scope included dewatering, pressure testing, visual inspections, metallography, Vickers micro-hardness, scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS), X-Ray Diffraction (XRD), and Optical Emission Spectroscopy (OES). The laboratory examinations identiﬁed one tube containing a through-wall crack. The most likely cause of the crack appeared to be OD-initiated caustic stress corrosion cracking (SCC). The presence of heavy deposits on the tube OD surface and heat tinting on the primary and secondary flow seals indicated boiling occurred near the tight radius region of the bundle. Once boiling occurred, caustic-forming species such as calcium phosphate deposited and concentrated on the tube OD surface. The literature indicates as caustic concentrations approach *20%, the conditions become favorable for caustic SCC to occur in austenitic stainless steels such as Type 316L.

Keywords Letdown cooler Type 316L stainless steel tubing stress corrosion cracking Caustic stress corrosion cracking

Intergranular

Introduction Oconee is a three-unit pressurized water reactor (PWR) plant with Babcock & Wilcox (B&W) 177FA nuclear steam supply systems (NSSS). The reactor coolant system (RCS) letdown water for each unit flows from the cold leg of one of the two J. Hyres (&) BWX Technologies, Inc., 2016 Mt. Athos Road, Lynchburg, VA 24504-5447, USA e-mail: [email protected] R. Thompson Duke Energy, 526 South Church Street, Charlotte, NC 28202, USA J. Batton Duke Energy, Oconee Nuclear Station, Seneca, SC 29672, USA © The Minerals Metals & Materials Society 2019 J.H. Jackson et al. (eds.), Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, The Minerals, Metals & Materials Series, https://doi.org/10.1007/978-3-030-04639-2_136

2015

2016

J. Hyres et al.

once-through steam generators and through two parallel letdown coolers to cool the water before entering the puriﬁcation and deborating demineralizers. Forty-two (42) failures of letdown coolers have occurred in B&W plants since 1977 (Refs. [1–3]). Of these, 27 (or 64%) of the failures have occurred at the Oconee plants. OD-initiated high cycle fatigue at tube-to-tube stitch welds was identiﬁed as the cause of early failures by the destructive examination of letdown coolers from Crystal River-3 and Three Mile Island-1 (TMI-1). A number of recommendations were developed and implemented at B&W plants to minimize or prevent additional tube failures by high cycle fatigue. For a while these recommendations appeared to be effective. Oconee reported no letdown cooler failures between 1997 and 2003. However, failures reoccurred from 2003 to 2014. Similar letdown coolers are in service at Davis-Besse and ANO-1. ANO-1 has experienced no failures to date; Davis-Besse reported a cooler failure in 2009. The cause of failure was not determined, but assumed to be due to high cycle fatigue. Due to the recent poor reliability of the letdown coolers at Oconee and since no letdown cooler from any B&W plant had been destructively examined since 1987, it was decided to destructively examine the archived Oconee 3A letdown cooler which had a known leak in order to determine the cause of the tube failure. Leakage was discovered in the 3A letdown cooler in July 2003. This letdown cooler was in service from 1994 until the EOC 22 Refueling Outage in spring 2006, when it was removed and replaced with another cooler.

Letdown Cooler Description and Metallurgy The 3A letdown cooler was manufactured by Graham Manufacturing Company in 1994. It is a compact counter flow heat exchanger having a Helicoil tube bundle design consisting of thirty (30) seamless Type 316L stainless steel tubes, each having a 19 mm (0.750 in) OD and 1.8 mm (0.072 in) wall thickness, and measuring approximately 18 m (60 feet) long. The tubes were procured in the solution annealed condition, i.e. heated at 1038 °C (1900 °F) followed by rapid quenching to below 427 °C (800 °F) in less than 3 min. When rolled into the bundle, each tube contains approximately nine (9) coils. Once coiled, the entire tube bundle is stress relieved at 1093 °C (2000 °F) for 32 min, followed by rapid cooling to below 427 ° C (800 °F). Each tube is tack-welded to a support bar on the inlet and outlet ends, and then formed at a 90° angle toward the inlet and outlet tubesheets. The tubes are roll expanded (2–5% wall reduction) nominally 3.8 cm (1.5 in) into the 4.8 cm (1.88 in) thick tubesheets and then seal welded. The bundle is encased within a 13 mm (0.5 in) thick carbon steel shell containing inlet and outlet ports for the primary and secondary fluids. The as-manufactured cooler weighs approximately 3000 kg (6600 lb). Primary inlet flow (RCS) is from the center of the bundle outward. Secondary side cooling water (Component Cooling) flows in the opposite direction, from the

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3

2017

outer periphery of the cooler, spiraling inward toward the center. A thin stainless steel baffle plate between the outer 2 and 3 coils of the tubes helps direct the cooling water flow inward toward the center. The cooling water exits the central cavity through the outlet nozzle located on the same side as the inlet nozzle.

Letdown Cooler Operating Conditions RCS letdown fluid enters the tubes nominally at 14.9 MPa (2155 psig) and 291 °C (555 °F). Treated Component Cooling water flows on the shell side. The operating conditions on the shell side of the letdown cooler are 71 °C (160 °F) measured at the cooler outlet and 0.48–0.62 MPa (70–90 psia) (pressure estimated in the cooler), which corresponds to a saturation temperature of 149–160 °C (300–320 °F). The RCS temperature is 291 °C (555 °F) at the cooler inlet and decreases as the fluid cools as it passes through the tubes. Localized boiling of the Component Cooling water (producing some superheated steam) occurs on the tube surface at the RCS inlet end of the letdown cooler where both the RCS temperature and the OD of the tube wall equal or exceed the saturation temperature at the shell side operating pressure. In this respect, letdown cooler design operating conditions in the B&W NSSS differ from other PWR and even BWR designs that have stepwise cooling and pressure reduction of the letdown water via a regenerative heat exchanger in order to prevent localized boiling conditions from occurring. From 1994, when the 3A letdown cooler went into service, until April 2003, the secondary side Component Cooling water chemistry was potassium chromate (100– 500 ppm as CrO4), sodium phosphate (100–300 lg/L as PO4), and pH *10 in accordance with Ref. [4]. In May 2003, chromate chemistry was replaced with molybdate (500–1000 lg/L MoO4)-azole (tolyltriazole) chemistry at pH 9.0–11 in accordance with Ref. [5]. Makeup water to the Component Cooling system is high purity demineralized water. After the letdown cooler leak developed, boric acid from RCS in-leakage was introduced into the Component Cooling water (secondary side) system until the leaking cooler was removed from service. The Component Cooling system operates at a lower pressure than the raw water system, which acts as its heat sink.

Results Receipt, Dewatering, and Pressure Testing Figure 1 is a receipt photograph taken of the cooler showing the primary side inlet and outlet manifolds. The tubes were dewatered prior to pressure testing, and then each tube was individually pressure tested with 206 kPa (30 psi) helium for

2018

J. Hyres et al.

Fig. 1 Receipt photograph showing the primary flow seal side of the letdown cooler. The primary inlet (near center) and outlet (toward bottom) manifolds are visible

Fig. 2 Photograph taken during the shell-side pressure test. A bubble formed at the end of tube #17, indicating a through-wall defect

*15 min to identify leaking tubes. One tube, identiﬁed as Tube #17 in the laboratory, failed the pressure test. The shell side was also pressurized with 206 kPa (30 psi) of helium. Soap bubbles were used to check the integrity of each tube, the tubesheet seal welds, and the weld joining the tubesheet to the shell wall. This test also conﬁrmed the leak in Tube #17 (see Fig. 2). No other leaks were identiﬁed.

Cooler Disassembly The letdown cooler shell was cut just above the primary side end plate and then removed to access the tube bundle. Figures 3, 4 and 5 show the tube bundle and its central cavity. White-to-brown deposits were found on the tubes in the tight radius region on the inlet ends. The deposits were most heavily concentrated on the tube intrados, which corresponds to the underside (i.e., 6:00 position) of the tubes in-service. Few or no deposits were present on the tube extrados. The leak in Tube

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3

2019

Fig. 3 The letdown cooler after shell removal. The cut location to isolate the tubes is indicated

Fig. 4 The central cavity near the inlet side of the tube bundle. The tube #17 leak location is indicated

Fig. 5 Higher magniﬁcation view of the tube #17 leak location. Cut locations for the destructive examinations are indicated

#17 was located in the deposit-covered region of the tubes as indicated in Fig. 5. Evidence of heat tinting was also found near the tight radius area of the tube bundle corresponding to the Tube #17 leak location (Figs. 6 and 7).

2020

J. Hyres et al.

Fig. 6 View of the primary flow seal after removing the tubes

Fig. 7 Higher magniﬁcation view of the tight radius region showing the most pronounced heat tinting

The bulk residual stresses were relatively low as evidenced by the tubes not “springing open” after they were cut at the outlet and inlet side support bars in order to remove them. The tube bundle stress relief treatment was therefore concluded to have been effective.

Visual Examinations on Leaking Tube Section Visual examination of the OD surface of the 5.1 cm (2 in) section of Tube #17 revealed three circumferential cracks in the tube intrados (see Figs. 8 and 9). Visual examination under higher magniﬁcation revealed the cracking to be jagged with a minor branch at one tip of the through-wall crack (see Fig. 10). IGA-like features were also found on the OD surface at the opposite end of the crack. This ﬁnding indicated corrosion was associated with the cracking, which likely initiated on the tube surface exposed to the Component Cooling water (see Fig. 11). The tube section was cut open clamshell-style to permit visual examination of the ID surface (see Fig. 12). One crack was found (see Fig. 13), which was the through-wall leak identiﬁed during pressure testing. Cuts were then made through the cracked region in order to produce a through-wall open crack specimen and two

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3

2021

Fig. 8 Initial cuts to isolate the through-wall crack region

Fig. 9 Resulting piece as viewed from the tube #17 intrados. Three cracks were visible in this area

Fig. 10 Higher magniﬁcation view of the tube #17 cracks

metallographic mounts, one through the through-wall crack tip and one through a secondary crack (see Fig. 14).

SEM/EDS Examinations High magniﬁcation Scanning Electron Microscopy/Energy Dispersive X-ray Spectroscopy (SEM/EDS) examinations were performed on the opened through-wall crack, OD surface deposits, and a cross section prepared through a

2022

J. Hyres et al.

Fig. 11 IGA-like features were observed at one of the crack tips

Fig. 12 Photograph taken after cutting the piece open clamshell-style to reveal the ID surface. The through-wall crack location is indicated

Fig. 13 Higher magniﬁcation view of the through-wall leak location

secondary crack. The secondary electron (SE) imaging technique was employed to characterize the open crack surface topography. The mounted sample was analyzed in the as-polished condition using backscattered electron imaging (BSE).

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3

2023

Fig. 14 Photograph showing the additional sections made through the leak location

Opened Through-Wall Crack Low magniﬁcation optical microscopy and low magniﬁcation SEM both revealed that the through-wall cracking was intergranular from OD to ID (see Figs. 15, 16, 17 and 18). Pitted features on many grain facets indicated corrosive attack occurred at some point after the crack formed. Secondary cracking was also noted in several locations, a characteristic typical of stress corrosion cracking (SCC) in austenitic stainless steels. EDS analysis of the heavily deposited region on the opened fracture surface showed the deposits were composed of major amounts of O, Cr, Fe, and Ni, with minor amounts of Mo, and trace amounts of Si, Ti, Cu, Zn, and Sn (see Fig. 19 for a typical EDS spectrum). No evidence of deleterious soluble species such as chloride was found in the areas examined. It should be noted that the escaping fluid through the crack would likely remove deposited soluble species from the crack surfaces.

Fig. 15 Photograph of the opened through-wall crack. Cracking is intergranular in nature

2024 Fig. 16 SEM image of the opened crack taken near the tube ID surface

Fig. 17 SEM image of the opened crack taken in the mid-wall region

Fig. 18 SEM image of the opened crack taken near the tube OD surface

J. Hyres et al.

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3

2025

Fig. 19 EDS spectrum results for a heavily deposited region on the opened through-wall crack

OD Deposits Deposits collected from the OD surface of Tube #17 in the through-wall crack region were afﬁxed to a specimen stub using double-sided carbon tape. EDS analyses of regions of interest found that the deposits consisted of major amounts of O and Ca, with minor amounts of Mg, Si, P, Fe, Cu, Zn, and Ba, along and trace amounts of Al, Cr, Ni, and Cu. The presence of Mg, Ca, and Si indicated in-leakage of raw water into the Component Cooling water system. The presence of phosphorus in the deposits was consistent with deposition occurring during the time period that CrO4–PO4 chemistry was employed. Barium was also identiﬁed during the 1987 letdown cooler examination (Ref. [1]). A deﬁnitive source for the barium and its potential impact on the tube cracking could not be conclusively determined.

Secondary Crack Cross Section An axial cross section through a secondary crack in Tube #17 was prepared for SEM/EDS examination by placing the specimen in conductive copper mounting material, and then grinding and polishing using standard techniques. The cross section was examined in the as-polished condition. Figure 20 shows the cracking clearly initiated on the OD surface and propagated toward the ID surface. Shallow penetrations were noted mostly on the OD surface, but a few were also present on the ID surface. Very minor surface cold work was evident on the OD surface in this region. The cracking contained several shallow branches, three of which were selected for higher magniﬁcation inspections. All three tips were intergranular in nature and exhibited a similar blunted morphology (see Fig. 21). Blunted crack tips are typically associated with inactive crack growth. The EDS results indicated the base metal contained major amounts of Cr, Fe, and Ni with minor amounts of Mn and Mo, as well as trace amounts of Mg, Si, V,

2026

J. Hyres et al.

Fig. 20 Secondary crack cross section showing crack propagation from OD (top) toward the ID

Fig. 21 Secondary crack tip cross section at higher magniﬁcation

and Cu. No evidence of deleterious species and/or de-alloying was identiﬁed within or adjacent to the cracking, or ahead of the crack tip within the grain boundaries.

Metallographic Examinations Metallographic cross sections were prepared through various regions of interest in the tubes to characterize the crack region, tubes adjacent to the leaking tube, a tube near the center of the tube bundle, and tubes adjacent to the primary and secondary flow seals. Features speciﬁc to the pluggable cooler design were also investigated. Results for two of the cross sections, the through-wall crack tip and a secondary crack, are presented in this paper.

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3

2027

Mounts were examined in the as-polished condition and after electrolytic etching with 10% oxalic acid (*6 V for *15 s) to reveal the material microstructure. The secondary crack cross section was also etched according to ASTM A 262 Practice A to assess material sensitization. This procedure involved oxalic etching at 0.5 Å for 90 s.

Through-Wall Crack Tip The through-wall crack tip region exhibited a few branches, but was generally straight and unbranched for much of its length (see Figs. 22 and 23). Secondary cracking was noted on the OD surface as well. Intergranular cracking and branching are evident near the OD, but the latter half of the crack was clearly straight and transgranular. There was also evidence of cold work on the OD surface, as evidenced by the presence of disturbed grains in this area.

Fig. 22 Micrograph of the cross section prepared through the tube #17 through-wall crack tip. OD surface is along top edge. Electrolytic 10% oxalic etch

Fig. 23 Same area as Fig. 24, using Differential Interference Contrast (DIC) illumination

2028

J. Hyres et al.

Secondary Crack The cross section of a secondary crack in Tube #17 is shown after standard oxalic etching in Fig. 24 and after ASTM A 262 etching in Fig. 25. The cracking extended nearly through-wall in this plane. The grain boundaries exhibited a step structure after the ASTM A 262 etch, indicating that the material microstructure was non-sensitized.

Vickers Microhardness Vickers microhardness (100 g load) traverses were taken on the polished cross section prepared from Tube #17. An OD-to-ID traverse indicated elevated hardness near the OD surface (285 HV) due to surface cold work. The bulk hardness for tube Fig. 24 Tube #17 secondary crack. Electrolytic 10% oxalic etch

Fig. 25 Same area as Fig. 24, after etching per ASTM A 262, Practice A. No evidence of ditching was noted at the grain boundaries

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3

2029

#17 was approximately 150–160 HV, compared to 140–150 HV for the other the other tubes tested. Elevated hardness generally increases a material’s susceptibility to SCC.

Optical Emission Spectroscopy (OES) OES Chemical analysis was performed on three tube specimens (tube #17 and two adjacent tubes, #6 and #16). The results were all within the speciﬁed limits for Type 316L tubing per ASTM A 213. Of speciﬁc interest was the carbon content, which was consistent for all three tubes (0.018 wt%) and within speciﬁcation for Type 316L stainless steel.

Analysis of OD Deposits A clean razor knife was used to scrape deposits from the OD surface of tube #17 and two adjacent tubes, #6 and #16. Powder X-ray diffraction (XRD) was performed to identify the crystalline phase compositions of the tube deposit samples removed from the OD surface of tube #6, tube #16, and tube #17. Approximately 0.01–0.02 grams of powder was used for the analysis of the three samples. Samples were prepared using a greased microscope slide due to the small amount of available deposits. The powder specimens were scanned by X-ray diffractometry, using a Philips X’Pert Diffractometer System, and the heights of selected peaks were measured. To minimize preferred orientation effects, several peaks from each phase were included. Empirically derived relative intensity factors were applied to the sums of the peak heights for each phase detected and the volume percent of the phases were normalized to give a total of 100%. Utilizing a copper target X-ray tube operating at 40 kV and 40 mA, the samples were scanned over an angular range from 5° to 90° 2h at a step size of 0.010° and a dwell time of 3.5 s per step. Cu Ka radiation was used. Qualitative results of the samples indicated calcium phosphate, Ca3(PO4)2, was the major constituent in all three specimens. Other compounds identiﬁed included: tri-calcium silicate (Ca3SiO5), silicon dioxide (SiO2), magnetite (Fe3O4), potassium, and chromium.

Discussion Pressure testing and subsequent destructive examinations revealed a single through-wall crack in one of the letdown cooler tubes, identiﬁed as tube #17 in the laboratory. The through-wall crack was located in a heavily-deposited region of the

2030

J. Hyres et al.

tube bundle central cavity near the primary inlet. Low magniﬁcation visual inspections revealed at least three cracks on the tube #17 OD surface; only one crack was visible on the tube ID surface. This observation indicated crack propagation was most likely OD to ID. The open crack SEM examinations showed the through-wall portion of the cracking was intergranular from OD to ID. Shallow pitting present on grain facets indicated chemical attack occurred at some point after the crack formed. Cross-sectional examination of the through-wall crack tip region revealed cracking was mixed intergranular/transgranular in nature, as evidenced by the relatively straight, non-branching nature of the crack. OD intergranular attack (IGA) and intergranular secondary cracking were also identiﬁed. Analysis of the deposits by XRD indicated a major constituent was calcium phosphate. The EDS data collected from the deposits indicated in-leakage of raw water as evidenced by the presence of Mg, Ca, and Si. No chlorine was detected on the though-wall open crack or the OD deposits located near the cracking. Based on the laboratory ﬁndings, the most likely cause of the cracking appears to be OD-initiated caustic SCC. The presence of heavy deposits on the tube OD surface and heat tinting on the primary and secondary flow seals indicate boiling occurred near the tight radius region of the bundle. Once boiling occurs, there is the potential for caustic-forming species such as calcium phosphate to deposit and concentrate. As the caustic concentration approaches *20%, the conditions become favorable for cracking/IGA to occur. For stress corrosion cracking to occur, three conditions must be present: (1) a susceptible material, (2) tensile stress, and (3) an environment known to cause stress corrosion cracking. Each of these conditions is discussed below.

Material Cracking through the failed tube was primarily intergranular with a minor amount of transgranular propagation near the crack tips. This crack morphology is consistent with the literature, which indicates caustic SCC can produce both types of cracking in stainless steels (Ref. [6]). Ref. [7] indicates SCC caused by concentrated caustic in stainless and low-alloy steels is often observed to be intergranular in nature. Conversely, chloride-induced SCC generally follows a transgranular path in non-sensitized, low carbon grades of austenitic stainless steels (Ref. [8]). This is because the low carbon grades are less likely to become sensitized, or chromium depleted, at the grain boundaries. (In contrast, chloride-induced SCC typically produces intergranular cracking in sensitized stainless steels.) Indeed, the ASTM A 262 Practice A test results revealed no evidence of sensitization in the tube material. Further, the tube material carbon content was 0.018% by weight, which is below the 0.02% solubility limit of carbon in austenite. Consequently, reducing the carbon

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3

2031

level to below 0.02% by weight is the recommended threshold to prevent sensitization of austenitic stainless steels by processes such as welding (Ref. [8]). The failed tube contained evidence of surface cold work on the OD, with a maximum measured hardness of 285HV. Surface cold work was not identiﬁed in other tubes tested. In addition, the bulk hardness of the failed tube was slightly higher (150–160HV) compared to the other tubes tested (140–150HV). Elevated hardness due to strain hardening has been linked to intergranular cracking in non-sensitized Type 304L and Type 316L austenitic stainless steels (Ref. [9]).

Tensile Stress Since the tubes are stress relieved after the coiling operation, the most likely source of tensile stress is operating stresses. Minimal “spring open” was noted during the tube sectioning operations, indicating minimal bulk residual stresses were present in the tube material and that the stress relief treatment was effective.

Environment Because of the low operating pressure of the letdown cooler shell side (0.48– 0.62 MPa, 70–90 psia) and the high RCS inlet temperature (291 °C, 555 °F) on the tube side, localized boiling will occur on the OD surface (Component Cooling water side) of the letdown cooler tubes until the RCS is cooled to less than the shell side saturation temperature. Non-volatile species in the Component Cooling water concentrate as the water boils until their solubility limits are exceeded and they precipitate on the tube surfaces. The tube deposits appeared porous based on visual examination. Evidence of corrosive attack was present on the tube OD surface underneath deposits which is consistent with attack after the deposits precipitated. Boiling continued to occur at the tube surface within the pores of the deposits, forming concentrated caustic solutions that could attack the tube. There was evidence of chemical attack in tubes located toward the primary flow seal end of the tube bundle. No attack was observed near the center of the bundle or at the opposite end of the bundle. This observation indicates the tubes toward the primary flow seal side of the tube bundle experienced the highest in-service temperatures and therefore were the most susceptible to corrosive attack. The apparent temperature gradient across the tube bundle may be attributable to the varying lengths of tubing from the inlet end to the tight radius region, i.e. the tube length is shorter on the primary flow seal side compared to the secondary flow seal side. The XRD results for the heavy tube deposits indicated they were largely composed of calcium phosphate, Ca3(PO4)2, with lesser amounts of tri-calcium silicate, Ca3SiO5, and silica, SiO2. The source of the phosphate is the sodium phosphate

2032 Table 1 Oconee nuclear station raw water chemistry

J. Hyres et al. Parameter

mg/L

mole/L

HCO3− SiO2 Na+ Cl− Ca+2 Mg+2 SO−2 4 K+ pH

9.70 7.36 1.75 1.13 1.28 0.62 1.62 0.62

0.159 0.123 0.0761 0.0319 0.0319 0.0255 0.0169 0.0159

pH

7.1

added for pH buffering of the Component Cooling water during the time period that potassium chromate was used as a corrosion inhibitor in the system. In-leakage of raw water into the Component Cooling water is the only plausible source of Ca, Mg, and Si (see Table 1). A review of historical Component Cooling water system chemistry data (Ref. [10]) did reveal speciﬁc raw water in-leakage events via Component Cooler tube leaks that occurred after the 3A letdown cooler was removed from service in 2006; however, no speciﬁc leak events were found in the reported chemistry data for the time period that the letdown cooler was in service. Tube-to-tubesheet joint leakage of the Component Coolers (the joints are not seal welded) is therefore the likely chronic source of Ca, Mg, and Si by low level raw water in-leakage. MULTEQ simulations (static model with steam removed) of Component Cooling water containing sodium phosphate and faulted with raw water of the chemistry shown in Table 1 do predict calcium phosphate as a major precipitate with lesser amounts of calcium silicate (in agreement with the XRD ﬁndings) and a caustic pH at system temperature. With the current molybdate-azole chemistry faulted with raw water, MULTEQ simulations still predict a caustic pH, but with silica and CaMg(SiO3)2 as the major precipitates. As the caustic concentration approaches *20%, the conditions become favorable for cracking and IGA to occur in austenitic stainless steels. According to Ref. [8], the 1970s witnessed increased interest in caustic SCC as a result of the use of austenitic alloys in heat-exchanger systems of water-cooled nuclear power plants. It is well known from conventional boiler technology that boiling and steam blanketing at heat-transfer surfaces give rise to very high local caustic concentrations. The data show that there is an inherent danger of caustic SCC in strong caustic solutions when the temperature approaches 100 °C (212 °F). Non-sensitized Type 304 stainless steel components have failed at caustic concentrations about 20% (NaOH) at 100 °C (212 °F). Alloys containing nickel concentrations of 30% by weight or higher exhibit increased resistance to caustic SCC (Ref. [6]). Reference [11] states: “Susceptibility of austenitic stainless steels to caustic SCC usually becomes a problem when the caustic concentration exceeds approximately 25% and temperatures are above 100 °C (212 °F).” A case history involving a Type 316L weld HAZ revealed cracking to be branching and intergranular. The lowest

Laboratory Analysis of a Leaking Letdown Cooler from Oconee Unit 3

2033

reported temperature for a caustic SCC failure of stainless steel with

Proceedings of the 18th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors

Recommend Stories

Idea Transcript

Helpful Links

Smile Life

Get in touch