Stainless Steels for Design Engineers (#05231G)

STAINLESS STEELS FOR DESIGN ENGINEERS

MICHAEL MCGUIRE

ASM International® Materials Park, Ohio 44073-0002 www.asminternational.org Copyright © 2008 by ASM International® All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the written permission of the copyright owner.

First printing, December 2008

Great care is taken in the compilation and production of this book, but it should be made clear that NO WAR- RANTIES, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MER- CHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, ARE GIVEN IN CONNECTION WITH THIS PUBLICATION. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM’s control, ASM assumes no liability or obligation in connection with any use of this information. No claim of any kind, whether as to products or information in this publication, and whether or not based on negligence, shall be greater in amount than the purchase price of this product or publication in respect of which damages are claimed. THE REMEDY HEREBY PROVIDED SHALL BE THE EXCLUSIVE AND SOLE REMEDY OF BUYER, AND IN NO EVENT SHALL EITHER PARTY BE LIABLE FOR SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES WHETHER OR NOT CAUSED BY OR RESULTING FROM THE NEGLIGENCE OF SUCH PARTY. As with any material, evaluation of the material under end-use conditions prior to speciﬁcation is essential. Therefore, speciﬁc testing under actual conditions is recommended. Nothing contained in this book shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this book shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement. Comments, criticisms, and suggestions are invited, and should be forwarded to ASM International. Prepared under the direction of the ASM International Technical Book Committee (2007–2008), Lichun L. Chen, Chair. ASM International staff who worked on this project include Scott Henry, Senior Manager of Product and Service Development; Steven R. Lampman, Technical Editor; Eileen De Guire, Associate Editor; Ann Britton, Editorial Assistant; Bonnie Sanders, Manager of Production; Madrid Tramble, Senior Production Coordinator; Diane Grubbs, Production Coordinator; Patty Conti, Production Coordinator; and Kathryn Muldoon, Production Assistant Library of Congress Control Number: 2008934669 ISBN-13: 978-0-87170-717-8 ISBN-10: 0-87170-717-9 SAN: 204-7586 ASM International® Materials Park, OH 44073-0002 www.asminternational.org Printed in the United States of America Contents

Preface ...... v

METALLURGY Chapter 1 Metallurgy...... 1

CORROSION AND OXIDATION Chapter 2 Corrosion Theory...... 11

Chapter 3 Corrosion Kinetics...... 19

Chapter 4 Corrosion Types...... 27

Chapter 5 Oxidation...... 57

STAINLESS STEEL ALLOYS Chapter 6 Austenitic Stainless Steels ...... 69

Chapter 7 Duplex Stainless Steels...... 91

Chapter 8 Ferritic Stainless Steels...... 109

Chapter 9 Martensitic Stainless Steels ...... 123

Chapter 10 Precipitation-Hardening Stainless Steels ...... 137

PROCESSING Chapter 11 Casting Alloys...... 147

Chapter 12 Melting, Casting, and Hot Processing...... 155

Chapter 13 Thermal Processing ...... 161 Chapter 14 Forming...... 173

Chapter 15 Machining ...... 181

Chapter 16 Surface Finishing...... 193

Chapter 17 Welding...... 201

APPLICATIONS Chapter 18 Architecture and Construction...... 213

Chapter 19 Automotive and Transportation Applications...... 225

Chapter 20 Commercial and Residential Applications ...... 233

Chapter 21 Marine Systems Applications...... 243

Chapter 22 Petroleum Industry Applications ...... 247

Chapter 23 Chemical and Process Industry Applications...... 257

Chapter 24 Pulp-and-Paper Industry Applications ...... 265

APPENDIXES Appendix 1 Compositions...... 269

Appendix 2 Physical and Mechanical Properties of Select Alloys...... 279

Appendix 3 Introduction to Thermo-Calc and Instructions for Accessing Free Demonstration ...... 281

Index ...... 285

iv Preface

The rate of growth of stainless steel has outpaced that of other metals and alloys, and by 2010 may surpass aluminum as the second most widely used metal after carbon steel. The 2007 world production of stainless steel was approximately 30,000,000 tons and has nearly doubled in the last ten years. This growth is occurring at the same time that the production of stainless steel continues to become more consolidated. One result of this is a more widespread need to understand stainless steel with fewer resources to provide that information. The concurrent technical evolution in stainless steel and increasing volatility of raw material prices has made it more important for the engineers and designers who use stainless steel to make sound technical judgments about which stainless steels to use and how to use them. This book provides design engineers with an up-to-date source of information at a level useful for both metallurgists and other engineers and technicians. It seeks to bridge the gap between the inter- net where much current, but raw information is available and scholarly books and journals that provide theory that is difficult to put into practice. The content of the book is selected for utility for the user of stainless steel. The first section gives elementary metallurgy and identification of constituents of stainless, the effects of alloying elements and a significant section on corrosion. A second section is oriented toward processes important to users of stainless steel. The third section is about each family of stainless alloys and includes the most recent additions that have come to the market. The fourth section deals in some depth with the major applications for stainless steel. This last part is presented without the promotional bias which is found in many steel producers’, alloy producers’, and trade associations’ literature. While a number of steel producers have provided assistance to the author, there has been no attempt to unfairly bias information in their favor. To the contrary, those producers responsible for generating factual, useful data for the user community are those who should benefit the most by books such as this. The author is particularly indebted to Allegheny Ludlum and John Grubb, and his many colleagues who assisted him, for technical assistance throughout the writing and to Carnegie Mellon University for their support. The author also wishes to thank Professor Srid- har Seetharaman at Carnegie Mellon University for his help in writing the corrosion chapter and others who helped: Roy Matway of CMU, Vittorio Boneschi of Centro-Inox; Paul Mason of Thermo- Calc; Bob Drab of Schmolz Bichenbach; Elisabeth Torsner and Chuck Turack Outukumpu, USA; Scott Balliett of Latrobe Steel; Jim Halliday and Fred Deuschle of Contrarian Metals Resources; Pro- fessors Tony DeArdo of Pitt and Gerhard Welsch of CWRU; the staffs of Centro-Inox, Euro-Inox, SSNA, The Nickel Institute; and the editorial staff at ASM International, Scott Henry, Eileen DeGuire, Charlie Moosbrugger and Steve Lampman. I would also like to thank the many members of my forum at Eng-tips.com who have contributed much collective knowledge and perspective to this book. ASM International is the society for materials engineers and scientists, a worldwide network dedicated to advancing industry, technology, and applications of metals and materials.

ASM International, Materials Park, Ohio, USA www.asminternational.org

Publication title Product code Stainless Steels for Design Engineers #05231G

To order products from ASM International: Online Visit www.asminternational.org/bookstore Telephone 1-800-336-5152 (US) or 1-440-338-5151 (Outside US) Fax 1-440-338-4634 Customer Service, ASM International Mail 9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA Email [email protected] American Technical Publishers Ltd. 27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX, In Europe United Kingdom Telephone: 01462 437933 (account holders), 01462 431525 (credit card) www.ameritech.co.uk Neutrino Inc. In Japan Takahashi Bldg., 44-3 Fuda 1-chome, Chofu-Shi, Tokyo 182 Japan Telephone: 81 (0) 424 84 5550

Terms of Use. This publication is being made available in PDF format as a benefit to members and customers of ASM International. You may download and print a copy of this publication for your personal use only. Other use and distribution is prohibited without the express written permission of ASM International.

No warranties, express or implied, including, without limitation, warranties of merchantability or fitness for a particular purpose, are given in connection with this publication. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this information. As with any material, evaluation of the material under end-use conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended.

Nothing contained in this publication shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this publication shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement. ASM International is the society for materials engineers and scientists, a worldwide network dedicated to advancing industry, technology, and applications of metals and materials.

ASM International, Materials Park, Ohio, USA www.asminternational.org

Publication title Product code Stainless Steels for Design Engineers #05231G

Nothing contained in this publication shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this publication shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement. Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 1-10 All rights reserved. DOI: 10.1361/ssde2008p001 www.asminternational.org

CHAPTER 1

Metallurgy

Summary phases coexisting simultaneously. Stainless steel is an exceptional alloy system in that it is COMPARED TO ALLOY STEELS, stainless not a dilute solution. Alloy steels may contain steels are chemically complex. The large number several percent of alloying elements, such as of alloying elements makes possible a larger carbon, manganese, nickel, molybdenum, range of possible phases or basic crystal struc- chromium, and silicon, in addition to the impu- tures. The large amount of the alloying elements rities sulfur, oxygen, and phosphorus. Alloy makes the deviation from the behavior of pure steels typically contain very small amounts of iron greater; consequently, the calculations that titanium, niobium, and aluminum. The total predict which phases will exist are more difficult. amount* of these alloying elements seldom ex- The three basic phases of stainless steels are fer- ceeds 5%. The same is true for most aluminum rite, austenite, and martensite. The wide variety alloys. In contrast, stainless steels contain no of alloys that exist is based on: less than about 11% chromium alone. Most stainless alloys have manganese, silicon, car- ¥ Combinations of these phases bon, and nickel in thermodynamically mean- ¥ Altering the composition of these phases ingful amounts as well as large concentrations ¥ Adding secondary phases for particular of nickel and/or molybdenum. purposes The result of the large number of alloying ele- Metallurgy, as discussed in this chapter, fo- ments in relatively high concentrations is that cuses on phases normally encountered in stain- stainless steel can have many stable phases con- less steels and their characteristics. In subsequent currently. In almost every case, having phases chapters on types of stainless steel, there are other than the principal one or two phases for more detailed treatments of the alloys made of which the alloy was designed is undesirable be- these phases and their properties. cause of the possibility of undesirable variations in mechanical or corrosion performance. The producer of stainless steel controls the chemical composition and thermomechanical processing, Introduction so that when the processor or end user receives the product it is usually in the correct condition. Most widely used alloy systems, such as car- However, subsequent processing or service con- bon steels, alloy steels, and aluminum alloys, are ditions may alter the carefully established phase relatively dilute solutions of several elements in structure. Therefore, it is necessary to discuss the parent matrix. Carbon and alloy steels, with the phases that can exist in stainless steel and very few exceptions, are principally of the mag- the conditions under which they form so that the netic body-centered cubic (bcc) phase or a enlightened user will know which phases to slightly distorted version of it. Aluminum alloys avoid and how to avoid them. share the face-centered cubic (fcc) structure of It is possible to use thermodynamics to calcu- pure aluminum. A given structure, which can late which phases may exist at a given tempera- have a certain range of compositions, is what is meant by a phase, just as a gas or liquid is a * All compositions are given in weight percent unless phase. In solid metals, there can be a number of stated otherwise. 2 / Stainless Steels for Design Engineers

ture for a given composition. It is not remotely not a small, pedantic point. Most stainless steels feasible, however, to give an adequate treatment are used in the metastable condition. For exam- of the thermodynamics required to do this. The ple, the common alloy 304 (also called 18-8) is topic alone requires a book. The necessary normally used in the fully austenitic condition. knowledge has been embedded in proprietary It would “rather” be partly ferritic, but the sub- computer programs that will be used instead. stitutional diffusion of chromium in austenite that is required to form a ferrite phase of a separate composition is so slow that it cannot occur Thermodynamics of Stainless Steel in terrestrial time frames. However, if energy is applied by mechanical shear, the austenite can Pure metals, from a practical viewpoint, are transform without diffusion to the lower free- either liquid or solid depending on temperature, energy martensite phase, a quasi-bcc structure with the possibility of some trivial small gas of lower free energy. vapor pressure. A law of thermodynamics is that The calculation of which phases exist under the number of possible condensed (i.e., solid) equilibrium conditions proves to be extraordi- phases equals the number of elemental con- narily difficult in complicated alloy systems. stituents plus one. The solid has a crystallo- This is because thermodynamic values can be graphic structure that may vary with tempera- measured accurately only in the liquid state, so ture. Many metals have a less-dense bcc the values for the solid state are extrapolations. structure at high temperature and transform to a Also, the interaction between elements is very denser fcc structure at lower temperatures. Iron important in nondilute alloys such as stainless does this. Iron has the curious characteristic of steel. Consequently, most published phase dia- transforming from fcc back to the low-density grams are experimentally derived. To determine bcc at still lower temperatures. This is a result of which phases exist at a given composition and the unpaired 3d orbital electrons (those that give temperature, a sample is made, equilibrated at rise to ferromagnetism) that are not given up as the appropriate temperature, and quenched to valence electrons, causing repulsive forces be- room temperature. It is assumed that the charac- tween atoms and requiring a more widely spaced teristic equilibrium phases have been frozen and structure. are then identified by various techniques for All thermodynamic properties are based on structure, composition, and the like. This impor- interatomic attractions. In metals, the metal tant work is obviously tedious and susceptible atoms give up valence electrons to the entire to experimental error and applies only to spe- mass. These electrons are of varying energy cific compositions. Any “what if” extrapolation states and highly mobile. They are responsible to a different alloy composition carries the risk for the ability of metals to conduct heat and of error. electricity well. The attraction, the strength of A practical tool has been developed that per- the bond, is proportional to the charge differ- mits phase diagrams to be calculated for arbitrary ence and distance. The attraction determines compositions. These are computer simulated, such macroscopic properties as melting temper- mathematical models that can perform the com- ature, density, and elastic modulus. plex thermodynamic calculations. To do this with In this book, the main concern with thermo- accuracy requires databases of thermodynamic dynamics is predicting which phases are present values. These values must be derived from com- both at equilibrium and in the quite frequent puter analysis of experimental phase equilibrium metastable state. The prediction involves calcu- diagrams. They are expensive to derive and vali- lating the free energy of the various possible date, and only a few exist. Hence, they are pro- phases. The phase with the lowest energy is prietary. In Appendix 3, a license to one such most favored, but others may have free energies program, Thermo-Calc, can be found. The ver- that permit them to exist. The difference be- sion has a reduced three-element capability but tween these two is that the equilibrium state, uses the same proprietary thermodynamic data- that of the lowest free energy, may require base of the full version. The program allows de- atomic rearrangements to occur for equilibrium termination of which phases can exist for any compositions to be reached on an atomic scale. composition and temperature. Whether the If diffusion is too sluggish for these rearrange- phases will form depends also on kinetic factors. ments to take place, the structure may retain the First, however, it is good to become familiar with prior metastable structure indefinitely. This is the principal phases found in stainless steel. Chapter 1: Metallurgy / 3

Phases and oxygen, have no influence on which phase is favored. Again, it must be emphasized that Ferrite the influence of an alloying element on structure has zero bearing on its influence on corrosion The basis of stainless alloys is, of course, resistance. iron. Iron, as stated, solidifies as a bcc alloy be- The elements that promote ferrite over austen- fore transforming to the denser fcc austenite at ite also have the effect, at still lower tempera- lower temperatures. At still lower temperatures, tures, of promoting intermetallic compounds it reverts to the bcc structure. It is accurate to generally composed of iron, chromium, and surmise from this that the free energy of both some of those alloying elements. These are dis- structures is close. Alloying elements that pro- cussed separately. mote one structure over the other can therefore Metals are effective solvents in both the liq- change which one predominates. The element uid and solid states. An important part of steel- that produces the ability to form the passive film making is refining the molten metal to remove that makes stainless corrosion resistant, the undesired impurities dissolved in it. The chromium, has the characteristic of stabilizing normal technique is to add elements that react the bcc structure. As chromium is added to iron, selectively with the targeted impurities to form the temperature range over which austenite is an immiscible reactant that can become part of stable grows smaller until, at about 12% the slag and physically separated from the re- chromium, ferrite is stable at all temperatures. fined alloy. This is done for the primary impuri- This is, coincidentally, the approximate level of ties oxygen and sulfur. A third common impu- chromium needed to keep alloys from rusting rity, phosphorus, is not so easily removed and under ambient conditions, but this effect is not must be excluded from raw materials to be kept related to whether the structure is bcc or fcc. under control. The iron-chromium phase diagram (Fig. 1) In stainless steel, carbon and nitrogen can be shows the composition and temperature regions detrimental impurities. Both are quite soluble where ferrite (a), martensite (α'), austenite (γ), in molten iron-chromium alloys and are fairly and sigma phase (σ) are stable. soluble in ferrite at high temperatures. This sol- While chromium is the principal ferrite-pro- ubility decreases exponentially with tempera- moting alloying element, other elements have ture so that it is essentially zero at room tem- similar effects, but none produces the quality of perature. These elements have small atomic stainlessness. Silicon, aluminum, molybdenum, sizes compared to iron and chromium and, tungsten, niobium, and titanium all favor ferrite. when dissolved, squeeze into interstitial sites Carbon, nitrogen, manganese, nickel, and cop- within the bcc matrix. Such interstitial solute per do not and expand the temperature range atoms profoundly distort the structure. They are over which austenite exists. Elements that are much more soluble in the fcc structure, which, insoluble in iron at austenite-forming tempera- while denser, has roomier interstitial spaces, so tures, such as the impurities phosphorus, sulfur, they stabilize that structure. To preserve the ferrite structure, carbon and nitrogen must be eliminated. There are additional reasons to eliminate carbon and nitrogen. During cooling as these elements become less and less soluble, they must precipitate. The most thermodynamically favorable form in which they can precipitate is as a compound of chromium, with which they are very reactive. This occurs at the grain boundaries, where nucleation is favored, and depletes those regions of chromium, rendering them less corrosion resistant. A second effect is a loss of toughness due to these precipitates. The diffusion rates of carbon and nitrogen in ferrite are too high to prevent this precipitation by quenching. Modern refining methods can reduce carbon Fig. 1 The iron chromium phase diagram. Courtesy of Thermo-Calc Software plus nitrogen to under 0.020%, but even this is 4 / Stainless Steels for Design Engineers

too high. So, to avoid the detrimental effects of The bcc structure of ferrite allows more chromium carbide and nitride formation in fer- rapid diffusion than does the fcc structure of rite, other benign carbides and nitrides such as austenite. This is true for both the interstitial those of titanium or niobium are allowed to form diffusion of the elements helium, boron, car- preferentially. This approach is called stabiliza- bon, nitrogen, and oxygen and the substitution and is used for most ferritic alloys today. The tional diffusion of all other elements. The rate older approach, as characterized by alloy 430, is of diffusion of all elements, both interstitial to permit chromium carbides and nitrides to form and substitutional, in ferrite is about two or but then to perform a subcritical anneal to reho- three orders of magnitude higher than in mogenize the chromium and coarsen the pre- austenite. The practical implication of this is cipitates so that they have only a small negative that precipitation reactions generally cannot be effect on mechanical properties. suppressed by quenching in ferrite if they in- Hydrogen and boron are other elements that volve interstitial elements, whereas they can can be interstitially dissolved in ferrite. Boron is be in austenite. Intermetallic phases can form normally found at levels of around 5 to 10 ppm. more rapidly in ferrite. This becomes an issue At higher levels, boron substitutes for carbon in only when total chromium plus molybdenum carbides. Hydrogen is soluble to several parts exceeds about 20%, above which the sigma per million by weight. It does not cause hydro- phase appears. This is thus only an issue for gen embrittlement in annealed ferrite. If the fer- superferritic (high chromium content) alloys or rite is cold worked, the solubility of hydrogen for the ferrite phase of duplex (ferrite-austen- increases as the defect structure accommodates ite) alloys. hydrogen atoms. In this condition, ferrite may The mechanical properties of the ferrite phase be embrittled by hydrogen, especially if it en- are discussed extensively in Chapter 8, “Ferritic ters the metal through corrosion processes like Stainless Steels.” Here, it is only necessary to pitting. This is one explanation of, and the most note that ferrite in stainless steel closely resem- likely explanation for, stress corrosion cracking. bles low-carbon steel in mechanical behavior. It While hydrogen is easily removed by argon shares the following characteristics: oxygen decarburization (AOD), assuming ab- ¥ A toughness transition that occurs around solutely dry blowing gases and additions are room temperature used, it can be picked up during pickling, weld- ¥ Notch sensitivity ing, or annealing as well as by corrosion. ¥ A yield point phenomenon All stainless alloys rely on having a uniform ¥ Pronounced crystallographic anisotropy of level of chromium and the other element, mechanical properties molybdenum, which assists in corrosion resist- ¥ High stacking fault energies and low work- ance, distributed throughout the matrix. If there hardening rates are locally low levels of these elements, localized resistance to corrosion is reduced, and lo- These issues are dealt with in the same way as calized corrosion can occur. This can occur by in carbon steel when these characteristics be- the precipitation of any phase that is richer in come an issue. The first two are controlled by chromium or other corrosion-resisting ele- reduction of interstitial levels and refining of ments. Because chromium is a reactive ele- grain size. The yield point is eliminated by ment, its success depends to a great degree on slight elongation by temper rolling or elimina- maintaining the homogeneity required for tion of interstitial carbon and nitrogen, whose proper corrosion-resistant performance. Incor- interaction with dislocations causes the yield rect thermal processing is the main way homo- point. The anisotropy is either utilized to advan- geneity can be lost. Stabilizing makes it much tage by maximizing it, as in the case of deep- easier to keep chromium from segregating in drawing alloys, or minimized by refining grain ferritic alloys. size and randomizing grain orientation by spe- A by-product of stabilization with titanium is cial thermomechanical processing. that oxygen and sulfur are also eliminated as Ferrite has a greater thermal conductivity compounds of titanium along with carbon and and lower thermal expansion than austenite. Its nitrogen. These impurity elements would other- strength decreases with temperature more than wise also precipitate as compounds containing that of austenite, but the good match in thermal some chromium, potentially depleting chromium expansion between the ferrite and its oxide in the vicinity of their precipitation. still makes it an excellent high-temperature Chapter 1: Metallurgy / 5

material. Ferrite has very nearly the same cor- basis. Their use is limited by their solubility and rosion resistance as austenite, but since ferrite their tendency to form precipitating compounds can hold no nitrogen in solution, it cannot ben- with chromium. Manganese acts largely through eﬁt from this element. In duplex alloys, the its ability to promote nitrogen solubility. Super- ferrite is generally the more corrosion resistant austenitic stainless steels, such as S34565, use 4 phase because it is richer in chromium and to 6 % manganese to permit nitrogen levels of molybdenum. 0.4 to 0.6% to be achieved, resulting in higher pitting corrosion resistance. Since all stainless steels contain principally Austenite iron and chromium, the addition of a substantial The second major constituent phase of the amount of austenitizing elements is necessary to stainless steel alloy system is austenite. Austen- transform the structure to austenite. As a rule of ite has an fcc atomic structure. The fcc structure thumb, iron alloys require about 17% chromium is common in many transition metals to the right and 11% nickel (or its equivalents) to remain of iron in the periodic table. As stated, the fcc austenitic at room temperature. One percent structure should be considered normal for metals nickel can be replaced by about 2% manganese well below their melting temperature as it is a as long as nitrogen is present to maintain the denser structure. The presence of the bcc struc- same phase stability. The omnipresent carbon ture relates to the unpaired 3d electrons, which and nitrogen have an effect 30 times that of provide ferromagnetism. Adding elements to nickel, so even in the small amounts in which iron that causes pairing of the 3d electrons di- they are normally present, they have a signiﬁ- minishes ferromagnetism and promotes the fcc cant effect. These stabilizing factors are mapped structure. Nickel and manganese are the most in the Schaeffler diagram of Fig. 2 (Ref 1), prominent alloying elements that do this, but the whose purpose is to predict the phase makeup interstitials carbon and nitrogen are the most of weld metal. Since welds solidify relatively powerful austenite stabilizers on a percentage rapidly, no carbides or intermetallic phases

Fig. 2 Schaeffler-Delong constitution diagram showing phases present in as-solidiﬁed stainless steels at room emperature as a function of composition demonstrating carbon and nitrogen contributions to nickel effects. Adapted from A.L. Schaeffler, Constitution Diagram for Stainless Steel Weld Metal, Met. Prog., Vol 56, Nov 1949, p 680–688; and W.T. Delong, A Modiﬁed Phases Diagram for Stainless Steel Weld Metals, Met. Prog., Vol 77, Feb 1960, p 98 6 / Stainless Steels for Design Engineers

form, and only ferrite, austenite, and martensite and then frozen in a state of supersaturation in will be present. Thus, they provide useful infor- the austenite when it forms on cooling. The sul- mation about the compositional effects on phase fur and oxygen then precipitate during cooling or development in nonequilibrium situations. The subsequent hot working as isolated inclusions. nickel equivalent (vertical axis) summarizes The interface between these inclusions and the how nitrogen, carbon, and other elements com- matrix is the locus of corrosion pit initiation, bine to create a nickel-like effect. The horizon- quite probably because of chromium depletion tal axis does the same for chromium and those occurring during and as a result of inclusion elements that have a similar effect. growth. When an alloy solidifies as austenite, In most common stainless steels, austenite is sulfur immediately segregates to the grain normally present in the metastable state, for ex- boundaries because of its low solubility in ample, the retained austenite in alloy steels. Those austenite, and it forms a low-strength film with a with carbon above 0.02% would eventually break low melting temperature. This causes poor hot down into austenite plus carbides, and those with workability and hot cracking of welds. less than about 30% chromium plus nickel will The diffusion rates in austenite are quite low form martensite if deformed sufficiently. But in compared to ferrite, so even interstitial elements the annealed state, the austenite in standard cannot move quickly enough to precipitate below austenitic stainless steels will remain indefinitely about 400 ¡C (750 ¡F). This permits carbon and as fully austenitic without precipitates unless nitrogen to exist in very high degrees of supersat- heated above 400 ¡C (750 ¡C) for protracted peri- uration if introduced below this temperature, as ods of time or deformed extensively. is done by various proprietary processes. The Interstitial elements are much more soluble in low diffusion rates restrict such colossally super- austenite than in ferrite. Of these, only nitrogen saturated zones to thin surface layers, but they is considered a beneficial alloying element. It can reach phenomenal hardness of over Rc 70. both strengthens and improves the pitting corro- The austenite structure does not discourage the sion resistance of austenite. Carbon has a paral- formation of intermetallic compounds such as lel effect, but its tendency to form chromium sigma, but it does, fortunately, make their forma- carbides limits its use and in fact leads to its tion very sluggish, as seen in Fig. 4. The differ- minimization in most alloys. Before the AOD ence of three orders of magnitude for carbide was developed and carbon levels in stainless formation reflects the difference between the dif- steels were higher, austenitic stainless steels fusion of carbon and that of substitutional ele- were sometimes stabilized by titanium or nio- ments. The formation of sigma in ferrite is about bium to counter the effects of carbon. Both car- 100 times faster than in austenite. Sigma is al- bon and nitrogen stabilize the austenite phase, most never seen in commercial 316 alloys. permitting lower levels of nickel to be used in austenitic alloys. Interstitial atoms of carbon and nitrogen distort the fcc lattice, causing it to expand about 1% linearly per 1 wt% of solute (Fig. 3) (Ref 2). This produces solid solution hardening of the austenite. The work hardening of austenite is increased by nitrogen. A third interstitial solute, hydrogen, produces the same effect but to a lesser degree. Austenite is not embrittled by hydrogen to the extent ferrite or martensite is, but hydrogen does raise its flow stress and hardness while lowering its work-hardening rate. Sulfur and oxygen are considered impurities because they form inclusions, usually chrome/ manganese silicates and sulfides. If present in sufficient amounts, sulfur and oxygen precipitate as primary inclusions before or during solidification. In most austenitic stainless alloys, the re- mainder of these elements are near saturation in the as-solidified ferrite at very high temperatures Fig. 3 Lattice expansions due to carbon. Source: Ref 2 Chapter 1: Metallurgy / 7

The mechanical properties of austenite are The physical properties of austenite com- quite different from those of ferrite. Austenite is pared to ferrite include lower thermal and elec- characterized by: trical conductivity and greater thermal expansion. It is also, of course, nonmagnetic. ¥ Low stacking fault energies leading to high work-hardening rates ¥ Good toughness even at very low tempera- Martensite tures Martensite is a phase that forms from the dif- ¥ Low notch sensitivity fusionless shear of austenite to a distorted cubic ¥ Lack of a sharp elastic limit or hexagonal structure. This transformation can ¥ Good high-temperature strength occur spontaneously on cooling or isothermally ¥ Fairly isotropic mechanical properties with externally applied deformation. It is essentially ferrite that has been formed with a super- While there is not a great deal of difference saturation of carbon. The resulting structure is in the yield strengths of austenitic and ferritic very fine and highly faulted, making it quite alloys of similar alloy levels, austenitic alloys hard. As in carbon steel, the hardness of the are more ductile, have high work-hardening martensite increases dramatically with intersti- rates, and therefore have higher tensile tial content because of the huge strain intersti- strengths. Austenite can be cold worked to ex- tial elements impose on the bcc lattice, distort- tremely high strengths, around a maximum of ing it into tetragonality. 2000 MPa (290 ksi). Chapter 3, “Austenitic Martensite in stainless steels is restricted to Stainless Steels,” gives a more thorough and alloy levels at which austenite can form at quantitative treatment of the mechanical prop- higher temperatures but at which the austenite is erties of austenite. unstable at ambient temperatures. This gives In duplex stainless steels, a secondary austen- martensite a fairly narrow composition range. γ ite, 2, can form from ferrite below 650 ¡C The lowest alloy level is that of the basic 12% (1200 ¡F). At this temperature, it has the same chromium steels with 0.1 to 0.2% carbon. The composition as the ferrite from which it forms most highly alloyed martensites are found in the and is called type 1. In the 650 to 800 ¡C (1200 precipitation-hardening grades. Thus, marten- to 1470 ¡F) range, a range that can be encoun- sitic stainless steels are inherently limited in γ δ tered in the heat-affected zone (HAZ) at / corrosion resistance to a level no better than a boundaries during welding, another type forms. 17 or 18% chromium alloy and often barely γ This so-called secondary austenite, 2, type 2, is qualify as stainless after the chromium tied up somewhat enriched in nickel over the ferrite as chromium carbide is recognized as not con- from which it forms but poorer in nitrogen than tributing to the corrosion resistance. the primary austenite, giving it poorer corrosion The as-formed martensite to the degree it has resistance. Secondary austenite can also coform significant carbon content is hard and requires γ δ with sigma as / grain boundaries are depleted tempering to give it adequate toughness. The of chromium. This secondary austenite is called tempering reaction is the precipitation of car- type 3 and is also poor in chromium. bon in the form of carbides with the concurrent loss of internal strain in the martensite lattice. The complexities of tempering require its discussion in detail to be found in Chapter 3, “Martensitic Stainless Steels.” It is worth noting, however, that all tempering involves carbide formation, thus losing some corrosion-fighting chromium. There are two forms of martensite, the ε, epsilon, and the α', alpha prime. Epsilon is formed in steels with low stacking fault energy, which are primarily the leaner austenitic alloys. Thus, it forms at cryogenic temperatures or by cold working. It appears in martensitic alloys of the precipitation-hardening type. It is nonmagnetic, Fig. 4 Precipitation kinetics in 316 stainless steel. Source: Ref 3 has a hexagonal close-packed (hcp) structure, 8 / Stainless Steels for Design Engineers

and is very difficult to identify microscopically. chromium or molybdenum, causing localized The a' martensite is the familiar magnetic vari- lower corrosion resistance. Intermetallic phases ety known in alloy steels that forms both by form by diffusion of substitutional alloying ele- quenching and by deformation. ments, which makes their precipitation slower The mechanical properties of stainless than that of carbides, but they can form in a martensite are parallel to those of alloy steels. matter of minutes in alloy-rich grades. Defor- The high quantity of alloying elements in mation, which enhances substitutional diffu- stainless give an extreme depth of hardening, sion, accelerates their formation. The principal so there is no concern with ancillary phases intermetallic phases are described next. such as bainite. The physical properties are Alpha Prime. Not to be confused with very close to those of ferrite of the same com- martensite, alpha prime is an ordered iron- position. chromium phase (i.e., iron and chromium atoms occupy speciﬁc, rather than random, sites on Intermetallic Phases two intersecting superlattices). This structure is quite brittle. It forms at relatively low tempera- The number of phases that can coexist in an tures, between 300 and 525 ¡C (570 and 980 alloy is proportional to the number of alloying ¡F). Before its true nature was understood, its elements in the alloy. Table 1 lists data on the presence was known through its causing the more common precipitates found in stainless phenomenon called 475 embrittlement, origi- steel. It is not surprising that stainless steel with nally called 885 ¡F embrittlement. This is some- iron, chromium, nickel, manganese, silicon, and times confused with temper embrittlement, often molybdenum, titanium, and niobium which occurs in the same temperature range but should have numerous ancillary phases. Inter- is caused by phosphide precipitation on prior metallic phases are normally hard and brittle. austenite grain boundaries of martensite. Alpha They can render the bulk alloy brittle when they prime precipitation can cause 475 embrittle- form along grain boundaries. The other concern ment in ferritic or duplex stainless steels and arising from intermetallic phase formation is the limits their use in this temperature range but not depletion from the surrounding matrix of at higher temperatures, at which the phase dissolves. This phase forms at chromium contents Table 1 Precipitated phases found in stainless as low as 15%, but fortunately it takes a rela- steels tively long time to form, on the order of hours, Precipitate Structure Parameter, A Composition so it will not occur inadvertently during thermal processing such as welding or annealing. NbC fcc(a) a = 4.47 NbC Sigma. Sigma is a brittle tetragonal phase NbN fcc a = 4.40 NbN richer in chromium and molybdenum than ei- TiC fcc a = 4.33 TiC ther the ferrite or austenite matrix around it. It TiN fcc a = 4.24 TiN forms preferentially at ferrite-austenite bound- Z-phase Tetragonal a = 3.037 c = 7.391 CrNbN aries in the temperature range 600 to 1000 ¡C (1110 to 1470 ¡F) in alloys with more than M C fcc a = 10.57Ð10.68 Cr Fe Mo C (e.g.) 23 6 16 5 2 about 18% chromium plus molybdenum. Its M23(C,B)6 fcc a = 10.57Ð10.68 Cr23(C,B)6 composition is sometimes given as (CrMo)35 M C Diamond a = 10.62Ð11.28 (FeCr) Mo C; 6 21 3 (FeNi) , but examination of the iron-chromium cubic Fe Nb C; M SiC 65 3 3 5 phase diagram shows that it is archetypically an M N Hexagonal a = 2.8 c = 4.4 Cr N 2 2 equiatomic iron chromium compound. It is MN Cubic a = 4.13Ð4.18 CrN strongly promoted by silicon and suppressed by nitrogen. Stabilized alloy grades show more Gamma fcc a = 3.59 Ni3(Al,Ti) prime rapid sigma formation than unstabilized alloy Sigma Tetragonal a = 8.80 c = 4.54 Fe, Ni, Cr, Mo grades (e.g., 347 versus 304). In unstabilized alloys the prior precipitation of carbides destabi- Laves Hexagonal a = 4.73 c = 7.72 Fe2Mo, Fe2Nb phase lizes austenite, leading to subsequent sigma for- Chi phase bcc(b) a = 8.807Ð8.878 Fe Cr Mo mation. This makes alloys like 310H, 36 12 10 essentially 25Cr-20Ni, especially prone to G-phase fcc a = 11.2 Ni16Nb6Si7, Ni16Ti6Si7 sigma formation. Sigma forms much more rapidly from ferrite (a) fcc, face-centered cubic. (b) bcc, body-centered cubic. than from austenite because of the 100-fold Chapter 1: Metallurgy / 9

higher diffusion rate of alloy elements in ferrite. The precipitation of the carbide from ferrite This makes it a much larger issue in superfer- occurs at grain boundaries, is extremely rapid, ritic and duplex alloys, which have high and cannot be suppressed by quenching. Less chromium and/or molybdenum levels. Chapter than 20 ppm carbon content is required to pre- 7, “Duplex Stainless Steels,” contains an in- vent its precipitation from ferrite, although up depth discussion of sigma. to 50 ppm can be effectively kept in solution by Chi. Chi, χ, is similar to sigma except it con- very vigorous quenching. From austenite, car- tains more molybdenum and less chromium and bide precipitation occurs below about 900 ¡C has a cubic structure. It can coexist with sigma (1650 ¡F) for carbon levels under 0.10% and at and forms in the same temperature range. It also 650 ¡C (1200 ¡F) for carbon levels below precipitates at ferrite-austenite boundaries and 0.03%. For practical purposes, precipitation has the same deleterious effects. ceases below 500 ¡C (930 ¡F) due to the slow- Laves Phase. The laves phase has the struc- ing diffusion of carbon. While carbon is essen- ture A2B where A is iron or chromium and B is tially insoluble in austenite at room tempera- molybdenum, niobium, titanium, or silicon. It ture, quenching can easily preserve up to 0.10% forms at 550 to 650 ¡C (1020 to 1200 ¡F) over in supersaturation, as is commonly seen in type the course of hours. Thus, although its effect 301 stainless. would be deleterious, it seldom becomes a prac- The carbide precipitation occurs first at grain tical problem. It is possible for it to form at tem- boundaries. The chromium that combines with peratures below sigma and above alpha prime, the carbon comes from the matrix in the imme- but the long times for formation make it rare. diate vicinity and therefore decreases the chromium content of that region, giving rise to the phenomenon of sensitization, which comes Carbides, Nitrides, Precipitation Hardening, from the original phrase “sensitization to inter- and Inclusions granular corrosion.” Nickel and molybdenum Carbon and nitrogen are very important in all decrease the solubility of carbon and thus accel- steels, but they take on a special significance in erate the precipitation. Nitrogen retards precipi- stainless steel because chromium, the essential tation. Cold work accelerates precipitation. The alloying element of stainless steel, reacts more carbide has a hardness of about Rc 72. This vigorously with carbon and nitrogen than iron makes the phase a useful constituent in wear does. Except for its role in hardening martensite resistance in martensitic alloys. and strengthening austenite at high tempera- In higher carbon grades such as the marten- tures, carbon is almost universally a detrimental sitic stainless alloys, additional, more carbon- impurity from a corrosion point of view and is rich, carbides may form. These include M7C3 minimized. Its beneficial effect on corrosion re- and M3C. The latter carbide forms during the sistance when it is in solution is negligible low-temperature tempering of martensite, while because so little of it can be held in solution. the former precipitates at higher temperatures. Nitrogen has a lesser tendency to form com- Stabilizing carbides are those that are formed pounds with chromium, so it is considered a by the intentional addition of elements such as beneficial alloying element in austenite but not titanium and niobium. These elements form car- in ferrite, in which it has essentially zero solu- bides of the type MC (metal carbide). The car- bility. Common carbide and nitride precipitating bon in these compounds may be replaced by ni- phases are also listed in Table 1. trogen or, in the case of titanium, sulfur. These

Carbides. M23C6 is the main carbide found in carbides form preferentially over chromium car- stainless steel. Its structure is orthorhombic, and bides and thus prevent sensitization. They pre- it contains both iron and chromium. It can form cipitate in both the liquid and solid states. In the at any temperature at which the host austenite or solid state, the precipitate normally forms within ferrite becomes saturated with carbon. It is grains. The Ti(CN) appears as a cube of gold mainly chromium carbide, but iron can substitute TiN surrounded by gray TiC. The Nb(C,N) is for chromium up to about 50%. Other elements, less regularly shaped. They affect mechanical such as tungsten, vanadium, and molybdenum, properties in ferrite both by their inﬂuence on can also dissolve in this carbide. The ratio of recrystallization and by their ability to act as nu- chromium to iron in the carbide increases with cleation sites for brittle fracture time and temperature, as chromium diffusion Nitrides. At low levels, nitrogen can substi- permits, up to a maximum of 4 or 5 to 1. tute for carbon in M23C6. At higher nitrogen 10 / Stainless Steels for Design Engineers

levels, Cr2N can form. This can occur in duplex tal as inclusions have been shown to be the initi- alloys if they are heated to a solution annealing ation sites for corrosion pits, which have been temperature at which the alloy has high solubil- linked to both their sulfur ions disrupting the ity for nitrogen. Cooling from these tempera- passive layer and their chromium content caus- tures can cause the excess nitrogen to precipi- ing slight local chromium depletion. tate as needles of Cr2N. Another nitride CrN can form in HAZs of welds. Precipitation-Hardening Phases. Phases that have a very similar lattice match to the par- Properties of Stainless Steels ent phase can precipitate coherently, that is, without changing the continuity of the crystal Physical and mechanical properties of repre- lattice. In these cases, the slight mismatch sentative stainless steel alloys are summarized causes a strain that can significantly restrict dis- in Appendix 2. Properties are also discussed in location movement and thereby strengthen the chapters specific to each alloy family. The matrix. One such precipitate is gamma prime, reader is referred to primary sources, such as an intermetallic, ordered, fcc phase with the company web sites, such as Ref 4 and 5. composition Ni3(AlTi). Copper forms the epsilon phase, essentially pure copper, which causes precipitation hardening. The secondary REFERENCES hardening of martensite due to the precipitation of molybdenum nitride or carbide is also a pre- 1. D.J. Kotecki, Welding of Stainless Steels, cipitation-hardening reaction. Welding, Brazing, and Soldering, Vol 6, Inclusions. Inclusions are principally oxides ASM Handbook, ASM International, 1993, and sulfides that form in the melt (type I), at the p 677Ð707 end of solidification (type II), or in the solid 2. G.E. Totten, M. Narazaki, R.R. Black- (type III). Type I inclusions are the largest and wood, and L.M. Jarvis, Failures Related to are globular. Except when they are deliberately Heat Treating Operations, Vol 11 ASM kept to improve machinability, they are physi- Handbook, ASM International, 2002, p cally removed by various steelmaking practices. 192Ð223 Type II inclusions form in interdendritic spaces 3. High Performance Stainless Steels, Refer- as the solubility of oxygen and sulfur drop on ence Book Series 11 021, Nickel Develop- solidification. Type III inclusions precipitate the ment Institute, p 16 remaining oxygen and sulfur, up to 100 ppm for 4. ASM Handbook, Vol 1, Properties and normal manganese-silicon killed stainless Selection, ASM International, 1990 steels, in the solid state either on preexisting in- 5. ASM Speciality Handbook, Stainless Steels, clusions or as micron-size particles. Inclusions ASM International, 1996 are mainly oxides and sulfides of silicon and manganese. If more reactive elements, such as aluminum or titanium, are present, their oxides SELECTED REFERENCES and sulfides can also be present. Sulfides and oxysulfides can be beneficial for ¥ D.J. Kotecki and T.A. Siewert, WRC 1992 machining as solid-state lubricants and chip Constitution Diagram, Welding Journal, Vol breakers. Otherwise, their presence is detrimen- 5, 1992, p 171sÐ178s Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 11-18 All rights reserved. DOI: 10.1361/ssde2008p011 www.asminternational.org

CHAPTER 2

Corrosion Theory

Summary Electrochemical Reactions

THIS CHAPTER INTRODUCES THE funda- In electrochemical reactions, charge is trans- mentals of electrochemical theory as it pertains ferred across interfaces of species of different to corrosion. Topics covered include an overview chemistry. Consider, for example, the reaction: of electrochemical reactions, Faraday’s law, the Nernst equation, galvanic versus electrochemical ++→+2+− 2224Fe() s O22 () g H O Fe OH (Eq 1) cells, and Pourbaix diagrams. The examples provided relate these fundamentals to the corrosion resistance of stainless steels. An inspection of this reaction suggests that three phases must be present for the reaction to proceed: an ion-conducting phase (water-based solution), a metallic phase (iron), and a gas g Introduction phase O2( ). Second, electrons have been transferred from the metallic phase, iron to O2 + Corrosion—the environmental degradation H2O. Figure 1(a) shows the arrangement of an of materials through electrochemical reac- experimental setup in which Reaction 1 could tions—is a key subject for more or less all proceed. classes of alloys that fall within the broad deﬁ- On the left, iron is allowed to dissolve nition of stainless steels because these alloys according to: were developed with the intention of prevent- +− ing corrosion. This chapter aims ﬁrst to provide 224Fe() s→+ Fe2 e (Eq 2) an introduction to the fundamentals of electrochemical theory as it pertains to corrosion. resulting in Fe2+ ions that dissolve in the Thermodynamics are presented in light of elec- water-based solution and electrons that are car- trochemical potentials as opposed to purely ried to the right side, where they participate in chemical ones. Chapter 3 introduces the formal the reaction: terms needed to describe electrode reaction kinetics. Chapter 4 describes the various forms of ++→−− OHOe22244 OH (Eq 3) corrosion and how they are related to alloy metallurgy, chemistry, and structure. Chapter 5 Inside the water-based solution, ions (Fe2+, focuses on oxidation. For an in-depth study of OHϪ, H+, or any others) migrate, thereby con- electrochemical kinetics and electroanalytical stituting a so-called ionic current. This current methods, Ref 1 is recommended. For a broader together with Reactions 2 and 3 and the trans- study of corrosion, the reader is referred to port of electrons from left to right form a closed texts by Jones (Ref 2), Uhlig and Revie (Ref 3), circuit called an electrochemical cell. The cell is and Fontana (Ref 4) and to ASM Handbook, made up of four parts: the two electrodes where Volume 13A (Ref 5). the charge transfer Reactions 2 and 3 take place 12 / Stainless Steels for Design Engineers

often are described as half cells, for example, − + Fe// O2 OH and Fe/ Fe 2 .

Faraday’s Law

If the cell in Fig. 1(a) was allowed to proceed and thermodynamics favored to proceed according to the direction in Reaction 1, then a current i will ﬂow from the anode to the cathode, and the amount of charge passed per unit time as a result of this current will be linked to the amount of iron dissolved per unit time or the amount of oxygen reacted per unit time by virtue of Eq 2 and 3. This is given by Faraday’s law:

nNF= it (Eq 4)

Here, i * t is the charge passed (in coulombs); N is the moles of consumed/produced specie (e.g., moles consumed iron in Reaction 2); n is the ratio of electrons to consumed/produced species, which in the case of Reaction 2 will be 2; and F is Faraday’s constant, which is essentially the charge in coulombs corresponding to 1 mole of electrons.

Fig. 1 Schematic illustration of (a) a differential aeration cell involving iron dissolution and (b) the same cell with a The Nernst Equation variable resistor and voltmeter Electrochemical reactions require a transfer (the anode and cathode, respectively), an elec- of charge; hence, there is a coupling between trolyte, and an electron pathway. It should be chemical and electrical energy. Consider the hy- noted that electrodes are interfaces that require pothetical setup in Fig. 1(a) with the addition of several phases to be in contact. Oxidation, Re- a variable resistor and a voltmeter, resulting in action 2, occurs at the anode and reduction, Re- the arrangement shown in Fig. 1(b). Thermody- action 3, occurs at the cathode. The electrolyte namically, the Gibbs free energy of the cell is is the medium through which the ions migrate; that of Reaction 1: in the case of corrosion reactions, this is most ΔΔ=− Δ commonly a water-based solution, but at high GHTS temperatures it could be a solid oxide. The ﬁnal ( )42( ) aa− 2+ (Eq 5) constituent of the electrochemical cell is a path- =+ΔGRT0 ln OH Fe aP way through which electrons can migrate from HO22 O the anode to the cathode. As a shorthand notation, electrochemical cells where ⌬G is the Gibbs free energy, H is the en- are written by separating components within a thalpy, S is the entropy, R is the gas constant, phase by a comma and separating phases by a and T is the absolute temperature. If this is neg- slash; gaseous species are written next to their ative, the reaction would be expected to proceed conducting electrode. For example, the cell de- spontaneously as written in Reaction 1. Let us scribed in Fig. 1(a) would be recorded as assume that this is the case. The thermal heat −+ Fe// O22 OH , Fe / Fe. This cell is an example produced by the system can be divided into two of a differential aeration corrosion cell, which is parts: the thermal heat produced by the cell Qt discussed later. Processes at a single electrode and the heat produced at the resistor QRes. QRes Chapter 2: Corrosion Theory / 13

in this case is heat, but in essence it represents Here, p and r are the concentrations of reac- i i α β the available energy or work, which in the case tant and products, respectively, and i and i are of a resistance is given by the product of charge the numbers that are needed to balance the reac- passed times potential difference. If the resist- tion stoichiometrically. In the case of Reaction ance approached infinity (R→ ), Reaction 1 1, Eq 10 would be: would proceed through infinitesimal steps and can be considered thermodynamically reversible. ⎛ ( )42( ) ⎞⎞ RT aa− 2+ In this case, the thermal heat produced by the EE=−0 ln ⎜ OH Fe ⎟ (Eq 11) rxn rxn ⎜ ⎟ 4FPaHO O cell is minimized and according to thermody- ⎝ 22 ⎠ ⌬ namics is given as Qt = Qrev = T S1. On the other hand, the net work gained QRes is maxi- If the emf according to Eq 11 is positive, this mized and constitutes the rest of the free energy: means that the free energy is negative (according to the Nernst equation); hence, the net reaction is QGHTS==−ΔΔ Δ Res (Eq 6) thermodynamically favored as it is written in Reaction 1. By inspection of Eq 11, it can be As mentioned, the energy dissipated through seen that it is the difference between two hypo- the resistance is charge passed times potential ()EE=−− E thetical half reactions, rxn O/OH Fe/Fe2+ difference, and in this case the potential differ- defined as: 2 ence is the reversible potential difference E; thus, in an absolute sense: ⎛ ( )4 ⎞ a − =−0 RT ⎜ OH ⎟⎟ EE−−ln (Eq 12) O/OH22 O/OH ⎜ ⎟ Δ = 4FPaHO O GnFE (Eq 7) ⎝ 22⎠

Here, n is the number of electrons passed per which corresponds to the reduction Reaction 3 atom of iron reacted, and F = 96,485 C per mole and: electrons, is Faraday’s constant. The reversible potential difference E represents the potential ⎛ 4 ⎞ RT (a ) difference between the two electrode reactions EE=−0 ln ⎜ Fe ⎟ (Eq 13) Fe2+ /Fe Fe2+ /Fe 4F ⎜ a ⎟⎟ (cathode and anode), and as such they are asso- ⎝ Fe2+ ⎠ ciated with Reaction 1 rather than a physical cell. The potential difference is referred to as which corresponds to the reverse of Reaction the electromotive force (emf) of the cell. It is 2, that is, if it was a reduction reaction. The also referred to as the open circuit potential be- potentials as written in Eq 12 and 13 are cause it is the potential measured by the volt- called reduction potentials, and because meter in Fig. 1(b) when a negligible current =− EErxn − E2+ has to be positive for the O/OH2 Fe/Fe flows. It is defined here as Erxn. By convention, reaction to be thermodynamically favored as this potential is positive for a spontaneous reac- written in Eq 1, the reduction potential E − tion (as opposed to the chemical free energy, O2 /OH has to be larger than EFe2+ /Fe . If it was not, then which is negative); hence, Eq 7 becomes: Reaction 1 would proceed in the reverse direction, which means that the electrode Reactions ΔGnFE=− rxn (Eq 8) 2 and 3 would be reversed and thus so would the anode and cathode of the cell. and if all elements have unit activities: It is useful to list reduction potentials for half- cell reactions, just as it is useful to list free en- ΔGnFE00=− (Eq 9) ergy data. However, half-cell potentials (like rxn any electrical potentials) cannot be measured in an absolute sense; only potential differences can Equation 8 is the Nernst equation. By virtue be measured. (EE=− E can be rxn O/OH–2+ Fe/Fe of Eq 8 and 9 and the expression for Gibbs free 2 measured because it is a difference.) Therefore, energy of a reaction (e.g., Eq 5), an expression half-cell potentials are measured with respect to for E is obtained: rxn a reference electrode. Reference electrodes are ααα constructed such that they have a stable potential; RT ⎛ ppp123 ...⎞ =−0 123 (Eq 10) this is discussed further in Chapter 3. A com- EErxn rxn ln ⎜ βββ ⎟ nF ⎝ 1233 ⎠ rrr123... mon reference electrode in aqueous solutions is 14 / Stainless Steels for Design Engineers the normal hydrogen electrode (NHE), also Galvanic versus Electrochemical Cells known as the standard hydrogen electrode (SHE), with a potential set (arbitrarily) as zero When reactions in a cell occur spontaneously at all temperatures. The NHE is schematically in the direction dictated by the open-circuit po- shown in Fig. 2. In shorthand notation, it is: tential of a cell that is positive ()E > 0 , a cur- ==+ rxn Pt/(HH2 a11 )/( a ), and the half-cell reac- rent flows as shown in Fig. 3(a). This is the case tion is: in environmentally caused electrochemical corrosion reactions. It also is the case in fuel cells 22HH+−+=e (Eq 14) 2 and batteries (under discharge), in which the Table 1 (Ref 6) lists half-cell reduction stan- current is used as electricity. These types of 0 cells are called galvanic cells, in which chemi- dard potentials (EOx/Re ) versus NHE that are a result of the emf of the following types of cells cal energy is converted to electrical energy. (for Reaction 2, as an example): Most of the discussion in the following chapters concerns these types of cells. In electrolytic Pt/(H a==11 )/( H++ a ),( Fe2 a = 1 )/ Fe cells (Fig. 3b), an imposed electrical potential 2 counters the “natural” cell potential to drive a reaction in a desired direction. These types of cells are used for many metallurgical processes, such as electroplating, electrorefining and elec- troextraction (e.g., the Hall-Heroult aluminum smelting cell), and for other applications, such as charging batteries. In the case of corrosion, the principle is used for protection against corrosion. In electrolytic cells, electrical energy is converted to chemical energy.

Table 1 Standard half-cell reduction potentials versus the normal hydrogen electrode

Standard half-cell reduction Reaction potential vs. NHE(a) (V)

Fe32+−+=e Fe + 0.771 ++=− – = OHO2 2442 e OHpH() 14 0.401 + +=− 22HHe 2 0.000 Ni2+−+=2e Ni –0.250 Fe2+ +=2e− Fe –0.447 Cr3+ +=3e− Cr –0.744 +=+−− = 22HO2 e H2 2 OHpH() 14 –0.828 Fig. 2 The normal hydrogen electrode (NHE) (a) NHE, normal hydrogen electrode. Source: Ref 6

Fig. 3 Schematic of (a) galvanic cell and (b) electrolysis cell Chapter 2: Corrosion Theory / 15

Corrosion Tendency ing to Eq 13, which, assuming a Fe2+ activity of 10−6 (this is an arbitrary value but is usually The tendency to corrode, that is, whether a taken to represent a low ion concentration), be- system consisting of anode, cathode, and elec- comes at room temperature (using Table 1 for trolyte can react thermodynamically, is deter- the standard potential): mined by evaluating Erxn . If this is positive, then there is thermodynamically a possibility for cor- =− − Ea2+ 0..log() 447 0 0295 2+ rosion. The rate of corrosion, which is in most Fe /Fe Fe = cases determined by corrosion kinetics, is dis- –0.624 V (vvs. NHE) (Eq 16) cussed in Chapter 3. Consider, for example, a case of iron in aerated water. Figure 1 (with Figure 4(a) shows a schematic plot of the two electrode Reactions 2 and 3) can be viewed as reduction potentials (Eq 15 and 16) versus pH. an idealized equivalent cell for this situation. It Because a spontaneous reaction requires Erxn should be noted, however, that the locations of to be positive, if the only pertinent reactions anode(s) and cathode(s) on the iron surface can- were Eq 2 and 3, this means that corrosion (due not be identiﬁed with ease. At room temperature to iron dissolution to Fe2+ and oxygen reduc- (298 K), 1 atm oxygen partial pressure, and tion) is possible when the line representing − using Table 1, Eq 12 can be written by assuming EO2/OH (Eq 15) lies above the line representing E + unit activity for water and unit activity coeffi- Fe2 / Fe (Eq 16). This is indicated by the region cient for OH−: shaded in gray in Fig. 4(a). Hydrogen reduction is another possible cathode reaction in water: Ep=+0.. 401 0 059 OH OOH/ – +− 2 +→ 22HHe 2 (Eq 17) =+0..() 401 0 059 14 −pH = 1..− .2270 059pH V (vs. NHE) (Eq 15) and its reduction potential is (using the deﬁni- tion of pH):

Here, the following deﬁnition of pH has been P 0 RT H2 used: pH = –log C , pOH = –log C and pH EE+ =−ln H+ OH− HH// HH+ 2 2 2 nF a + pOH = 14. Similarly, the iron dissolution Re- H+ action 2 will have a reduction potential accord- =−0 0. 059pH Vvs ..NHE (Eq 18)

Fig. 4 Reduction potential versus pH for iron and (a) oxygen gas reduction and (b) hydrogen ion reduction 16 / Stainless Steels for Design Engineers

= − + Figure 4(b) shows the condition in which corro- Fe(OH)22 HFe O +H , sion under deaerated conditions (due to iron dis- =+ 2+ pH 14.log 30 (a − ) (Eq 19b) solution to Fe and hydrogen ion reduction) is HFeO2 possible as a gray shaded region. In Fig. 4(a) and (b), the regions where iron is stable are denoted as immunity (corresponding to immu- 3+ += + nity from corrosion). When comparing these Fe3 H2 O Fe(OH)3 +3H , = two ﬁgures, it is noteworthy that hydrogen ions pH 1.613–(1/3) log(a 3+ ) (Eq 19c) are able to cause corrosion only under relatively Fee low pH conditions, whereas oxygen gas is able to corrode iron in the entire pH range. Since these are independent of potential, they will appear as vertical lines (see lines 19a to 19c in Fig. 5a). The following pH-independent electrochemi- The Construction of Pourbaix Diagrams cal reactions need to be considered, and they will result in horizontal lines (Fig. 5a): Figures 4(a) and (b) are types of phase diagrams that show the stable phases in an area bounded by pH and potential. In reality, several Fe2+ +→2e− Fe, electrochemical and chemical reactions need to Ea=−0..log 447 + 0 0295 ( )) be considered when constructing these types of Fe2+ /Fe Fe2+ (Eq 20a) diagrams. Each reaction is represented by a line. In the case of iron, the following chemical reactions will have to be considered (the pH de- HFeO2−−+= H O Fe( OH) + 2e , pendency of these reactions is listed next to 2 3 E =−08.110− 0.log 0591 (a ) them [Ref 7] and since they are not electro- Fe(OH)/ HFeO2– HFeO– chemical, they are evaluated from the equilib- 3 2 rium constants): (Eq 20b)

++ The following electrochemical reactions will Fe2 +=22 H OFe() OH + H , 2 2 depend on pH and thus will be sloped depend-

pH =−665..log 05 (a + ) (Eq 19a) Fe2 ing on this dependence (Fig. 5a).

Fig. 5 Pourbaix diagram for iron. (a) Schematic matching Eq 19 to 21 in text to lines. (b) Actual complete diagram. Source: Ref 7 Chapter 2: Corrosion Theory / 17

Fe+=22 H O Fe( OH) ++ H+ 2e− , 2 2 E =−0. 0470 −− 0 0591. pH (Eq 21a) Fe(OH)2 / Fe

+=−− +++ Fe232 H2 O HFeO2 H e , =− E - 0..8 495 0 0886pH HFeO2 /Fe + 0.log 0295 (a ) (Eq 21b) – HFeO2

Fe2+ +=33 H O Fe( OH) ++ H+−e , 2 3 = − E 2+ 10. 557 0. 1773pH Fe(OH)3 / Fe − 0.log 0591 (a ) (Eq 21c) Fe2+

+= +++− Fe(OH)23 H2 O Fe(OH) H e , Fig. 6 Pourbaix diagram for chromium in water. Source: Ref 8 E = 00271..− 0 0591pH (Eq 21d) Fe(OH)32 /Fe(OH)

For the pH-dependent reactions (chemical reduction are able to cause corrosion through and electrochemical), one can readily label the the entire pH region. Unfortunately, Fe-OH cor- regions depending on what iron species increas- rosion products are generally not passivating. ing pH favors. If iron would be an anode and Iron or carbon steel alloys are therefore not par- the tendency to corrode were to be evaluated, ticularly corrosion resistant in water solutions. then the reduction potential for a possible cath- Figure 6 shows the Pourbaix diagram for ode reaction would be placed on this diagram. If chromium (Ref 8). While chromium oxidizes this point were to be, for example, in A in Fig. even more readily than iron, it forms Cr2O3 over 5(a), this means that the reduction potential for a signiﬁcantly large region that is of relevance this assumed cathode lies below any reduction to pH values in water solutions. Since Cr2O3 is potential of iron, and hence under these condi- protective, it prevents further corrosion. When chromium is added to iron as an alloying ele- tions iron is immune (since Erxn is negative). In fact any Fe2+ ions present could plate as iron. ment, it corrodes selectively due to its low re- On the other hand, if the reduction potential of duction potential, but this means that it also pro- the assumed cathode reaction were to lie in tects the iron alloy due to the properties of point B, then there is a tendency to dissolve iron Cr2O3. This is the basic design principle behind 2+ iron-chromium-based stainless steels. to Fe since Erxn is positive. Finally, if the reduction potential of the assumed cathode was at point C, corrosion would occur, resulting in

Fe(OH)3, but when oxides or hydroxides are REFERENCES formed there is a possibility that this product could form a solid protective layer that kineti- 1. A.J. Bard and L.R. Faulkner, Electrochemi- cally hinders further corrosion. These types of cal Methods: Fundamentals and Applica- diagrams are called Pourbaix diagrams. Figure tions, 2nd ed., Wiley, 2001 5(b) shows the Pourbaix diagram for iron over- 2. D.A. Jones, Principles and Prevention of laid with the common cathode reactions in Corrosion, 2nd ed., Prentice Hall, 1996 water, Eq 15 and 18 (Ref 8). The ionic activity 3. H.H. Uhlig and R.W. Revie, Corrosion and was previously arbitrarily set at 10–6, but from Corrosion Control: An Introduction to Cor- the Pourbaix diagram it can be seen that changes rosion Science and Engineering, 3rd ed., in ion activity do not have dramatic effects on Wiley, 1985 the boundaries. It can be seen that both the 4. M.G. Fontana, Corrosion Engineering, 3rd oxygen gas reduction reaction and hydrogen ion ed., McGraw-Hill, 1986 18 / Stainless Steels for Design Engineers

5. ASM Handbook, Vol 13A, Corrosion: Fun- 8. S.A. Bradford, Corrosion Control, 2nd ed., damentals, Testing, and Protection, S.D. CASTI Publishing, Inc., 2001, p 41 Cramer and B.S. Covino Jr., Ed., ASM International, 2003 6. Handbook of Chemistry and Physics, 71st SELECTED REFERENCE ed., CRC Press, 1991 7. D.A. Jones, Principles and Prevention of • M. Pourbaix, Atlas of Electrochemical Equi- Corrosion, 2nd ed., Prentice Hall, 1996, p. 59 libria in Aqueous Solutions, NACE, 1974 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 19-25 All rights reserved. DOI: 10.1361/ssde2008p019 www.asminternational.org

CHAPTER 3

Corrosion Kinetics

Summary the penetration due to dissolution of element i becomes: CORROSION INVOLVES chemical reactions with equilibrium that is known through ⋅ tMi thermodynamics. In practice, the rate at which rj= (Eq 1) nF⋅⋅ρ corrosion reactions occur is the most important i consideration. This chapter deals with corrosion kinetics, which allows us to understand the The penetration rates for iron and various rates of corrosion. stainless steels are listed in Table 1 (Ref 1) in units of mils (0.001 in.) per year, or mpy. In the

case of alloys, the ratio of Mi /ni is computed as Introduction an equivalent weight (EW) according to:

Consider the differential aeration cell discussed in the Chapter 2 on corrosion theory, Fe/ 1 EW = O2/OH–, Fe2+/Fe. If the thermodynamic condi- fn ∑ ii (Eq 2) tions favor electrochemical corrosion of iron, M = i that is, EErxn –2– E+is positive, then OOH2 // FeFe a net corrosion current i will ﬂow, resulting in iron dissolution and consumption of oxygen gas where fi, ni, and Mi are the weight fraction, va- according to the net reaction, 2Fe (S) + O2 + lence, and molar mass of element i, →2 2+ – 2H2O Fe + 4OH . The magnitude of this respectively. current will determine the rate or iron dissolu- As stated, the amount of corroded (dis- tion according to Faraday’s law, which was solved) iron is determined by the current i, and introduced in Chapter 2: nNF = it. Because n = the magnitude of this current is determined by 2 and F = 95,485 C per mole electrons, the moles of dissolved iron are given as a function of time as N = i* t/(2* 95,485). Practically, this can be readily converted to lost mass m, which Table 1 Penetration rates for a current of 1 µA/cm2 (mpy) in the case of iron loss becomes m = N* MFe = Density, Penetration MFe* i * t/(2* 95,485), or thickness reduction r, which in the case of iron becomes r = M * i* Alloy Element/oxidation state g/cm3 EW(a) rate, mpy ␳ ␳ Fe Fe Fe/2 7.87 27.92 0.46 t/(2* 95,485* A* Fe). Here, MFe and Fe are 304 Fe/2,Cr/3,Ni/2 7.9 25.12 0.41 molar mass and density of iron, respectively. It 321 Fe/2,Cr/3,Ni/2 7.9 25.13 0.41 is often the thickness loss (referred to as 309 Fe/2,Cr/3,Ni/2 7.9 24.62 0.41 316 Fe/2,Cr/3,Ni/2,Mo/3 8.0 25.50 0.41 penetration per unit time) that is useful; 430 Fe/2,Cr/3 7.7 25.30 0.42 therefore, i/A is often replaced by j, which is 446 Fe/2,Cr/3 7.6 24.22 0.41 deﬁned as current density and has the units 20Cb3 Fe/2,Cr/3,Mo/3,Cu/1 7.97 23.98 0.39 amperes/square meters. A general equation of (a) Equivalent weight. Source: Ref 1 20 / Stainless Steels for Design Engineers

4 RT ⎛ (a ) ⎞ EE=−0 ln⎜ Fe ⎟ Fe2+ // Fe Fe2+ Fe 4F a ⎝ Fe2+ ⎠

When a cell is not under open circuit (i.e., a net current passes through it), the cathode and anode potentials deviate from the half-cell potentials, and the electrode states are then defined as being polarized. The polarization is quantified as overpotentials ␩, which are defined by the deviation from the equilibrium half-cell potentials, η = that is, for the cathode,c EEcathode – O/OH and η = 2 for the anode, EEanode – . Effectively, the a Fe2 + Fe Fig. 1 Schematic illustration of a differential aeration cell overpotential reduces the activation energy for involving iron dissolution. Kinetic steps: (1) electrode the electrode Reactions 3 and 4. In the case of a reactions, (2) ion conduction, (3) electron conduction reduction reaction at a cathode, such as Reac- tion 3, the overpotential is negative, and driving the corrosion potential. The corrosion potential an electrode toward a lower potential drives is determined by the reaction potential (which electrons from the electrode into the solution, was discussed in Chapter 2) and the kinetics of resulting in a net cathodic current ic at this elec- the various steps involved in completing the trode. Similarly, at the anode the overpotential electrochemical circuit depicted in Fig. 1. is positive, which results in electrons that are fa- These involve: (a) electrode reactions at the vored to be removed from the solution and cathode and anode, (b) conduction of ions in transferred into the electrode, thus producing a the electrolyte, and (c) conduction of electrons net anodic current ia. from the anode to the cathode. The conduction If the magnitudes of cathode and anode polar- of electrons is generally not a problem in stain- ization are large, as would be expected in a gal- less steels because the corroding metal (iron) vanic cell, the relation between each electrode and scale (Cr2O3) provide an easy path for elec- current/current density and overpotential is trons. The other two kinetic processes are dis- given by the following equations (for a thor- cussed briefly in this chapter. ough derivation of the current overpotential equation, Ref 2 is recommended): The Butler-Volmer Equation ⎛ i CtO (,)0 αηnF ⎞ j ==c j 2 exp⎜− c ⎟ (Eq 5) For the case study 2Fe (S) + O2 +2H2O c 0,c ⎝ ⎠ Ac C * RT → + – O2 2Fe2 + 4OH , the cathode and anode reactions are: and ++→−− OHO22244e OH (Eq 3) i ⎛ ()1− αηnF ⎞ j ==a j exp⎜ a ⎟ (Eq 6) a A 0,a ⎝ RT ⎠ and a

224Fe()se→+ Fe 2+− (Eq 4) where the jo,i terms are the exchange current densities and represent the equally large cath- The Nernst equation predicts an open circuit ode and anode currents at equilibrium (zero potential of EE= – E overpotential) at the electrodes. The exchange rxn OOH//–2 FeFe+ 2 current densities are a measure of the electrocat- where alytic ability of the surface to promote/demote ⎛ 4 ⎞ (a ) the electrode charge transfer reactions; as such, =−0 RT ⎜ OH– ⎟⎟ EE– – ln they can vary over many orders of magnitude O/OH2 O/OH2 ⎜ ⎟ 4F aPHO O ⎝ 22⎠ depending on the surface chemistry and structure and on electrode reaction. The α-terms are and fractions that deﬁne the amount to which the Chapter 3: Corrosion Kinetics / 21

activation energies are lowered. They do not electrode, and the rate of cathode reaction will have to be the same for the anode and the cath- depend on how rapidly oxygen molecules dif- ode, but due to the uncertainty in evaluating fuse to the electrode/electrolyte interface. As a them, they are often taken as 0.5. The concen- limiting case, when the oxygen concentration is tration terms represent the ratios between the re- actually zero at the interface, the corrosion cur- actant concentration at the electrode/electrolyte rent can, through Faraday’s law, be coupled to interface and bulk, which could deviate from the steady-state flux of diffusive oxygen supply unity as a result of consumption/production of through a boundary layer δ. This limiting case species at the interface. In an iron-based alloy, current is called the limiting current (iL or jL) this ratio for the anode would be close to unity and can be expressed as: because the reactant is iron itself, and no concentration gradient would be expected as a * DnFCOO result of the corrosion reactions. When a net jj== 22 (Eq 11) corr l δ corrosion current flows, icorr = ia = ic. If the cathode and anode areas are assumed to be equal, then j = j = j and Eq 5 and 6 can be rewrit- In a nonlimiting case, the corresponding corr βa c α β equation would be: ten (using c = 2.3RT/( nF) and a = 2.3RT/ ((1 Ϫ α)nF) as: * − DnFCOOO((,)) C0 t = 2 22 (Eq 12) Ct(,)0 jcorr j ,0 c O δ ηβ=+log β log 2 (Eq 7) cc j c C corr O* 2 Combining Eq 11 and 12, one obtains: and Ct(,)0 j O2 corr =−1 (Eq 13) jcorr ηβ= log (Eq 8) C * jl aa O2 j ,0 a Thus, Eq 7 can be written: Tafel Regime: Electrode-Kinetics Control. If the electrode charge transfer reactions are j ⎛ j ⎞ rate limiting, the supply of oxygen to the reac- ηβ=+−log,0 c β log⎜1 corr ⎟ (Eq 14) cc c ⎝ ⎠ tion site would be rapid enough to maintain a jcorr ji concentration at the electrode close to that of the bulk. In this case, Eq 7 and 8 would both re- The slower the diffusion (small d), the lower sult in a linear dependence of the overpotentials the limiting current and thus a larger contribu- versus log j : tion from the mass-transfer-dependent second corr term on the overpotential. ηβ=− β (Eq 9) cclogjj,0 cc log corr Migration and Ionic Diffusion and The ionic transport in the electrolyte phase, φ ηβ=−logjj β log (Eq 10) the flux of an ion i under an electric field aa corra,0 a across a distance L, can be shown to be: Mass Transfer Control. In Eq 7, the term: ∂c zF Δφ JD=− ii− Dc (Eq 15) Ct(,)0 ii∂ ii O2 x RT L C O* 2 In an electrolyte with many different ions, an stands for the ratio of oxygen gas concentration ion current through an area A can be computed at the electrode/electrolyte interface and the by multiplying Eq 15 with zi * A and summing concentration in the bulk, sufficiently far away the contribution from all ions: from the interface. If the electrode reaction ki- 2 ∂Cx+ () netics are very fast, the depletion of oxygen will i FA iFAzD= ∑ + + lead toward a zero oxygen concentration at the i i ∂x RT 22 / Stainless Steels for Design Engineers

and the respective equations describing the × 2 Δφ overpotentials will be: ∑ zCDiii / L (Eq 16)

j ⎛ j ⎞ ηβ=+−log,0 c β log⎜1 corr ⎟ (Eq 21) Because the ﬁrst term is important only at the cc j c ⎝ j ⎠ regions near the electrodes (where consump- corr i tion/creation of species occur), the current in the majority region of the electrolyte can be es- and timated as: j ηβ= log corr (Eq 22) FA2 aa j i ≅ ∑ zCD2 Δφ / L (Eq 17) ,0 a RT iii Now, the potentials of anode and cathode Using Ohm’s law (R = U/i), the electrolyte re- when current is ﬂowing are in each case the sistance can be computed as: equilibrium potential plus overpotential, that is:

2 P ⎛ FA ⎞ RT H = 2 EE=−0 ln 2 +η RLelectrolyte / ⎜ ∑ zCDi i i ⎟ (Eq 18) cathode HH+ / 2 c ⎝ ⎠ 2 nF a RT H+

j The resistivities of some test solutions are =− ++β ,0 c H0 059pH.00 c log shown in Table 2. jcorr

⎛ j ⎞ +−β log⎜1 corr ⎟ (Eq 23) Mixed Potential Theory and c ⎝ j ⎠ Polarization Diagrams i

Viewing the electrochemical cell as an electrical circuit, Kirchoff’s law can be used to Ea=−0..log 447 − 0 0295 ( ) + η design a so-called polarization diagram. Con- anode Fe2+ a sider, as a case study, a steel corroding under j deaerated conditions, in a water solution, as =−0 624. +β llog corr (Eq 24) a j shown in Fig. 2. Assume that the pH is such that 0,a a passive layer does not form (see the discussion of Pourbaix diagrams in Chapter 2). The A polarization diagram is now constructed by cathode and anode reactions, respectively, are: plotting the anode and cathode potentials versus

log jcorr. Strictly speaking, to close the circuit, +−+→ 22HHe 2 (Eq 19)

Fe()se→+ Fe 2+−2 (Eq 20)

Table 2 Test solution resistivity

Ratio by Resistivity, Test solution volume ohm-cm Natural Seawater ... 25 Fresh (tap) water adjusted 28:1 500 with seawater Fresh (tap) water adjusted 68:l 1,000 with seawater Fresh (tap) water adjusted 950:1 3,000 with seawater Deionized water adjusted 21:10 10,000 with fresh (tap) water Fig. 2 Schematic polarization diagram Chapter 3: Corrosion Kinetics / 23 the potential drop across the electrolyte needs to Passivation be included, which simply equals icorr * Relectrolyte (the electrolyte resistance is evaluated from Eq Theory. In Chapter 2, it was identified 18); however, in many cases, this term can be through the Pourbaix diagrams that there were neglected. A schematic polarization diagram is conditions under which an alloy could be shown in Fig. 3. The anode polarization is linear passive. In the case of stainless steels, the range with decade current as predicted by Eq 24 be- of pH and other conditions under which this cause the overpotential has only a Tafel regime would occur has been increased thanks to the and no mass transfer dependence. On the other chromium content, which readily forms a Cr2O3 hand, the cathode polarization deviates from the scale. In general, a passive layer constituted of Tafel behavior as a result of the effect of the adsorbed molecules or thin oxide/hydroxide mass transfer (hydrogen ion supply), dependent layers decreases the corrosion current. Re- on the limiting current in Eq 23. It is notewor- searchers (Ref 3) have reported that the con- thy that the cell shown in Fig. 1 does not have a stituents of the passive film are alpha Cr2O3 and macroscopic anode and cathode. Different mi- Cr(OH)3nH2O. The structure is reported to be a croscopic regions on the surface are assumed to nanocrystalline spinel, epitaxial to the surface. act as cathodes and anodes, and in the lack of The grain size may decrease with increasing more detailed knowledge, the cathode and chromium content. This protection by anode areas are assumed to be equal. The over- chromium requires a threshold level of 11 to all mixed potential of the surface would be at a 12% chromium. corrosion potential Ecorr, defined in Fig. 2. Effect on Polarization Diagrams. The polar- In effect, the corrosion current resulting from ization diagram for a passive alloy is quite dif- the cell depends on the equilibrium half-cell po- ferent from those discussed for active alloys. A tentials (E and E ), the Tafel slopes (β schematic of a typical polarization curve is β cathode anode c and a), the exchange current densities (jo,c and shown in Fig. 5. When a passive alloy is anodi- jo,a), and any limiting current density (jl). Figure cally polarized, it initially behaves like an active 3 shows schematically how decreasing any of the alloy (i.e., with a Tafel slope, etc., as the pas- Tafel slopes and increasing an exchange current sive layer is building up). The building up is density increases the corrosion rate. The effect of actually a selective dissolution of iron, which the electrolyte resistance has been ignored; that causes a greater remaining surface concentra- is, corrosion current is where the two polarization of chromium and other alloying elements. tion curves intersect. Figure 4 shows the effect of Once the passive layer is formed and offers increased mass transfer, which would result in an protection against further dissolution, the po- increase in the limiting current. In an active (non- tential-decade current relation drops to lower passive) alloy, this results in an increased corro- currents. This happens at potentials beyond the sion current up to a point. passivation potential Epp. At some high enough

Fig. 3 Corrosion rate and the effect of (a) Tafel slope and (b) exchange current density 24 / Stainless Steels for Design Engineers

polarization level, the passive layer breaks j down, and the metal becomes active again; this =− 059pH.00 ++β log ,0 c (Eq 25) c j region is called the transpassive regime. The corr design of a structure involving a passive metal should aim at forming a corrosion cell in which This will result in a straight line as shown in the cathode polarization curve intersects the an- Fig. 6, which will be shifted vertically depend- odic one in the passive regime. ing on the pH. The dashed circles indicate the Consider, for example, an alloy that exhibits intersection between anode and cathode polar- the behavior shown in Fig. 6 in deaerated ization curves that would yield the corrosion acidic solutions with different pH. If mass current. At a sufficiently high pH (= pH1), the transfer limitations due to hydrogen ion supply alloy is clearly not optimal because intersection are neglected, then the cathode polarization is occurs in the active regime, and the passive given by: properties are not utilized. This is what occurs when a reducing acid is too strong for a given stainless steel, such as with concentrated hy-

RT PH drochloric acid. At pH = pH2, on the other EE=−0 ln 2 +η cathode HH+ / 2 c hand, a low-corrosion current is obtained as a 2 nF a H+ result of intersection at the passive regime. This is the benevolent case when stainless steel is correctly matched to the environment, and low rates of uniform corrosion occur. Finally, at

pH = pH3, the resulting corrosion current is again high as a result of intersection occurring at the transpassive region. This could occur with some stainless steels exposed to a very strong alkali solution. Using a similar argu- ment, the readers can themselves deduce the effects of cathode exchange current density and Tafel slopes. In the discussion of active anode polarization, it was found that increasing the transport rate of cathode reactants through agitation, for example, would increase the corrosion rate up to a point but beyond that have no further effect (see Fig. 4). In the case of a passive/active behavior, the effect of mass Fig. 4 Effect of increasing the limiting current by, for example, increased agitation in the electrolyte. Beyond the dashed line, increasing the limiting current would have no further effect

Fig. 5 Schematic of a passive anode polarization curve Fig. 6 Effect of cathode polarization Chapter 3: Corrosion Kinetics / 25

is the actual passivation; the iron removal is really a chemical cleaning operation, which happens to be called passivation. During the production of stainless steel, after a final anneal another version of passivation is carried out. The oxide from annealing in air is dissolved by a strong mixture of nitric and hydrofluoric acids, which does not allow passivation. This treatment, called pickling, removes by dissolution both the oxide layer and the chromium-depleted layer below the oxide formed during annealing. The depleted layer can extend a number of microns in depth and would seriously degrade corrosion resistance if not removed (Ref 4). This is then followed by a straight nitric acid immersion, Fig. 7 Effect of mass transport which ensures complete passivity. This is the procedure that should be performed on the ox- transport is somewhat different, as shown ides formed during welding if full corrosion re- schematically in Fig. 7. sistance is to be restored. Simply removing the Increasing mass transport, such that the limit- oxide through mechanical means leaves a ing current increases, results initially in an in- chromium-depleted layer that corrodes more creased corrosion current (e.g., increasing jl readily than is expected of the alloy. from 1 to 2). It should be noted that there are several intercepts possible (both in the active and passive regime), but assuming there are REFERENCES defects present, it is likely that there will be corrosion corresponding to the higher current. In- 1. D.A. Jones, Principles and Prevention of creasing the limiting current beyond the knee Corrosion, 2nd ed., Prentice Hall, 1996 Electrochemi- corresponding to Epp, however, results in a drop 2. A.J. Bard and L.R. Faulkner, in the current because now the only corrosion cal Methods: Fundamentals and Applica- potential possible is at the intersection in the tions, 2nd ed., Wiley, 2001 passive regime. This is the case for jL3. 3. M.P. Ryan et al., Critical Factors in Focal- In the normal use of stainless steel, achieving ized Corrosion, Proc. Electrochem. Soc., passivity takes on several forms. What is often Vol 150, 2003, p 583–594 called passivation is actually a cleaning process 4. J. Grubb and J. Maurer, “Corrosion of the in which contaminants, such as tramp iron, are Microstructure of a 6% Molybdenum Stain- removed from the surface. Dilute nitric acid is less Steel with Performance in a Highly Ag- an excellent vehicle to achieve this. This gressive Test Medium,” paper 300 pre- medium has the additional benefit of forming a sented at Corrosion 95, NACE passive film on an active stainless surface. This International, 1995 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 27-56 All rights reserved. DOI: 10.1361/ssde2008p027 www.asminternational.org

CHAPTER 4

Corrosion Types

Summary plex in their behavior because the influence of processing and alloying variables changes the STAINLESS STEEL is unusual among alloy ability of this layer to form and remain stable in systems in that its corrosion resistance derives the face of aggressive environments. The be- from the passivating ability of a minor con- havior of stainless steel is further affected by its stituent, chromium. Thus, while stainless steels microstructural complexity. Stainless steel al- can be made to be essentially immune to corro- loys may have many constituent elements and sion in many environments, it can also experi- many thermodynamically possible phases, and ence various debilitating forms of localized cor- none of these are necessarily uniform in their rosion, which stem from the failure of this composition. Yet, it is the composition of the passive film. This chapter explores the behavior alloy in contact with the specific environment at of stainless steel in media that promote uniform any microscopic point that determines the cor- corrosion and the various mechanisms of local- rosion resistance of that particular point. ized corrosion, such as pitting and crevice corrosion. Uniform Corrosion

Introduction When all parts of a corroding surface have equal access to the corroding atmosphere and To most designers, the most recognized char- the structure of the corroding metal is relatively acteristic of stainless steel is corrosion resist- uniform, a uniform thinning of the material is ance. Stainless, unlike noble metals such as expected. Stainless steels are materials of choice gold, does not obtain its excellent corrosion re- because, by virtue of their passive behavior, they sistance from inertness. Instead, it is the reactiv- show very low rates of uniform corrosion in ity of chromium that allows the surface layer of many environments. The metallurgy and pro- corrosion product to become sufficiently adher- cessing of a particular grade are designed to pro- ent and impenetrable, which effectively stops vide passivity in a given environment. The envi- further corrosion by isolating the base material ronment can be too aggressive to allow passivity from the environment. This resistance to corro- to be maintained either by being too reducing, as sion is called passive behavior or passivity. with some acid media, so that passivating Other metals, such as aluminum and titanium, species cannot form or by being too oxidizing so form similar layers and also exhibit passivity. that the oxidized species that normally affect The important difference in the case of stainless passivity are no longer stable. The former is steel is that chromium is still a minor con- called dissolution in the active state, while the stituent, never more than 30% by weight, some- latter is termed transpassive dissolution. times little more than 10%. How much Intelligent design and knowledge of the envi- chromium there is and how uniformly it is dis- ronmental variables for a stainless steel compo- tributed have a profound effect on corrosion re- nent ensure that the alloy is used in the passive sistance by virtue of its ability to concentrate state, at which uniform corrosion occurs at a into the surface ﬁlm. Stainless steels are com- very low rate. 28 / Stainless Steels for Design Engineers

Among the important media with which we sive, or transpassive; Fig. 1 illustrates the effect encounter uniform, but acceptably controlled, of redox potential on a solution. corrosion in stainless steel are atmospheric and Certain anions have strong effects in media marine environments and chemical environ- through their well-known, if not well under- ments such as sulfuric acid, phosphoric acid, ni- stood, disruption of the passive film. Halides are tric acid, strong bases, and organic acids, such well known for this effect, but sulfides are as acetic and formic. active. These anions seem to intervene in the Pickling is an example of controlled, acceler- adsorption of the hydroxyl ions. In acid media, ated uniform corrosion. This is typically done these anions accelerate uniform corrosion, with 10 to 20% hot sulfuric acid or a mixture of while in neutral media they may result in local- hydrofluoric and nitric acids. ized corrosion. Anions that form soluble complexes with elements in stainless, such as amines, formates, or acetates, can also disrupt Environmental Variables Influencing the stability of the passive film and thus pro- Uniform Corrosion mote active corrosion. The corrosion of stainless steels is usually the Of the physical variables of the environment, result of contact with an electrolyte, allowing a it should be obvious that temperature is para- complex set of partial electrochemical reac- mount since all the reactions are thermally acti- tions, which may occur sequentially or concur- vated. Increasing temperature may speed the rently. The corrosion rate depends on the formation of the passive film when thermody- current exchanged between the negative and namic conditions are favorable, but in general positive electrode (anode and cathode). These one expects increasing temperature to increase may be on a macroscopic or microscopic level. corrosive attack. Access to passivating species, The main consideration is normally ionic trans- such as oxygen, is important in establishing and por tthrough the passive film, which after all maintaining passivity. is what makes stainless so effective against Increased diffusion of reacting species in the corrosion. liquid will normally accelerate the partial reac- The chemical parameters that influence the tions, but if passivity is stable, the rate-limiting media with respect to uniform corrosion rate are transport through the passive film will not be af- the acidity and the oxidation-reduction (redox) fected. Therefore, increasing the flow rate of a potential of the electrolytic medium, both of corrosive fluid does not automatically acceler- which act through their influence on the stabil- ate corrosion. The reduction of concentration ity of the passive film, rendering it active, pas- gradients can be beneficial against localized

Fig. 1 Reduction potential versus pH for iron and (a) oxygen gas reduction and (b) hydrogen ionreduction Chapter 4: Corrosion Types / 29

corrosion, and flow can bring to the surface an stituent of sigma phase, which it promotes, increased supply of passivating species. however, it combines with chromium. If this Increased flow rate in a fluid medium is dele- happens at relatively low temperatures, the sur- terious if it induces mechanical damage to the rounding matrix is depleted of both chromium passive film by erosion, abrasion, or cavitation. and molybdenum, and the corrosion resistance These are complex mechanisms, but it should in that region is diminished. Nitrogen also is ef- be apparent that the success of a stainless steel fective when in solid solution in austenite but to a given flow condition will depend mainly on can precipitate as a chromium nitride under cer- its ability to form and re-form its passive film. tain conditions and cause depletion of the re- Somewhat counterintuitively, thinner passive maining matrix. Local structure and composi- films are more protective than thicker films tion are paramount. This becomes more among stainless alloys. The tenacity of the thin important to localized corrosion, as discussed passive films on stainless (and titanium) make later, but it should be remembered in examining these alloys quite resistant to flow-accelerated uniform corrosion because corrosion will cease corrosion, as contrasted to copper and alu- to be uniform when composition becomes minum alloys, which have soft, thick corrosion nonuniform. product films. The compositional material variables that influence uniform corrosion are not exactly the same as those that will be seen to influence lo- Material Variables calized corrosion. The foremost element is, of Stainless steels have a great variety of alloy- course, chromium. Researchers (Ref 1) have re- ing elements and microstructures. As a general- ported that the constituents of the passive film ization, we can say that corrosion resistance is a are alpha Cr2O3 and Cr(OH)3nH2O. The struc- function of composition rather than structure. ture is reported to be a nanocrystalline and epi- Then, we must quickly add the qualifiers to this taxial to the surface. The grain size may de- statement. On an undisrupted, stress-free sur- crease with increasing chromium content. This face, local composition does quite precisely de- protection by chromium requires a threshold termine corrosion resistance. But, stainless level of 11 to 12% chromium. This threshold steels are seldom homogeneous or at thermody- has been attributed most convincingly to the namic equilibrium. Impurities such as oxygen minimum chromium content that permits and sulfur are usually present, mainly as inclu- chromium atoms on surface sites to be linked by sions since they have diminishingly small solu- adsorbed oxygen atoms (Ref 2). In any event, bility at room temperature. At high tempera- the mechanism by which this thin, several- tures after solidification, as in welds, they can nanometer-thick, film forms is the subject of on- be present in supersaturation, ready to precipi- going debate, but we do know that it is enriched tate as inclusions that alter local composition. in chromium, and that it is thinner for higher The tendency of carbon and nitrogen to form chromium alloys. The critical current density j, precipitates is controlled by diffusion rates, as measured during polarization, is also smaller which if elevated by increasing temperature can as chromium content increases (Fig. 2). This is cause debilitating, composition-altering precipi- consistent with the lower dissolution of noncon- tation. The even more slowly diffusing substitu- tributing elements required to achieve a critical tional alloying elements, such as chromium, surface chromium concentration. Increases in molybdenum, and nickel, have strong tenden- chromium can also be seen to lower the current cies to form phases that disturb their uniformity density in the passive region. This is manifest in in the austenite or ferrite matrix in which they alloy performance as a reduction in the uniform are intended to work. So, any discussion of the corrosion rate in a given medium. From an elec- influence of alloying element on corrosion re- trochemical point of view, this is explained as a sistance of a phase like austenite or ferrite must manifestation of the stability of the Cr(OH)3 recognize that alloying elements exert their ef- nH2O. fect when they are in solution in that phase. The The role of molybdenum is less clear. The ob- same element may under some conditions not served action of molybdenum is to greatly re- be in solution and have a contrary effect. An ex- duce the critical current density required for ample is molybdenum, which is obviously a passivation. This is also seen as accelerating the great enhancer of corrosion resistance when in formation of the passive films and as increasing solid solution. When it precipitates as a con- the resistance of the alloy to depassivation at 30 / Stainless Steels for Design Engineers

Fig. 3 Influence of alloying element on corrosion rate as explained by the effect on polarization.Source: Ref 6 Fig. 2 Schematic illustration of polarization behavior for a passive alloy with and without pitting occurring lower pH. The role of molybdenum is not to enrich in the passive film itself, although it can be found in the film. Its potency is far more than its presence can take into account. Pure molybdenum is itself not passive. Its action does not appear to be via a product of reaction. Instead, it seems to reduce the dissolution rate of elements other than iron, which would promote a surface richer in chromium (Ref 3). The action of molybdenum as an alloying element is complicated by the fact that molybdate ions are known to impede pit growth as a separate effect from Fig. 4 Influence of alloying elements on uniform corrosion their action within the alloy matrix (Ref 4). rate in 20% sodium chloride solution with carbon Copper has a similarly complicated effect, with dioxide pressure of 20 MPa. Source: Ref 7 copper ions gettering sulfide ions and redeposit- ing as metallic copper (Ref 5). chromium in the passive layer and to decrease Nickel also lowers the critical current density active dissolution of noniron alloying elements, for passivation without contributing directly to thereby promoting both the formation and sta- the passive film’s stability. This also may be the bility of the passive film. A summary of the result of the stronger bond between nickel and known major alloying effects in acidic chloride chromium reducing the anodic dissolution rate media is shown in Fig. 3 in acidic chlorides. Al- of the alloy by permitting the anodic enriching loying elements provide benefits in the part of of the surface by selection iron dissolution. the chart where they appear (Ref 6). From this, Nickel does not actively help passive film for- it can be seen that chromium, molybdenum, mation and can actually hinder film stability in nickel, copper, and nitrogen all assist in the ac- highly acidic/oxidizing environments. tive region, while chromium, molybdenum, and Nitrogen, however, appears to be more like nitrogen expand the region of passivity and di- molybdenum in its effect. While nickel and cop- minish the corrosion current. per provide no benefit to the stability of the pas- An example of the influence of these alloying sive film once it is formed, both nitrogen and elements on the uniform corrosion rate of stain- molybdenum do, and to a degree that cannot be less steels in a sodium chloride/carbon dioxide explained by their presence in the film. This environment is shown in Fig. 4 (Ref 7). Note may then relate to their thermodynamic action the alloying composition is measured by a within the alloy itself. Molybdenum and nitro- crevice corrosion index (CCI), which is dis- gen act both to enhance the enrichment of cussed in the section Localized Corrosion. Chapter 4: Corrosion Types / 31

Fig. 5 Corrosion table for stainless steels and titanium in sulfuric acid plus copper sulfate. Corrosion rate legend: 0, < 0.1 mm/yr (corrosion resistant); 1, 0.1–1 mm/yr (useful in certain circumstances); 2, > 1.0 mm/yr (material not recommended). Source: Ref 8; see source for interpretation of data. Courtesy of Outukumpu Stainless

Unfortunately, it cannot be assumed that this less steels in a great number of environments relationship is true for other environments, al- can be obtained. Figure 5 shows an example of though other empirical relationships exist or can one such table. Many of the isocorrosion charts be generated. Because the inﬂuence of alloying in this book are reprinted from this source, element varies with environment, we need to http://www.outokumpu.com/applications/docu- discuss some of the more commonly encoun- ments/start.asp (Ref 8). tered severe environments. These tables are supplemented by isocorrosion diagrams such as that shown in Fig. 6. These diagrams show constant corrosion behav- Corrosion in Acids and Bases ior under varying environmental conditions such The examples discussed in Chapter 3, “Cor- as temperature and solution composition. This rosion Kinetics,” refer mostly to corrosion in information is available to guide the designer in these aqueous solutions, in which the slow thin- selecting appropriate steels for various environ- ning rate of the chosen alloy can be determined ments, and it is highly recommended that it be through the mixed potential theory and polar- used. Free sites tend to promote proprietary al- ization diagrams. In the case of stainless steels, loys, as these charts suggest. The serious engi- the alloy chemistry is chosen such that the neer will consult multiple sources and unbiased passive-active behavior favors corrosion in the sources before making alloy decisions. passive regime. The corrosion rate of the vari- The inﬂuence of alloying element is by no ous stainless steels in the myriad possible envi- means the same in all environments. So, while it ronments has been measured in probably all is useful and necessary to have these experi- practical cases. These data can be obtained from mental data, it is also helpful to understand the a number of sources, such as the National Asso- peculiarities of some of the major alloy-envi- ciation of Corrosion Engineers (NACE) and ronment pairings. ASM Handbook volumes. None is more accessi- Sulfuric Acid. Stainless steels require more ble than the Web site of Outukumpu, which than a minimum amount of alloying to resist sul- contains a “Steel Professional Tool,” a lookup furic acid. Straight 16% chromium grades such table in which the corrosion rate of many stain- as 430 fare poorly, while the nickel-containing 32 / Stainless Steels for Design Engineers

Fig. 6 Isocorrosion curves for 17-12-2.5 stainless steel and titanium in sulfuric acid plus copper sulfate.Source: Ref 8. Courtesy of Outukumpu Stainless

304 has more than an order of magnitude better corrosion rate in either dilute or concentrated sulfuric acid at ambient temperatures. Figure 7 (Ref 9) shows the isocorrosion rate curves for several common alloys. Alloying with molybdenum is also very effective, as is alloying with copper. If passivity cannot be established, increasing chromium content actually increases corrosion rate. The corrosion behavior of sulfuric acid varies greatly with concentration. At low concentrations, sulfuric is a classic reducing acid. It dis- sociates in water to create hydrated hydrogen + ions (H3O ) that release hydrogen gas bubbles during the corrosion reaction. As the acid concentration increases, the solutions become more corrosive, and progressively more highly alloyed stainless steels are required to provide adequate corrosion resistance. At about 50% acid, only the most highly alloyed stainless alloys (alloy 20, AL-6XN, C-276, etc.) can provide acceptable corrosion rates, and even these alloys are restricted to use at near ambient temperatures. As acid concentration increases beyond 50%, the solution begins to show oxidizing behavior. At acid concentrations above 80%, nickel-molybdenum-copper-bearing stainless steels begin to exhibit useful corrosion resistance. In the 93 to 98% sulfuric acid concentration range, carbon steel can be used to hold sul- Fig. 7 Isocorrosion rates of various stainless steels in sulfuric furic acid at ambient temperatures, although acid. Source: Ref 9 stainless steels provide better performance at elevated temperatures or if ﬂow-erosion can occur. resistant high-chromium (type 310S) and high- In the 96 to 100% sulfuric acid concentration silicon (MECS ZeCor UNS S38815 and Sand- range, at elevated temperatures, the oxidizing vik SX S32615) stainless steels are frequently character is quite pronounced, and oxidation- used, especially in sulfuric acid-manufacturing Chapter 4: Corrosion Types / 33

Fig. 8 Inﬂuence of alloying element on corrosion rate in contaminated sulfuric acid. Source: Ref 11

non-molybdenum-bearing alloys. Their superiority in deaerated solutions is much less marked. Studies (Ref 10) have shown that in sulfuric acid molybdenum is highly enriched in the passive film, and when molybdenum is an alloy, chromium also is enriched. This is a manifestation of selective dissolution of other elements in the matrix. Oxidizing impurities, such as ferrous ions, act like aeration to diminish the corrosive attack, but reducing impurities such as halides have an extremely negative effect, as the corrosion tables will show. These effects are not linear and underscore the value of these tables. The uniform corrosion rate in contaminated Fig. 9 Isocorrosion curves for various alloys in sulfuric acid sulfuric acid may be more important than in pure acid since this represents a potentially equipment. Sulfuric acid-containing dissolved likely failure mode because contamination is a sulfur trioxide is called oleum, and such solu- constant hazard. tions are often identified as sulfuric acid of Figure 8 shows the corrosion rate of various greater than 100% concentration. High- alloys in sulfuric acid contaminated with chlo- chromium stainless steels (i.e., type 310S) are rides and iron. These researchers (Ref 11) found among the very few materials that exhibit corro- that the resistance to attack correlated to the sion resistance in oleum. (See MTI Materials alloy content by the formula shown. Selector Volume 3—Sulfuric Acid at www.mti- Figures 9 and 10 show how isocorrosion rates global.org for more information.) vary with alloy and contamination level. Aeration has a major influence on corrosion Hydrochloric acid is very destructive of the rates because oxygen stabilizes the passive film. passive film on stainless. An alloy like 304 is Molybdenum-alloyed stainless has dramatically not suitable even in a deaerated 1% HCl solu- lower corrosion rates in aerated solutions than tion at room temperature. Chromium additions 34 / Stainless Steels for Design Engineers

Fig. 12 Isocorrosion curves for austenitic AL-6XN (UNS N08367) and 904L (UNS N08394) stainless steels Fig. 10 Isocorrosion curves for various alloys in sulfuric acid in hydrochloric acid. Source: Ref 12 with chlorides

Nitric acid is strongly oxidizing. This actually promotes the passive film formation; consequently, even low-chromium alloys remain passive at all concentrations at ambient temperature (see Fig. 14) (Ref 8). The addition of molybdenum, which is so generally helpful, is deleterious in this case because it forms soluble compounds. It is useful to keep carbon, silicon, and phosphorus as low as possible. Silicon is unusual in that normal levels (0.4 to 1.0 %) are worst, with very low (0.05%) or very high levels (4.0%) beneficial. The low levels of silicon contents of these alloys are useful for their action in minimizing grain boundary seg- Fig. 11 Isocorrosion curves for various stainless steels in hydrochloric acid. Source: Ref 8. Courtesy of Out- regation, which is the usual locus of attack. okumpu Stainless High silicon levels contribute to a general protective silica surface layer in concentrated acid, which augments the true passive layer. This superiority appears above the azeotropic composi- are only modestly helpful, while nickel, copper, tion of about 67%, which is a common com- and molybdenum are more beneficial. Stainless mercial concentration, as shown in Fig. 15 for steels are not good materials for contact with high-silicon austenitic stainless steels (Ref 13). hydrochloric acid. Figures 11 (Ref 8) and 12 Phosphoric Acid. This oxidizing acid be- (Ref 12) show how even the most highly al- haves more like sulfuric acid in that simple loyed grades can withstand only dilute concen- iron-chromium alloys have only moderate re- trations and low temperatures. sistance to uniform corrosion in them, while al- While a stainless steel vessel may not be in- loying with molybdenum and copper produces tended to be used for hydrochloric acid, resist- major improvements. This can be seen in Fig. 16, ance to lesser amounts of chlorides is important in which alloys with increasing nickel (18-10) because of the possibility that an acidic environ- show clear benefits over a chromium-molybde- ment may be contaminated with chlorides. When num alloy (18-2), and the added molybdenum this is a possibility, then proper alloy selection in 317 (17-14-4) is better, while 904L with must guard against it. Figure 13 shows a correla- nickel, molybdenum, and copper is even better tion between alloy content and resistance to the (Ref 8). general corrosion (GI) by sulfuric acid contami- In the commercial production of phosphoric nated with hydrochloric acid (Ref 11). acid, halide impurities may be present, in which Chapter 4: Corrosion Types / 35

Fig. 13 Inﬂuence of alloy content on corrosion rate in hydrochloric acid

case alloys with higher molybdenum, chromium, copper, and nitrogen may be required. Organic Acids. The weakly dissociating organic acids are normally not aggressive against stainless steels. The exceptional dangerous environments are those that include high temperature and the presence of chloride contamination. It should be noted that in formic acid, which does dissociate more strongly, nickel is detrimental. This phenomenon is also seen in the production of urea via the intermediary ammonium carbamate. The difficulty lies in the high- temperature solubility of nickel complexes and is best addressed by the use of ferritic or duplex alloys. Fig. 14 Isocorrosion curve for nitric acid. Courtesy of Outokumpu Stainless Alloying with molybdenum seems to provide the greatest resistance to uniform corrosion in strong organic acids, as illustrated in Fig. 17 (Ref 8). If halides are present in organic acids and liberated by contact with water, then pH and chloride concentration will govern the corrosive attack, which could then become nonuniform. Strong Bases. In strong bases, the stainless steels are generally quite resistant to uniform corrosion. Straight chromium (17%) alloys are usable at any concentration up to 50 ¡C. Adding molybdenum and nickel does little to further improve performance as the underlying resistance is due to chromium. Increasing chromium levels provide increased resistance. Attack when it does occur can be manifested as grain Fig. 15 Corrosion behavior of high-silicon alloys in concen- boundary attack. Figure 18 shows isocorrosion trated nitric acid. Courtesy of Outokumpu Stainless curves for sodium hydroxide (Ref 8). 36 / Stainless Steels for Design Engineers

Fig. 16 Isocorrosion curves in phosphoric acid: (a) 0.1 mm/yr for various stainless steels; (b) 0.1 mm/yr for titanium and 17-12-2.5 stainless steel. Courtesy of Outokumpu Stainless

Fig. 17 Isocorrosion curves in organic acids: (a) acetic acid; (b) formic acid. Source: Ref 8. Courtesy of Outokumpu Stainless

In the pulp-and-paper industry, chemical pulping is called the kraft or sulfate process. In the presence of sulfur, nickel can be quite detrimental, and ferritic or duplex alloys are preferred. This again is caused by the solubility of nickel complexes formed in the presence of sulfur-containing compounds. This can be seen in Fig. 19, which shows a 26-1 (chromium- molybdenum) alloy signiﬁcantly outperforming higher alloys that contain nickel and molybdenum (Ref 12).

Atmospheric Corrosion Atmospheric corrosion is an example of uni- Fig. 18 Isocorrosion curves for various materials in sodium hydroxide. SCC, stress corrosion cracking. Courtesy form corrosion that occurs when a thin layer of of Outokumpu Stainless water condenses on a metal surface and as such Chapter 4: Corrosion Types / 37

2) that atmospheric corrosion can in effect be prevented. The most deleterious impurity in the atmosphere for stainless is the chloride ion. Chlorides are pervasive. Borne from oceans by normal cli- matological processes, they are found far inland. In many colder climates, they are also seen in high concentrations from road salts. Without washing or the natural rinsing by rain, surface chloride concentrations can become very high. Thus, the rules of corrosion of aerated aqueous solutions are followed by stainless with respect to atmospheric corrosion. The difficulty is accurately estimating the solution that constitutes the aqueous solution. Much experience has shown that if coastal and road salt effects are minimal, then 18% chromium alloys such as 304 experi- Fig. 19 Corrosion rates of various alloys in simulated evaporator liquid. Source: Ref 12 ence such negligible visible corrosion that they can be used for exposed, unrinsed architectural depends on humidity, temperature, and other at- purposes. If the same alloy is used in an unrinsed mospheric conditions. The rate of corrosion coastal environment, red rust stain will occur. measured as defined in Chapter 3, “Corrosion This is the corrosion product from metastable Kinetics,” as dissolution r of element i as and possibly stable pitting. In Japan, where coastal conditions prevail throughout, much re- tM⋅ search has been done that has shown that a pit- rj= i ting resistance equivalent number (PREN) of 25 nF⋅⋅ρ i is necessary for freedom from corrosion (i.e., zero pitting) (Ref 14). This contrasts to a re- where j is current density, t is time in seconds, quirement of about PREN 35 to resist pitting in M is molar mass, n is valence, F is the faraday seawater. Pitting is a form of localized corro- constant, and ρ is the density), and therefore has sion, and PREN is an index to pitting resistance. two contradicting effects of temperature. In These concepts are examined in the next section. general, temperature increases the exchange current density and transport properties and thus the kinetic rates involved in corrosion. On the Localized Corrosion other hand, increasing temperatures may re- Localized corrosion is in general more duce the concentration of dissolved oxygen in damaging from a structural integrity point of the electrolyte and eventually will dry the sur- view than uniform corrosion since the corro- face and thus limit the electrochemical corrosion current is limited to a small area and the sion due to the access to an electrolyte. In steels, penetration distance is large. Often, localized the corrosion products are (a) an outermost corrosion involves a large-area cathode and a layer of porous rust (FeOOH) characterized by small-area anode, which means that for a given low water content but easy access to oxygen corrosion current, the corrosion current density and (b) an inner layer of magnetite (Fe2O3) in at the anode is very large. In localized corro- which pores are filled with water. The access of sion, unlike uniform corrosion, the anode and oxygen to the bare metal limits the cathode re- cathode are clearly identifiable locations, and duction reaction rate, and in relatively pure at- the reason that certain structural features as- mospheres, the corrosion rate is negligible due sume the roles of cathode and anode can be the protective nature of the oxide. However, sul- used to categorize and exemplify different fur dioxide impurities in the atmosphere react cases. Interestingly, the cathode and anodes, with water to form sulfuric acid, which tends to while identifiable, can vary across scales, that dissolve the protective oxide. In the case of is, from distinct macroscopic components or stainless steels, the passive region is extended parts to microstructural features. due to chromium, to a wide enough region in In Chapter 2, the tendency for corrosion terms of pH (see Pourbaix diagrams in Chapter was introduced as a positive value for an 38 / Stainless Steels for Design Engineers

electrochemical cell potential (Erxn) correspon- arises in which access of oxygen is not the same ding to a spontaneous electrochemical reaction to different areas of a sample. In effect, this re- forming a galvanic cell. Erxn is obtained as: sults in that the cathode reaction:

ααα −− RT ⎛ ppp123 ...⎞ OHO++→244e OH (Eq 3) =−0 123 22 EErxn rxn ln ⎜ βββ ⎟ (Eq 1) nF ⎝ 1233 ⎠ rrr123... is limited from proceeding in some areas but not others. This gives rise to a differential aeration Here, p and r are the concentrations of reac- cell. For example, consider Fig. 20(a), in which i i α β a metal is partially immersed in water. Trans- tant and products, respectively, and i and i are the numbers that are needed to balance the reac- port distance of oxygen increases with depth; the limiting current would then vary with depth, tion stoichiometrically. Any time that Erxn is positive, there is thermodynamically a tendency such as at locations 1 and 2, and cathode polar- for an electrochemical reaction, in our case a ization curves as a result of this are schemati- corrosion reaction. The rate of corrosion, as dis- cally plotted in Fig. 20(b). Near the surface of cussed in Chapter 3, is dependent on the polar- the water, where oxygen is readily replenished, ization behavior. passivation is likely to be fast, and thus anodic iron dissolution is slow. This region assumes the role of the cathode, and reaction 1 occurs. Suffi- Dissimilar Metals and Differential ciently far away from the surface, if there are Aeration Cells regions where passivation is incomplete (e.g., surface defects or scratches) or has broken The case of dissimilar metals and differential down as a result of, for example, Cl− (see sec- aeration cells is perhaps more important in action on pitting), repassivation does not readily tive alloys than for stainless steels, which are occur since oxygen transport is too slow. These generally passive, and occurs when two regions become anodes where the following re- metals/alloys are in contact that have elements action occurs: in them that are dissimilar in the electromotive +− force (emf) series (see Chapter 2) and there is an 224Fe()se→+ Fe 2 (Eq 4) electrolyte present. For example, if nickel and iron pipes are connected and water flows though them containing some traces of Ni2+ ions, then: The distance at which this occurs is balanced by being large enough to limit the rate of ⎛ ⎞ oxygen transport but not too long to be a + 00RT Fe2 EE=−−++ E ln ⎜⎜ ⎟ rxn Ni22// Ni Fe Fe (Eq 2) strongly influenced by ion transport that is 2F ⎝ a + ⎠ Ni2 needed to complete the electrochemical cell. Resulting corrosion currents are shown in Fig. In this case, the corrosion tendency is prima- 20b. This type of degradation is called water- rily caused by the first two terms on the right line corrosion. side of Eq 2, the dissimilarity in the standard Crevice Corrosion. In stainless, the more half-cell reduction potentials: significant occurrence of this type of cell occurs when a crevice, from whatever cause, exists, 00 EE++−=−+=0... 250 0 447 0 197V Ni22// Ni Fe Fe and reactions within the crevice or pit cause the accumulation of iron ions by: This tendency is caused by the galvanic dis- 2+− similarity between the metals. This is normally 224Fe()se→+ Fe (Eq 5) important for alloys joined to stainless that are themselves less noble. Less-noble alloys, such The regions adjacent to the drop that main- as carbon steel, can fail rapidly if coupled to tained their passive layer and have access to stainless. A classic example is the use of carbon oxygen act as cathodes where the oxygen reduc- steel fasteners for joining stainless sheets. Dif- tion reaction takes place: ferent stainless steel alloys have minor differ- ++→−− ences when passive, but if the environment is OHO22244e OH (Eq 6) such that one alloy is active while another is passive, then the galvanic differential could be This reaction maintains an alkali solution. As harmfully large. In many cases, a situation a result of the geometry, Fe2+ ions remain and Chapter 4: Corrosion Types / 39

Fig. 20 Schematic illustration of (a) sample partially immersed in water; (b) resulting polarization behavior for two different passivating alloys (A and B polarization curves) enrich in the water-filled pit; to maintain charge Pitting corrosion is important to designers be- neutrality, ClÐ migrates into the pit. This causes cause it is corrosion under conditions at which the following reaction: corrosion may not have been anticipated. Thus, it is both a materials selection and an environ- Fe2+−++→22 H O Cl Fe( OH) + 2 HCl (Eq 7) 2 2 mental control problem. Its consequences may be only cosmetic, such as on a building or appli- which has several consequences: (a) Hy- ance facade, or potentially catastrophic, such as drochloric acid further acidifies the pit and if leaks of toxic materials were to result from increases the rate of iron dissolution since de- perforation. Stainless steels are designed to be creasing pH increases cathode half-cell poten- passive, and localized corrosion is the local loss tial, which increases corrosion rate (see polar- of passivity. Whether the consequences are ization diagram construction in Chapter 3, major or not, it is always undesirable, and good “Corrosion Kinetics.”) (b) The formation of design allows it to be avoided. porous Fe(OH)2 further helps to isolate the pit, What do we know for certain about pitting? thereby separating anode and cathode regions in We know quite a lot, really. Experts now con- the differential aeration cell. (c) The presence of clude that since the early 1970s the local chem- ClÐ prevents repassivation. istry of pitting has been understood (Ref15). As a result of increased acidification, the dis- The greatest contributions to this field have solution rate becomes autocatalytic, and as a re- been electrochemical studies. The tools of elec- sult the pit grows in depth. trochemistry have been especially successful in At the outside, the reaction: elucidating the mechanism involved in pit growth and pit stability (Ref 16). The local en- 22Fe( OH) ++ O H O → Fe( OH) (Eq 8) 2 22 3 vironment within pits has been sufficiently measured and correlated with cavity geometry further consolidates the isolation of the pit and that some experts can say, “In a sense, all pitting impedes the ingress of oxygen. is crevice corrosion” (Ref 15). This is to say that the electrochemistry of cavities such as pits and crevices is quite similar and has been well mod- Pitting Corrosion eled. These same tools, however, have been Pitting corrosion is the most intensely studied much less successful in clarifying the mecha- and debated form of corrosion of stainless steel. nism of pit initiation, which is still the subject 40 / Stainless Steels for Design Engineers

of debate, possibly indicating that the root Inclusions. The question of what causes the causes are more metallurgical than electro- initial dissolution that causes both stable and chemical. metastable pits focuses on inclusions, which Figure 21 depicts a polarization curve for most authorities (Ref 18) have concluded are stainless steel in a chloride-containing solution. associated in some way with pit initiation. In Pitting occurs in the zone in which passivity is the absence of inclusions, metastable pitting expected. As potential increases, small spikes in events are not noted, and the potential at which corrosion current occur. These spikes measure pitting occurs is the beginning of the transpas- local dissolution, called metastable pitting. sive regime. Some such sites complete their dissolution and What are the typical inclusions in stainless repassivate, while others continue to grow as steel? Inclusions in steel are normally the stable pits. The potential at which stable pitting residue of normal deoxidation and desulfuriza- occurs is the pitting potential, while metastable tion taken during steel refining usually done in pitting can occur at much lower potentials. Pit- an argon oxygen decarburization (AOD). After ting events, stable or not, cause the generation removal of the carbon, the subsequent objective of iron ions and local pH reduction. To the ex- is to remove or render less harmful the dis- tent these remain concentrated in a small vol- solved oxygen and sulfur, which if left in solu- ume, they will affect subsequent events. The tion would later precipitate as low-melting- dissolution during metastable pitting is located point iron compounds that would make the steel at the matrix-inclusion interface. Different re- fragile and unworkable at high temperatures. searchers assume dissolution of the inclusion, Inclusions in stainless steel are typically oxides while others assume dissolution of the matrix. and sulfides. A key point to understand when The dissolution parameters, as measured by considering inclusions as initiation sites for pit- current transients, depend on variables not of ting is that inclusions are not simply inert debris the inclusion chemistry but of the matrix com- but precipitates that are seeking thermodynamic position, notably molybdenum and nitrogen equilibrium with the steel in which they have levels (Ref 17), which is in keeping with the re- previously been dissolved. The reactions in duction in dissolution of the matrix that these stainless steel differ thermodynamically from alloying elements confer. those in carbon steel because of the presence of high chromium concentrations. This lowers the activity of oxygen and sulfur, making them more soluble, as Table 1 indicates (Ref 19). It also alters the efficiency of deoxidizing elements. Aluminum is a powerful deoxidant in carbon steel but is less effective in stainless, while titanium becomes a stronger deoxidizer in stainless. Their effect on sulfur is similar to that on oxygen. The bottom line is that oxygen and sulfur are generally removed by silicon/manganese deoxidation, but that this process occurs in both the liquid and solid states. That it carries over significantly into the solid state means that diffu- Fig. 21 Schematic of a passive anode polarization curve sion has a major role in determining if equilib-

Table 1 Typical values of activities and activity coefficients in liquid steels: activities in the 1 mass % solution: ai = fi . %i

Metal Al C Mn P S Si Ti H N O Cr Ni

Carbon steel, %i . . . 0.05 0.45 0.02 0.01 0.3 0.05 ...... 1600 ¡C fi 1.05 1.06 1.0 1.1 1.0 1.15 0.93 1.0 0.97 0.85 ......

ai . . . 0.053 0.45 0.022 0.01 0.345 0.046 ......

Stainless steel, %i . . . 0.05 0.45 0.02 0.01 0.3 0.05 ...... 18 8 1600 ¡C fi 3.6 0.49 1.0 0.32 0.66 1.24 9.4 0.93 0.17 0.21 0.97 1.0

ai . . . 0.025 0.45 0.006 0.007 0.372 0.47 ...... 17.5 8.0 Chapter 4: Corrosion Types / 41

rium reactions occur and whether they go to the liquid. At high sulfur and manganese concen- completion. We will see that they do not. trations, some manganese sulfides can precipitate Oxide inclusions also are common. They are during solidification interdendritically, while nor- formed as the products of the reactions of sili- mal alloys with less than 100 ppm of sulfur form con and manganese with dissolved oxygen. The their inclusions after solidification. The distinc- thermodynamics of the reactions determine at tion is important because precipitation in the liq- any time how much oxygen can be dissolved in uid state permits rapid diffusion, which results in the steel at equilibrium. That equilibrium is eas- the most thermodynamically favorable species, ily achieved in the molten state, in which diffu- manganese sulfide, to form. It may, and often sion is very rapid, but achieved more slowly does, nucleate on a preexisting inclusion, such as once the material has solidified. The inclusions silicate present from the deoxidation process. In in the solid state grow by the diffusion of oxy- austenitic steels, manganese is generally present gen to inclusion sites, where it precipitates as an at a level of around 1.5% as a deoxidant and as a oxide of silicon or manganese to the extent that substitute for some nickel. An inclusion formed these are locally present or of chromium when in the molten metal does not cause alloy deple- its local concentration (or more properly, its action around it. One that forms or grows in the tivity) makes it more favorable. These oxides solid state does cause depletion of the elements are often the nucleation sites for manganese sul- that are precipitating, causing its growth. fide inclusions. If manganese is lowered to very low levels, Sulfur is a very surface active impurity that as- the supersaturation of sulfides is pushed to a sists in weld penetration in stainless by virtue of lower temperature, at which lower diffusion its effect on weld pool circulation. Otherwise, it rates hinder or prevent the precipitation. Thus, is a detrimental impurity, forming low-melting low-manganese alloys can be free of manganese oxysulfides that diminish hot workability. Man- sulfide inclusions even at somewhat high sulfur ganese is a strong sulfide former, and it is the levels. Such alloys have elevated resistance to main agent used to tie up sulfur. Manganese sul- pit initiation. Lower manganese levels also ther- fide precipitates as an inclusion as a function of modynamically reduce the chromium sulfide co- manganese and sulfur concentrations and tem- precipitation in inclusions, lowering chromium perature. Inclusions form not only in the molten depletion around manganese sulfide/chromium metal but also in the solidified metal. The solubil- sulfide inclusions. ity, which is high in the liquid state, decreases on Elements more effective than silicon and solidification, as seen in Fig. 22. Only resulfur- manganese are now in use for deoxidation and ized free-machining stainless steels have suffi- desulfurization. These include aluminum, cal- cient sulfur to precipitate manganese sulfide in cium, cerium, and other rare earth metals (REMs), and titanium. The action of calcium is notable. In a well-deoxidized and well-stirred melt and with a basic slag, calcium dissolved in the metal will react with dissolved sulfur to form calcium sulfide, which will be incorporated into the slag phase. Aluminum, while a potent deoxidizer, is less effective directly in desulfurization, but it can act indirectly by reducing a small amount of Ca2+ in the slag, allowing the formation of calcium sulfide. Tita- nium can sequester some sulfur as titanium carbosulfide precipitates. The greatest amount of sulfur removal is obtained by the addition of cerium or other REMs, usually in the form of the alloy mischmetal. These reactive elements typically form oxysulfide particles in the melt that may be trapped in the slag before metal solidification. Oxygen is normally dissolved in solidifying stainless steel, also at amounts in the neighbor- Fig. 22 Pseudo-binary-phase diagram for iron and sulfur at 1.8% manganese and 18% chromium hood of 100 ppm depending on deoxidation 42 / Stainless Steels for Design Engineers

methods. Inclusions based on oxygen and sulfur solidification as opposed to austenite first, AF), formed in the liquid or during solidification are as is almost always the case with commercial relatively large, greater than 1μ. As the alloy alloys, more sulfide precipitation happens in the cools after solidification, precipitation continues solid state, pitting resistance is lowered propor- since sulfur and oxygen are decreasingly solu- tionately to the sulfur level (Ref 22), and there ble with temperature, to virtually nil at room is little negative effect from solidification segre- temperature. This causes existing inclusions to gation, as is shown in Fig. 23 and 24 (Ref 23). grow and new ones to nucleate. This precipita- Solidification can also occur in a mixed ferritic- tion is similar to that which carbon undergoes in austenitic mode, in which case each microstruc- stainless, except carbon is generally not super- tural component behaves according to the chart. saturated until below 1200 ¡C at the highest in The ratio of chromium and chromium-like ele- most alloys, whereas sulfur and oxygen are nor- ments molybdenum and silicon to nickel and mally near saturation even at freezing or almost nickel-like elements carbon, nitrogen, man- always when the solidifying ferrite transforms ganese determines the mode of solidification. It to austenite. Thus, inclusions grow via diffusion of oxygen and sulfur, which, as interstitials, diffuse much more rapidly than the silicon or manganese with which they have the greatest thermodynamic affinity. But precipitate they must, even if the silicon and manganese in the vicinity of their inclusion are exhausted. Thus, inclusions can grow with chromium substituting for either silicon or manganese as the precipitating partner for oxygen and sulfur. The inclusion growth necessarily depletes the surrounding region of reactants, silicon, manganese, and chromium (Ref 20). Inclusions thus formed are nonequilibrium in nature, and thermal cycles of steel production are rarely sufficient for the equilibrium to be attained. The chromium enrichment of such inclusions and corresponding chromium depletion of surrounding regions has Fig. 23 Influence of sulfur level on pitting resistance of been measured (Ref 21) and corresponds to the unannealed welds for different solidification modes. depletion seen next to chromium carbide pre- Source: Ref 23 cipitates at grain boundaries in sensitized alloys. These zones are altered in size and shape by thermomechanical processing in wrought alloys but exist fairly undistorted in welds. Hot rolling and cold rolling followed by annealing elongate manganese sulfide inclusions and flatten them, allowing depleted zones around the inclusion in the reduced dimension to be more rapidly homogenized during annealing. Thus, wrought material has better pitting resistance than cast or welded material. Inclusions that precipitate from the liquid, as is more the case for alloys solidifying in an austenitic mode, are at equilibrium with the surrounding matrix by virtue of the faster diffusion in liquids, do little to diminish the chromium content around them, and have a small effect on lowering pitting resistance. Pitting resistance is still affected to a degree by alloy depletion due to solidification Fig. 24 Influence of sulfur level on pitting resistance of welds without homogenizing anneal. FA, ferrite forming segregation. However, if the alloy solidifies in a first on solidification as opposed to austenite first, AF. Source: ferritic mode (FA, i.e., ferrite forming first on Ref 23 Chapter 4: Corrosion Types / 43

can also be altered by freezing rate. Faster cool- stress corrosion cracking (SCC). There have ing favors austenitic solidification. been numerous proposed mechanisms for the Long-term annealing of welds has shown breakdown of a passive film in chloride-con- that sufficient time and temperature to achieve taining media; these have been summarized in some rehomogenization the alloy result in bet- other publications (Ref 23). These hypotheses ter pitting resistance (Ref 24), approaching that deal with how a passive film on a homogeneous of the wrought alloy. Examination of the de- surface could break down. They include: creasing solubility of sulfur in stainless in Fig. ¥ Adsorption of chloride ions 22 indicates that the precipitation of sulfides ¥ that cause chromium depletion occurs in delta Penetration of the passive film by chloride ferrite on freezing when sulfur exceeds 0.007% ions ¥ Film breakdown by electrostriction and in austenite when sulfur exceeds 0.003%. ¥ Oxygen behaves in a parallel manner and is Formation of stable metallic chlorides ¥ Coalescence of cationic vacancies usually present in sufficient quantities, about ¥ 0.01% in manganese/silicon deoxidized steels, Random localized thinning of the passive film to cause the same phenomenon. This funda- ¥ mentally is due to the high ratios of the diffu- Local variations in the composition of the sivities of oxygen and sulfur to chromium, corrosive medium which are about 10,000 and 680, respectively. By and large, these mechanisms presuppose Whenever fast-diffusing elements such as oxy- a stainless steel surface that is homogeneously gen, sulfur, carbon, and nitrogen, which have a passive and try to explain the observed inho- strong affinity for chromium and a solubility mogeneous behavior of the passive film. How- that decreases strongly with temperature, are ever, since it is clear that the surface is not ho- present in steel, their precipitation will result in mogeneous, especially with regard to the some degree of chromium depletion around the passive film, these hypotheses are not neces- precipitation site because chromium diffuses sary to explain the behavior of everyday stain- too slowly to be replenished. less steels, which unfortunately have abundant The low chromium around inclusions is a suf- inclusions and chemical inhomogeneities capa- ficient condition for the local dissolution meas- ble of locally diminishing the integrity of the ured as metastable pitting, and if the depletion passive film. More research in understanding zone shape and size are favorable, then stable the exact nature of the inhomogeneity of stain- pitting would ensue. less steel surfaces is necessary for a complete Certain other types of inclusions/precipitates understanding of pit nucleation and therefore are less harmful in this regard. Titanium, for in- prevention. stance, which is often added to form carbides Pitting Resistance. Pitting has been exten- and nitrides, also forms sulfides and oxides sively correlated with environment and compo- more strongly than manganese and therefore sitional variables. The most well-known and does so at higher temperatures. Such precipi- useful correlations are between the PREN and tates have a much lower tendency to allow the critical pitting temperature (CPT) and by ex- chromium to join in the precipitation since the tension to the pitting potential. higher the temperature of precipitation the more For austenitic alloys: that diffusion allows the more favorable reaction to occur. Rare earths also behave the same PREN = % Cr + 3.3 % Mo + 30 % N (Eq 9) way. Metastable pitting is diminished by the presence of these elements. For ferritic alloys, which hold no nitrogen in The initiation of pitting is also affected by solution: stress and inclusion orientation (Ref 25), which PREN = % Cr + 3.3 % Mo (Eq 10) the researchers correlated to the dimensions of the inclusion-derived cavity being able to sus- For duplex alloys, which have two phases, tain a sufficiently low pH due to iron dissolu- neither of which matches the bulk composition: tion to maintain stable pitting. The influence of PREN = % Cr + 3.3 % Mo + 16 % N (Eq 11) stress was to cause cracking at otherwise unfa- vorably shaped inclusions, which then provided These equations are useful, if approximate, a crevice capable of sustaining stable pitting. and their correlation is shown in Fig. 25 (Ref This will be relevant to later discussions of 26) They do not include tungsten, which, if 44 / Stainless Steels for Design Engineers

alloy behavior predictions. It is noteworthy that the elements copper and nickel, which are beneficial against uniform corrosion and which slow the growth of pits by this same action, do not contribute to increasing the resistance to the onset of pitting. This is another manifestation of pitting initiated by the local stability of the passive film, which is primarily a function of local chromium content. Molybdenum and nickel thus seem to bolster local chromium content in the passive film. Nitrogen seems to act by con- centrating at the passive film-alloy interface rather than by buffering the solution by ammonia formation, which has been proposed (Ref 36). Research on very pure sputtered films of iron-chromium alloys have demonstrated that both titanium and niobium in solution diminish active dissolution, assist repassivation, and improve pitting resistance (Ref 29). In most practical cases, these elements are not found in solution because of their affinity for oxygen, sulfur, Fig. 25 Variation of critical pitting temperature with pit- carbon, and nitrogen, with which they form ting resistance equivalent number (PREN) of austenitic steels in water plus 6% FeCl . Source: Ref 26 compounds. e It should also be noted that the critical PREN values vary with crystallographic structure. Fer- ritic alloys require somewhat lower PREN values to exhibit similar pitting resistance as austenitic alloys of somewhat higher PREN. While pitting is of great theoretical and practical interest, there are significant problems in actually conducting good tests. Monitoring of the electrochemical potential during the test is considered mandatory by most researchers. How is a metallic sample suspended in a solution without creating any crevices and without exposure at the liquid-gas interface? The development of the flooded gasket technique (used in ASTM G150) was a milestone, but it also has Fig. 26 Differential variation of critical pitting temperature some problems—most notably the potential for of several stainless steel alloys for unwelded dilution of the test solution, especially during wrought and welded material. Source: Ref 13 prolonged testing. FeCl3 testing benefits from the fact that the solution creates a reproducible present, has half the effectiveness of molybde- positive potential. num. They neglect carbon, which seldom varies While the PREN approximates the pitting re- enough to have a visible effect but has been sistance of an alloy, there is a standard test by shown when in colossal supersaturation to have which the CPT is measured. Pitting in a given a factor of about 10, not unlike nitrogen, an- medium capable of causing pitting does not other interstitial that it resembles in solution occur below a temperature that is characteristic thermodynamically (Ref 27). It also does not in- of the medium and the material, with the myriad clude the negative influence of elements such as exceptions of stress state, surface finish, mi- sulfur. Likewise, the equations cannot deal with crostructure, etc. The most commonly used test inhomogeneity issues, so welded alloys have media are the unacidified 10% FeCl3, which is different CPTs for the same PREN (Fig. 26) used in the ASTM G 48 practice B, and the 3.5% (Ref 13). These equations are all-other-things- NaCl solution of the ASTM G 150. The latter, if being-equal equations and are useful for gross modified to 0.1N NaCl, allows the ECPT, the Chapter 4: Corrosion Types / 45

electrochemical pitting potential, of lower alloys Cl = %Cr + 4.1%Mo + 27%N (Eq 12) such as 304 to be measured (Ref 30). Since a crevice has a preexisting favorable geometry for pit growth, any pitting event, Crevice Corrosion metastable or stable, can initiate ongoing In the case of pitting, the geometry that crevice corrosion. Crevices are thus incubators makes up the pit is essential in creating the dif- for corrosion triggered by metastable pitting ferential aeration cell and to cause the autocat- events. The dissolution of iron during passiva- alytic dissolution process. In many cases, a tion itself as well as the differential oxygen cell geometry that retains and acidifies water is al- created by the crevice contribute to the process. ready present in crevices in different types of It is logical to think that alloying the elements structures such as gaskets, under faulted coat- that contribute to lowering the critical current ings, under bolt or screw heads, etc. Crevice density for passivation and the uniform corro- corrosion occurs because zones have restricted sion rate, such as nickel, would reduce the cre- access of reactants and restricted exit of corro- ation of the reactants that start the crevice corrosion products. It is especially the inhibition of sion process, but this presumed effect is not the cathodic reaction inside the crevice by the strong enough to be reflected in this actual be- dearth of oxygen, which sets up a more aggres- havior Eq 12 represents, although it is generally sive environment within the crevice than with- acknowledged that austenitic steels perform bet- out. The interior reactions become increasingly ter than ferritic steels in the absence of molyb- anodic, and the aggressiveness of the environ- denum. Materials are characterized as having a ment can reach a threshold at which active cor- critical depassivation pH. If crevice conditions rosion occurs, while the situation exterior to the are such that the reactions over time allow the crevice is safely passive. Crevice corrosion oc- pH to be reduced to this level, then active corro- curs at lower temperature than pitting in the sion will begin within the crevice. Thus, passive same environments, so it is a greater danger in film stability seems to be the critical factor that sense. rather than corrosion rate after initiation. The relationship between the alloy content, Preventing Crevice Corrosion. The coun- given as the crevice corrosion resistance equiv- termeasures against crevice corrosion are ca- alent number (CCREN), and critical crevice thodic protection, design, maintenance, and, of corrosion temperature (CCT), shown in Fig. 27 course, alloy selection. Designing to avoid (Ref 11), is similar to that of PREN (PI) to CPT crevices should include maximizing the volume except for the molybdenum factor being more of unavoidable crevices, engineering flow to en- important: hance transport in and out of crevices, and

Fig. 27 Variation of critical crevice corrosion temperature with alloy content 46 / Stainless Steels for Design Engineers

avoiding stagnation. Any maintenance or design chromium. In the various grades of stainless procedure that prevents formation of deposits is steels, there are many intermetallic phases that are beneficial. Welds are particularly vulnerable thermodynamically stable but kinetically slow to surface sites, so any combination of welds and precipitate that are enriched in chromium. An ex- crevices or crevices caused by poor weld geom- ample of such a phase is chromium carbide (Fe, etry must be avoided. S32205 is a benchmark Cr)23C6. These phases tend to form at grain alloy of sorts. It has just sufficient alloying to boundaries where nucleation is favored, resulting resist pitting in seawater, but it is susceptible to in a depletion of chromium in the adjacent re- crevice corrosion. gions, as shown in Fig. 28. Thus, the chromium- As a practical matter, crevices are almost im- depleted regions near the grain boundaries are possible to eliminate. Threaded fasteners and sensitized in that they behave as active anodes joints represent severe crevices and should be compared to the larger interior of the grains that avoided in aggressive environments if possible. are still passive. In an aerated corrosive environ- Gasketed joints are another severe crevice loca- ment, the smaller chromium-depleted nonpassive tion, and their usage should be curtailed to the anodes dissolve, whereas the larger cathodes re- minimum practical extent. In these situations, duce oxygen, resulting in a localized corrosion judicious use of very expensive, highly corrosion along grain boundaries. Any heat-treating or resistant materials is justified. The use of smooth welding procedure of stainless steels should thus welded joints is thus generally preferred. In a be tailored to avoid sensitization. more general consideration, deposition and When a stainless steel is heat treated, there is fouling create crevice sites, and design and op- a risk that the unwanted phases may form, de- erational controls to preclude the formation of pending on the time-temperature history and deposits and the prompt removal of sludge and precipitation kinetics of the unwanted phase. the like are necessary. But in some situations, Figure 29 shows schematically the temperature such as marine exposures, biofouling will create versus time due to welding and the resulting crevice sites. This fouling may be macroscopic, sensitization. Figure 29 shows a TTT (time- such as from shellfish and barnacles, or it may temperature-transformation) curve for precipi- be microscopic. Microscopic biofouling causes tation of the unwanted phase. Near the weld the special form of crevice corrosion called mi- (A), the time spent in the temperature region crobiologically influenced corrosion (MIC) dis- where precipitation occurs is too short, whereas cussed in a separate section). far away from the weld (C) the temperature experienced is too low. At location B, there is, Sensitization/Grain Boundary Corrosion however, a risk for sensitization. The maintenance of a passive layer in a wide Austenitic. Sensitization can occur at any range of pH conditions in stainless steels is de- temperature at which carbon is supersaturated pendent on the alloying elements, primarily in an alloy. Current austenitic stainless steels have carbon levels of under 0.10% normally

Fig. 28 Schematic illustration of sensitization due to Fig. 29 Schematic illustration of how a heat treatment re- chromium-rich precipitates that deplete adjacent lates to sensitization due to precipitation kinetics. regions of chromium. GB, grain boundary TTT, time-temperature-transformation Chapter 4: Corrosion Types / 47

and under 0.03% for low-carbon L grades. activity of carbon and make alloys more suscep- Thus, normal grades sensitize below around tible. Nitrogen lowers the tendency for carbide 800 ¡C. The supersaturation increases with formation and slows sensitization. Nonthermo- decreasing temperature, but below about 500 ¡C dynamic effects are those of austenite grain size diffusion of carbon is too slow for carbon to and prior cold work. Decreasing grain size and move to grain boundaries and cause the damag- therefore increasing grain boundary surface ing combination with chromium that causes area decreases the amount of precipitate per unit sensitization. Low-carbon grades avoid sensiti- area of grain boundary and therefore the amount zation because they are not sufficiently super- of chromium depletion per unit area. Cold work saturated at temperatures at which carbon is accelerates diffusion and makes precipitation mobile enough to diffuse to grain boundaries. more rapid, thus aggravating sensitization. Ferritic. Another situation exists in ferritic The thermodynamic affinity tool can be used stainless steels, in which carbon is much less to prevent chromium carbide formation in an- soluble but is much more mobile. Annealing other way. Introducing alloying elements that over 900 ¡C can put enough carbon in solution to combine with carbon more strongly and rapidly cause sensitization even at the lowest carbon lev- than chromium can exhaust the supply of car- els attainable in an AOD and even at the fastest bon available to precipitate as chromium car- possible quench rates. The damaging chromium bide. There are a number of candidate elements, depletion caused by this very rapid precipitation zirconium, vanadium, tantalum, niobium, and can be undone by a simple rehomogenization titanium, most prominently. Of these, the diffu- anneal of the remaining chromium. This is theo- sivity and affinity for carbon of niobium and ti- retically possible with austenitic alloys also, but tanium make them the best for this purpose. the diffusion rates of chromium in austenite as Each forms stable carbides at much higher tem- so slow that it is impractical in most real cases. peratures than chromium, starving chromium of Duplex steels have a subtle near immunity to sufficient carbon to form damaging precipitates. carbide sensitization. While they are typically The caveat with titanium is that it forms oxides, low carbon anyway, the carbides that do form sulﬁdes, and nitrides preferentially to carbides. do so at ferrite-austenite grain boundaries. Here, Therefore, sufficient quantities must be used to chromium is consumed from the chromium-rich accommodate the prior formation of these ferrite phase, leaving the austenite intact. Their phases. Niobium tends more toward carbide large grain boundary area keeps carbide con- than nitride formation but is a weaker carbide centration per unit area low, and the fast diffu- former than titanium. The solubility products of sion in the ferrite keeps austenite from becom- these precipitation reactions are: ing depleted. However, the rapid formation of 6780 intermetallic phases at the ferrite-austenite in- log [Ti ][ C ]=−297 . (Eq 13) terfaces can lead to a rapid loss of corrosion re- T sistance and a severe loss of toughness if exposure to temperatures within the intermetallic precipitation range is not controlled. 9350 log[Nb ][ C ]=−455 . (Eq 14) Martensitic steels are quenched as austenite T to and through the Ms temperature without time for carbon to precipitate in austenite. The car- These equations follow the form of the gen- bon in the martensite can precipitate and cause eral equation for precipitation reactions: sensitization if reheated to the 300 to 700 ¡C region. Fortunately, heating to above 700 ¡C re- log[MX ][ ]=− AHRT / (Eq 15) homogenizes the chromium and eliminates sensitization. in which A is a constant, H is the heat of disso- Effect of Alloying. Besides determining lution, R is the gas constant, and T is the ab- basic phase structure, alloying plays a role in solute temperature. If the amount of titanium or susceptibility to sensitization. Those elements niobium is stoichiometrically sufficient, no car- that reduce the tendency of chromium carbides bon will form chromium carbides under equilib- to form also reduce the susceptibility to sensiti- rium conditions. It is possible to defeat the sta- zation. This is a purely thermodynamic effect. bilization reactions by quenching the alloys Molybdenum, silicon, and nickel promote car- from temperatures at which titanium carbide or bide formation by increasing the thermodynamic niobium carbide is dissociated. If free carbon is 48 / Stainless Steels for Design Engineers

left free in the matrix by quenching, then on re- layer occur, and region II, where the protective heating it may form carbides with the most lo- layer is not fully developed, suggesting an ap- cally accessible favorable element, such as preciable electrochemical effect. The latter is a chromium, rather than the most thermodynami- zone that exists in alloys that have zones of cally favorable element, which would be tita- chromium depletion. nium or niobium. This can occur when a stabi- Stress corrosion cracking has always been lized alloy such as 321 is welded. A zone away among the most controversial subjects among from the weld may experience a high enough metallurgists and electrochemists. The debate temperature to put carbon into solution and then centers on whether the critical mechanism is cool just rapidly enough to not form only the dissolution or fracture, and if a fracture, by what equilibrium titanium carbide but also Cr23C6 at mechanism. Is the cracking zone locally soft- grain boundaries, causing the type of sensitiza- ened, locally hardened, transformed, to a more tion called knife-line attack. This problem has brittle phase or embrittled by hydrogen? As of nearly ceased to exist as modern 321 has low this writing, there is no general agreement on levels of carbon and nitrogen for economic rea- which type mechanism is the fundamental sons; this effectively precludes this chromium cause, but there is room for convergence. Obvi- carbide precipitation in most cases. ously, elements of many may come into play. It Welding. Many of the most severe problems is likely, as in most prolonged arguments, that of sensitization arise when stainless steels are no hypothesis is completely correct. We will try welded to carbon or low-alloy steels. In these to fairly set out what is known and agreed on as situations, construction code rules usually re- fact and then present researchers’ views in an quire that the carbon steel component be given a unbiased manner, but since we concern our- stress relief annealing (SRA) treatment. Such selves only with stainless steel, no attempt is SRA treatments are typically in the sensitization made to address an all-encompassing theory. temperature range for austenitic stainless steels. Crack Initiation. In stainless steels, cracks Use of low-carbon or stabilized grades is neces- can be seen to initiate at surface defects and ir- sary in such cases. Even then, use of the lowest regularities. In stainless steel, it must be agreed allowable temperature SRA treatment for the by all that the preponderant initiation site is a shortest allowable time is preferred. corrosion pit or, in some cases, a crevice. Inter- granular corrosion sites, as are seen in sensitized material, can also provide the conditions Corrosion Combined with for SCC initiation. The interrelationship be- Fatigue or Fracture tween pits and SCC cracks has been studied (Ref 25). Stress lowers the anodic potential at which pitting occurs and permits metastable pits Environmentally induced failure occurs when brittle failure under tensile mechanical loading occurs at a lower stress when a material is subjected to a corrosive environment than what would happen in a noncorrosive environment. This introduces us to what is perhaps the most controversial technical subject in all of stainless steel research, SCC.

Stress Corrosion Cracking The key cause for SCC is the cooperating effects of tensile stress and a corrosive environment. Such cases can be identiﬁed in most alloy systems, and even pure metals, which were thought to be more or less immune, also have had cases of SCC reported. In passive metals, two sensitive potential regions for the occurrence of SCC have been identiﬁed and are shown in Fig. 30: region I, where pitting and breakdown of the passive Fig. 30 Zones of susceptibility to stress corrosion cracking Chapter 4: Corrosion Types / 49

to become stable via the generation of cracks. environmental variables. The most common ex- Cracks, once formed, presumably have favor- ample is that of sensitized 304 in high-tempera- able geometry to duplicate pit internal chemical ture water or caustic media. The relevance of reactions and must be considered to be de- this to the normal case of stainless steels must scribed by the models that apply to pits and be questioned since, by deﬁnition, the sensitized crevices. grain boundaries themselves can be depleted of The stress at which SCC initiates has a chromium to a degree they are not stainless and threshold, which has been reported as between have a much less stable austenitic structure,

25 and 50% of the yield strength in austenitic having their martensite start temperature Ms, stainless steel. raised by the loss of chromium. The temperature at which SCC is initiated Material Variables. Martensitic stainless ranges from ambient to under 100 ¡C for steels and martensitic precipitation hardened martensitic materials, while austenitic alloys stainless steels are quite susceptible to SCC. begin their sensitivity above room temperature This susceptibility increases with hardness, and increase in susceptibility with increasing yield strength, and embrittling heat treatments. temperature. The ferritic steels, while consid- They will crack at threshold stresses equal to ered nearly immune to SCC, have their maxi- 50% of yield strength. These alloys can be tem- mum susceptibility in the same range as marten- pered at sufficiently high temperatures that they sitic steels. In environments of mixed chlorides become soft and tough enough to have very and sulfides, however, SCC can occur in all good resistance. types of stainless at room temperature. This has Ferritic stainless steels of low and medium been seen in the SCC of austenitic stainless chromium are generally not susceptible to SCC. steel in swimming pool environments, in which Ferritic alloys, which can have a martensitic chloride ions can condense on the stressed steel structure, should be considered martensitic for and cause pitting and SCC. SCC purposes. If purely ferritic alloys are al- Cracks propagate very slowly below spe- loyed with copper, molybdenum, and nickel, cific certain stress intensity levels, but once that they can become susceptible. The presence of intensity is reached, they have a plateau rate α'( or high-temperature embrittlement also in- that is fairly constant until the stress level at creases susceptibility, as does cold work. which catastrophic failure occurs at very high Despite the controversy surrounding the propagation rates. Rates of crack propagation mechanism of SCC in austenitic stainless steels, are exponentially increased by increasing tem- there is almost complete agreement that SCC of perature. The crack propagation rate has been body-centered cubic (bcc) stainless steels, seen across a range of alloys to be linearly pro- martensitic, ferritic, and pH is simply a mani- portional to the average current density that festation of HE, with hydrogen provided by ei- alloy experiences when its surface is strained, ther anodic (e.g., active corrosion within a pit) indicating that reactions at the crack tip are or cathodic reactions. strain sensitive, and overall rate limiting, but Duplex stainless steels have low susceptibil- not necessarily the mechanism of cracking. ity to SCC. Their dual-phase microstructure en- Crack growth is discontinuous with individual sures that under conditions that crack austenite, steps of growth many times the average rate, ferrite remains as a crack-arresting phase, while which is similar to that seen with gaseous hy- under conditions that cause SCC in highly al- drogen embrittlement (HE). The crack growth loyed ferrite, the austenite is a crack arrester. gives off acoustic emissions as cracking steps A second explanation of the resistance of du- occur. These steps of growth are brittle and are plex alloys to SCC is that their two phases have seen as facets on fractographs with cleavages different corrosion potentials, and that the corresponding to crystallographic planes. The mixed potential that arises because they are in crack facets match with high perfection, show- intimate contact is outside the potential range ing almost no evidence of plastic deformation for SCC on either phase. This fits with the re- or dissolution. sistance to SCC of wrought alloys with a lamel- The propagation path may be intergranular or lar structure and the lesser resistance of cast al- transgranular. Grain boundary propagation in loys that lack that structure. stainless steels usually corresponds to condi- Austenitic stainless steels are the type of stain- tions under which grain boundaries are less cor- less steel generally associated with SCC, and rosion resistant because of either material or they vary in their degree of susceptibility to 50 / Stainless Steels for Design Engineers

SCC. All other things being equal, alloying which are probably the greater cause of resist- elements that delay or prevent localized corro- ance to SCC. sion do the same to delay SCC. This is simply Environmental Variables. There are three the delay of initiation. However, if pitting can be key types of environments in which SCC occurs delayed indefinitely, then SCC can also, assum- in stainless: ing, of course, more harmful localized corrosion, ¥ such as that due to intergranular chromium de- Chloride-containing solutions ¥ Caustic solutions pletion, is not occurring. Molybdenum, which ¥ we already know helps prevent pitting and Polythionate and thiosulfate solutions crevice corrosion, also increases the threshold The cases of polythionate and thiosulfate solu- stress for SCC, as shown in Fig. 31 (Ref 31). tions are industrially important but can be ade- But, if metastable or stable pitting is occurring, quately explained as simply the stress-assisted the threshold stress has been reached, and the intergranular corrosion of sensitized material. temperature is sufficient, then SCC will proceed. Hot caustic solutions are aggressive against It is mitigated by material variables such as cold stainless steels. Certain combinations of concen- work and by alloying elements that increase trations, temperature, impurity, and dissolved austenite stability. Many publications cite nickel oxygen can cause SCC as well as other undesir- as beneficial in enhancing resistance to SCC, able corrosive attack. Resistance to general cor- often referring to the data from Fig. 32. However, rosion is proportional to nickel content, but fer- its role seems mainly to be as an austenite stabi- ritics and duplex alloys are less prone to SCC. lizer and as a retarder of active corrosion. The 304 has been reported to have no meaningful minimum in the curve corresponds to the nickel threshold stress for SCC in hot caustic solutions, level at which the structure is entirely austenitic, leading one to question whether such a failure but least stably so. Lower nickel levels produce should even be classified with SCC of the typi- better immunity through the duplex structure, cal chloride-induced type or belong with the pre- while higher levels promote austenite stability vious polythionate and thiosulfate solutions. and correspond to alloys having more alloying Chloride containing environments are the elements, such as chromium and molybdenum, main ones that induce SCC. Water can cause

Fig. 31 Inﬂuence of molybdenum on resistance to stress Fig. 32 Variation of resistance to stress corrosion cracking corrosion cracking (SCC) in austenitic steels with nickel (and other) content and structure Chapter 4: Corrosion Types / 51

SCC at sufficiently high temperatures (i.e., does advance discontinuously, and after each above 100 ¡C) if there are even very low com- advance there is fresh surface, which comes bined concentrations of chloride (greater than into equilibrium with the solution within the 0.1 ppm) and oxygen (greater than 0.1 ppm) crack. So, any experiment, such as that shown dissolved (see Fig. 33) (Ref 32). in Fig. 34 (Ref 33), that tests crack propagation Failure times decrease exponentially with de- against electrochemical events will absolutely creasing chloride content. Crack growth rate in- support this model. creases by a factor of ten with each 30 ¡C rise in It is axiomatic that films must rupture and re- temperature. Decreasing pH lowers the temper- form as cracks advance discontinuously. The ature at which SCC occurs in a given time. weakness of this model is that it does not pro- Mechanisms. There have been many mecha- vide a mechanism for brittle fracture, and the nisms proposed for SCC in stainless steel. We very brittle features of transgranular SCC frac- focus only on those that address the failure in ture surfaces do not show any supporting evi- chloride-containing media, the main concern for dence of dissolution. Research (Ref 34) show- users of stainless steel. ing that metal dissolution at the crack tip is The models that have found some support isotropic rather than crystallographically ori- are: ented make dissolution models incapable of being reconciled with the crystallographic frac- ¥ Slip dissolution ture surface facets. ¥ Adsorption-enhanced plasticity Adsorption-induced brittleness, also known ¥ Adsorption-induced brittleness as stress-sorption, looks to the parallels be- ¥ Hydrogen embrittlement tween liquid metal embrittlement and SCC to Slip dissolution (anodic dissolution) was the explain the mechanism of SCC as the action of earliest proposed model for SCC. It simply pro- adsorbed species weakening atomic bonds on poses that at a crack tip a passive film forms, and the crack tip surface. If the action is on the sur- after time it fractures by an unspecified mecha- face, however, the mechanism cannot produce nism. The fresh active surface may or may not the observed discontinuous, brittle cracks which repassivate, after which the process repeats itself. characterize SCC. Only in alloys such as Fe-3Si The strength of this model is that it actually are steps small enough to make this mechanism does describe what is happening. The crack plausible.

Fig. 33 Variation of susceptibility to stress corrosion cracking (SCC) with media oxygen and chloride content for 304 stainless steel. Source: Ref 32 52 / Stainless Steels for Design Engineers

Fig. 34 Crack propagation rates of various metals plotted versus current density. Source: Ref 33

Adsorption-enhanced plasticity/hydrogen em- actually take place, but none is speciﬁc enough brittlement encompasses a number of models to have been tested by critical experiments to that observe that adsorbed species enter the lat- prove or disprove it. tice in the vicinity of the crack tip and then It has been demonstrated that hydrogen is ab- cause failure by one of several mechanisms: sorbed into the material at the crack tip. The main question is whether it causes damage by ¥ Dealloying and porosity creating porosity, altering dislocation mobility, ¥ Adsorption-induced brittleness or causing lattice decohesion. There is support ¥ Coalescence of voids formed by cross slip for each. enhanced by the adsorbed species It has been observed that where SCC occurs Since hydrogen is the only species that is pro- there is a large concentration of vacancies. This duced in quantity and is capable of diffusing into has led to speculation that porosity is a weaken- the lattice, HE is implicit in all these models. ing mechanism responsible for SCC (Ref 35). It All of the models have support in that they has been proposed and supported by calcula- have some experimental observations that show tions that hydrogen lowers the energy required that the phenomena they propose as causal or vacancy formation. The lowest energy Chapter 4: Corrosion Types / 53

configuration is calculated as two hydrogen atoms per vacancy. This pairing of hydrogen solute atoms to dislocations is very reasonable given the major distortion the interstitial hydrogen causes to the lattice, so there is no basis to challenge the enhanced vacancy formation. Whether the effect is large enough to cause failures has not been demonstrated. The largest measurable effect of hydrogen has been a slight acceleration of stress relaxation in martensite. The relevance of hydrogen-induced vacancy ag- glomeration as the principal cause of failure must be considered questionable until some further critical experiments link the vacancies to the observed instances of failure quantitatively, and Fig. 35 Stress-strain curve for single crystals of stable austenitic stainless steel with and without hydrogen. more important, to show how this mechanism Source: Ref 36 could account for the temperature and stress dependence observed. The major influence of va- in Fig. 35 (Ref 36). All these interstitials strain cancy formation due to hydrogen may be to en- the lattice and therefore harden in proportion to hance the volume expansion due to hydrogen. their atomic size. Hydrogen, as the smallest of Adsorption-induced brittleness, also known them, has about half the distorting effect and as stress sorption, looks to the parallels be- half the hardening effect. But, its small size tween liquid metal embrittlement and SCC to makes it mobile at ambient temperatures, so it explain the mechanism of SCC as the action of can diffuse to sites where it can alter mechani- adsorbed species weakening atomic bonds on cal properties. the crack tip surface. If the action is on the sur- But, while hydrogen causes dislocation mo- face, however, the mechanism cannot produce tion and lower work hardening, it does not the observed discontinuous, brittle cracks that weaken austenite, so this theory by itself cannot characterize SCC. Only in alloys such as Fe- account for the role of hydrogen in SCC and, by 3Si are steps small enough to make this mech- inference, in HE in the more general case. anism plausible. In stainless steels, there The quandary of hydrogen finally having seems to be nothing to support this proposed been shown to have a clear effect on mechanical mechanism. properties but having that not account for either Adsorption-enhanced plasticity has become SCC or HE may be put to rest by the additional known recently as HELP or hydrogen-enhanced observations of hydrogen’s role as a lattice dis- localized plasticity. The underlying mechanism torter (Ref 37). While not formalized as a pro- at work in this model is the hydrogen-induced posed hypothesis for SCC, the role of hydrogen shielding between microstructural defects. This as a generator of very high stresses has been has been observed distinctly in single crystals of pointed out as a factor that cannot be neglected austenitic stainless alloys. The Cottrell atmos- when evaluating other proposed mechanisms. phere of hydrogen around dislocations causes Hydrogen has been shown to distort the lattice mutual repulsion, causing strain to be localized in proportion to its concentration. The effect is on certain slip systems. This has been observed not small, accounting for about 1% strain per to occur and has caused deformation to become 0.1% concentration by weight, as shown in concentrated in Luders bands in austenitic al- Fig. 36 (Ref 38). At hydrogen levels of over loys, which of course do not show such behav- 1000 ppm, which are thought to exist around ior without hydrogen (Ref 36). This also pro- growing SCC crack tips, there could therefore duces ε-martensite in austenitic alloys, which be hydrogen concentration gradients capable of would be considered stable without hydrogen producing additional tri- or biaxial stresses on and deformation. the matrix ahead of the crack tip that may ap- This theory encounters a problem, however, proach the yield stress and account for some or with the fact that the same studies showed that all of the difference between the normal fracture hydrogen actually strengthens the matrix by toughness KI and the KISCC, that for SCC. This solid solution hardening. It acts in much the also precludes the necessity of hypothesizing same way as carbon and nitrogen do, as shown hydrogen-induced phase changes, although 54 / Stainless Steels for Design Engineers

Fig. 36 Dilation of austenite due to hydrogen in solution. Source: Ref 38 were they to exist, they would result in the same volved distinguishing among the same mecha- lattice expansion. In both cases, the failure nisms, namely: would occur at a region ahead of the crack tip ¥ Decohesion and beyond the highest hydrogen concentration, ¥ which is what is observed to occur. The growth Enhanced local plasticity ¥ Adsorption embrittlement of this stress over time with increasing hydro- ¥ gen-producing corrosion would account for the Void coalescence observed kinetics, locus, and stress dependence The identiﬁcation of the operative mechanism of SCC. for HE involves again distinguishing what role If nothing else, the main models for SCC and each of the above contributes to HE in a given the experimental results on which they are situation since all are known to be real metallur- based should be reexamined in view of the fact gical phenomena. that the stresses induced by hydrogen are not The main difference between HE and SCC in negligible and, in fact, may account for much of stainless steel is that HE is limited to ferrite, the observed SCC behavior of stainless steels. which is hardened by cold work or alloying, and The next few years may ﬁnally see the resolu- martensite. Austenite is somewhat diminished tion of the lengthy debate over the causes of in ductility by hydrogen, but not subject to the SCC. If it comes, it will be from critical experi- completely brittle, discontinuous cracking of ments, which can quantitatively differentiate bcc stainless. The observations that make a among the above effects and measure the con- given model plausible as a mechanism for SCC tribution of each. lack traction for the same materials in HE. It hard to envision enhanced plasticity involved in the completely brittle fracture of high-strength Hydrogen Embrittlement martensitic stainless steels, whereas void coa- Like SCC, there has been debate about HE lescence by vacancy creation seems more likely that has produced more heat than light. This in- to account for the observed behavior. Chapter 4: Corrosion Types / 55

The resolution of mechanism here also must Biocorrosion and Microbiologically account for the contribution of hydrogen-in- Induced Corrosion duced stress as well as hydrogen effects on mechanical processes, especially since the ob- There are many cases for which biological or- served susceptibility to HE is proportional ganisms contribute to initiating or enhancing hardness, therefore to the amount of hydrogen a rates of corrosion. This can occur in natural en- given material can hold both in normal intersti- vironments such as ground or seawater as well tial solution and the amount it can trap at lattice as domestic and industrial environments such as defects (Ref 39), especially the dislocations the nuclear and chemical processing industries, within the plastic zone at the crack tip, which for example. This is called biocorrosion or provide enhanced hydrogen solubility where it MIC, microbiologically induced corrosion. can aggravate the applied crack opening with a The bacteria that are known to influence cor- wedge effect from hydrogen dilation. rosion can be sorted as aerobic bacteria that lie in aerated water and anaerobic bacteria. Among Corrosion Fatigue the anaerobic bacteria that are known to [16] Just like SCC, corrosion fatigue causes brittle (Ref 40) affect stainless steels can be counted: failure under a combined environment of corro- Desulfibrio and Desulfotomaculum. Both of sion and a tensile stress component. The stress, these are so-called sulfate-reducing bacteria however, is cyclic and in a test of stress versus (SRB), which means that they promote the reac- number of cycles (S vs. N), failure will occur at a tion: lower N under the corrosive environment. The SO22−−→+ S4 O (Eq 16) cracks are transgranular, and the collaborative ef- 4 fect of corrosion and fatigue is that corrosion accelerates the plastic deformation that accompa- which in turn accelerates the cathode reaction: nies the evolution of extrusions and intrusions. In corrosion fatigue, an obvious pit corrosion 22HH=−+→e 2 (Eq 17) site may not be necessary because of the combined action of cyclic stresses and the environ- Aerobic bacteria flourish under oxygen (Ref ment. However, an initiation site that is the 40). Examples are the iron-oxidizing Gal- weakest link in a combined mechanical and lionella and Sphaerotilus, which increase the metallurgical sense will be the initiation point anode dissolution reaction: after which conditions that may not cause SCC can help propagate fatigue cracking at lower Fe→= Fe 2+−2e (Eq 18) stresses than would be expected in more benign environments and in environments that may not by converting the ferrous iron-ion product cause SCC or pitting under static loads. (Fe2+) to less soluble ferric (Fe3+). Due to this, The importance of the environmental interac- macroscopic so-called tubercules form that can tion is reflected in the sensitivity to frequency of cause crevice-type shelters where differential stress application. High-frequency loading aeration and pit initiation can occur. gives less time for corrosive attack and brings Countering MIC with biocides can cause crack propagation rates down closer to those in problems in manganese-containing waters. Oxi- air. In some materials, crack propagation rates dizing biocides, such as ozone, chlorine, or per- are elevated above those in air at all stress lev- oxide, can cause manganese to be oxidized to els, while in others a threshold stress intensity manganese dioxide. The precipitated deposits of must be reached before an acceleration is noted. manganese dioxide can accelerate pitting corro- Some materials show a combination of both. sion even in low-chloride waters in which al- The first case seems to be merely fatigue as- loys such as 316 would otherwise be safe from sisted by corrosion, while the last two seem to pitting attack. indicate an SCCÐtype behavior. Biocorrosion is most commonly encountered The same uncertainties that cloud our under- in ambient aqueous environments, which are the standing of SCC necessarily disguise the pre- environments in which most microorganisms cise mechanism of corrosion fatigue, which have evolved to thrive. So, it tends to be a prob- must be viewed as a combination of SCC and lem for the medium-alloyed steels such as 304 fatigue. and 316, which are used in these environments. 56 / Stainless Steels for Design Engineers

The more chemically or thermally hostile envi- 15. R.C. Newman, Corrosion, Dec 2001, ronments in which higher alloyed grades are p 1030Ð1041 used are also hostile to bioorganisms and thus 16. N.J. Laycock and R.C. Newman, Corros. minimize the problem. Sci., Vol 39, 1997, p 1771 The development of microbiological consortia 17. Y. Kobyashi, S.Virtanen, and H. Bohni, allow anaerobes to flourish under biofilms that Proc. Electrochem. Soc., 1999, p 533Ð540 form in an aerated environment. These represent 18. Z. Szlarska-Smialowska, Pitting Corrosion a differential aeration cell that acts just like a se- of Metals, NACE, Houston, TX, 1986 vere crevice. Also, the action of microbes in 19. E.T. Turkdogan, Fundamentals of Steelmak- raising the corrosion potential is key to under- ing, Institute of Materials, 1996 standing why natural seawater is so much more 20. H.S. Kim and H. Lee, Met. Trans. A, corrosive than sterile sodium chloride or syn- Vol 32A, June 2001, p 1519 thetic seawater solutions. And, macrofouling or- 21. M.P. Ryan, D.E. Williams, et al., Nature, ganisms are important. They create crevices and Vol 415, Feb 2002, p 770Ð777 sites where microfouling can start early. At the 22. A.J. Grekula et al., Corrosion, 40, 1984, same time, they are sources of turbulence in p 569 flowing systems, and this turbulence can cause 23. Stainless Steels Les Editions de physique, 1993 flow erosion in copper materials, making use of 24. N. Suutala and M.Kurkela, Stainless Steel stainless steels more attractive. ‘84, Metals Institute, 1985, p 240Ð247 25. T.Suter, E.G. Webb, H. Bohni, and REFERENCES R.C. Alkire, J. Electrochem. Soc., Vol 148 (No. 5), 2001, B174ÐB185 1. M.P. Ryan et al., Critical Factors in Local- 26. M.O. Spiedel, Stainless Steels ‘87, Institute ized Corrosion, Proc. Electrochem Soc., Vol of Metals, London, 1988, p 247Ð252 150, 2003, p 284Ð294 27. Y. Cao, F. Ernst, and G.M. Michal, Acta 2. K. Sieradski and R.C. Newman, J. Elec- Mater., Vol 51, 2003, p 4171. trochem. Soc., Vol 133, 1986, p 1980 28. G. Lothongkum et al., Corros. Sci., Vol 48, 3. L. Brewer, Science, Vol 161, 1968, p 115 2006, p 137Ð153 4. W.J. Tobler and S.Virtanen, Critical Factors 29. S. Fujimoto , Sci. Technol. Adv. Mater., Vol in Localized Corrosion, Proc. Electrochem 5, 2004, p 195Ð200 Soc., 2003, p 583Ð594 30. J.D. Fritz, J.F. Grubb, B.W. Parks, and 5. B. Baroux et al., Corros. Sci., Vol 47 (No. C.P. Stinner, Stainless Steel World, KCI, 5), 2005, p 1097Ð1117 P01488, 2001 6. http://www.alleghenyludlum.com/pages/ 31. M.O. Spiedel, Met. Trans. A, Vol 12A, products/xq/asp/T.1/qx/productCategory. 1981, p 779 html 32. A.J. Sedricks, Corrosion of Stainless Steels, 7. K. Kimura et al., High Cr Stainless OCTG Wiley, 1979, p 158 with High Strength and Superior Corrosion 33. R.N. Parkins, Br. Corros. J., Vol 14, 1979, p 5 Resistance, JFE Technical Report 7, Jan 2006 34. S. Tahtinen, H. Hahhinen, and T. Hakkarainen, 8. http://www.outokumpu.com/applications/ Stainless ‘84, Metals Institute, 1985, p 143Ð148 documents/start.asp 35. M. Nagumo et al., Met. Trans. A, Vol 32A, 9. J.E. Truman, Corrosion: Metal/Environ- Feb 2001, p 332 ment Interaction, Vol 1, Newness-Butter- 36. H. Hanninen et al., Hydrogen Effects on worths, 1976, p 352 Materials Behavior, TMS, 2003, p 201Ð210 10. J.P. Audouard, Stainless Steels, Les Editions 37. V.J. Gadgil, Scr. Metal., Vol 28, 1993, de physique, 1993, p 268 p 1489Ð1494 11. H. Mimura et al., Nippon Steel Tech. Report 38. M. Hoelzel et al., Mater. Sci. Eng. A, 90, July 2004, p 94Ð99 Vol 384, 2004, p 255Ð261 12. http://www.alleghenyludlum.com/ludlum/ 39. B.G. Pound, Hydrogen Effects on Materials Documents/AL-6XN_sourcebook.pdf Behavior, TMS, 2003, p 93Ð103 13. http://www.alleghenyludlum.com/ludlum/ 40. S.C. Dexter, Microbiologically Influenced Documents/al610_611.pdf Corrosion, Corrosion: Fundamentals, Test- 14. F. Tagashi et al., Kawasaki Technical Re- ing, and Protection, Vol 13A, ASM Hand- port 31, 1994 books, 2003, p 398Ð413 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 57-68 All rights reserved. DOI: 10.1361/ssde2008p057 www.asminternational.org

CHAPTER 5

Oxidation

Summary is negative. In Eq 1, the free energy G is decreased by a lower nobility of the metal (or a higher activ- STAINLESS STEEL, often considered mainly ity a of a metallic alloying element), a lower tem- as a corrosion-resisting material, plays an impor- perature T, and a higher partial pressure P of the tant role as a heat-resisting material. This is oxidizing gas according to: partly due to its ability to retain strength at higher ⎛ ⎞ temperatures at which many otherwise useful al- aMeX ΔΔGGRT=+ln ⎜ y ⎟ loying systems, such as aluminum, copper, and 0 ⎜ 05. y ⎟ (Eq 2) ⎝ aPMX ⎠ even titanium, soften. Stainless steel retains 2 strength and has excellent oxidation resistance from room temperature to nearly 1000 °C, at In the case of alloy oxidation, for which tem- which other economical alternatives are lacking. peratures are high enough to form mixed oxides or spinels, the activities of the oxide species also need to be considered. Δ The standard Gibbs free energy G0 is often Introduction presented in Richardson-Jeffes (Gibbs free energy-temperature) diagrams such as the one High-temperature oxidation is a form of envi- shown in Fig. 1 (Ref 4). ronmental degradation of metals and alloys that It is evident from Fig. 1 that the major alloying results from the following chemical reaction in element in stainless steels, chromium, forms a which metal atoms M react with gaseous oxidants: thermodynamically signiﬁcantly more stable +→ oxide, Cr2O3, than those of the base alloy iron Ms()05 . yXg2 () MXy (Eq 1) (FeO, Fe3O4, and Fe2O3) or the major ternary el- Due to the high temperatures involved, these ement nickel (NiO), and to a great extent, the reactions are generally rapid and thus are a con- chromium content determines the oxidation be- cern for high-temperature applications such as havior of stainless steels. components for power generation. The elec- The Effect of Chromium. The oxidation of tronegative gaseous oxidant X could be sulfur, multicomponent alloys is a complex process chlorine, etc., but the discussion here mainly is from both thermodynamic and kinetic points of limited to oxidation by oxygen or water vapor view. A range of oxides may form with various (in the latter case, hydrogen would be added as degrees of thermodynamic stabilities and stoi- a product in Eq 1. For a thorough study of oxi- chiometries (including complex ones with dif- dation, referred to Ref 1 to 3). ferent cations), and there might be degrees of solubilities of oxides in one another. Kinetics of their growth is complex because metal solute diffusion in the metal phases varies, as do metal Thermodynamics of Oxidation and oxygen ion mobilities in the different oxide phases. As discussed in Chapter 2, Corrosion Theory, a Birks, Meier, and Pettit distinguished between reaction will be possible when the net free energy two basic types of behavior: (a) a noble matrix 58 / Stainless Steels for Design Engineers

Fig. 1 Standard Gibbs free energy of formation of some metal oxides as a function of temperature. Source: Ref 4

metal with less-noble alloying elements and (b) spinels Fe1.5Cr1.5O4 (with a solid solubility with both matrix element and alloying elements are Fe3O4) and FeCr2O4 form. The progressive nonnoble. The concept of nobility is decided by change in oxidation behavior as chromium is the thermodynamic conditions of Eq 2; that is, added to iron has been described in the litera- an element for which the free energy defined by ture (Ref 1). Eq 2 is negative is nonnoble. We first discuss the At lower chromium contents and above a min- more common case (b), in which oxidation takes imum temperature, an iron-chromium alloy place under significantly oxidizing conditions, would behave as pure iron, where FeO would such as air. In such a situation, it can be seen form next to the metal, then gradually Fe3O4 and from Fig. 1, that the matrix iron is nonnoble and Fe2O3 would form toward the gas as oxygen po- so are many of the solutes (chromium, molybde- tential increases. Isolated pockets of spinel may num, aluminum, silicon, manganese, etc.). form within the FeO layer. The oxidation of iron A high-temperature Fe-Cr-O phase diagram is proceeds predominantly due to the rapid ionic 2+ shown in Fig. 2. It can be seen here that Fe2O3 diffusion of Fe cations on the FeO layer, and Cr2O3 are soluble in each other, and that the which leads to growth of this layer. If chromium Chapter 5: Oxidation / 59

Fig. 3 Parabolic rate constants for the growth of several oxides. Source: Ref 6

Here, m is the added mass, A is the area exposed to the oxidizing atmosphere, t is the time exposed, and km is the parabolic rate constant. The subscript “m” is added here to denote that the reaction is measured as added mass (it can also be defined for oxide thickness X). As Fig. 2 The iron-chromium-oxygen phase diagram at 1300 °C. shown in Fig. 3, the parabolic rate constant for Source: Ref 5 the oxides of chromium, silicon, and aluminum are low compared to others, and this is the rea- in the base alloy is increased, the spinel pockets son that these elements are used as alloying ele- 2+ increase, and the mobility of Fe decreases. As ments to reduce oxidation rates for alloys in chromium content is increased further, a mixed- high-temperature applications. spinel scale is formed. Iron diffusion through the The solubility of Fe3O4 in the spinel will mixed spinel is significant, and thus the scale is eventually result in continued iron oxide for- not yet protective. As chromium content is in- mation, as the iron-chromium system is not an creased further, an outer layer of Cr2O3 is optimal basis for high-temperature oxidation formed, and the oxidation behavior becomes resistance, although it might be an option from similar to that of chromium. A chromium limit an economical standpoint compared to other of roughly 20% is needed to achieve a perma- alloy systems such as superalloys. The high nent Cr2O3 scale. This amount decreases if mobility of both iron and manganese in the nickel is added. While the thermodynamic driv- spinel structure is also an important factor. ing force is important, the chromium content of The major stainless steels used for oxidation the alloy can override because the supply of resistance fall into two categories: the ferritic chromium to the interface becomes dominant. stainless steels and the austenitic. Table 1 lists At chromium contents less than about 16 wt%, some of the more significant alloys commonly the oxidation rate is influenced by the rate of encountered in applications for which oxida- supply of chromium from the alloys beneath the tion resistance is paramount. The value of fer- oxide. Above 16%, the supply of chromium is ritic alloys (such as 409, 439, and 446) is that fast enough that chromium gradients are low they are relatively inexpensive, and that they enough that instead transport in the oxide layer have a thermal expansion coefficient that is controls the rate. The rate of oxidation then fol- closer to that of the oxide than do austenitic allows a so-called parabolic law (this is explained loys (such as 302B, 309, and 310). This gives in the next section), by which the mass change them an advantage in cyclic oxidation applica- per unit area due to oxidation (incorporation of tions even though their strength at high temper- oxygen) is given by: atures does not rival that of austenitic alloys. The ferritic stainless steels are the most widely ( )2 = mA/ ktm (Eq 3) used alloys based on their low cost, which has 60 / Stainless Steels for Design Engineers

Table 1 Oxidation-resisting grades of stainless steel in common use

Composition, %

UNS Name C N Cr Ni Mn Si Ti Nb Other S40900 409 0.08 10.5–11.75 0.5 1 1 6X(C + N) to ...... 1.10 11Cr-Cb(a) 0.01 0.015 11.35 0.2 0.25 1.3 . . . 0.35 . . . 12SR(a) 0.02 0.015 12 ...... 0.3 0.6 1.2 Al S43935 439 0.07 0.04 17.0–19.0 0.5 1 1 0.20 + 4X(C + ...... N) to 0.75 18Cr-Cb(a) 0.02 . . . 18 . . . 0.3 0.45 0.25 0.55 . . . 18SR(a) 0.015 . . . 17.3 0.25 0.3 . . . 0.25 . . . 1.7 Al 4742(a) 0.08 . . . 18 . . . 0.7 ...... 1.0 Al S44600 446 0.2 0.25 23.0–27.0 0.6 1.5 1 ...... S30215 302B 0.15 . . . 17.0–19.0 8.0–10.0 2 2.0–3.0 ...... S30415 153MA 0.04–0.06 0.12–0.18 18.0–19.0 9.0–10.0 0.8 1.0–2.0 ...... 0.04 Ce S38150 253MA 0.05–0.10 0.14–0.20 20.0–22.0 10.0–12.0 0.8 1.4–2.0 ...... 0.04 Ce S30900 309 0.2 . . . 22.0–24.0 12.0–15.0 2 0.75 ...... S31000 310 0.25 . . . 24.0–26.0 19.0–22.0 2 1 ...... Note: All compositions include Fe as balance. Single values are maximum, unless otherwise speciﬁc (a) Indicates typical analysis made them the standard alloys for automotive scale-gas surface. Similarly, the electric charge exhaust systems. can be carried by either n-type (electrons) or p- type (electron holes) electronic defects. The case will be determined by the equilibrium de- Transient Oxidation fect structure of the oxide, which depends on temperature and oxygen partial pressure. In the The oxidation of a clean metal surface on ex- case of Cr2O3, chromium cations are the pre- posure to an oxidizing environment will ini- dominantly mobile defects (Fig. 4b) as a result tially lead to all the nonnoble components of the of a very small degree of deviation from stoi- alloy being oxidized together, forming mixed chiometry in the cation lattice, that is, Cr2-5O3, oxides having composition similar to the base leading to metal deﬁciency. The contribution of alloy. In stainless steels, these initial oxides are chromia grain boundary diffusion is large and typically Fe-Cr-Ni-Mn mixed oxides. This is probably dominates the process at temperatures called transient oxidation. As these oxides of interest. The defect can be described as an in- thicken, the partial pressure of oxygen at the teraction with oxygen, at high oxygen poten- scale-metal interface falls until only the most tials, through Kroger-Vink notations as: reactive element present in high concentration can be oxidized. For stainless steels, this means 3 =+''' 3 x +• OO2 ()gVCr O 3 h that a layer of Cr2O3 is eventually established in 2 2 contact with the alloy.

−3 33 ⎡ ''' ⎤ 3 =→34/ ⎡ ''' ⎤ = 2 4 ⎣VpKpCr ⎦ 2 O ⎣ VCr ⎦ KKP1 2 o (Eq 4) The Electrochemical Nature of 22 Oxidation The electron holes that form as charge-com- pensating defects serve as the “electron lead” in Once an inner scale of Cr O is formed, as 2 3 the electrochemical cell in Fig. 4b. The free en- shown in Fig. 4(a), the oxidizing gas is reduced ergy of Reaction 4 determines the concentration at the gas-scale interface, and the chromium is of mobile defects and thus the diffusion coeffi- oxidized at the metal-scale interface. The Cr O 2 3 cient and electrochemical mobility (Be) of the scale serves as both electrolyte, through which cation according to: ions are transported, and electron lead, through which electronic defects are transported. In ⎛ ΔG ⎞ principle, either or both metal or oxygen ions = ⎡ ⎤ γυΛ2 ⋅−m DV33++⎣ ⎦ exp⎜ ⎟ can migrate. If oxygen ion mobility dominates, Cr Cr ⎝ RT ⎠ then the scale would continue to grow at the ⎛ ΔH ⎞ oxide-metal interface, whereas if chromium ion = ⎡ ⎤⎤⋅⋅−m ⎣V 3+ ⎦ constan t exp⎜ ⎟ (Eq 5) mobility dominates, the oxide will grow at the Cr ⎝ RT ⎠ Chapter 5: Oxidation / 61

Fig. 4 Metal with oxide scale. (a) A protective scale that prevents gas access. (b) Schematic of electrochemical oxidation through a protective oxide scale that serves as electrolyte and electron lead. The case is for mobile cations

effects are ignored (i.e., scales are relatively 3FD + e = Cr3 B 3+ (Eq 6) thick compared to range of space charge ef- Cr RT fects). As a case study, let us assume that the mobile ion defect is cations due to metal vacan- where zi is the ion charge, F is Faraday’s con- cies in the scale. stant (96,457 C.eq–1). Compared to wustite The molar ﬂux J (moles/m.s) of a particle i in (FeO), the equilibrium constant of Eq 4 is quite an electrolyte subjected to an electrochemical low, resulting in a low degree of nonstoichiome- potential gradient was shown to be: δ try in Cr2–δO3 compared to Fe1–δO (where can be as large as 0.05), and thus the transport of ∂+(μφzF ) 3+ JcB=− ii (Eq 7) Cr through its scale is much slower than the iii ∂x transport of Fe2+ through FeO and thus the difference in parabolic rate constants in Fig. 3. where zi is the ion charge, F is Faraday’s constant (96,457 C/gram equivalent), and φ is the electric ﬁeld (V). The electronic or ionic con- Kinetics and Oxidation Rates: Wagner’s ductivity κ in an electrolyte can be computed Theory through:

The parabolic oxidation rate was introduced κ= 2 2 FzBc∑ iii (Eq 8) without explanation in Eq 3. It was ﬁrst described in terms of oxide defect structure and re- The contribution of a given ion specie type or sulting transport properties by Wagner (Ref 6), electron defect type to this conductivity is de- and the theory is explained in most of the mono- noted as the partial conductivity and computed graphs on oxidation, such as Chapter 3 in Ref 3 as: and Chapter 4 in Ref 1. This treatment follows the derivation in Ref 1. Consider a general case, 2 κ = Fz2 Bc as shown in Fig. 4 under the assumptions that iiii (Eq 9) (a) the scale is compact and adherent, (b) electrode reactions are rapid enough to be in equi- Inserting Eq 9 into Eq 7 yields: librium at the interface and surface, (c) nonstoi- chiometry is small and uniform throughout the κ ∂+(μφzF ) J =− i ii scale (i.e., defects are in thermal equilibrium i 2 (Eq 10) Fz2 ∂x throughout the scale), and (d) double-layer i 62 / Stainless Steels for Design Engineers

Now, if the mobile particles are a single type assumed, that is, a linear drop across X, at all of metal cations (e.g., Cr3+) and electrons, then times: two fluxes are present: dx = k κ ∂+(μφzF ) (Eq 18) J =− c cc dt Xt() c 2 (Eq 11) 2 ∂x Fzc The constant k is the parabolic rate constant. and A mass balance can be written where the flux of cations for a period of time dt is equated to the κ ∂+(μφzF ) amount of metal being accumulated as cations J =− e ee e 2 (E q 12) 2 ∂x inside the scale of thickness dx: Fze ⋅= JdtCdxcC (Eq 19) Electrical neutrality requires that: By combining Eq 18 and 19 and inserting Eq JZ+= JZ 0 (Eq 13) cc ee 16 for the flux, an expression for the parabolic rate constant is obtained: And at the oxide-scale/metal interface, the anode reaction is in equilibrium, that is: μ' M z + − 1 κκ MM=+c zeand therefore, k = ce dμ c 22 ∫ κκ+ M (Eq 20) zFCcCμ'' ce M

μμ=+z μ Mcce (Eq 14) If the mobility and thus partial conductivity of electrons is significantly higher than that of Combining Eq 11 to 14, the potential gradient the ions (a reasonable assumption), then Eq 20 is eliminated, and the chemical potential gradi- can be simplified as: ents can be replaced by the metal (M) potential gradient, and the following equation results: μ' 1 M k = ∫ κμd (Eq 21) κκ ∂μ zFC22 cM J =− ce M cCμM c 22(κκ+ ) ∂ (Eq 15) zFccex Since from diffusion theory we know that To obtain an explicit function for the cation Dc = BcRT, and inserting this in Eq 9, one flux, Eq 15 needs to be integrated after variable obtains separation, keeping in mind that conductivities 2 D and metal chemical potential may vary within κ = Fz2 c c the scale. Integrating from the gas-scale surface ccRT c μ μ′′ (x = 0, M = M) to the scale-metal interface μ μ′′ (x = X, M = M) one obtains. and inserting this into Eq 21 results in:

μ" μ' 1 M κκ 1 M J =− ce dμ k = Ddμ c 22 ∫ κκ+ M (Eq 16) ∫ cM (Eq 22) zFxc μ' ce RT μ'' M M

Now, if the growth of the oxide scale is con- It is often more convenient to express Eq 22 trolled by the ﬂux of cations: in terms of oxygen potentials rather than the metal potentials. It was assumed at the onset of dx ∂C this analysis that the deviation from stoichiom- ∝=−JDC (Eq 17) dt CC∂x etry is small and constant throughout the scale. Therefore, the oxide potential is constant and: The concentration drop across the scale is z constant since interface and surface reactions c μμμ+= = OMMO constant (Eq 23) are at equilibrium. If quasi steady state is 4 22zC / Chapter 5: Oxidation / 63

And thus, ddRTPdμμ=∝−ln . Therefore, O OMe thickness loss due to evaporation, which is de- 2 2 scribed by a ﬁrst-order reaction kinetics expres- Eq 22 can be written: sion with rate constant ke. Thus, the thickness ln P'' P'' change becomes: O2 O2 D ∝=c kDdP∫ cOln ∫∫ dPO 2 2 (Eq 24) dx k ' ' PO ln PO PO 2 =−k 2 2 dt Xt() e (Eq 27)

In the case of Cr2O3, if bulk diffusion is dom- inating, Eq 4 and 5 inserted into the diffusion This results in a so-called paralinear (as op- coefficient in Eq 24 for C = Cr3+, results in: posed to parabolic) rate for the oxide thickening (Fig. 6), and at a critical oxide scale thickness X( the rate of thinning due to evaporation equals P'' P'' O2 34/ O2 the rate of thickening due to oxidation. In Eq P − k ∝=∫∫dP P14/ dP 27, this means that dx/dt = 0; consequently, X = O2 O2 ' P ' P O2 P O2 O2 k/ke. While at first this seems to suggest that it 3 ⎡ 34/ 34/ ⎤ does not affect the oxidation process in that the = (PP'' − ( ' (Eq 25) ⎢ OO) ) ⎥ rate of oxidation does not increase, the forma- 4 ⎣ 22⎦ tion and evaporation of chromium oxides results in greater chromium consumption in the Thus, the parabolic rate constant would be pre- alloy compared to what would be the case if dicted to vary with the power of three-quarters of evaporation did not occur. As a result of evapo- the external oxygen partial pressure. ration losses, stainless steels that depend on a Grain boundary diffusion has however been protective chromium oxide layer are limited in identified to be important in the case of Cr3+ use to temperatures up to 900 to 1000 °C. transport (Ref 8). The observed growth rate of The presence of water vapor promotes the Cr2O3 polycrystalline films is far too fast to be formation of even more volatile oxyhydroxides accounted for by bulk diffusion of chromium (e.g., CrO2(OH)2 ) (Ref 11, 12). ions; instead, grain boundary diffusion would be expected to dominate (Ref 9). Spalling and Cracking of the Scale

The Volatile Nature of Cr O At elevated temperatures or during tempera- 2 3 ture cycling, there are multiple ways in which At high enough temperatures and high stresses can develop that may crack and blister enough oxygen partial pressures, the formation the scale, rendering it nonprotective. The differ- of a gaseous hexavalent chromium oxide CrO * ent causes of stress generation are described in 3 Chapter 5 in Ref 1. So-called growth stresses could lead to thinning of the Cr2O3 scale according to the following reaction: arise due to changes caused by the oxidation process itself. These include differences in lat- 3 tice mismatch, alloy depletion in the metal, Cr O()sg+= O ()2 CrO () g (Eq 26) 232 2 3 point-defect gradients in scales containing oxides such as FeO, with large deviation from sto- Figure 5 shows the vapor pressure of the su- ichiometry, recrystallization, and volume differ- peroxide as a function of temperature and par- ences between the oxide and metal. The last is perhaps the most commonly mentioned and is tial pressure of O2. The effect of this reaction on the oxidation ki- characterized by the Pilling-Bedworth ratio, ab- netics can be described as follows: The thick- breviated as PBR (Ref 13). ness change described through the parabolic V V Oxide rate constant in Eq 18 is corrected for by the PBR ==Oxide m Metal ν (Eq 28) VMetal Vm *

* Hexavalent chromium is now considered a human carcinogen and is rigorously regulated by both the Occupa- Here, the subscript m stands for molar volume, tional Safety and Health Administration (OSHA) and the and υ is the number of metal atoms needed to U.S. Environmental Protection Agency (EPA). form a stoichiometric unit of the oxide (in the 64 / Stainless Steels for Design Engineers

Fig. 5 Chromium-oxygen system species volatility as a function of temperature and oxygen pressure. Source: Ref 10

υ case of Cr2O3, is 2, and in the case of FeO, it is 1). When PBR is greater than 1, then the oxide is expected to be in compression and is likely to be protective, whereas if it is less than 1, the oxide is in tension and thus nonprotective. There are, however, many exceptions to this, partly because the stress state often depends more on the mechanisms and conditions of the oxidation process rather than the properties of metal and oxides. Thermal, stresses are caused by differences in thermal expansion between the oxide and metal, and the stresses generated in oxide scales can be Fig. 6 Schematic of paralinear oxidation as a result of evapo- estimated: ration of chromium superoxide

−−ET(αα)Δ σ = Oxide Oxide Metal Δ Oxide (Eq 29) ity, t is thickness, and T is the temperature ⎛ t E ⎞ α (112− ν )⎜ + Oxiide Oxide ⎟ change. In general, is larger for the metal than p ⎝ ⎠ tEMetal Metal the oxide; thus, during cooling the stresses are expected to be compressive and during heating The equation is written for a case shown tensile. Thermal stresses can cause spalling of schematically in Fig. 7, where both sides on a the protective oxide layer, and it is most severe metal undergo oxidation. Here, σ is the stress, under cyclic conditions. υ p is Poisson’s ratio (it has been assumed that Effect of Silicon, Aluminum, and Molyb- there is no mismatch), α is the coefficient of denum. Due to concerns about the cost of thermal expansion, E is the modulus of elastic- chromium and its (former) classiﬁcation as a Chapter 5: Oxidation / 65

internal Cr2O3 layer forms and thickens. Even- tually, the surface aluminum-oxide layer flakes off. Also, due to the low oxygen potential needed to form Al2O3, internal oxidation may result below the metal-scale interface in alloys tox tm tox in which formation of a continuous alumina scale film does not occur. Molybdenum is suggested to strain the lattice due to its larger size and consequently in- Oxide Metal Oxide crease the rate of bulk diffusion of elements (Ref 11), which can enhance the rate of initial Cr2O3 formation. Molybdenum is usually considered detrimental for oxidation resistance. Molybdenum normally forms MoO2 oxide, but Fig. 7 Schematic of a cross section of oxidized sample indi- this can oxidize further to form the low-melt- cating dimensions in Eq 29 for predicting thermal ing and volatile MoO3. If the MoO3 evapo- stresses rates, there is little problem, but if its volatilization is inhibited by low atmosphere circulation, liquid MoO3 can accumulate and strategic material, there were efforts to try to dissolve the protective Cr2O3 scale, leading to substitute less-expensive elements such as alu- catastrophic oxidation. minum and silicon that also are known to form Effect of Rare Earth Additions. Cerium, lan- protective layers (Ref 14), even though they thanum, and yttrium additions are known to im- have significant metallurgical and mechanical prove oxidation resistance of high temperature drawbacks. nickel- and iron-based alloys (Ref 6). Rare earth Silicon additions of 1.5 wt% or more have the additions have been suggested to have a multi- effect of forming a continuous amorphous sub- tude of beneficial effects, such as reducing the surface layer in iron-silicon and Fe-Cr-Si alloys growth kinetics of Cr2O3 scales, stabilizing that is relatively impervious to transport of ions. Cr2O3 scales at lower chromium levels, increas- The mechanism for the evolution of such a layer ing adhesion, and preventing spalling of the is as follows: (1) The more readily available oxide scale during thermal cycling. The expla- iron or chromium first forms a surface layer, nation for any of this does not seem clear, but and this causes an enrichment of silicon at the some hypotheses have been suggested. The ef- oxide-metal interface. (2) As sufficient silicon is fect on the growth kinetics could be because the enriched, the SiO2 layer is formed. It has been reactive element ions collect at grain boundaries reported that alloys with chromium content as and block fast path diffusion. low as 6 wt% and silicon content of 1.5 wt% The improved adherence could be because perform in terms of oxidation as well as com- these elements getter tramp elements such as mercial stainless steels. Also, an addition of 4 sulfur and suppress void formation at the inter- wt% Si to a Fe14wt%Cr14wt%Ni alloy resulted face. Furthermore, they might form so-called in a 200-fold reduction in weight gain at 900 °C oxide pegs at the interface (Ref 16). The pre- However, this SiO2 layer seems to promote cise role of the rare earth additions to Cr2O3 oxide spalling, especially in cyclic service. oxide protection and the mechanism by which Aluminum forms a very stable thin outer they are incorporated into the scale during the layer of Al2O3 that initially reduces the oxida- surface treatment processes remain unknown. tion rate. Alumina is among the most stable and An understanding of these fundamental issues defect-free oxides, giving it an extremely low would help to develop optimum alloy diffusion rate. In consequence, if sufficient alu- chemistries for selected high-temperature and - minum is present to maintain the protective alu- pressure applications and to further develop the mina scale, the aluminum-bearing alloys pro- surface infusion process. The lack of funda- vide the greatest oxidation resistance attainable mental understanding of how rare earths im- in engineering alloys. If, however, the alu- prove oxidation resistance has not stopped the minum content is not sufficient to force alumina development of several alloys that benefit from scale formation at the scale-metal interface, an the effect. 66 / Stainless Steels for Design Engineers

Oxidation Under Less-Oxidizing quasi-steady-state situation as shown in Fig. 9, Atmospheres the ﬂux can be written: dm NNX − S N S The two types of alloy oxidation behaviors, ==− O O = O JOOD DO (Eq 30) (a) a noble matrix metal with less-noble alloy- dt VXm VXm ing elements and (b) both matrix element and alloying elements are nonnoble, were men- Within the depth X, the oxygen solubility is tioned. When designing against oxidizing envi- such that Eq 2 is negative enough that Cr2O3 ronments, case b is perhaps the most relevant, forms. Beyond X, it is not. At the distance X, the and most of the discussion has been devoted to oxygen concentration is negligibly low com- this. However, during annealing for microstruc- pared to the surface composition, which is in tural control, steels are exposed to furnace gases equilibrium with the gas phase. The molar vol- at high temperatures that have relatively low ume Vm is used to obtain the ﬂux in units of oxygen or steam contents, for which case a will moles per square meter. Within the layer 0 < x < apply, that is the atmosphere does not cause iron X, all the chromium is assumed to be oxidized; (or nickel) to oxidize but chromium (and alu- therefore, the accumulated mass due to oxygen minum, silicon, molybdenum, etc.) does. addition: For simplicity, assume a binary system A-B of “noble” iron and “reactive” chromium. In o 3 0 NX NXυ Cr this case, depending on the concentration of m ==B 2 (Eq 31) the reactive element and atmosphere, the Vm Vm oxide of the reactive element can, in principle, form either on the surface (as has been dis- Differentiating Eq 31 with time, equating to cussed so far) or internally as discrete oxide Eq 3, and separating variables results in: particles in a metal matrix through oxygen diffusion into the metal. Both cases are shown NDS NDS XdX ==O O dt O O (Eq 32) schematically in Fig. 8. Let us discuss the υN O 3 conditions that promote one or the other of B N 0 2 Cr these by starting with a situation in which (a) no surface oxide exists and (b) the oxygen atmosphere is such that the solubility of oxygen Integrating Eq 32 from x = 0 to x = X results in within a distance X is enough to thermodynam- an expression for the internal oxidation depth X: ically render Cr O stable according to Eq 2 2 3 ⎛ ⎞ 12/ but none of the iron oxides. The derivation is 12/ ⎛ 22NDS ⎞ ⎜ NDS ⎟⎟ done in terms of both a generic system A-B X = ⎜ O O t⎟ = ⎜ O O t⎟ ⎝ υ 0 ⎠ 3 (Eq 33) causing an oxide BOν and for iron-chromium N B ⎜ 0 ⎟ ⎝ NCr ⎠ causing Cr2O3. Assume for a start that the DO 2 >> DCr, and thus while oxygen diffuses into the alloy, chromium does not counterdiffuse. The flux of oxygen inward into the metal is then the cause of increased mass. Assuming a

Fig. 8 Schematic of two cases in a less-oxidizing atmosphere. Fig. 9 Quasi-steady-state approximation of the moving (a) Adsorption of oxygen leading to internal oxidation boundary problem of internal oxidation. Counterdif- and (b) external oxidation as the B element migrates. fusion of B is assumed to be negligible Chapter 5: Oxidation / 67

This expression predicts a parabolic dependence of X with time, just as Eq 3 did for the external oxidation. For a more rigorous derivation (without assuming quasi steady state), refer to Appendix B in Ref 1. Now, let us see what causes this to transition into an external scale. It was assumed in the derivation of Eq 33 that counterdiffusion of chromium does not occur. When considering Eq 33, is clear that the rate of penetration of the internal oxidation front will 0 decrease with (a) increasing NB , (b) decreasing S NO , and (c) decreasing DO. If DCr was not negligible, there would be a gradual change in oxida- 0 S tion morphology if the ratio (NB DB)/(NO DO). Gradually, if the ratio were increased there would be a slowing of the inward penetration of the internal oxide front and an enrichment of Fig. 10 Temperature dependence of metal dusting of iron. Source: Ref 18 BOυ in the internally oxidized zone. Wagner developed a model (Ref 16), based on that at some trogen absorption, but if the oxygen is depleted point, when the volume fraction of BOυ versus volume metal in the internally oxidized zone before all surfaces are oxidized, the remaining reaches a critical value g*, there is a transition material can be rapidly nitrided by the residual, from internal to external oxidation; speciﬁcally, essentially pure, nitrogen atmosphere. this happens when:

12/ ⎡ π * ⎤ REFERENCES 0 g S DV N > ⎢ N OM⎥ (Eq 34) BO2ν DV ⎣ BOx⎦ 1. N. Birks, G.H. Meier, and F.S. Pettit, Intro- duction to the High-Temperature Oxidation If more than one reactive elements were pres- of Metals, 2nd ed., 2006, Cambridge Uni- ent (such as is the case in stainless steels in versity Press, New York which aluminum, silicon, molybdenum, nio- 2. P. Kofstad, High Temperature Corrosion, bium, etc., may be present), this will decrease 1988, Elsevier Applied Science, London the inward ﬂux of oxygen, and thus the transi- 3. K. Hauffe, Oxidation of Metals, Plenum tion to external may occur at a lower solute Press, New York, 1965 0 (CCr) concentration than what is predicted by 4. F.D. Richardson and J.H.E. Jeffes, J. Iron Eq 34. Steel Inst., Vol 160, 1948, p 261 Metal Dusting. Under reducing conditions, 5. C. Wagner and K. Grünewald, Z. Phys. in products of combustion atmospheres, oxida- Chem., Vol 40B, 1938, p 455 tion and carburization may occur simultane- 6. J.H. Park, W.E. King, N.L. Peterson, and ously and at a higher rate than exhibited in pure S.J. Rothman, The Effect of Reactive Ele- oxidation. Under even more reducing condi- ment on Self-Diffusion in Cr2O3, Norman tions, the condition called metal dusting may L. Peterson Memorial Symposium, Oxida- occur. Metal dusting is often characterized by tion of Metals and Associated Mass Trans- the generation of large, smooth pits that look as port, edited by M.A. Dayananda, S.I. Roth- if metal had been scooped from the surface. The man, and W.E. King, AIME, Warrendale, underlying phenomenon is the formation of PA, 1998, p 103–107 metal carbides, which manifests itself as the 7. C. Wagner and K. Grünewald, Z. Phys. breakup of bulk metal to metal powder. This oc- Chem., Vol 40B, 1938, p 455 curs at temperatures at which the carbide is 8. D. Caplan and G.I. Sproule, Effect of Oxide most stable (Fig. 10). Grain Structure on the High Temperature During oxidation in air, if large surface area is Oxidation of Cr, Oxid. Met., Vol 9, 1975, present under conditions of restricted air supply, p 459–472 oxygen can be depleted to the point that oxida- 9. B.B. Ebbinghaus, Combust. Flame, Vol 93, tion essentially ceases. Oxide ﬁlms inhibit ni- 1993, p 119–137 68 / Stainless Steels for Design Engineers

10. K. Hilpert et al., JECS, Vol 143/11, 1996, p of Metals and Associated Mass Transport, 3642–3647 edited by M.A. Dayananda, S.I. Rothman, 11. N.B. Pilling and R.E. Bedworth, J. Inst. and W.E. King, AIME, Warrendale, PA, Met., Vol 29, 1923, p 529 1998, p 323–340 12. J.K. Tien and J.M. Davidson, Oxide Spalla- 14. E.J. Felten, J. Electrochem. Soc., Vol 108, tion Mechanisms, Stress Effects and the Ox- 1961, p 490 idation of Metals, ed. J.V. Cathcart, AIME, 15. C. Wagner, J. Electrochem. Soc., Vol 103, New York, 1975, p 200 1956, p 571 13. J. Rawers, Understanding the Oxidation 16. C.M. Chun, J.D. Mumford, and T.A. Rama- Protection of Fe-Cr-Si Alloys, Norman L. narayanan, J. Electrochem. Soc., Vol 147, Peterson Memorial Symposium, Oxidation 2000, p 3680 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 69-90 All rights reserved. DOI: 10.1361/ssde2008p069 www.asminternational.org

CHAPTER 6

Austenitic Stainless Steels

Summary because their greater thermal expansion coefficient tends to cause the protective oxide AUSTENITIC STAINLESS STEELS are the coating to spall. most common and familiar types of stainless 2. They can experience stress corrosion crack- steel. They are most easily recognized as non- ing (SCC) if used in an environment to which magnetic. They are extremely formable and they have insufficient corrosion resistance. weldable, and they can be successfully used 3. The fatigue endurance limit is only about from cryogenic temperatures to the red-hot tem- 30% of the tensile strength (vs. ~50 to 60% peratures of furnaces and jet engines. They con- for ferritic stainless steels). This, combined tain between about 16 and 25% chromium, and with their high thermal expansion coeffi- they can also contain nitrogen in solution, both cients, makes them especially susceptible to of which contribute to their high corrosion re- thermal fatigue. sistance. Were it not for the cost of the nickel However, the risks of these limitations can be that helps stabilize their austenitic structure, avoidable by taking proper precautions. these alloys would be used even more widely.

Introduction Alloy Families in Perspective

Austenitic stainless steels have many advan- The fundamental criterion in the selection of tages from a metallurgical point of view. They a stainless steel is generally that it can survive can be made soft enough (i.e., with a yield with virtually no corrosion in the environment strength about 200 MPa) to be easily formed by in which it is to be used. Good engineering the same tools that work with carbon steel, but practice sometimes requires that materials be they can also be made incredibly strong by cold selected for sufficient, but ﬁnite, service life. work, up to yield strengths of over 2000 MPa This is especially true for high-temperature (290 ksi). Their austenitic (fcc, face-centered service, for which creep and oxidation lead to cubic) structure is very tough and ductile down limited life for all materials. The choice among to absolute zero. They also do not lose their the stainless steels that can be used in that envi- strength at elevated temperatures as rapidly as ronment is then based on the alloy from which ferritic (bcc, body-centered cubic) iron base al- the component can be produced at the lowest loys. The least corrosion-resistant versions can cost, including maintenance, over the intended withstand the normal corrosive attack of the service life. The ferritic stainless steels are less everyday environment that people experience, expensive for the same corrosion resistance but while the most corrosion-resistant grades can sometimes are found lacking because of: even withstand boiling seawater. ¥ If these alloys were to have any relative Lack of toughness, as is the case at subambi- weaknesses, they would be: ent temperatures or in thicknesses greater than about 1.5 mm 1. Austenitic stainless steels are less resistant ¥ Lack of great ductility, speciﬁcally if more to cyclic oxidation than are ferritic grades than about 30% elongation is needed 70 / Stainless Steels for Design Engineers

¥ Susceptibility to high-temperature embrit- The austenitic alloys can have compositions tling phases when moderately alloyed anywhere in the portion of the Delong diagram labeled austenite shown in Fig. 1 (Ref 1). This The less-expensive martensitic grades are diagram was designed to show which phases are used instead of austenitic when high strength present in alloys in the as-solidified condition, and hardness are better achieved by heat treat- such as found in welds. Thus it also applies to ing rather than by cold work, and mechanical castings and continuously cast products. As a properties are more important than corrosion re- practical matter of castability, the composition sistance. This is also the case for the more ex- of most commercial alloys falls along the zone pensive PH grades, which can achieve corrosion of several percent ferrite as cast. The salient fea- resistance only matching the least corrosion re- ture of austenitic alloys is that as chromium and sistant of the austenitic alloys. molybdenum are increased to increase specific Duplex grades match austenitic grades in cor- properties, usually corrosion resistance, nickel rosion resistance and have higher strength in the or other austenite stabilizers must be added if annealed condition but present the designer the austenitic structure is to be preserved. with challenges with regard to embrittling The traditional way of displaying the phases that can form with prolonged exposure austenitic stainless steels is to present 302 as a to elevated temperatures and only moderate base. Figure 2 shows one such diagram. Dia- ductility like the ferritic alloys. grams such as these treat alloys as an evolution- So, the austenitic grades are the most com- ary family tree and subtly mislead. Many alloys monly used grades of stainless mainly because, were pushed toward obsolescence because of in many instances, they provide very predictable advances in processing. For instance, 321 was levels of corrosion resistance with excellent me- developed as an alloy in which the detrimental chanical properties. Using them wisely can save effects of carbon were negated by addition of ti- the design engineer significant costs in his or her tanium. The widespread adoption of the argon product. They are a user-friendly metal alloy oxygen decarburization (AOD) in the 1970s with life-cycle cost of fully manufactured prod- made this alloy unnecessary, except for special ucts lower than many other materials. circumstances, since carbon could be cheaply

Fig. 1 Schaeffler-Delong stainless steels constitution diagram. Adapted from Ref 1, 2 Chapter 6: Austenitic Stainless Steels / 71

Fig. 2 The austenitic stainless family. Source: Ref 3 removed routinely. Likewise, 302 gave way to nickel equivalents (manganese, nitrogen, carbon, the lower-carbon 304, for which the even lower- etc.) must also be added in matching amounts, carbon 304L is commonly substituted and du- austenite stability is also increased. If molybde- ally certified to qualify as either grade. While num, a chromium equivalent, is added, corro- low carbon prevents sensitization, stabilized sion resistance but not oxidation resistance is grades may still be preferred for special applica- enhanced. And, if nitrogen is the austenite stabi- tions such as type 321 in aerospace and type lizer added to balance increases chromium or 347 in refinery service. Similar inertia keeps the molybdenum, then corrosion resistance is also higher-nickel 300 series as the de facto standard increased. With small exceptions, that is the when the more cost-efficient high-manganese rationale of austenitic grade design. Silicon 200 series is the logical basic grade. The rele- is used as an alloy to promote oxidation resist- vant types of austenitic alloys can nonetheless ance and resistance to corrosion by oxidizing be rationalized with this diagram. acids. Copper is used to promote resistance to As chromium is added, oxidation resistance sulfuric acid. Rare earths make a more stably and corrosion resistance increase. Because oxidation-resisting scale. Niobium increases 72 / Stainless Steels for Design Engineers creep resistance. Sulfur and selenium increase Lean Alloys machinability. In this chapter, austenitic alloys are classified Lean austenitic alloys constitute the largest into three groups: portion of all stainless steel produced. These are ¥ Lean alloys, such as 201 and 301, are gener- principally 201, 301, and 304. Alloys with less ally used when high strength or high forma- than 20% chromium and 14% nickel fall into bility is the main objective since the lower, this unofficial category. Since they are stainless, yet tailorable, austenite stability of these al- it is generally taken for granted that these alloys loys gives a great range of work-hardening will not corrode, and these alloys have suffi- rates and great ductility. Richer alloys, such cient corrosion resistance to be used in any in- as 305, with minimal work hardening are the door or outdoor environment, excluding coastal. high-alloy, lowest work-hardening rate These grades are easily weldable and formable grade for this purpose. The general-purpose and can be given many attractive and useful sur- alloy 304 is within this group. face finishes, so they are very much general- ¥ Chromium nickel alloys when the objective purpose alloys. Table 1 lists some typical com- is high temperature oxidation resistance. positions of the most commonly used lean This can be enhanced by silicon and rare austenitic alloys. These typical compositions earths. If the application requires high-tem- vary with end use, raw material cost factors, and perature strength, carbon, nitrogen, niobium, the preference of a given manufacturer. The and molybdenum can be added. 302B, 309, compositions of standard alloys are often fine- 310, 347, and various proprietary alloys are tuned to the intended end use. In this table, the found in this group. word drawing indicates higher nickel for lower ¥ Chromium, molybdenum, nickel, and nitro- work hardening, while tubing indicates alloys gen alloys when corrosion resistance is the with higher sulfur to facilitate gas tungsten arc main objective. Alloys such as silicon and welding (GTAW) penetration. Tensile indicates copper are added for resistance to specific lower alloy levels to increase the work-harden- environments. This group includes 316L, ing rate for material that is intended to be used 317L, 904L, and many proprietary grades. in the cold-worked, high-strength condition. 316L is included in its most common tubing end Wrought alloys generally have cast counter- use chemistry even though it is a corrosion-re- parts that differ primarily in silicon content. sisting alloy because it is so pervasively used as Versions that require enhanced machinability a service center sheet item. have a high content of controlled inclusions, The main difference among the lean sulfides, or oxysulfides, which improve machin- austenitic alloys lies in their work-hardening ability at the expense of corrosion resistance. rate: the leaner the alloy, the lower the austenite Carbon is kept below 0.03% and designated an stability. As unstable alloys are deformed, they L grade when prolonged heating due to multi- transform from austenite to the much harder pass welding of heavy section (greater than martensite. This increases the work-hardening about 2 mm) or when welds requiring a post- rate and enhances ductility since it delays the weld stress relief are anticipated. onset of necking since greater localized

Table 1 Typical compositions of the most commonly used lean austenitic alloys

Alloy Designation C N Cr Ni Mo Mn Si Other Other Other 201 S20100 0.08 0.07 16.3 4.5 0.2 7.1 0.45 0.001 S 0.03 P 0.2 Cu 201 drawing S220100 0.08 0.07 16.9 5.4 0.02 7.1 0.5 0.001 S 0.30 P 0.6 Cu 201LN S20153 0.02 0.13 16.3 4.5 0.2 7.1 0.45 0.001 S 0.03 P 0.5 Cu 301 tensile S30100 0.08 0.4 16.6 6.8 0.2 1.0 0.45 0.001 S 0.03 P 0.3 Cu 301 drawing S30100 0.08 0.04 17.4 7.4 0.02 1.7 0.45 0.007 S 0.03 P 0.6 Cu 303 S30300 ...... 304 S30400 0.05 0.05 18.3 8.1 0.3 1.8 0.45 0.001 S 0.03 P 0.3 Cu 304 drawing S30400 0.05 0.04 18.4 8.6 0.3 1.8 0.45 0.001 S 0.03 P 0.3 Cu 304 extra drawing S30400 0.06 0.04 18.3 9.1 0.3 1.8 0.45 0.001 S 0.030 P 0.4 Cu 304L tubing S30403 0.02 0.09 18.3 8.1 0.3 1.8 0.45 0.013 S 0.030 P 0.4 Ci 305 S30500 0.05 0.02 18.8 12.1 0.2 0.8 0.60 0.001 S 0.02 P 0.2 Cu 321 S32100 0.05 0.01 17.7 9.1 0.03 1.0 0.45 0.001 S 0.03 P 0.4 Ti 316L S31603 0.02 0.0 16.4 10.5 2.1 1.8 0.50 0.010 S 0.03 P 0.4 Cu Chapter 6: Austenitic Stainless Steels / 73

deformation is more than offset by greater lo- such as occurs when they are sensitized or when calized strain hardening. solute segregation occurs, as from welding, then These grades are best viewed as a continuum the equation applies on a microscopic scale. with a lower boundary at 16%Cr-6%Ni and an Sensitized zones (i.e., the regions near grain upper boundary at 19%Cr-12%Ni. This repre- boundaries where chromium carbides have presents the range from minimum to maximum cipitated) will have a much higher tendency to austenite stability. Since that is the main distinc- transform to martensite. Figures 3(a) and (b) tion within this grade family, let us examine its show the changes in phase structure as a func- basis. tion of composition over ranges that encompass Martensite and Austenite. Stability. The these alloys. formation of martensite at room temperature Martensite can be present in two different may be thermodynamically possible, but the forms. The α′-form is the bcc magnetic form, driving force for its formation may be insuffi- while ε is a nonmagnetic, hcp (hexagonal close- cient for it to form spontaneously. However, packed) version. The formation of ε versus α′ is since martensite forms from unstable austenite related to the stacking fault energy of the alloy, by a diffusionless shear mechanism, it can occur which is given by (Ref 6): if that shear is provided mechanically by exter- 300 -2 0 nal forces. This happens during deformation, Y SF (mJ m ) = Y SF + 1.59Ni Ð 1.34Mn and the degree to which it occurs varies with + 0.06Mn2 Ð 1.75Cr + 0.01Cr2 composition according to (Ref 4): + 15.21Mo Ð 5.59Si ° Ð60.69(C + 1.2N)1/2 Md30 ( C) = 551 Ð 462(%C + %N) Ð 9.2(%Si) Ð 8.1(%Mn) Ð 13.7(%Cr) + 26.27(C + 1.2N) 1/2 Ð 29(%Ni + Cu) Ð 18.5(%Mo) (Cr + Mn + Mo) 1/2 Ð 68(%Nb) Ð 1.42 (GS Ð 8) (Eq 1) + 0.61[Ni¥(Cr + Mn)] (Eq 2) This is the temperature at which 50% of the Epsilon martensite formation is favored in austenite transforms to martensite with 30% alloys of lower stacking fault energy. The fcc true strain (Ref 5). It should be noted that even structures deform by slip between (111) elements that are chromium equivalents in pro- planes. Viewed from these planes, the structure moting ferrite are austenite stabilizers in that is a series of ABCABC atom arrangements. they impede martensite formation. This temper- Slip between planes can result in an ature is the common index of austenite stability. ABCA/CAB structure. This so-called stacking This regression analysis was generated for ho- fault generates an hcp structure. With lower mogeneous alloys. If alloys are inhomogeneous, stacking fault energies, these are more readily

Fig. 3 (a) Iron-chromium phase diagram at 8% nickel; (b) iron-nickel phase diagram at 18% chromium 74 / Stainless Steels for Design Engineers

Fig. 4 Variation of martensite formation with temperature and true strain for 304. Source: Ref 7

formed, and ε predominates. The stacking fault and tensile strength, respectively, are reported can also be viewed as two partial dislocations (Ref 10) to follow the equations: with the material between them faulted. These =++ partial dislocations, when generated in abun- YS().[.(%)(%) MPa 15 4 4 4 23CN 32 dance, cannot readily slip past one another ++024.(%)Cr 09 .94(%Mo ) and thus pile up, increasing work-hardening ++13.(%)Si 12 .(%) V rates. ++029.(%).(%)WNb 26 As in carbon and alloy steels, the martensite +++17.(%)Ti 082 . (%) Al transformation can take place simply by cooling, but in the lean austenitic alloys the temper- + 016.(%Ferrite ) atures are well below ambient. The more stable + 046.(d −1112// ) (Eq 3) alloys do not transform even with cryogenic treatment. Figure 4 shows the variation of TS().[(%)(%) MPa =++15 4 29 35CN 55 martensite formation with temperature and true ++ + strain for 304. Martensite formed in these alloys 24.(%)Si 011( . (%Ni )12 . (% Mo ) is quite stable and does not revert until heated +++50.(%Nb ) 30 .(%) Ti 12 .(%) Al well above the temperatures (Fig. 5) at which it + 0014.(%Ferrite )+ 082 .(d −12/ ) (Eq 4) was formed. The carbon levels of austenitic stainless steels are always relatively low, so In each case, d is the grain diameter in mil- strain-induced martensite is self-tempering and limeters. not brittle. Another researcher (Ref 11) gave the rela- Martensite has been found to form in unstable tionships as: austenite due to the electrochemically induced supersaturation by hydrogen (Ref 9). Under YS()MP a =+120 210N + 0. 02 + 2 Mn + 2 Cr conditions of cathodic charging, superﬁcial lay- +++14Mo 10 Cu (. 6 115− 0.) 054δδ ers were found to transform to ε under condi- −12/ ++((.))735Nd + 02 (Eq 5) tions of intense hydrostatic compression. Dur- ing subsequent outgassing, α′ was found to form due to reversals in the stress state. Marten- TS =+470 600(.)N + 0 2 site thus formed is, of course, susceptible to +++14Mo 1. 5δ 8d −12/ (Eq 6) hydrogen embrittlement. Mechanical Properties. The tensile proper- Again, d is grain diameter in millimeters, and ties in the annealed state not surprisingly relate δ is percent ferrite. The claimed accuracy for well to composition. The 0.2% yield strength the latter set of equations is 20 MPa and is said Chapter 6: Austenitic Stainless Steels / 75

Fig. 6 Variation of impact strength with temperature for (a) austenitic, (b) duplex, and (c) ferritic stainless steels form more stable dislocation arrays that break loose at a higher and distinct yield point. The tensile properties of austenitic stainless steels with unstable austenite, that is, those with Md30 temperatures (Eq 1) near room temperature, are very strain rate dependent. This is sim- Fig. 5 Reversion of martensite formed by cold work. Source: Ref 8 ply due to the inﬂuence of adiabatic heating during testing increasing the stability of the to apply to both austenitics and duplex stainless austenite. Tests run under constant temperature steels, but clearly the tensile strength relation- conditions, either by slow strain rates or use of ship must break down for leaner alloys, such as heat sinks, produce lower tensile strengths. 301, in which tensile strength increases with de- Thus, reported tensile strengths should not be creasing alloy content because of the effect of taken as an absolute value but a result that can increasing alloying causing less transformation be signiﬁcantly changed by changes in testing to martensite, which inarguably produces higher procedure, even with accepted norms and tensile strengths in austenitic stainless steels. standards. Equation 3 must also be favored over Eq 5 in Highly cold-worked austenitic stainless that it accounts for carbon explicitly. steels are often used for their robust mechanical One other hardening mechanism is possible properties. Few metallic materials can match in austenitic stainless steels, and that is precipi- the very high strengths they can achieve. Very tation hardening. Most precipitation-hardening lean 301 can be cold worked to yield strengths stainless steels are unstable austenite, which is on the order of 2000 MPa because of its unsta- transformed to martensite before the precipita- ble austenite transforming to martensite. When tion hardening takes place. One commercial cold worked to lower degrees, it can provide alloy, A-286, is entirely austenitic and employs very high strength while keeping impressive the precipitation within the austenite matrix of ductility. Ni3 (titanium, aluminum) for strengthening. Austenitic stainless steels have exceptional This is dealt with in a separate section. toughness. The ambient temperature impact Austenitic stainless steels do not have a clear strength of austenitic stainless steels is quite yield point but can begin to deform at as little as high. This is not surprising in view of their high 40% of the yield strength. As a rule of thumb, tensile strengths and high elongations. What is behavior at less than half the yield strength is most remarkable is the absence of a transition considered fully elastic and stresses below two- temperature, which characterizes ferritic and thirds of the yield strength produce negligible martensitic materials. Figure 6 shows impact plastic deformation. This quasi-elastic behavior strength of the various stainless steel types ver- is a consequence of the many active slip sys- sus temperature. This again is due to the multi- tems in the fcc structure. Even highly cold- plicity of slip systems in the fcc structure and worked material exhibits this phenomenon, al- the fact that they do not require thermal activa- though stress-relieving cold-worked material tion. This makes the austenitic stainless steels, will cause dislocations to “lock in place” and especially the 200 series, the optimal cryogenic 76 / Stainless Steels for Design Engineers

material, surpassing the 9% nickel martensitic tion in austenite, and diffusion rates are suffi- steels in cost, toughness, and, of course, corro- cient for carbon and chromium to segregate into sion resistance. precipitates. The solubility of carbon in austen- Precipitation of Carbides and Nitrides. ite is over 0.4% at solidiﬁcation but decreases Carbon is normally considered as an undesir- greatly with decreasing temperature. The solu- able impurity in austenitic stainless steel. While bility is given by (Ref 12): it stabilizes the austenite structure, it has a great 6272 thermodynamic affinity for chromium. Because log (C ppm ) =−7771 (Eq 7) T ()¡K of this affinity, chromium carbides, M23C6, form whenever carbon reaches levels of supersatura- The equilibrium diagram for carbon in a basic 18%Cr10%Ni alloy is shown in Fig. 7. At room temperature, very little carbon is soluble in austenite; even the 0.03% of L grades is mostly in a supersaturated solution. The absence of carbides in austenitic stainless is due to the slow diffusion of carbon and the even slower diffusion of chromium in austenite. At a carbon level of 0.06%, which is found in most 304, supersaturation is reached below about 850 °C. Below this temperature, supersaturation increases exponentially, while diffusion decreases exponentially. This results in precipitation rates that vary with temperature and carbon level as shown in Fig. 8. At these temperatures, grain boundary diffusion is much more rapid than bulk diffusion, and grain boundaries provide excellent nucleation sites, so precipitation occurs along grain boundaries. And, because carbon diffuses several orders of magnitude more rapidly than chromium, carbon diffuses to and combines with chromium essentially in situ, depleting the Fig. 7 Carbon solubility in 18–10 austenitic stainless. Source: Ref 13 grain boundaries of chromium in solution.

Fig. 8 The precipitation rates for Cr23C6 as a function of carbon content Chapter 6: Austenitic Stainless Steels / 77

Fig. 10 Variation of carbide precipitation locus with time. Source: Ref 16 Much longer term heat treatment is required to eliminate these depleted zones by rehomogenization of slowly diffusing chromium than the short time required to form them. This is very evident for carbides, but also true for oxides. Underneath chromium-rich oxide scales is a layer depleted in chromium and lower in corrosion resistance. This is why not only scale from welding must be removed, but also the underlying chromium-depleted zone. Other precipitation processes that give rise to chromium depletion are α and χ and the solid- state precipitation of oxides, nitrides, and sulfides. Chromium precipitates that form in the liquid alloy do not cause depletion of chromium Fig. 9 Depletion of chromium from the austenite near grain boundaries due to chromium carbide precipitation. locally because no chromium gradients are set Source: Ref 14 up around them during precipitation as diffusion in the liquid is very rapid. Thus, primary Figure 9 shows that the local chromium de- carbides, oxides, and sulfides are not per se pletion is such that the chromium level can be- harmful to corrosion resistance. But, if the same come low enough that it has not even enough to compounds form and grow in the solid state, be stainless and certainly much lower corrosion chromium depletion occurs (Ref 15). resistance than the surrounding area. This zone, Alloying elements can have a major influence because it is lower in chromium, also has very on carbide precipitation by their influence on the unstable austenite and is quite prone to marten- solubility of carbon in austenite. Molybdenum site formation. Figure 10 shows how the locus and nickel accelerate the precipitation by dimin- of precipitation changes with time and tempera- ishing the solubility of carbon. Chromium and ture. Carbon relatively far from grain bound- nitrogen increase the solubility of carbon and aries in the interior of grains remains in super- thus retard and diminish precipitation. Nitrogen saturation until much longer times and much is especially useful in this regard (Fig. 11). greater supersaturation since bulk diffusion is Increasing austenite grain size accelerates required for the nucleation and growth of these precipitation, as does cold work, especially in precipitates. the interior of grains, where diffusion is en- The key observation is that any solid-state hanced by increased defect density. precipitation of a chromium-rich precipitate Nitrogen is much more soluble than carbon will necessarily cause local chromium depletion and does not give rise to sensitization phenom- and a resulting loss of corrosion resistance. ena as does carbon even though Cr2N can be a 78 / Stainless Steels for Design Engineers

Fig. 11 Delay in carbide precipitation induced by nitrogen level. Source: Ref 17 stable phase when the solubility limit is exceeded. The solubility is over 0.15% in austenite, so its precipitation seldom has the possibility of occurring, but it does become an issue in Fig. 12 Variation of hardness with depth and therefore car- ferritic stainless steels in this regard, for which bon content in colossal supersaturation solubility is much lower. Manganese and chromium increase the solubility of nitrogen in to ignore the titanium-consuming capacity of austenite. oxygen and sulfur unless they have been mini- Stabilization. Before carbon was easily low- mized by refining, which can be done quite ered to harmless levels, it was found that adding readily. more powerful carbide formers than chromium Even if sufficient titanium or niobium is pres- could preclude the precipitation of chromium ent to combine with all carbon, kinetic consid- carbides. Titanium and niobium are the most erations may result in that not occurring. High useful elements in this regard. They form car- temperatures, such as encountered in welding, bides with solubility that follows the following dissociate carbides. If quenched from this state, equation type: carbon can be free to form Cr23C6 if it is reheated to temperatures above 500 °C. log [MX ] [ ] =+ A −H (Eq 8) Carbon has always been considered totally RT undesirable from a corrosion point of view because of its tendency to form chromium car- For titanium carbide and niobium carbide, the bides. Recently, however, new processes have respective solubilities are: been developed to supersaturate carbon in 6780 austenite below the temperatures at which it has log [Ti ] [ C ]=−297 . (Eq 9) sufficient mobility to form carbides. This so- T called colossally supersaturated austenite results in very high hardness (Fig. 12) and corro- 9350 log [Nb ] [ C ]=−455 . (Eq 10) sion resistance over limited depths. From this, T however, we can see that carbon, like nitrogen, is actually beneficial to corrosion resistance in Oxides and sulfides are more energetically fa- solid solution, although this is not observed at vorable than are carbides and nitrides of these normal concentrations. It is possible to see that metals. Thus, any additions made to form car- if it could be kept in solution it would be appro- bides must be sufficient to account for the prior priate to give it a factor of around 10 in the pit- formation of these compounds. Nitrogen also ting resistance equivalent number (PREN) competes with carbon for available titanium or equation: niobium. Thus, for successful gettering of all carbon, there must be sufficient titanium or nio- PREN=+%.(%) Cr33 Mo bium to combine stoichiometrically with all ++ these species present. 30(%NC ) 10 (% ) (Eq 11) This requires in rough terms that titanium exceed four times the carbon plus nitrogen, or that This is consistent with the similar thermody- niobium exceed eight times, after accounting namic interaction coefficients that carbon and for the oxygen and sulfur. It would be a mistake nitrogen share with regard to chromium. Chapter 6: Austenitic Stainless Steels / 79

High-Temperature Alloys

The austenitic stainless steels can have an exceptional combination of strength and corrosion resistance at temperatures above 500 °C. They are often called on to resist attack by oxygen, sulfur, carburizing, nitriding, halogens, and molten salts. Austenitic stainless steels are the most creep resistant of the stainless steels. Alloying with carbon, nitrogen, and niobium produces the greatest strength at elevated temperatures. Refer to the properties database for comparisons among the grades. This discussion concentrates on their resistance to high- temperature environments, which is their salient characteristic. Oxidation Resistance. Resistance to oxidation comes from the protective Cr2O3 scale that forms on the surface of the material. Above about 18% chromium, a continuous scale forms. The scale acts as a barrier to oxygen and greatly slows further oxidation of metal below the scale. Below the composition at which complete Cr O coverage occurs, the ﬁlm will also Fig. 13 Variation of parabolic oxidation rate with chromium 2 3 level and temperature. Source: Ref 18 contain the less-protective spinel FeCr2O4. The the less-protective FeCr O , and the scale Cr2O3 scale is more protective because it better 2 4 restricts the diffusion of oxygen to the interface growth rate will increase beyond the parabolic between the scale and the base metal, which is relationship, leading to breakaway oxidation. where the oxidation reaction occurs. As the The breakaway temperature increases with in- oxide grows, the path to the interface lengthens, creasing chromium level. and the rate of oxidation slows. This generates The austenitic alloys beneﬁt over the ferritic the parabolic-type oxide growth rate that char- alloys from the presence of nickel. For a given acterizes these alloys. The rate of oxide growth chromium level, oxidation rates decrease with is expressed simply as: increasing nickel content. Figures 14, 15, and 16 display this relationship. The optimal range Rkt= (Eq 12) for the iron base stainless steels, shown in Fig. 14, is reached by the commercial alloy 310, in which the rate is in units of mass gained per with 25Cr-20Ni composition. unit of surface area and time. This rate is a Alloying can alter the oxidation-resisting per- strong function primarily of chromium level, as formance of the austenitic stainless steels. Some can be seen in Fig. 13. elements form more protective oxide layers

The rate increases exponentially with temper- than Cr2O3. Aluminum and silicon are most use- ature since diffusion governs this phenomenon. ful in this regard. Aluminum forms a layer of

The rate drops dramatically as chromium Al2O3 that is more restrictive to oxygen diffu- reaches the concentration necessary to generate sion than is Cr2O3, as does silicon through the the protective Cr2O3 layer. Above this suffi- formation of SiO2. The alloys 302B, 153MA, ciency level, further increases in chromium are and 253MA all use elevated silicon levels. Alu- not as beneficial; they mainly provide a reser- minum’s strong ferrite-promoting tendency re- voir of chromium to re-form the Cr2O3. stricts its utility in austenitic grades, however. As long as chromium can diffuse to the inter- While the gains from using under 3% silicon face at a sufficient rate to satisfy the incoming are impressive, rare earths can yield even flux of oxygen, the parabolic rate is maintained. greater benefits from mere trace additions. If there is insufficient chromium flux, then the 153MA (UNS S30415) is a variation on 304 oxygen penetrates beyond the interface and using silicon and cerium. Cerium appears to act forms Cr2O3 in situ. The oxide will change to at the metal-scale interface such that the oxides 80 / Stainless Steels for Design Engineers

Fig. 14 Inﬂuence of nickel on oxidation of iron-chromium alloys. Source: Ref 19 Fig. 16 Corrosion rates for various stainless steels and nickel base alloys. Source: Ref 20

Fig. 15 Isooxidation curves. Source: Ref 20 formed are thinner, tougher, more adherent, and ritic alloys, they stress their scale more during more protective, although there is no consensus thermal cycling. This can fracture the scale, on the mechanism. Figure 17 shows the im- causing spalling and rapid subsequent oxidation provement quantitatively (Ref 21). of the underlying metal. This serious perform- Because austenitic stainless steels have a ance ﬂaw also is remedied by rare earths, as greater thermal expansion coefficient than fer- shown in Fig. 18 (Ref 21). Chapter 6: Austenitic Stainless Steels / 81

Fig. 17 Comparison of rare earth-alloyed stainless alloys to conventional stainless alloys, 4833 = 309S, 4845 = 310S. Source: Ref 21

The formation of an oxide on stainless steel should be understood to imply de facto the depletion of chromium from the underlying metal surface. Whether the scale is formed in service, during heat treating, or during welding the surface, once the oxide is removed, there will be less chromium than the bulk alloy, often by a very significant amount, and therefore the corrosion resistant will be less. This is why welds must have not only their heat tint removed, but also the underlying metal which is depleted in chromium, to a depth on the order of 10 μ (Ref 24). Fig. 18 310S (4845) versus rare earth-modified 253MA in Microstructure can also affect oxidation re- cyclic oxidation. Source: Ref 21 sistance. As a generalization, it can be said that changes that promote the diffusion of chromium While the mechanism by which rare earths assist in the formation of a protective scale and make the scale tougher and more adherent are improve oxidation resistance. Thus, cold work vague, their effect in making austenitic alloys and finer grain size are positive factors via their much better at resisting high-temperature oxida- enhancement of chromium diffusion. tion, especially cyclic oxidation, are undeniable. At the very highest temperatures, Cr2O3 can be Alloying elements can also be detrimental to further oxidized to CrO3, which has significant oxidation resistance. Manganese, an even more vapor pressure above about 1100 °C. The com- potent oxide former than chromium, forms a positions of some of the main high-temperature manganese-chromium spinel that is less protec- austenitic alloys mentioned here are given in tive than the Cr2O3. Molybdenum and tungsten, Table 2. which are refractory metals and are beneficial to Other Environments. The most common ad- aqueous corrosion resistance, form volatile, low- dition species that aggravates high-temperature melting oxides (MoO3 and WO3) that promote oxidation is water vapor. At 10%, water vapor catastrophic oxidation (Ref 22, 23). Vanadium will increase oxidation by a factor of ten. It acts also forms an oxide, V O , which melts at 660 by increasing the porosity of the oxide scale and ° 2 5 C and can also cause catastrophic oxidation. by promoting formation of the volatile CrO2 82 / Stainless Steels for Design Engineers

(OH)2 species. As a rule of thumb, maximum In lean austenitic alloys used in high-temper- service temperatures should be reduced by 50 to ature, 600 to 1000 °C service, formation times 100 °C in the presence of steam. are relatively long, on the order of 100 h or Halogens can attack oxide scales and cause more. In richer alloys, such as 310, times can be their degradation or volatilization. as short as several hours. If the temperature at Carburization and nitriding are best prevented which the alloy is to be used is one in this tem- by an oxide layer that forms at very low oxygen perature range, then some σ is a foregone con- partial pressures as chromium and silicon con- clusion, and while σ will have little detrimental tents are increased. Austenitic alloys have no effect on short-term properties at these tempera- advantage over ferritic in this regard. tures, long-term properties such as creep, stress Intermetallic Phases. Transition elements rupture, and especially rupture ductility are de- may combine to form intermetallic phases in graded by σ. For alloys, σ is an even greater which the formula can vary from B4A to A4B. concern as these are prone to its formation and Table 3 lists the most common secondary can inadvertently form some during processing. phases encountered in austenitic stainless steels, If such alloys are intended for use near room i.e., apart from austenite and ferrite. Sigma for- temperature, then their toughness will be seri- mation is retarded by nitrogen, so alloys such as ously reduced by the brittle σ, which will form 153MA are less prone to it. Lower chromium ﬁrst at triple points and then throughout grain and higher nickel are beneﬁcial. Silicon and boundaries. Because of this morphology, just a aluminum are detrimental, as is molybdenum. few percent intermetallic phase can cause The most relevant is σ. It can contain as little toughness to decrease by an order of magnitude. as four iron to one chromium or molybdenum in High-Temperature Mechanical Properties. a tetragonal structure. Thus, it can exist in many Above about 500 °C, yield strength is less ap- conventional austenitic alloys. Other relevant propriate than creep strength in assessing the phases are χ and Laves. The greatest risk from adequacy of an austenitic stainless steel for these phases is the loss of room temperature structural purposes. The resistance of a material toughness, followed by some loss of corrosion to creep is generally measured by the creep rup- resistance. ture strength, which is the stress that causes

Table 2 Notable high-temperature austenitic alloys

Alloy Designation C N Cr Ni Mo Mn Si Other Max temp, °C 302B S30215 0.15 0.07 17.8 8.1 . . . 1.8 2.5 . . . 950 304H S30409 0.08 0.08 18.8 8.1 . . . 1.8 0.50 . . . 820 321H S32109 0.06 0.03 17.8 9.1 . . . 1.8 0.50 0.6 Ti 820 153MA S30415 0.05 0.15 18.5 9.5 . . . 0.6 1.3 0.05 Ce 1000 309S S30909 0.08 0.07 23.0 12.2 . . . 1.7 0.50 . . . 1040 309Si DIN 1.4828 0.08 0.07 19.8 11.1 . . . 1.8 2.0 . . . 1040 253MA S30815 0.08 0.17 21.0 10.5 . . . 0.6 1.5 1.0 Al, 0.05 Ce 1100 310S S31008 0.05 0.03 24.6 19.2 . . . 1.6 0.60 . . . 1090 353MA S35315 0.05 0.15 25.0 35.0 . . . 1.5 . . . 0.05 Ce 1200+ 330 S33000 0.06 . . . 18.0 35.0 . . . 1.7 0.90 . . . 1200 332Mo S35125 0.04 0.04 21.0 34.5 2.4 1.1 0.40 0.40 Nb 1200

Table 3 Secondary phases in austenitic stainless steel

Precipitate Structure Parameter, (Å) Composition NbC fcc(a) a = 4.47 NbC NbN fcc a = 4.40 NbN TiC fcc a = 4.33 TiC TiN fcc a = 4.24 TiN Z-phase Tetragonal a = 3.037, c = 7.391 CrNbN

M23C6 fcc a = 10.57Ð10.68 Cr16Fe5 Mo2C (e.g.)

M6C Diamond cubic a = 10.62Ð11.28 (FeCr)21Mo3 C; Fe3Nb3C; M5SiC σ-phase Tetragonal a = 8.80, c = 4.54 Fe,Ni,Cr,Mo

Laves phase Hexagonal a = 4.73, c = 7.72 Fe2Mo, Fe2Nb χ -phase bcc(b) a = 8.807Ð8.878 Fe36Cr12 Mo10

G-phase fcc a = 11.2 Ni16Nb6 Si7, Ni16Ti6 Si7 (a) fcc, face-centered cubic; (b) body-centered cubic Chapter 6: Austenitic Stainless Steels / 83

rupture after 10,000 or 100,000 h. If deforma- preferred. Substitutional elements have limited tion is a greater concern, however, the creep de- effect, but interstitial solid solution elements, formation strength, that is, the stress that results such as carbon and nitrogen, are quite useful. in a strain of 1% after 10,000 or 100,000 h, can Nitrogen is the better addition in this regard, be used as a basis for design. plus it has the collateral beneﬁt of strongly Cold work and precipitates tend to be ineffec- retarding intermetallic phase precipitation. tive strengtheners at temperatures that produce Figures 19 to 22 compare mechanical properties solution annealing and precipitate coarsening of the major high-temperature austenitic alloys (overaging). Thus, solid solution hardening is (Ref 25).

Fig. 19 Charpy V toughness after 200 hr aging

Fig. 20 Relative 100,000-h creep strength 84 / Stainless Steels for Design Engineers

Fig. 21 100,000-h creep rupture strength

Fig. 22 High-temperature, short-time yield strength

Corrosion-Resistant Austenitic Alloys temperature, thus giving the common perception that they are superior in corrosion resistance. Stainless steels are almost always chosen at The main advantage austenitic alloys have is least in part for their corrosion resistance. In their ability to utilize the powerful and inexpen- normal atmospheric conditions, alloys with sive alloying element nitrogen. That is the key more than 10.5% chromium do not rust. aspect of the more modern austenitic stainless Austenitic alloys require higher levels than steels that have come into use in the last two this to stabilize the austenitic structure at room decades. Chapter 6: Austenitic Stainless Steels / 85

The ion that is most aggressive against stain- The main factors in the resistance of austenitic less steel is one of the most pervasive in our alloys to pitting attack are generally given by: environment. The chloride ion is found in abundance over the entire earth. It is, of PREN = %Cr + 3.3(%Mo) + 30(%N) (Eq 13) course, in seawater, but also in the rain, on roads, in food, and even in our bodies. Chlo- Pitting resistance is measured by ASTM G 48 rides destabilize the passive film. If the condi- (practice C) in which the lowest temperature at tions of chloride concentration, temperature, which pitting occurs in a 6% FeCl3 is measured. and acidity are sufficiently aggressive to break This is the critical pitting temperature, CPT. The down the film, then active corrosion ensues. If relationship between PREN and CPT is shown this is highly localized because of a local in Fig. 23 (Ref 26). weakness in the passive film, then pitting oc- The function of chromium in the passive film curs. Such a pit may be unstable and repassi- is intuitively clear. As the chromium content of vate, or it may grow without limit. Other an alloy increases, the ready reservoir of halides have the same effect, but they are less chromium to form the chromium-rich layer is fa- ubiquitous. cilitated. The roles of molybdenum and nitrogen Because of the specific virulence of the chlo- are subtler and are still subject to controversy. It ride ion and because of its universal presence, is the subject of much research, which has been corrosion-resistant austenitic stainless steels all summarized in reviews. The obvious paradox is look like they were designed to resist chloride- how can elements that are not active in the pas- pitting attack. Pitting in stainless steels is in sive film be so effective in maintaining its in- most instances the threshold level of corrosion. tegrity. We do know that the essential chromium Crevice corrosion is, however, more severe and in the matrix of stainless steels is quite reactive usually design limiting vis-à-vis corrosion. and will form compounds with carbon, oxygen, Crevices can exist not only in deliberate joints sulfur, and other transition elements. When it but also under environmental deposits, paint does, it is no longer effective as a passive film films, weld splatter, etc. There are other envi- former. The regions from which the chromium ronments in which the resistance follows differ- diffused to form the chromium-rich phase will ent rules, such as oxidizing acids, bases, and or- be poor in chromium unless subjected to a ganic acids, but these are best regarded as lengthy homogenizing anneal. Most theories of exceptions. pitting founder at the start because they assume

Fig. 23 Critical pitting temperature versus pitting resistance equivalent number (PREN); SUS 329J4L = S31260, YUS 270 = S31254. Source: Ref 26 86 / Stainless Steels for Design Engineers

a homogeneous passive film, which is an impos- The discovery that nitrogen was beneficial sible goal in reality. The chapter on corrosion against corrosion permitted a breakthrough in deals with this topic in more depth. alloy development by the 1980s. Nitrogen was The passive layer is extremely thin compared increased to around 0.20% from nominal levels to oxide layers. It is on the order of 1 to 10 nm of 0.05%. This was found to increase PREN by thick. Its formation does not cause chromium another five units, but more importantly, also depletion beneath it, as oxide layers do. As the suppressed σ formation to times that permitted alloy content of chromium and molybdenum in- thicker sections to be welded without embrittle- crease, the film is thinner, and the current den- ment. Research into the thermodynamics of sity required to form the film is correspondingly nitrogen in austenite showed that manganese in- reduced. creased the solubility of nitrogen appreciably. The corrosion-resistant austenitic stainless This permitted even higher levels of total alloy- steel grades range from 316 to the various high- ing to be achieved. This was exploited in the molybdenum, high-nitrogen alloys commercial- alloys UNS S34565 and S32654, which contain ized in the last ten years, the most notable of 3 to 6% manganese and about 0.50% nitrogen. which are listed in Table 4 with their typical The PRENs of these alloys are over 50, which analyses. gives them a critical pitting temperature around Early grades were based on alloying with 100 °C. Table 5 lists the performance of the var- chromium and molybdenum with sufficient ious popular corrosion-resistant austenitic stain- nickel added to preserve the austenitic structure. less steels. Each of these elements facilitates the formation The advances are quite appreciable and made of the passive film and reduces the corrosion stainless steel a viable material for many appli- rate in the active state. Further experimentation cations for which previously nickel base or tita- showed that molybdenum was not beneficial nium alloys had been required. under highly oxidizing conditions, but that sili- The obvious question in view of the success con was helpful under such conditions. Copper of the use of high manganese levels in conjunc- was beneficial against sulfuric acid. tion with high nitrogen levels in the most highly Alloy development came in stages. First, 317 alloyed austenitic stainless steels is when this was the most significant corrosion-resistant approach will be used for the medium-alloyed alloy. Then, more chromium and molybdenum austenitics to make alloys superior to 316, 317, were added, and the class of alloys known col- and 904 without the high levels of nickel and lectively as the 6%Mo alloys was commercial- molybdenum that render these alloys so expen- ized. Allegheny’s AL-6Xª is representative of sive. It does not take much imagination to envi- this group. With PRENs of around 40, these al- sion alloys equal to 904L in PREN with less loys were resistant to seawater at ambient but nickel and molybdenum than 316L that would not at elevated temperatures. This left a great be almost totally resistant to intermetallic phase deal wanting in corrosion resistance. These al- precipitation and have much greater resistance loys were also very difficult to process, at least to SCC because of higher austenite stability. in part because they rapidly formed brittle grain The same case could be made for a 317-equiva- boundary σ-phases in as little as several minutes lent alloy in corrosion resistance with less than at some temperatures. This limited chromium 7% nickel and 0.5% molybdenum, essentially a and molybdenum levels to a total of about 30%. 301 in alloy cost. In view of these trends in

Table 4 Typical compositions of corrosion-resistant austenitic stainless steels

Alloy Designation C N Cr Ni Mo Mn Si Other Other 316L S31603 0.02 0.03 16.4 10.5 2.1 1.8 0.5 ...... 316Ti S31635 0.02 0.03 16.4 10.5 2.1 1.8 0.5 0.40 . . . 317L S31703 0.02 0.06 18.4 12.5 3.1 1.7 0.5 ...... 317LM S31725 0.02 0.06 18.4 13.7 4.1 1.7 0.5 ...... 904L N80904 0.02 0.06 19.5 24.0 4.1 1.7 0.5 1.3 Cu . . . JS700 N08700 0.02 0.06 19.5 25.0 4.4 1.7 0.5 0.4 Cu 0.3 Nb 254SMO S31254 0.02 0.20 20.0 18.0 6.1 0.8 0.4 0.8 Cu . . . 4565 S34565 0.01 0.45 24.0 18.0 4.5 6.0 ...... 654SMO S32654 0.01 0.50 24.0 22.0 7.2 3.0 . . . 0.5 Cu . . . AL6-XN N08367 0.02 0.22 20.5 24.0 6.2 0.4 0.4 0.2 Cu . . . Al6-XN Plus N08367 0.02 0.24 21.8 25.3 6.7 0.3 0.4 0.2 Cu . . . Chapter 6: Austenitic Stainless Steels / 87

alloy development, the future use of 316 and thermomechanical history. As a rule of thumb, 317 should be numbered. There is no justifica- the environment to initiate SCC must be suffi- tion for the use of scarce and expensive re- ciently severe to cause localized corrosive at- sources such as nickel and molybdenum when tack. The most dangerous situation is one in cheap, abundant replacements like manganese which the expectation of pitting is marginal. and nitrogen are available. The use of 316 as a The mechanism of SCC has been a subject of standard alloy should in the future be eroded by intense academic controversy for many years. more cost-effective alloys such as the lean du- Because of this, much of the literature has fo- plex alloys like 2003. cused on arguing a case rather than clarifying The same environments that cause pitting the phenomenon. What can be said about SCC corrosion also cause crevice corrosion. A in austenitic stainless steels with consensus? crevice is a volume in and out of which diffusion is restricted to a degree that corrosion prod- ¥ Risk of SCC is low at room temperature and increases exponentially with temperature. ucts accumulate and cause the contained envi- ¥ ronment to become increasingly aggressive in SCC is preceded by localized corrosive at- pH and [ClÐ]. Conditions that are below the tack, which has an incubation time, and then proceeds in a discontinuous manner. threshold for pitting can cause crevice corro- ¥ sion. The critical temperature for crevice corro- Fracture may be transgranular, intergranular, or both. It is almost entirely lacking in plas- sion is also measured in 6% FeCl3 (ASTM G-48 B or D). It is the lowest temperature at which tic deformation with little, if any, metal loss. crevice corrosion occurs. This temperature, the ¥ Alloying or treatments that delay localized CCT (critical crevice temperature), is lower attack or stabilize austenite can delay SCC than the CPT. GTAW a wrought alloy also low- up to the point of virtual immunity. ers the CPT to about the level of the CCT. The ¥ SCC is aggravated by increased chloride reason for this lowering of resistance to local- concentration and acidity, but also exists in ized attack has been thought to be related to caustic environments. alloy depletion caused either by dendrite coring ¥ Stress must exceed a threshold for a given during solidification or chromium depletion set of conditions for SCC to occur. around inclusions. The relation to crevices ¥ Anodic or cathodic polarization may prevent would thus seem to be that surfaces contain nu- SCC under conditions at which it would oth- merous flaws with respect to corrosion resist- erwise be expected, or it may cause SCC in ance, which, while not capable of sustaining pit- environments in which it would not other- ting, can in a crevice dissolve and alter the wise occur. environment sufficiently that the new harsher Austenitic stainless steels are not alone in their environment can generally destabilize the pas- susceptibility to SCC. All stainless steels suffer sive film and proceed autocatalytically. from SCC under some set of conditions of envi- Stress corrosion cracking is the bane of ronment and material thermomechanical history. austenitic stainless steels. SCC occurs when The key is to choose an alloy that is resistant there is both a tensile stress of a sufficient mag- under the conditions of use. To a first approxima- nitude and a sufficiently aggressive environ- tion, this means using an alloy that will not pit ment. The threshold stress varies with alloy and under the conditions of use, then designing its

Table 5 Corrosion resistance ratings of various austenitic stainless steels, using 30 factor for nitrogen

Alloy Designation PREN(a) CPT oC CCT oC 316L S31603 24 15 Ð3 316Ti S31635 23 15 Ð3 317L S31703 30 25 0 317LMN S31726 34 30 4 904L N80904 35 40 15 JS700 N08700 36 43 15 254SMO S31254 46 73 38 4565 S34565 53 90 50 654SMO S32654 63 105 75 AL6-XN N08367 50 78 43 AL-6XN Plus 50 min 95 60 (a) PREN, pitting resistance equivalent number. 88 / Stainless Steels for Design Engineers

use to be below the threshold stress completes a Nitric acid is not particularly aggressive sound design approach if residual stresses can be against stainless steels. Resistance to it is pro- accurately known. Otherwise, assuming that the portional to chromium content. So, attack, metal will have residual stresses equal to 100% should it take place, is preferably at grain of the yield strength is the prudent engineering boundaries, where segregation of elements such approach. Figure 24 shows threshold stress for a as carbon, phosphorus, and silicon can lower number of alloys. chromium locally. These elements are kept low Special Corrosive Environments. Knowl- for nitric acid service. Pitting is not a risk. Stan- edge of the ability of the various stainless steels dard usage is: to resist specific environments is essential to the design process. This information is extensive ¥ Below 50% concentrations and below 100 ° since it must correlate many environments and C, 304L and 17% Cr ferritics are used. temperatures for many materials. Hence, refer ¥ Around the 65% aziotrope, 310 is most re- to the publications of organizations such as the sistant, especially a low carbon version, but National Association of Corrosion Engineers 304 NAG with low carbon, phosphorus, and (NACE) or to the Web sites of companies such silicon is more often used. as Allegheny Ludlum or Outukumpu, where ¥ For 98% solutions or for lower concentra- such information is available freely. The more tions that contain other stronger oxidizers, reputable producers will give assistance on spe- alloys with 4% Si, 18% Cr, and 15% Ni or cific questions. Engineering forums on the In- 5% Si, 17% Cr, and 17% Ni (UNS S30600 ternet, such as www.eng-tips.com, should also and S30601) have been developed. be considered a resource. The following discus- Hydrochloric acid, not surprisingly, is quite sion presents just the principles of the resistance aggressive against stainless steel. It is very ef- of austenitic stainless steels to specific, more fective in destabilizing the passive film. Thus, common environments. resistance to hydrochloric acid is simply an ex- Sulfuric acid is common, aggressive, and must treme case of resistance to pitting in chlorides be contained. Figure 25 shows the isocorrosion with resistance given by Eq 12. Only the most curves for several alloys in pure sulfuric acid. highly alloyed austenitic alloys, such as AL- Alloy 20, 904L, and alloy 825 were devel- 6XN¨, should be considered and then under oped specifically for sulfuric acid service. Each conditions that are tolerable, such as those contains several percent copper that, while not shown in Fig. 26. beneficial against pitting, concentrates in the Strong bases such as NaOH and KOH are not passive film and diminishes general corrosion. especially aggressive against stainless. The 17% Molybdenum and tungsten are also very benefi- chromium alloys can be used up to 50 °C, while cial for resistance to sulfuric acid. Phosphoric 304L can be used to 90 °C. As is the case with acid is similar to sulfuric acid in its effect on nitric acid, chromium and nickel are beneficial, austenitics but somewhat less aggressive. while molybdenum is counterproductive. The 25% chromium alloys such as 310 or an equal

Fig. 24 Threshold stress for stress corrosion cracking (SCC) for various alloys. Source: Ref 27 Fig. 25 Isocorrosion in pure sulfuric acid. Source: Ref 28 Chapter 6: Austenitic Stainless Steels / 89

contributing causes. Powder injection-molded stainless components often have porosity that is generally spherical. When exposed to the surface, such pores act as crevices and lower the pitting potential also. All of these factors are operative and can act in unison. Very fine abrasive polishing causes little residual stress and has very minimal crevice creation. Thus, mirror-type polished finishes do not degrade corrosion resistance, but they do not enhance corrosion resistance as does elec- tropolished mirror finishes, which remove chromium-depleted sites, which can initiate Fig. 26 Resistance to hydrochloric acid. Source: Ref 28 pitting. chromium duplex can be used to 150 °C, above REFERENCES which temperature nickel base alloys are required. High-chromium ferritic stainless steels 1. A.L. Schaeffler, Constitution Diagram for are also very good choices. Stainless Steel Weld Metal, Met. Prog., Organic acids are generally less aggressive Vol 56, Nov 1949, p 680Ð688 against stainless than are mineral acids since 2. W.T. Delong, A Modified Phases Diagram they are less dissociated in solution. They be- for Stainless Steel Weld Metals, Met. Prog., come hazardous when they contain chloride Vol 77, Feb 1960, p 98 ions, at high temperatures, or when they disso- 3. Design Guidelines for Selection and Use of ciate strongly, such as with formic acid. Stainless Steel, SSINA,1998, p 3 Because of the large number of organic com- 4. K.-J. Blom, “Press Formability of pounds that may be considered, refer to the var- Stainless Steels,” paper presented at Stain- ious corrosion tables. less steels ‘77 Surface Finish. The corrosion resistance of 5. F.B. Pickering, “Physical Metallurgical De- austenitic stainless steels is quite dependent on velopments in Stainless Steel,” paper pre- surface condition, as are other stainless steels. sented at Stainless Steel ‘84, Goteborg Treatments that enhance surface concentrations 6. Q.-X. Dai et al., Chin. Phys., Vol 11, 2002, p of beneficial elements or remove detrimental 596Ð600, doi:10.1088/1009-1963/11/6/315 constituents can greatly alter performance. 7. Aciers Inoxidables, Les Editions de Oxide formation depletes surface chromium, so Physique, Les Ulis, Paris, 1993, p 564 strong pickling or electropolishing of the 8. Aciers Inoxidables, Les Editions de descaled surface is especially important. Studies Physique, Les Ulis, Paris, 1993, p 565 have shown chromium depletion of a maximum 9. P. Marshall, Austenitic Stainless Steels, Mi- of 6% extending 10 μ before reaching bulk crostructure and Mechanical Properties, chromium levels. This is equivalent to the de- Elsevier, 1984 pletion seen in sensitization. The increase in at- 10. Aciers Inoxidables, Les Editions de tack rate from this depletion is huge. A 1000- Physique, Les Ulis, Paris, 1993, p 579 fold increase in weight loss in the ASTM G 48 11. H. Nordberg, Mechanical Properties of B test has been seen by a superficial loss of 6% Austenitic and Duplex Stainless Steels, In- chromium. novation in Stainless Steels ‘93 (Firenze), Likewise, surface abrasion, especially coarse 1993, p 2.217 abrasion, has a major detrimental effect. The 12. Aciers Inoxidables, Les Editions de 120 grit #3 finish often seen on stainless reduces Physique, Les Ulis, Paris, 1993, p 566 pitting resistance by as much as the equivalent 13. S.J. Rosenberg and C.R. Irish, J Res. Nat. of 5 PREN, that is, equal to a reduction in Bar. Stand., Vol 48, 1952, p 40 chromium content of 5%. Rolled finishes are 14. Aciers Inoxidables, Les Editions de much preferred. The mechanism for this has not Physique, Les Ulis, Paris, 1993, p 410 been clearly established; exposure of MnS in- 15. M. McGuire, “A Diffusion Model for the clusions, the microcrevices abrasion produces, Influence of Oxygen and Sulfur on the Non- and residual stress have been cited as possible Equilibrium Distribution of Chromium in 90 / Stainless Steels for Design Engineers

Austenitic Stainless Steel Welds and Slabs,” 23. N. J. Grant, Accelerated Oxidation of Met- paper presented at Proceedings MS&T als at High Temperature, Trans. ASM, Vol ‘04,2004 44, 1961, p 128Ð137 16. R. Stickler and A. Vinckier, Trans. ASM, 24. J.F. Grubb, paper 04291 presented at Corro- Vol 54, 1961, p 362 sion 2004, NACE, 2004, p 1Ð15 17. Aciers Inoxidables, Les Editions de Phy- 25. ACOM Files, High Temperature Stainless sique, Les Ulis, Paris, 1993, p 568 Steels, www.outukumpu.com 18. Aciers Inoxidables, Les Editions de 26. J. Okamoto et al., A Super-Austenitic Stain- Physique, Les Ulis, Paris, 1993, p 448 less Steel for Tubing and Piping Applica- 19. Aciers Inoxidables, Les Editions de Phy- tions, Nippon Steel Technical Report 90, sique, Les Ulis, Paris, 1993., p 453 July 2004 20. Aciers Inoxidables, Les Editions de 27. ACOM Files, High Temperature Stainless Physique, Les Ulis, Paris, 1993, p 454 Steels, www.outukumpu.com 21. www.outukumpu.com 28. Allegheny Technologies, “AL6-XN¨ 22. W.C. Leslie, Mechanism of Rapid Oxida- Alloy” tion at High Temperature, Trans. ASM, Vol 41, 1958, p 1213Ð1219 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 91-107 All rights reserved. DOI: 10.1361/ssde2008p091 www.asminternational.org

CHAPTER 7

Duplex Stainless Steels

Summary they are exceptional materials. Whether these duplex alloys will grow to the full extent of THE NEWEST FAMILY of stainless steels is their potential depends on several factors: the duplex alloys. The mixture of ferrite and austenite in their structure gives them higher ¥ Will high nickel and molybdenum prices be strength than either phase by itself. Duplex al- sufficient motivation to drive designers to loys have at least 20% chromium, so they are explore alternatives to traditional austenitic considered as highly corrosion-resistant alloys grades? but not high-temperature alloys because of em- ¥ Will producers overcome their inhibition to brittling phases. Their low nickel content makes aggressively market these grades through them more economical than austenitic alloys of their cost-saving potential? the same level of corrosion resistance, espe- ¥ Will producers perfect the techniques to pro- cially when their greater strength can be utilized duce these grades reliably so that their avail- to reduce the amount of material required. They ability is unquestioned? should largely replace alloys such as 316L ¥ Will design codes change to correctly reﬂect and317L in the future. the duplex materials’ higher ratio of yield strength to tensile strength? Why are there such issues with a family of al- Introduction loys that has been successfully used for 20 years? The concept of duplex stainless steels is simple: Duplex stainless steels are the newest and islands of austenite in a continuous matrix of fastest-growing alloy group in the stainless steel highly alloyed ferrite, as seen in Fig. 1. This family. They are called duplex because at room temperature they consist of two phases, ferrite and austenite. Discovered in the 1920s, they languished in a suboptimized and underutilized state until recently. They possess excellent strength, toughness, and corrosion resistance. They also display exceptional resistance to stress corrosion cracking (SCC) and corrosion fatigue. The leaner grades, such as 2304, correspond to 316L in corrosion resistance but have double the yield strength, while the higher alloy grades like 2507 compete with the 6% molybdenum superaustenitics in corrosion resistance while still possessing much greater strength. Their limitations lie in their lack of cryogenic toughness and their inability to withstand temperatures much above 300 ¡C without forming embrittling phases. But between Ð100 and 300 ¡C Fig. 1 Wrought 2205 duplex microstructure 92 / Stainless Steels for Design Engineers

combination in principle offers high strength be- in the ferrite matrix, and each phase would have cause of the possibility of refining the dual-phase equal corrosion resistance despite having differ- grain structure and thereby raising yield strength ent compositions. It took a long time for that to according to the Hall-Petch relationship as well be accomplished. as by solid solution hardening, especially with Figure 2 shows a simple Fe-Cr-Ni constitu- nitrogen. In addition, the absence of a continuous tional diagram (Ref 1). The salient points are austenite phase provides relief from SCC by hav- that the typical successful alloys nearly bisect ing any propagation of cracks in austenite ar- the two-phase field for austenite and ferrite. It is rested by the ferrite phase. also obvious that the composition of the austen- The optimization of the alloy system had to ite and the ferrite must be quite different. wait for two events, both related to nitrogen. Ferrite contains a great deal more chromium First, the control of nitrogen in the refining by than austenite; hence, its pitting corrosion re- the argon oxygen decarburization (AOD) process sistance contribution from chromium is much allowed the control nitrogen content up to the greater than the resistance of the austenite be- solubility limit. Second, the understanding of cause in duplex grades: the thermodynamics of the alloy system became understood and reduced to a computer model. PREN = %Cr + 3.3 × %Mo + 16 × %N (Eq 1) At this point, the alloys developed over the first 50 years of development became obsolete, and If one were to add molybdenum to increase new grades with higher nitrogen vastly improved pitting resistance, it would preferentially parti- the performance and user friendliness. Why this tion to the ferrite, further exacerbating the dif- was so important can be seen by studying the ferential between the two phases. structure of these alloys. This is where nitrogen saves the day. Addi- tions of nitrogen concentrate nearly entirely in the austenite. This lowers the activity of Structure and Alloy Design chromium and thereby effectively attracts more chromium to the austenite phase than would The ideal structure of a duplex grade would otherwise be present. This stabilizes the austen- be a stable 50-to-50 ratio of austenite to ferrite ite, keeping the ratio of ferrite to austenite more at all temperatures at which it is to be used with- nearly constant with temperature. The pitting out other phases. The austenite would be islands resistance of the austenite increases signifi-

Fig. 2 The Fe-Cr-Ni phase diagrams. The shaded area results from nitrogen additions Chapter 7: Duplex Stainless Steels / 93

cantly to approximately that of the ferrite. In ad- (PREN) of about 35 and ﬁlls a niche in corrosion dition, the nitrogen solid solution strengthens resistance where austenitics and ferritics are the austenite and retards the formation of inter- lacking, greater than 317L stainless, PREN = 30, metallic phases, which is not bad for an element and below the 6% molybdenum grades, such as that costs nothing. AL-6XN alloy, with PRENs of around 45. Fer- ThermoCalc, developed by the Swedish Royal ritics have a gap between 442 (18Cr-2Mo) and Academy, has been an especially valuable tool in the super ferritics (28Cr-4Mo). By varying the helping us understand and design better duplex chromium, nickel, and molybdenum, leaner al- stainless steels. Without being able to computer loys can be devised that save cost based on re- model the thermodynamics of the system, it duced molybdenum and nickel. Conversely, would be impossible to project the partitioning of more corrosion-resistant alloys with higher the various potential alloying elements. Figure 3 PRENs can also be mapped, such as Fig. 4, with shows isopleth diagrams for a basic 2205 compo- the same diagrams varying nickel for a higher sition in which nickel is varied. molybdenum level. This composition includes The 2205 is the workhorse grade of duplex. It the important 2507 alloy. has a pitting resistance equivalent number Partitioning of elements (Fig. 5) between austenite and ferrite is an important issue. The partitioning tendency is a strong function of temperature. Figure 6 shows that as temperature

Fig. 3 The iron-nickel diagram for 22% Cr, 3% Mo, 0.15% N Fig. 5 Partitioning tendencies of various elements between ferrite and austenite. Source: Ref 2

Fig. 4 The iron-nickel diagram for 25% Cr, 4% Mo, and 0.25% N: N is a nitride, χ is chi, σ is sigma, α is fer- Fig. 6 Variation of partitioning ratio with temperature. rite, and γ is austenite Source: Ref 2 94 / Stainless Steels for Design Engineers

increases, the partitioning diminishes until at cleation and diffusion. The areas around the just above 1300 ¡C it approaches unity for all newly formed σ are naturally somewhat di- normal substitution-alloying elements (Ref 2). minished in chromium and molybdenum, so For nitrogen, however, the tendency is to in- the alloy’s resistance to localized corrosion is creasingly segregate to austenite as temperature compromised also. increases (Ref 2). Figure 9 shows the TTT (time-temperature- A danger in these alloys is that austenite transformation) diagram for various high-alloy formed from ferrite on heating, such as during stainless steel, including austenitic, ferritic, and welding or annealing, will contain only the low duplex. Alloys of all structures, ferritic, amount of nitrogen that was in the ferrite from austenitic, and duplex, with high chromium and which it was formed, until diffusion can restore molybdenum encounter the σ problem fairly equilibrium. If the heating time does not permit equally and in proportion to their alloy content this, this so-called secondary austenite will (Ref 2). This is the reason that the use of nitro- have low nitrogen and therefore low pitting gen instead of molybdenum is so beneﬁcial to corrosion resistance, as shown in Fig. 7(e). Ni- the leaner alloys, not just in cost, but for the trogen alters the phase stability, making austen- major reduction in rate of formation of sigma. ite stable to higher temperatures. This helps Figure 10 shows the large reduction in σ forma- keep welds from becoming excessively ferritic tion enjoyed by the lean alloy AL 2003™ mate- and disturbing the desirable 50-to-50 ratio of rial compared to the higher molybdenum 2205 austenite to ferrite. Secondary austenite with alloy (Allegheny Ludlum). low nitrogen is remedied by diffusion if the There are other intermetallic phases in addi- phase forms at higher temperatures at which tion to σ. They include χ, R, π, and τ. These are diffusion of nitrogen can rehomogenize the ni- of more research than practical interest because trogen level. σ, with its bad consequences, forms sooner and A crucial aspect of alloy design in the duplex in greater quantity under the same conditions alloys involves the avoidance of unwanted compared to the others. phases. The duplex stainless steels have all the Carbides and nitrides can also form in duplex potential problems with embrittling phases of alloys. The nitride Cr2N can form when satu- the ferritic and austenitic stainless steels com- rated ferrite is quenched from a high tempera- bined since they contain both as phases. Ferrite ture, as can occur in the welding process. It is forms two main embrittling phases, α′ and σ. possible that this would result in nearby The α′ is generally believed to be a result of chromium depletion and a decrease in corrosion the miscibility gap that exists in the iron- resistance. Carbide formation does not as easily chromium system, by which ferrite undergoes cause chromium depletion in duplex alloys be- spinodal decomposition into the iron-rich α, cause the precipitation at the ferrite-austenite normal ferrite, and the chromium-rich α′, grain boundary does not deplete the austenite as which is a brittle ordered alloy. Higher levels greatly in chromium locally because of the of chromium or the presence of copper or neighboring ferrite having a much higher diffu- molybdenum exacerbate this reaction, which sivity for chromium. The point is generally has a formation that follows an Arrhenius-type moot since all modern duplex grades contain curve with a maximum at around 400 ¡C. Fig- less than 0.030% carbon. ure 8 shows the α′ formation kinetics for ﬁve Table 1 lists the duplex grades currently duplex alloys (Ref 2). While duplex grades available commercially. Figure 7 shows a series have good oxidation resistance and high-tem- of duplex photomicrographs. perature strength, the α′ problem restricts their use to below about 315 ¡C. Ferrite and austenite both form intermetallic phases, of which the most prominent and dangerous is σ, a Mechanical Properties tetragonal phase richer in chromium and molybdenum than the ferrite from which it In many ways, the duplex stainless alloys forms. It is brittle and forms at grain bound- represent a best of both worlds in combining aries, so its precipitation has the immediate ef- traits from the austenitic and ferritic alloys. fect of lowering toughness. Cold work acceler- They offer high as-annealed strength with ates the precipitation process by up to an order good toughness and ductility. Table 1 lists of magnitude by virtue of its dual effect on nu- the major grades of duplex stainless steels; Chapter 7: Duplex Stainless Steels / 95

Fig. 7 (a) As-cast duplex structure, austenite in a ferrite matrix. (b) 2205 annealed; austenite phase contains twins. (c) 2507 as-welded; weld is highly ferritic because of rapid cooling rate. (d) Same weld as (c) after homogenization anneal. (e) 7-Mo Plus with ( (dark areas) that has induced the formation of secondary austenite (arrows)

Table 2 lists typical and minimum properties The most striking and unexpected character- for the major duplex alloys and those of some istic of the duplex grades is their high yield comparable ferritic and austenitic alloys for strength, more than double that of comparable comparison. austenitic grades. 96 / Stainless Steels for Design Engineers

Fig. 8 Kinetics of ( formation

The strength of the duplex grades is driven by the strength of the continuous ferrite phase. It owes its strengthening primarily to: ¥ Solid solution hardening by nickel, molybdenum, chromium, copper, and manganese ¥ Interstitial solid solution hardening by carbon and nitrogen ¥ Strengthening by grain reﬁnement These components have been related to the mechanical properties by the following equations (Ref 3):

Fig. 9 Sigma formation kinetics at various alloy levels =+ + RNp02. 120 210 0. 02 ++++21410()Mn Cr Mo Cu +(.6155054− .)δδ −12/ ++((.))735N + 002d (Eq 2)

Rp1.0 = Rp0.2 + 40 ± 9 (Eq 3)

=+ + Rm 470 600(.)N 0 02 − +++14Mo 1. 5δ 8d 12/ (Eq 4)

where δ is the ferrite content in percent, d is the lamellar spacing, and results are in megapas- cals. The inﬂuence of nitrogen is interesting in that at lower levels (e.g., below 0.1% nitrogen) Fig. 10 Delay in ( precipitation in lean duplex 2003 austenite is the weaker phase, but additional ni- Chapter 7: Duplex Stainless Steels / 97

Table 1 Duplex compositions

UNS Name C N Cr Ni Mo Mn Si Cu W P S S32900 329 0.08 . . . 23.0Ð28.0 2.5Ð5.0 1.0Ð2.0 1.0 0.75 ...... 0.040 0.030 S31200 44LN 0.03 0.14Ð0.20 24.0Ð26.0 5.5Ð6.0 1.2Ð2.0 2.0 1.0 ...... 0.040 0.030 S31260 DP3 0.03 0.10Ð0.30 24.0Ð26.0 5.5Ð7.5 2.5Ð3.5 1.0 0.75 0.2Ð0.8 0.1Ð0.5 0.030 0.030 S31500 3RE60 0.30 0.05Ð0.10 18.0Ð19.0 4.25Ð5.25 2.5Ð3.0 1.2Ð2.0 1.4Ð2.0 ...... 0.030 0.030 S31830 2205(old) 0.03 0.08Ð0.20 21.0Ð23.0 2.5Ð3.5 2.5Ð3.5 2.0 1.0 ...... 0.030 0.020 S32001 19 D 0.03 0.05Ð0.17 19.5Ð21.5 1.0Ð3.0 4.0Ð6.0 1.0 ...... 0.040 0.030 S32003 2003 0.03 0.14Ð0.20 19.5Ð21.0 3.0Ð4.0 1.5Ð2.0 2.0 1.0 ...... 0.040 0.030 S32101 2101 0.04 0.20Ð0.25 21.0Ð22.0 1.35Ð1.70 0.1Ð0.8 4.0Ð6.0 1.0 0.1Ð0.8 . . . 0.040 0.030 S32205 2205 0.03 0.14Ð0.20 22.0Ð23.0 4.5Ð6.5 3.0Ð3.5 1.0 2.0 ...... 0.030 0.020 S32304 2304 0.03 0.05Ð0.20 21.5Ð23.5 3.0Ð5.0 2.5 1.0 0.05Ð0.6 . . . 0.040 0.040 S32520 Uranus 0.03 0.20Ð0.35 24.0Ð26.0 5.5Ð8.0 3.0Ð5.0 1.5 0.8 0.5Ð3.0 . . . 0.035 0.020 52N+ S32550 255 0.04 0.10Ð0.25 24.0Ð27.0 6.0Ð8.0 2.9Ð3.9 1.5 1.0 1.5Ð3.0 . . . 0.040 0.030 S32750 2507 0.03 0.20Ð0.30 24.0Ð26.0 6.0Ð8.0 3.0Ð5.0 1.2 0.8 0.5 . . . 0.035 0.020 S32760 Zeron 100 0.03 0.20Ð0.30 24.0Ð26.0 6.0Ð8.0 3.0Ð5.0 1.0 1.0 0.5Ð1.0 0.5Ð1.0 0.030 0.010 S32906 2906 0.03 0.30Ð0.40 28.0Ð30.0 5.8Ð7.5 1.5Ð2.6 0.8Ð1.5 0.5 0.8 . . . 0.030 0.030 S32950 7ÐMo Plus 0.03 0.15Ð0.35 26.0Ð29.0 3.5Ð5.2 1.0Ð2.5 2.0 0.6 ...... 0.035 0.010 S39274 DP3W 0.03 0.24Ð0.32 24.0Ð26.0 6.0Ð8.0Ð 2.5Ð3.5 1.0 0.8 0.2Ð0.8 1.5Ð2.5 0.030 0.020 S39277 AF 918 0.025 0.23Ð0.33 24.0Ð26.0 6.5Ð8.0 3.0Ð4.0 0.8 0.8 1.2Ð2.0 0.8Ð1.2 0.030 0.020

Table 2 Duplex mechanical properties

Charpy-V Grade Name Rp0.2 Rm A5 HB RC Ð40 ¡C, J S31200 44LN 450 690 25 293 31 . . . S31260 DP3 485 690 20 290 31 . . . S31830 2205(old) 450 62 25 293 31 . . . S32003 2003 450 620 25 290 30 40 S32001 19D 450 640 25 290 25 . . . S32101 2101 450 650 25 290 32 40 S32205 2205(new) 460 640 25 290 32 40 S32304 2304 400 600 25 290 31 40 S32520 Uranus 52N+ 550 770 25 . . . 28 . . . S32550 Ferralium 550 760 15 302 32 . . . S32750 2507 550 795 15 310 32 40 S32760 Zeron 100 550 750 25 270 ...... S32960 7-Mo Plus 485 690 15 293 32 . . . trogen strengthens the austenite so that above above 300 ¡C is not recommended, no higher- 0.2% nitrogen, the austenite becomes the temperature properties are shown. Since they stronger phase. have a ductile-to-brittle transition, they also are The two phases are elongated parallel to the not well suited to cryogenic use. major strain axis of working such as from hot or Impact Strength. Toughness is a significant cold rolling. As working increases, the mi- consideration when using duplex alloys to re- crostructure and properties become increasingly place the extremely tough austenitic alloys. anisotropic, with the austenite taking on a (110) Duplex alloy low-temperature toughness is in- [223] texture and the ferrite (100) [011] to (211) termediate to that of ferritic and austenitic al- [011] (Ref 2). loys. This having been said, it should be noted Because the ferrite phase controls mechani- that the duplex alloys can have excellent tough- cal properties, the dependence of these proper- ness levels, such as 100 J at Ð100 ¡C in the so- ties on temperature is significant since flow in lution-annealed condition. As would be ex- body-centered cubic (bcc) structures is ther- pected, toughness improves with decreasing mally activated. Figure 11 shows the variation grain size and deteriorates with cold work. The of yield and tensile strengths of various grades most deleterious effect on toughness comes along with that of austenite and ferrite of simi- from the precipitation of intermetallic phases, lar composition. Because use of these alloys such as α′ and σ, which cause a sharp decrease 98 / Stainless Steels for Design Engineers

Fig. 11 Variation of ferrite, austenite, and duplex with temperature. Source: Ref 4

in toughness level and a concurrent increase in transition temperature. The combined effect of cold work and α′ can be seen in Fig. 12. Lean alloys such as 2001, 2003, and 2101 have a much slower rate of formation of α′ and are much less at risk for loss of toughness from exposure in the 300 to 600 ¡C range, as was shown in Fig. 11. Fatigue. Fatigue tests on duplex stainless steels indicate that they possess a fatigue limit of about 50% of the yield strength when tested in air (Ref 4). The ratio of the fatigue strength in a hostile environment to that in air is a useful measure of the complementary strong points of the duplex grades (i.e., strength and corrosion resistance). Figure 13 shows that ratio for various alloys plotted versus their PREN. As an alloy’s resistance to corrosive attack increases, its fatigue limit in a given environment approaches that in air, indicating simply that corrosion plays an increasingly small role in fatigue crack propagation as corrosion resistance increases. While this is intuitively reasonable, it is not diminished because the duplex reward the user with a higher level of yield strength and fatigue strength in air, so the net useful strength Fig. 12 Increase in transition temperature with α′ formation under cyclic loading is much greater than that of with aging for (a) annealed 2705 and (b) cold- equivalent-PREN austenitic alloys. worked 2205. Source: Ref 4 Chapter 7: Duplex Stainless Steels / 99

Fig. 13 Inﬂuence of pitting resistance equivalent number (PREN) to fatigue strength in NaCl solution versus in air. Source: Ref 2

Forming and Machining Heavily formed sections should be fully annealed, not just stress relieved, whenever The higher strength and lower ductility of the there is a potential for SCC in the service duplex grades compared to austenitics gives environment. them correspondingly less ability to be cold formed. Duplex alloys have sufficient ductility to be cold drawn; they behave like ferritics or Corrosion Resistance austenitics of similar alloy level. This, however, is an alloy level at which excellent formability Because duplex alloys are made up of two is seldom expected. Nevertheless, duplex alloys phases, ferrite and austenite, each must carry its can be cold formed like austenitic alloys. Oper- own weight in resisting corrosion. Early alloys ations such as bending, drawing, and pressing that were lacking in nitrogen generally had a fer- can readily be performed. Bend radii should be rite phase that, because of the greater partition- at least twice sheet thickness. Tubing can be ex- ing of the chromium and molybdenum to the fer- panded into tube sheets, but care must be taken rite, had higher corrosion resistance than the to produce tight roller-expanded joints. Tubing austenite. As nitrogen is added, it enriches the bend radii should be at least twice tubing out- austenite phase preferentially until the corrosion side diameter (OD). resistance of the austenite phase reaches that of 100 / Stainless Steels for Design Engineers

the ferrite. This approach is common to all more the aggressiveness of these media is enhanced recently developed alloys starting with the revi- by contamination. sion of 2205 from UNS S31803 to S32205, Sulfuric Acid. Figure 14 shows the behavior which has primarily higher nitrogen. The net re- of S32304 compared to 304 and 316 in sulfuric sult is a type of alloy that has most of the highly acid. Figure 15 shows additional, more highly desirable corrosion resistance characteristics of alloyed duplex grades. The use of copper as an superferritic grades without their limiting lack of alloying element in S32550 (1.5%) and S32760 mechanical properties, mainly toughness. (0.5%) gives them much better performance than the otherwise similar S32750. In real-life situations, such as seen in ﬂue gas General Corrosion desulfurization, sulfuric acid can be contami- The duplex alloys offer important advantages nated with chlorides. While this contamination in performance over the austenitic grades in a is deadly to 316 and 317, it has only a minor ef- number of signiﬁcant aggressive media, includ- fect on the copper-alloyed duplexes (Fig. 16). ing sulfuric acid, hydrochloric acid, sodium hy- Hydrochloric Acid. Historically stainless droxide, phosphoric acid, and organic acids. steels have had their poorest performance when This performance extends to situations in which confronted by hydrochloric acid. Here again, the

Fig. 14 The 0.1 mm isocorrosion curves. Source: Ref 5 Fig. 15 The 0.1 mm isocorrosion curves. Source: Ref 5

Fig. 16 Isocorrosion (0.1 mm/yr) performances of several austenitic and duplex alloys. Source: Ref 6 Chapter 7: Duplex Stainless Steels / 101

copper/tungsten-alloyed duplexes show excep- considered for use with nitric acid, and even tionally good performance, as seen in Fig. 17. then no advantage can be claimed. This extends the usefulness of stainless steels to Sodium Hydroxide. Much of the older pub- an environment that had previously been off lished data on the behavior of stainless steel has limits. Indeed, the duplex stainless steels in gen- seemed to promote the notion that higher nickel eral can be said to be relatively indifferent to the levels were beneficial in strong bases. There pH of chloride solutions and are affected rather seems now to be little to support that notion. by the chloride concentration and temperature. Figures 18 and 19 clearly indicate, respectively, Nitric Acid. It is fairly well known and ac- that the duplex alloys with their relatively low cepted that resistance to nitric acid, which was nickel levels significantly outperform the higher one of the first uses of stainless steel, depends nickel 304L and 316L, with performance im- almost entirely on the chromium content. proving with increasing chromium content. The Molybdenum, in all other instances a very bene- advantage is magnified when the environment is ficial alloying element, has a strongly negative contaminated with chlorides, as is the case of influence on resistance to this highly oxidizing the white liquors of kraft digesters. acid. Consequently, only the leanest-molybde- Phosphoric Acid. While pure phosphoric num duplex alloys, such as S32304, should be acid is not a very corrosive medium for stainless

Fig. 17 Isocorrosion (0.1 mm/yr) performance of duplex in HCl compared to 316L. Source: Ref 6

Fig. 19 Corrosion rates in white liquors plus chlorides. Fig. 18 Corrosion rates in boiling NaOH. Source: Ref 7 Source: Ref 8 102 / Stainless Steels for Design Engineers

Fig. 20 Minimum temperatures for wet phosphoric acid (WPA) with an isocorrosion rate of 0.127 mm/yr. Source: Ref 9

In combinations of acetic and formic acid, the superiority of duplex alloys is quite evident, as seen in Fig. 22. S32750 shows virtual immunity, while in mixtures contaminated with halides its performance ranks very closely to expensive nickel-based superalloys such as N06625 and N06455. Even the lower alloyed S32205 can offer an order of magnitude improvement over S31703 in hot contaminated acetic acid.

Pitting Corrosion The different analysis of the two main phases in duplex alloys means that each has its own pitting resistance equivalent number, PREN. The Fig. 21 Isocorrosion (0.1 mm/yr) performances of various alloys. Source: Ref 9 ferrite phase has the relationship common to ferritic grades: steel, contaminants again can render it so. PREN + %Cr + 3.3%Mo (Eq 5) Halides are particularly common and aggressive contaminants. Figure 20 shows the substantial while the austenite obeys the more familiar: improvement in performance of the duplex al- PREN + %Cr + 3.3%Mo + 30%N (Eq 6) loys over 316L when contaminants are present. The duplex grades partition these critical ele- Performance again improves with increasing ments in such a way that the overall PREN of chromium, molybdenum, and nitrogen levels. most alloys comes out to be approximately Eq 1. Duplex alloys perform partic- Organic Acids. If one has the actual analysis of each phase, ularly well in organic acids and have an excellent then the proper relationship to use is Eq 2. These record in industrial plants. In acetic acid, 304L relationships are incomplete in that they only ad- handles lower temperatures and concentrations. dress the major alloying elements. Tungsten has Alloys such as S32205 perform well. In half the value of molybdenum and is frequently formic acid, the most aggressive organic acid, included: S32750 is resistant at all concentrations almost to the boiling point, outperforming even tita- PREN = %Cr + 3.3(%Mo + 0.5 × %W) + 16%N nium (see Fig. 21). (Eq 7) Chapter 7: Duplex Stainless Steels / 103

Fig. 22 Corrosion rates for various alloys in 50% acetic plus formic acid, boiling. Source: Ref 10

If nonwrought material is involved, as in as- temperatures. This austenite, which forms from cast and welded alloys, these relationships ferrite, has very little nitrogen, which clearly greatly overstate PREN. This is because non- lowers its pitting corrosion resistance. equilibrium-diminished concentrations of The duplex alloys stand up very well in com- chromium are often found around precipitates, parison to corresponding superaustenitic alloys. especially (manganese, chromium) S inclusions Figure 23 shows how CPT varies with PREN. (Ref 11, 12) and because of lower alloy content This ranking is not always linear, as Fig. 24 locally due to solidification segregation, princi- shows, with pitting potential dropping fairly pally of molybdenum. This is most significant rapidly with temperature and at different rates in welded tubing, which can have higher sulfur for different alloys. Figures 25 and 26 show the levels to increase weld penetration. Tube welds influence of pH and chloride concentration, re- can be reequilibrated by high-temperature an- spectively. In 3% NaCl (Fig. 26), the rankings nealing, but field girth welds will show dimin- show a minor variation with pH and a rational ished corrosion resistance if unannealed. So, relationship to alloy content. The influence of untreated welds can have PREN’s 5 to 15 lower chloride concentration is strong over a wide than the parent alloy, which equates to the local- range of concentrations. ized lowering of chromium levels. The critical These tests are best for judging relative per- pitting temperature (CPT) of welds often de- formance of alloys and must be used cautiously creases to near the critical crevice corrosion when extrapolating lab results to service per- temperature (CCT) of the parent metal. formance. The degree to which short-term tests, The precipitation of chromium- or molybde- whether potentiostatic or strictly immersion, re- num-rich second (third, in this case) phases, flect long-term performance has not been well such as σ or α′ inevitably results in diminish- established. ment of these key alloying elements in the region surrounding the precipitate, which will make it more prone to localized corrosion. Crevice Corrosion This can also occur when secondary austen- Crevices exist both by design and inadver- ite forms during the heating of alloys to high tently. Crevices are occluded volumes of liquid 104 / Stainless Steels for Design Engineers

Fig. 23 Critical pitting temperature in seawater measured potentiostatically versus pitting resistance equivalent number (PREN). Source: Ref 13

Fig. 24 Variation of pitting potential with temperature. SCE, saturated calomel electrode. Source: Ref 14 in which oxygen and corrosion products reach The difference increases with total alloy con- levels quite different from the exterior environ- tent, as can be seen in Fig. 27. Interestingly, the ment and become highly corrosive. Thus, the difference is approximately the same as is the tighter the crevice is, the greater the restriction difference in CPT between the wrought alloy of diffusion between the crevice and the bulk and the welded alloy. and therefore the greater the chance of crevice corrosion occurring. An alloy’s susceptibility to crevice corrosion is proportional to its resist- Stress Corrosion Cracking ance to pitting corrosion under the same condi- Stress corrosion cracking (SCC) has long tions. The CCT is lower than the CPT by about been the Achilles’ heel of stainless steels. Only 10 to 30 ¡C. soft ferritic stainless steels are immune to it. It Chapter 7: Duplex Stainless Steels / 105

occurs at temperatures and in environments where stainless would be a perfect material if only it did not stress corrosion crack. The arrival of duplex stainless steels has to a very large degree ameliorated, if not solved, that problem. SCC is unfortunately poorly understood. Like pitting, whose initiation mechanism has not been identified, SCC has both its initiation and propagation mechanisms still open to debate. But the duplex alloys have good strength mainly through fine grain size and solid solution hardening, which seems to avoid the hydrogen- trapping dislocation types that seem to be associated with hydrogen failures. So, while we can- Fig. 25 Variation of critical pitting temperature (CPT) with not state the mechanism for SCC, we can map pH. Source: Ref 14 out the conditions under which duplex alloys are susceptible to SCC. The major environmental factors that affect SCC are chloride concentration and temperature. Figure 28 shows the remarkable advantage the duplex alloys have over the comparable austenitic alloys with regard to the temperatures at which they may be used without SCC. The duplex alloys in this regard are governed in their behavior by their ferrite matrix, through which cracks must propagate (Ref 16). Ferritic stainless steels are known for their resistance to SCC in the annealed condition. The advantage of the duplex lies in their composite-type microstructure with the crack- arresting austenite phase and the toughening fine grain structure. The duplex alloys show a higher threshold stress for SCC as a percentage of their yield strength (Fig. 29) than austenitic alloys. This is in spite of their higher yield strength, again giving these alloys more usable strength. Fig. 26 Critical pitting temperature (CPT) as a function of NaCl concentration. SCE, saturated calomel In ferrite, SCC susceptibility is a maximum electrode. Source: Ref 5 below 100 ¡C, while in austenite susceptibility appears to begin around 50 ¡C and increase monotonically with temperature. The temperature at which SCC occurs at the fastest rate increases with nickel content. This is also characteristic of ferritic and martensitic materials and mirrors their hydrogen embrittlement behavior. H2S also accelerates failure in chloride environments (Fig. 30), and cold work accelerates failure and lowers threshold stress values. While duplex alloys behave in many regards like ferritic alloys in their SCC or hydrogen embrittlement response, they do not have the same relationship between susceptibility and bulk hardness. Other ferritic and martensitic alloys display pronounced susceptibility to Fig. 27 Critical crevice temperature (CCT) and critical pitting temperature (CPT). Source: Ref 15 these failures modes when their hardness 106 / Stainless Steels for Design Engineers

Fig. 28 Stress corrosion cracking (SCC) in neutral aerated NaCl. Testing duration 1000 hr. Source: Ref 5

Fig. 30 Suggested chloride and pH limits for cold-worked duplex alloys. Source: Ref 17

exceeds Rc 22. The duplex alloys have annealed hardness over Rc 30 without being in danger. This probably simply indicates that hardness as a measure of susceptibility is valid only insofar as it reﬂects a certain yield strength threshold as it does in tempered martensite and is not valid for ferrite/austenite composite structures. Thus, it is very important to understand duplex SCC behavior as a sepa- Fig. 29 Constant load stress corrosion cracking (SCC) tests in rate study and not interpret it in terms of aerated MgCl2 at 150 °C. Source: Ref 5 austenitic or martensitic SCC. Chapter 7: Duplex Stainless Steels / 107

REFERENCES 9. Avesta Sheffield, Corrosion Handbook for Stainless Steels, 1994 1. P. Lacombe, B. Baroux, and G. Beranger, 10. B. Walden et al., Stainless Steel ’93, Stainless Steels, Les Editions de Physique, Florence, AIM, 1993, p 3.47 2003 11. M.F. McGuire, MS&T Conf. Proc., 2004, p 2. R.N. Gunn, Duplex Stainless Steels, Abing- 831Ð846 ton Publishing, 1997, p 28 12. M. Ryan, D. Williams, R. Chater, B. 3. H. Nordberg H, Innovation of Stainless Hutton, and D. McPhail, Why Stainless Steel, Conf. Proc., AIM, Florence, 1994, p Steel Corrodes, Nature, Vol 412, p 770 2.217Ð2.229 13. C.V. Roscoe et al., Duplex Stainless Steels 4. Charles, Duplex Stainless Steels ’91, Vol 1, ’86, The Hague, Nederlands Instituut voor Beaune, Les Editions de Physique, 1991, p Lasteckniek, 1986, p 126Ð135 3Ð48 14. J.M. Drugli et al., Paper 270 presented at 5. S. Bernhardsson, Duplex Stainless Steels Corrosion ’90, Las Vegas, NACE, 1990 ’91, Vol 1, Beaune, Les Editions de 15. S. Bernhardsson, Paper 164 presented at Physique, 1991, p 137Ð150 Corrosion ’90, Las Vegas, NACE, 1990 6. J. Nichols J, 12th International Corrosion 16. T. Kudo, H. Tsuge, and A. Seki, Stainless Congress, Houston, NACE, p 1237 Steel ’87, The Institute of Metals, 1988, p 7. E.-M. Horn, Werkstoffe und Korrosion, Vol 168Ð175 42, 1991, p 511Ð519 17. R. Francis, Duplex Stainless Steels ’94, Vol 8. J.P. Audouard, Stainless Steel Europe, April 3, Glasgow, TWI, 1994, paper KIV 1992, p 45 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 109-122 All rights reserved. DOI: 10.1361/ssde2008p109 www.asminternational.org

CHAPTER 8

Ferritic Stainless Steels

Summary technology was documented long before AOD was invented (Ref 1). It was not until carbon THE FERRITIC STAINLESS STEELS are and nitrogen levels were brought down to AOD the lowest-cost highly corrosion- and oxidation- levels that it became truly practical for ferritic resisting alloys in existence. They are useful alloys. The level of carbon plus nitrogen was mainly as light-gauge sheet since their tough- lowered from around 0.10% to around 0.04%, ness drops off rapidly for heavier sections. Even and less-expensive high-carbon ferrochromium as they have grown in use more than any other could be used instead of expensive low-carbon type of stainless, they could still economically versions. Thus, there exist two types of ferritics: displace the popular but expensive 304 for the early high-carbon types such as 430, 434, many routine applications. 436, and 446 and the more modern stabilized alloys led by 409 and 439. The older, unstabilized grades are not always Introduction fully ferritic. Their carbon levels cause them to form some high-temperature austenite, which Ferritic stainless steels are simplest, lowest- transforms to martensite if quenched. This cost stainless steels. In their minimal form, they makes their welds brittle. To be used, they are contain simply enough chromium to overcome normally in the annealed condition, which re- their inherent level of carbon impurity and hit quires a lengthy subcritical anneal to avoid the 11% chromium in solution required for martensite and to evenly distribute chromium “stainlessness.” Early in the 20th century, 430 after all carbides have stably formed. The newer came into being, and the attainable levels of stabilized alloys behave as if they are interstitial carbon removal required 16% chromium for free. They are ferritic at all temperatures (ex- this to occur. So much extra chromium was re- cluding for the moment the possibility of extra- quired because during annealing, to develop the neous phases such as (α' and σ) and can be eas- fully ferritic structure, carbon combines with ily welded without fear of unwanted phases. chromium, rendering it useless as a corrosion Stabilization does not preclude excessive grain ﬁghter. In October 1967, the ﬁrst commercial growth in the fusion or heat-affected zone use of argon oxygen decarburization (AOD) (HAZ) of welds, which can render them brittle. changed the world for ferritic stainless steel. The mechanical properties of ferritic stainless This process, in which argon and oxygen are steels appear similar to austenitics strengthwise, blown through the molten metal to selectively but they lack the ductility of austenitics, and remove carbon without removing chromium they are limited at low temperatures by brittle- (described in detail elsewhere in this book), re- ness and at high temperatures by softness. duced the carbon plus nitrogen levels suffi- The lower thermal expansion coefficient of ciently that their effect could be nearly negated ferritics makes their scale more compatible with by small additions of titanium or niobium, the base alloy and provides them with a lesser which combine strongly with carbon and nitro- tendency to spall. This makes them excellent gen and effectively remove them from solution. for high-temperature applications with thermal This process is called stabilization, and the cycles, provided their strength is adequate. 110 / Stainless Steels for Design Engineers

The corrosion resistance of ferritics is ham- A similar predecessor was 410S, a low-carbon pered by their inability to utilize nitrogen. The version of 410 to which some understabilizing absence of nickel, which characterizes these al- amount of titanium is added but that still re- loys, is not a problem since nickel adds little to quires annealing for full ferritic properties. The corrosion resistance. The titanium stabilization key issue of the 11% chromium ferritics is how of the modern alloys has quite a beneficial effect to deal with carbon and nitrogen. The 405 and since titanium is a powerful deoxidizer and 410S take the approach of minimizing it and live desulfurizer, both of which can cause local with annealing. The 409 uses full titanium stabi- chromium depletion and pitting. Ferritics, more- lization. The hidden problem with using only ti- over, are essentially free from stress-corrosion tanium is that unless nitrogen levels are made cracking (SCC) since they are below the thresh- very low, the amount of titanium required to old hardness for hydrogen embrittlement in combine with it can reach levels at which the body-centered cubic (bcc) ferrous alloys. There first TiN precipitates in the molten metal. This are a few exceptions. slaggy precipitate agglomerates, causing casting The main attraction of ferritic stainless steels problems and surface defects. This gave 409 a over austenitics is their cost. The old compari- reputation as a grade unsuitable for applications son of 430 versus 304 is a bit unfair since 304 is that required good appearance because the tita- richer in chromium. A fair comparison might be nium streaks were difficult to avoid and greatly between 439 and 304. The corrosion resistance highlighted by polishing. This has largely been of these two alloys is barely distinguishable overcome by better refining techniques to reduce under normal ambient conditions. They are both carbon plus nitrogen to levels below 0.02% and very formable and weldable. The vast majority the use of dual stabilization by titanium and nio- of the objects made commercially from 304 bium; 468 (UNS S40930) is such an alloy. could be switched to 439 with no adverse con- The historical archetype of ferritic stainless sequence. But, if nickel is selling for $7 per steels was 430, which has existed since the pound, then the total cost of 304 versus 439 is 1920s and is still widely used. Its drawbacks are doubled by its presence. No design engineer can lack of weldability, relatively poor corrosion re- afford to ignore this level of incentive to learn sistance because so much of its chromium is to use ferritic stainless steels. tied up as carbides, and modest formability. The new archetype for this medium-chromium level is 439. With 17% chromium and single (439) or Ferritic Stainless Alloys dual stabilization (468), this alloy overcomes the problems of 430 and can readily replace 304 The ferritic stainless alloys generally group in in most applications with significant cost sav- low (10.5 to 12.0%), medium (16 to 19%), and ings. In North America, 439 is mainly used as a high (greater than 25%) chromium. They can be higher-temperature automotive exhaust alloy, stabilized or not. These distinctions are some- but in Europe 430Ti is used extensively in more what imposed after the fact. Rather than giving visually challenging applications, such as appli- them an order that they truly do not possess, the ances. There, it is generally used instead of 439 most significant alloys are all listed in Table 1 whenever the part can be designed to be formed with their compositions. from it. The low-chromium ferritic stainless steels Now, 434 and 436 are little used as their his- began with the development of MF-1, the prede- torical application in automotive trim finds little cessor of 409, in the 1960s. Its excellent corro- place in today’s automotive styling. A modern sion resistance, compared to carbon steel; rela- offshoot of these alloys, which are basically tively low cost; good welding; and formability molybdenum enhanced 430, is 444. This alloy permitted it to replace aluminized carbon steel has roughly the corrosion resistance of 316L but and cast iron in automotive exhaust systems, is fully resistant to SCC in the welded or an- opening up what eventually became the largest nealed condition. This makes it especially useful single market for stainless steel. It was made for applications such as hot water heaters, heat possible by the very low carbon plus nitrogen exchangers, and food- and beverage-processing levels the AOD process provided and the use of equipment. stabilization. Thus, 409 was an improvement on Both the nominally 11 and 18% chromium 405 in which aluminum performed a quasi stabi- alloys are sometimes modified to enhance lization, and low carbon suppressed martensite. their high-temperature strength or oxidation Chapter 8: Ferritic Stainless Steels / 111 Other ...... 0.15 Al Al 0.25 Al 0.15 Se 1 Al 0.10Ð0.30 Al 1.2 Al 3 Al 4 Al 1 Al 0.15 Al 1.7 Nb ...... 0.17 0.35 0.6 0.55 Ti ...... 0.20 + 4x(C+N) to 1.10 Ti + Nb: 0.20 + 4x(C + N) to 1.10 6x(C + N) to 0.75 6x(C + N) to 0.5 8x(C + N) to 0.15Ð0.50 8x(C + N) 0.8 + 8x(C+N) Ti + Nb 10xC to 0.75 Nb 0.3 0.4 0.4 . . . 0.20 + 4x(C+N) to 1.10 0.20Ti+Nb + 4x(C+N) to 0.75 0.35 0.25 0.25 . S ...... min 0.03 0.01 0.04 0.03 0.3 0.15 0.03 0.03 0.03 0.001 0.03 0.03 0.45 0.02 0.02 0.01 0.03 0.06 P ...... 0.04 0.03 0.45 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.02 0.04 0.04 0.04 0.06 0.06 0.04 0.04 Mo ...... (continued) Si . . 1 1 1 1 1 1 1 0.45 1 1 1 1 1 1 1 1.3 0.03 0.03 1 1 1 1 0.45 1.3 . Mn 1 1 0.035 1 1.25 1 1 1 1 1 1 0.75 1 1 1 0.25 0.035 0.7 1 1.25 1 0.35 0.3 0.3 0.7 Ni ...... 0.5 0.5 0.5 0.5Ð1 0.75 0.75 1 0.5 0.5 0.2 0.5 0.6 0.5 0.05 0.5 0.2 0.25 Cr 11.5Ð14.5 11.35 10.5Ð11.75 18.0Ð20.0 17.3 12.0Ð13.0 10.5Ð11.75 10.5Ð11.7 10.5Ð11.75 10.5Ð11.7 12 13 13 14.0Ð16.0 16.0Ð18.0 16.0Ð18.0 16.0Ð19.5 17.0Ð19.0 17.0Ð19 18 18 10.5Ð11.7 10.5Ð11.7 13.5 16.0Ð18.0 17.5 N ...... 0.02 0.03 0.04 0.02 0.015 0.04 0.03 0.03 0.03 0.015 0.01 C 0.12 0.03 0.05 0.08 0.03 0.02 0.06 0.03 0.02 0.025 0.12 0.12 0.1 0.03 0.01 0.02 0.015 0.08 0.08 0.03 0.02 0.01 0.025 0.08 0.12 0.07 typical typical typical typical alloys typical typical typical typical S40910 S40940 AK alloy AK alloy S43000 S43020 S46800 AK alloy S40900 S40920 AK alloy S40975 alloy ATI alloy ATI S42900 S43023 S43036 S43932 AK ATI, AK alloy Outukumpu Designation S40500 S40930 Outukumpu S43035 AK alloy 439 HP 439 ultraform 18 Cr-Cb Table 1Table stainless compositions Ferritic Alloy 405 400 409 409 409 409 ultraform 409Cb 409Ni Cr-Cb 11 12 SR Alfa II 4724 429 430 430F 430Ti 439 439LT 18SR 466 Alfa I 430Se 468 4742 112 / Stainless Steels for Design Engineers Other ...... 0.60 Al 0.1 REM 1.5 Al 0.2 Cu 0.5 Cu + Ni ...... Nb 0.5Ð0.20 10x(C + N) 9xC 0.3Ð1.0 Nb + 5xC 0.7 Ta . . Ti ...... Ti + Nb: 0.20 + 4x(C + N) to 0.80 0.02 Ti + Nb: 0.20 + 4x(C +Ti N) to 0.80 + Nb: 0.20 + 4x(C +Ti N) to 0.80 + Nb: 0.20 + 4x(C + N) to 0.80 0.1-0.6 8x(C + N) min S . . 0.001 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.015 0.001 0.02 P . . 0.02 0.02 0.04 0.04 0.04 0.02 0.04 0.02 0.04 0.04 0.04 0.04 0.04 Mo ...... 3.5Ð4.5 2.5Ð3.5 3.5Ð4.5 0.75Ð1.25 0.75Ð1.25 1.2 0.75Ð1.25 0.75Ð1.25 Si 1.5 1 1 0.4 0.3 1 1 1 0.4 1.4 0.4 0.75 1 0.75 1 Mn 1 0.3 0.4 0.3 1 1 1 1 0.2 1 0.7 1 1 1 1.5 . Ni . . . 0.25 0.6 0.3 0.3 0.5 4 1 3.5Ð4.5 1.5Ð3.5 0.5 0.6 Cr 18.0Ð23.0 17.5Ð19.5 20 22 25.0Ð27.5 26.0Ð30 16.0Ð18.0 16.0Ð18.8 17.5Ð18.5 17.3 24 24.5Ð26.0 25.0Ð27.0 28.0Ð30 23.0Ð27.0 N ...... 0.015 0.035 0.035 0.25 0.015 0.035 C 0.2 0.01 0.12 0.12 0.03 0.01 0.08 0.03 0.01 0.025 0.025 0.025 0.2 0.5 0.025 typical typical typical typical Designation S43400 S43600 S44200 alloy ATI alloy ATI Outukumpu alloy ATI S44627 S44635 S44660 S44735 S44600 Cast alloy S44100 S44400 (continued) 441, 4509, 430J1L 442 444, YUS 190-EM 433 Monit Sea-cure 29-4C CC-50 Table 1Table stainless compositions Ferritic Alloy 434 436 436S 4762 453 E-Brite, 26-1 446 Chapter 8: Ferritic Stainless Steels / 113

resistance. Again, the driving force has been changers and extensively in condensing por- the requirements of the hot end of exhaust tions of high-efficiency residential furnaces. systems (e.g., exhaust manifolds). Alloying The lower-alloyed Seacure had a slight tough- with niobium and molybdenum adds to the ness advantage that permitted it to be used at high-temperature strength, while additions of wall thicknesses of 1/16 in. when AL29-4C¨ chromium, silicon, and aluminum increase ox- was too brittle. As with other ferritics, these al- idation resistance. There exists an array of loys are generally only suitably tough when proprietary alloys as shown in Table 1; these used in thin section size (i.e., less than several are usually developed for specific automotive millimeters). needs and employ all or some of these alloy- It is difficult to say ferritic stainless steels are ing variations. The use of silicon and alu- underutilized since they account for about half minum decreases formability and can acc- the world’s production of stainless, but there are elerate (formation, so their use involves many applications in which more expensive trade-offs. austenitic stainless steels are used needlessly. Alloys with more than 20% chromium are Ferritic stainless steels are a viable alternative used specifically for high-oxidation or corrosion to nickel-bearing austenitics when thickness is 2 resistance. Despite the relative lack of high- mm or less and drawing and bending instead of temperature strength, these alloys are particu- stretch forming is permitted. There are many larly useful because of their high-oxidation re- applications where the longer corrosion life of sistance, which they derive from the tight low-chromium ferritics should economically re- adherence of their oxide scale. The close match place carbon steel, as they have in automotive between the thermal expansion coefficient of exhaust systems. There are no technical barriers the scale and the alloy prevents spallation of the to obtaining these savings; design engineers oxide, which would lead to breakaway oxida- need to learn how to use these alloys. tion. This was the purpose of the earliest high- chromium ferritic stainless, 446. The performance of 446 has been exceeded by lower alloyed Metallurgy of Ferritic Stainless Steels grades, such as the aluminum-alloyed ferritics. A prime example of the state of the art is 453, Chromium stabilizes the ferritic structure at which has not only 22% chromium and 0.6% high temperatures. Thus, above about 11% aluminum but also rare earths in trace amounts chromium, austenite does not exist at any tem- (i.e., 0.1%). perature in pure iron chromium alloys, as seen As in austenitic alloys, rare earths act as very in Fig. 1. powerful oxide and sulfide formers that concen- However, iron-chromium alloys devoid of trate at the metal-oxide interface and stabilize it, carbon are not practical, so early metallurgists again preventing spallation. This type of alloys saw the diagram shown in Fig. 2 with the level finds use in high-temperature applications such as planar oxide fuel cells. The high-chromium alloys, when used for corrosion resistance, are usually called superferritics. In the 1960s, E-Brite¨ was developed. To obtain high toughness, it was vacuum refined to very low carbon plus nitrogen levels. It was followed by the more capable 29-4¨. Later, this alloy was stabilized and became the still-popular AL 29-4C¨. (E-Brite now has a new life as a fuel cell material based on its oxidation resistance and very low thermal expansion coefficient.) These alloys saw success as replacements for 316L when SCC was a problem. This alloy and its close neighbor SeaCure¨ are used primarily in tubing where corrosion resistance is most important. It was developed for welded condenser tubing where seawater or brackish water is involved. It is also used in heat ex- Fig. 1 Iron-chromium phase diagram from Thermocalc 114 / Stainless Steels for Design Engineers

of carbon at 0.20%, which represented the pu- sufficiently long homogenization anneal at a rity level attainable in arc furnace refining. Car- low enough temperature that carbon and nitro- bon is essentially insoluble in ferrite at ambient gen have very little solubility. This is standard temperatures, and carbides of chromium and in the processing of unstabilized ferritic stain- iron will form to the extent carbon is available. less steels, such as 430. Rapid cooling of unsta- Since carbon diffuses interstitially much more bilized alloys causes carbon and nitrogen to pre- rapidly than chromium can substitutionally, cipitate within grains. This severely embrittles chromium is combined in situ, especially along the material and does not avoid sensitization. grain boundaries, which are fast-diffusion This is called high-temperature embrittlement paths. This locally depletes chromium, and the because it comes from putting carbon and nitro- alloy is sensitized. This can be eliminated by a gen into solution at a high temperature and then causing it to precipitate in a harmful manner. These alloys were only ferritic at room temperature if they were given a subcritical anneal to transform austenite to ferrite. Otherwise, at room temperature they would be ferrite plus martensite. There are alloys that are intended to use a mixed ferrite/martensite structure, but they are treated later as a variation from the normal ferritic alloys. The introduction of AOD refining permitted much lower levels of carbon, as seen in Fig. 3, opening the door for fully ferritic stainless steels. Carbon and nitrogen added together produce about the same effect as carbon alone. So, unstabilized fully ferritic alloys are not feasible below 20% chromium without extreme refining techniques, such as electron beam refining, which are not commercially viable for low-cost Fig. 2 Iron-chromium phase diagram at 0.20% carbon alloys. Thus, nearly all modern ferritic alloys

Fig. 3 Iron-chromium diagram at low carbon levels Source: Ref 2 Chapter 8: Ferritic Stainless Steels / 115

are “stabilized.” This means that a strong car- factors prevented TiC formation. The latter ef- bide former such as titanium or niobium is fect was real in early austenitic alloys, such as added in sufficient quantity to combine with all 321, leading to knife-line corrosion attack after the carbon plus nitrogen, removing them from welding, but does not exist in low interstitial solution. ferritic alloys, which have much greater diffu- These reactions are simply: sion rates than austenitic alloys. But, since carbon mobility is quite high, it is not practical to Ti + C = TiC (Eq 1) quench alloys quickly enough to prevent carbide precipitation as is possible in austenitics Ti + N + TiN (Eq 2) (detailed in the Chapter 6, “Austenitic Stainless Nb + C = NbC (Eq 3) Steels”). Figure 4 shows the time-temperature- transformation (TTT) curve for an unstabilized Nb + N = NbN (Eq 4) 430-type alloy with carbon plus nitrogen of Titanium is the stronger getter for carbon and 0.08% (Ref 3). nitrogen. The thermodynamic driving force for Stabilization causes nonchromium carbides to carbide and nitride formation is given by form at high temperatures, precluding chromium carbide precipitation. The net effect is that −7700 modern stabilized ferritic alloys behave as inter- ()()Ti C = + 275 . (Eq 5) T stitial free and can be mapped using the pure iron-chromium diagram shown in Fig. 1. The rate of diffusion of carbon in ferrite is −15790 around 100 times greater than that of carbon in ()()Ti N = + 540 . (Eq 6) T austenite. The solubility of carbon in ferrite is vastly lower than it is in austenite. Because of It must be noted that titanium has an even these factors, the heat treatments to avoid sensi- higher affinity for oxygen and sulfur than for tization are essentially reversed. Carbon in carbon, so that the removal of carbon from so- austenite can be retained in supersaturation for lution is preceded by the removal of oxygen, extended periods of time. This is why austenitic nitrogen, and sulfur in that order. This will be L grades do not sensitize even though they are seen to have a major influence on corrosion re- slightly supersaturated. Sensitization occurs sistance as the MnS inclusions generally asso- at higher levels of carbon by prolonged heating ciated with the initiation of pitting are not at 600 to 850 ¡C. In ferritics, carbon cannot found in titanium-stabilized grades of nor- be kept in supersaturation even by the most mally low sulfur. In practice, the removal of rapid quenching, and sensitization is alleviated oxygen begins in the molten state with the for- by prolonged heating in the 600 to 850 ¡C mation of titanium sulfide and nitride and next range to allow chromium to equalize where in the molten or solid state, depending on con- carbide precipitation has previously made it centrations. It is desirable to keep sulfur and inhomogeneous. nitrogen low enough that precipitation is in the solid state so that precipitates do not agglom- erate and cause large primary inclusions that become unsightly surface defects. TiCS forms in the solid state if sulfur is present; if not, TiC forms. Essentially all carbon is removed from solution below 1250 ¡C if carbon and nitrogen are kept as low as possible and a stoichiometric amount of titanium is available (i.e., greater than about four times the carbon plus nitrogen). The stabilization formula in various specifications is more than four times the carbon plus nitrogen because experimentally it has been found that sometimes understabilization occurs. This is due to the influence of oxygen and sulfur having prior compound formation with Fig. 4 430 time-temperature-transformation (TTT) curve. K, the titanium and less importantly that kinetic carbideSource: Ref 3 116 / Stainless Steels for Design Engineers

Ferritic alloys, like austenitic alloys, can form temperature with the same structure as ferrite intermetallic phases. The most prominent is σ, but with the chromium and iron atoms in an or- which can be seen to form in higher-chromium dered bcc matrix in which iron and chromium stainless steels (i.e., those with chromium plus occupy sites equivalent to two interlocking sim- molybdenum of 20% or more). Formation of σ ple cubic matrices. Because the lattice so occurs when such alloys are held between 500 closely matches that of ferrite, the precipitate is and 800 ¡C; it is a hard, brittle tetragonal phase coherent and causes hardening. The α' embrit- with equal parts iron and chromium. Thus, its tlement causes an extreme loss of toughness as formation causes chromium depletion of the ad- well as hardening. It also causes a loss in corro- joining ferrite. Formation requires substitutional sion resistance via the chromium depletion of diffusion of chromium so is slower to form than that part of the matrix that surrenders chromium carbides, minutes rather than seconds. Since to the α'-phase. cold work enhances substitutional diffusion, it Figures 5 and 6 show the hardening effect of accelerates σ formation. The σ forms preferen- α' and the resulting loss of toughness, respec- tially along grain boundaries for diffusion rea- tively (Ref 6). sons, and this causes it to have a major embrittling effect. The σ may be redissolved by solution annealing, but regaining full homo- Mechanical Behavior geneity is not immediate. Another embrittling phenomenon is the for- Ferritic stainless steels are quite similar in mation of α'. This was named 885 ¡F or 475 ¡C their mechanical behavior to carbon steel. The embrittlement before its cause was understood. main influence of chromium is to produce some Before the nature of α' was known, it was con- solid solution hardening. Let us review the fused with temper embrittlement, which occurs strengthening mechanisms of bcc iron. Pure in martensitic alloys at the same temperature. iron is an extremely soft material with a yield Temper embrittlement is the segregation of strength well under 10,000 psi. This softness is phosphorus to prior austenitic grain boundaries not seen in practice because steel is never pure. and does not occur in fully ferritic alloys. The α' Carbon has an extremely powerful effect on is the ordered equiatomic chromium iron phase hardening, as does nitrogen. that forms by spinodal decomposition; it has the The influence of substitutional alloying ele- same composition as σ but exists at a lower ments is also quite significant. According to

Fig. 5 Inﬂuence of α' formation on hardnessSource: Ref 4 Chapter 8: Ferritic Stainless Steels / 117

Fig. 6 Inﬂuence of α' formation on toughness Source: Ref 5

Paxton (Ref 7), the misfit of solute atoms causes lattice strains proportional to the amount dissolved and provides strengthening through the lattice friction term. This mechanism also increases the impact transition temperature unfa- vorably. Elements that produce a refining of grain size are the exception to this general rule in carbon steel, but the lack of an austenite-to-ferrite transformation in stabilized ferritic stainless steels negates this benefit for them. Figure 7 shows that fairly common ingredients and impurities have strong hardening effects (Ref 6). Man- ganese and silicon are normally deoxidizers, but in titanium-stabilized alloys, titanium takes over the deoxidizing role so their presence can be lim- Fig. 7 Influence of substitutional elements on hardness of ited. Phosphorus is virtually impossible to refine iron alloys from stainless steel, so its presence at around 0.02% is normally a given unless low-phospho- The softest ferritic stainless alloys are the rus raw materials are used as a starting point. 409 variations made for highly formed exhaust The worst toughness-inhibiting effects come system components. They contain as little from interstitial elements to grain boundaries: manganese, silicon, nickel, and other substitu- oxygen, carbon, and nitrogen. The effect of car- tional elements as possible and have a mini- bon plus nitrogen on transition temperature is mum of carbon plus nitrogen, so that the re- profound, as seen in Fig. 8 (Ref 8). sulting precipitate fraction after titanium Stabilizing removes the interstitial carbon and addition is as low as possible. To maximize nitrogen, along with oxygen and sulfur, from softness and formability, titanium and niobium solution. This does not produce a major soften- in excess of that required for stabilization must ing, however, because the precipitate itself has a also be minimized as they will cause solid so- hardening effect. lution hardening. 118 / Stainless Steels for Design Engineers

Fig. 9 Corrosion of titanium-stabilized 29% Cr plus 4% Mo alloys in ASTM A 763 Y test. Source: Ref 11

Fig. 8 Inﬂuence of interstitial carbon and nitrogen on toughness transition temperature Source: Ref 8

Stabilization

Stabilization is essential to ferritic stainless steels to avoid the precipitation of grain bound- Fig. 10 Corrosion of niobium-stabilized 29% Cr plus 4% Mo ary carbides. Combined carbon plus nitrogen alloys in ASTM A 763 Y test. Source: Ref 11 levels below 100 ppm are necessary to avoid both sensitization and embrittlement, but with- As titanium and niobium are added to alloys, out proper heat treatment even alloys of this pu- their corrosion resistance is improved (Figs. 9 rity can incur debilitating loss of toughness due and 10) (Ref 11). Maximum improvement in to carbide and nitride precipitates (Ref 9). These corrosion resistance levels off once full stabi- levels are not economically attainable for com- lization is reached. Excess amounts of the stabi- mercial alloys, so stabilization is the correct en- lizing elements have negligible effect, but tita- gineering answer. nium-stabilized alloys have a lower rate of Stabilization is generally considered as the corrosion than niobium-stabilized alloys. This is simple gettering of carbon and nitrogen by a probably due to titanium’s ability to eliminate suitable carbide and nitride former. It was not sulfur and oxygen from solution. known until about 1980 just what the mecha- Toughness improves for niobium-stabilized nisms of embrittlement were in the ferritic alloys up through full stabilization and then be- stainless steels, however. The distinguishing of gins to decline. This is a result of excess stabi- α' from those related to interstitials and their lizing alloy acting as a solid solution hardener stabilizers (Ref 10) permitted stabilizing ele- and therefore a toughness reducer. This tough- ments to be optimized. ness reduction is more pronounced with tita- Titanium combines with carbon and nitrogen nium, which is a stronger solid solution hard- stoichiometrically by: ener (Figs. 11 and 12) (Ref 11). The upshot of this understanding was the in- Ti=×+434 C. × N (Eq 7) troduction of dual stabilization, through which both weld and base metal toughness and corro- Niobium requires a greater weight percentage: sion resistance are optimized. The same study recommended that dual stabilization follow the =×+× Nb77.. C 66 N (Eq 8) following formula: Chapter 8: Ferritic Stainless Steels / 119

()()Ti+≥×+ Nb6 C N (Eq 9)

The toughness of these alloys has a broad optimum that takes advantage of the corrosion-resisting benefits of titanium (Fig. 13) (Allegheny Ludlum). Other strong carbide formers such as zirconium and vanadium are ineffective stabilizers because their mobility at the temperatures at which they are thermodynamically capable of forming sufficiently large percentages of car- Fig. 11 Charpy V-notch impact ductile to brittle transition bides and nitrides is too low to rid the matrix of temperature (DBTT) of titanium-stabilized 29%Cr these elements. They also have too great a ten- plus 4%Mo alloys test. Source: Ref 11 dency to form intermetallic compounds. Toughness in ferritic stainless steels is a major consideration. If ferritic alloys enjoyed the same toughness as austenitic alloys, there would be few instances when the use of the much more expensive nickel-bearing grades would be justified. Because stabilized alloys are ferritic at all temperatures, there is no automatic grain-refining transformation as exists in carbon steel. If grains grow large from annealing at high temperatures or welding, then the transition temperature increases. Section size also has an effect. Stabilized ferritic stainless steels are seldom used in thicknesses of Fig. 12 Charpy V-notch impact ductile to brittle transition over several millimeters because of decreasing temperature (DBTT) of niobium-stabilized 29%Cr toughness. Figure 14 shows how transition plus 4%Mo alloys test. Source: Ref 11

Fig. 13 Toughness of dual-stabilized low-alloy ferritic stainless. AL 466 is recognized as S40930 120 / Stainless Steels for Design Engineers

tion. But, the anisotropy does result in remarkable drawing characteristics, with ferritic stainless steels with elongations in tensile tests in the mid-30% range being nearly equal to austenitic stainless steels with over 50% elongation. The measure of anisotropy is the Lankford ratio. It is expressed as:

rrr++2 R = 04590 (Eq 10) 4

When this expression equals 1, then a material is isotropic. As the value increases from 1, the drawability increases, as measured by the Fig. 14 Change in transition temperature with thickness for 29Cr-4Mo-2Ni alloy. Source: Ref 12 limiting drawing ratio (LDR), the ratio of the diameter of a disk to that of the deepest cylinder temperature can increase with thickness (Ref into which it can be drawn. The ferritic stainless 12). This effect is due simply to stress states steels in sheet form have LDRs of around 2.2 transitioning from biaxial to the more embrit- compared to 2.0 for 304. tling triaxial with increasing thickness. The good formability of ferritic stainless steels has some drawbacks. They are subject to ridging, which is the formation of visible ridges Texture and Anisotropy parallel to the direction of elongation. This is an artifact of texture in the material. A combination The deformation of ferritic bcc materials is of careful chemistry design and thermomechan- characterized by limited slip systems, high ical processing is required to keep it under con- stacking fault energy, and lattice anisotropy. trol. The approach centers on variables that in- So, when ferritic stainless are deformed, dis- crease stored energy from deformation to locations tend not to dissociate as they do in promote recrystallization over recovery during austenitic stainless steels. The lack of dissoci- annealing. ation of dislocations encourages cross slip. The ferritic stainless steels even carry for- This minimizes dislocation tangles and work ward some of the preferred grain orientation hardening. that come from initial solidification when When ferritic stainless steels are deformed, growth of dendrites is along preferred crystallo- certain crystallographic slip systems predomi- graphic directions. Hot working merely reori- nate, so that large deformations mechanically ents these similarly oriented grains en masse. bring different grains via rotation into closer Without phase changes or enough stored energy crystallographic alignment. This preferred de- to provoke full recrystallization, randomness of formation along easier slip planes results grain orientation is never achieved. macroscopically in overall mechanical proper- Titanium-stabilized steels show more texture ties varying with direction with respect to the and recovery versus recrystallization than do prior deformation. Thus, ferritic stainless niobium-stabilized alloys. This is because tita- steels, like low-carbon steels, have pronounced nium carbides and nitrides form at higher tem- mechanical anisotropy. This is manifest in perature and are therefore coarser. They thus their deep drawing characteristics. Heavily present less obstruction to dislocation motion cold-rolled and annealed ferritic stainless than finer niobium precipitates. Furthermore, steels draw quite well. They resist thinning. niobium precipitates tend to dissociate to a When elongated, they contract in the width di- greater degree than those of titanium. This puts rection while keeping virtually the same thick- niobium in solution during hot working where it ness. This same phenomenon means that they can interact with dislocations. Thus, alloys at cannot be stretch formed since plain strain least partially stabilized with niobium can quickly results in fracture because of the re- achieve greater recrystallization, which can sistance to deformation in the thickness direc- translate to finer grain size and less anisotropy. Chapter 8: Ferritic Stainless Steels / 121

Boron additions to ferritic stainless steels chromium and molybdenum. Likewise, other result in the formation of grain boundary car- elements can have a negative effect. Any ele- bides, M23(C, B)6. If added to titanium-stabi- ment that can combine with chromium or lized steels, the carbides form on preexisting molybdenum can detract from corrosion resist- TiN particles and result in coarser overall pre- ance by their removal of these essential ele- cipitate arrays since finer, lower-temperature ments from solution. The most notorious of precipitating TiC or TiCS precipitates are at these is carbon, whose tendency to form least partially precluded. The net result is chromium carbides causes areas around such coarser grain size and no major improvement carbides to be partially depleted of chromium. in mechanical properties over the use of tita- However, nitrogen, oxygen, and sulfur can also nium alone (Ref 13). Additions of boron to form chromium compounds and cause localized niobium-stabilized steels does cause finer pre- loss of corrosion resistance. Manganese sul- cipitates and grain size than would niobium fides, for instance, are almost always seen to be alone (Ref 14). the locus of pitting corrosion (Ref 15). More careful examination has shown that such sulfides grow in the solid state as chromium/man- High-Temperature Properties ganese sulfides and deplete their very close sur- roundings of chromium, inviting corrosion to begin at the inclusion-matrix interface, where High-temperature mechanical properties of chromium levels in solution are reduced (Ref ferritic stainless steels are often important to 16). their successful use because their oxidation re- Other factors that lead to loss of localized cor- sistance is excellent and better than austenitics, rosion resistance are the formation of but their high-temperature strength is lower chromium-rich phases such as α' and σ. Either than that of austenitics. This has led to consider- of these with about 50% chromium will cause able development of high-temperature proper- adjoining ferrite to have lower chromium levels. ties, primarily for the automotive market. Re- Because ferrite has a non-close-packed struc- search has determined that high-temperature ture, diffusion rates, both substitutional and in- strength and creep resistance are best served by terstitial, are about two orders of magnitude stabilizing grain size and having niobium in higher than in austenite. That means that any solid solution. Adding titanium to niobium-sta- deleterious chromium-depleting reaction can bilized steels stabilizes the type of carbide, es- happen more rapidly. Alloys cannot be pecially preventing the formation of the coarse quenched rapidly enough to forestall sensitiza- M C, whose growth decreases strength. The 6 tion, the precipitation of chromium carbides relatively high insolubility of TiC causes this. that depletes grain boundary regions of Niobium is concurrently made available for chromium. Instead, carbon must be neutralized high-temperature solid solution strengthening. by stabilization, or the chromium depletion must be removed by homogenization in long- box anneals. Note that the latter technique is Corrosion and Oxidation Resistance also possible in austenitics but would require annealing for excessively long times, 102 h or Corrosion resistance is chemistry dependent so. rather than structure dependent, so ferritic stain- The ferritic stainless steels are valued for less steels behave just as do other stainless their resistance to SCC. Even in environments steels of the same crucial alloy content. that cause pitting, the normal initiation step for The main alloying elements that provide re- SCC, annealed ferritic stainless steels do not sistance to localized corrosion, general corro- undergo SCC as long as alloying elements such sion, and crevice corrosion are chromium, as nickel, copper, and cobalt are kept below molybdenum, and nitrogen. Since nitrogen is 0.5% in aggregate. Cold work sufficient to raise essentially insoluble in ferrite, it cannot con- their hardness above Rc 20 to 22 can make them tribute to the corrosion resistance of ferritic susceptible to both SCC and its cousin, hydro- stainless steels as it can in austenite. Other al- gen embrittlement. The more highly alloyed su- loying elements, such as copper and nickel, can perferritic alloys are even susceptible to hydro- add to corrosion resistance in special cases, but gen embrittlement in the annealed condition they are of secondary importance compared to (Ref 17). As with martensitic stainless steels, 122 / Stainless Steels for Design Engineers

this susceptibility is a maximum near room tem- 10. J. Grubb, R. Wright, and P. Farrar, “Micro- perature and declines with increasing tempera- mechanisms of Brittle Fracture in Titanium- ture, as opposed to austenitics, which see their Stabilized Stainless Steels,” Special Publi- maximum susceptibility above room tempera- cation 706, ASTM, 1980 ture. This limits these alloys’ ability to employ 11. J. Grubb, Stabilization of High-Chromium cathodic protection safely to Ð0.80 Vsce, at Ferritic Stainless Steels, Proc. Int. Conf. which point corrosion in seawater is, if not Stainless Steels, ISIJ, Chiba, 1991 eliminated, reduced to very low levels (Ref 18). 12. M.A. Streicher, Stainless Steel ‘77, p 27 13. E. El-Kashif, K. Asakura, T. Koseki, and K. Shibata, ISIJ Int., Vol 44, 2004, p REFERENCES 1568Ð1575 14. N. Fujita, K. Ohmura, E. Sato, and A. 1. F.M. Beckett and R. Franks, Trans AIME, Yamamoto, Nippon Technical Report 71, Vol 113, 1934, p 126Ð143 Oct 1996 2. Stainless Steel, Les Editions de Physiques, 15. T. Suter, E. Webb, H. Bohni, and R. Alkire, 1992, p 483 Pit Initiation in I M NaCl With and Without 3. Stainless Steels, Les Editions de Physique, Mechanical Stress, J. Electrochem. Soc., 2003 Vol 148 (No. 5), 2001, B174 4. H.D. Newell, High Chromium Irons, Met. 16. M. Ryan, D. Williams, R. Chater, B. Hutton, Prog., April 1947, p 617Ð626 and D. McPhail, Why Stainless Steel Cor- 5. P.J. Grobner, The 885 ¡C (475 ¡C) Embrit- rodes, Nature, Vol 412, 2002, p 770 tlement of Ferritic Stainless Steels, Metall. 17. J. Grubb, “Hydrogen Embrittlement of Su- Trans., Vol 4, 1973, p 251Ð260 perferritic Stainless Steels,” paper presented 6. Handbook of Stainless Steels, Peckner and at 1984 ASM Int’l Conference on New De- Bernstein, McGraw Hill, 1977, p 5Ð9, 5Ð12 velopments in Stainless Steel Technology, 7. H.W. Paxton, Alloying, ASM, 1998, p 213 Detroit, September 1984 8. H. Abo et al., Stainless Steel ‘77 18. J. Grubb and J. Maurer, “Use of Cathodic 9. J. Grubb and R. Wright, The Role of C and Protection With Superferritic Stainless N in the Brittle Fracture of Fe-26Cr, Met. Steels in Seawater,” paper presented at Cor- Trans. A, Vol 10A, Sept 1979, p 1247Ð1255 rosion 84, New Orleans, April 1984 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 123-135 All rights reserved. DOI: 10.1361/ssde2008p123 www.asminternational.org

CHAPTER 9

Martensitic Stainless Steels

Summary which restricts the temperature and composition ranges over which it is possible to obtain a fully THE SMALLEST CATEGORY of stainless austenitic structure from which to form marten- steels in usage volume is the martensitic stain- site. The presence of ferrite in a martensitic less steels. This is mainly because these alloys structure is detrimental to strength, hardness, are limited in corrosion resistance because of and toughness. Ferrite can appear in the as-cast the necessity of keeping alloy levels low to pro- structure and be formed during austenitizing or duce the martensite structure. Even so, they fill tempering. All the usual concerns inherent in an important niche as a strong, hard, and tough any martensitic alloys are still present; temper alloy of fairly good corrosion resistance and as embrittlement, retained austenite, etc. a strong, stable, high-temperature alloy. Martensitic stainless steels are the most marginally corrosion resistant of all the stainless alloys. The requirement that they be fully austeni- Introduction tizable limits the amount of corrosion-resisting chromium and molybdenum they can contain. Nearly 100 years ago cutlery was first sold in Much of the carbon in them detracts from the ef- Great Britain with a composition of 13% fective chromium content by forming chromium chromium and 0.25% carbon. This was the first carbides. In addition they are always susceptible commercial use of stainless steel and cutlery to stress corrosion cracking (SCC) when their with the same basic analysis is still sold today. hardness exceeds about Rc 22. These limitations The useful alloys of martensitic stainless steel combine to make their excellent properties us- contain from roughly 11 to 18% chromium and able in only mild environments compared to up to 1.0% carbon. Relatively small amounts of other stainless steels. Their high strength and nickel, molybdenum, tungsten, vanadium, and hardness for their relatively low cost ensure their niobium are also added at times for specific pur- place as a very useful engineering material. poses explained in this chapter. Those marten- Table 1 lists the most significant of the sitic stainless steels in which elements such as martensitic stainless steel alloys. The reader copper and titanium are added to produce addi- should be aware that some alloys which are tional hardening through precipitation are dis- quite similar are discussed primarily in other cussed in Chapter 4, “Corrosion Types.” chapters dealing with specifically PH stainless The designers and engineers already familiar steels or primarily ferritic stainless steels. The with martensitic carbon and alloy steels will distinction between martensitic stainless steels find nothing confusing about martensitic stain- and some other stainless alloy families is some- less steels. There is no aspect of martensitic times vague. Nearly all the precipitation-hard- steels that does not apply directly to stainless ening stainless steels are used in the martensitic martensitic steels. The additional concerns one state, but their special hardening mechanism of must have with stainless martensite relate precipitation within a martensitic matrix causes mainly to those that are due to the strong ferri- them to be categorized separately somewhat ar- tizing influence of chromium. Chromium bitrarily. By this conventional logic, some of the strongly promotes the formation of ferrite, martensitic alloys containing molybdenum or 124 / Stainless Steels for Design Engineers ...... Other 0.30 Cu 0.15-0.30V 0.040 N 0.15 V ...... Other 0.06-0.12 2.00 0.03 0.05-0.30 0.030 max 0.15 min 2.50-3.50 W 0.15 min Se 0.75-1.25 W 1.5 Cu -3.00 Cu N Nb N Se N Ni ...... 1.5 1.25-2.50 4.0- 7 0.50-1.00 3.50-4.50 1.00-2.00 0.5 5 4.5 1.5 3.50-5.50 1.80-2.20 4 Mo ...... -3 0.6 0.75-1.25 0.30-0.70 0.30-0.70 0.80-1.20 1.4 1.5- 2 0.50-1.00 0.65 0.6 0.6 1.00 1 2 Cr 11.5-13.5 11.5-13.5 11.5- 11.5- 11.5-13.5 11.5-14.0 11.0-13.5 -14 13.5 13.5 10.5- 12.5 12.0-14.0 12.0-14.0 12.0-14.0 12.0-14.0 14.0-16.0 13.0-14.0 12.7 10.5- 12.5 12.0- 15 12.0- 14 14.5 12.0-14.0 12.0-14.0 12.0 13 13 Si . . . . . 1 1 1 1 1 0.5 1 1 0.75 0.30-0.60 1 1 0.5 1 1 0.6 0.6 1 1 (continued) S . . . . 0.03 0.03 0.03 0.03 0.03 0.15-0.30 0.03 0.03 0.03 0.03 0.01 0.03 0.03 0.03 0.005 0.03 0.06 0.15 min 0.06 0.03 . Mn 1 1 1.5 1 1.25 0.5 1 0.50-1.00 1 1 1.45 1 1 1.5 0.5-1.0 0.50-1.00 1.25 1 1.25 1.25 1 0.45 0.45 C 0.15 max 0.03 0.18 max 0.030 max 0.15 max 0.15 max 0.15-0.20 0.5 0.20-0.25 0.06 max 0.08-0.20 0.03 0.15 max 0.08 0.15 max 0.05 0.05 max 0.15 min 0.15 min 0.15 min 0.50-0.55 0.30 max 0.025 0.025 Form Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought UNS . . S42200 S41000 S41040 S41003 S41400 S41425 S41500 S42023 S42400 S42500 S40300 S41003 S41008 S41600 S41623 S41800 S42000 S42020 DIN 1.4116 Nominal 1.4116 Nominal JFE JFE Nominal Nippon Nominal Alloy 410 410 416 418 4116 420F 420FSe 422 Trinamet HP13Cr-2 Table 1Table Compositions (wt%) of martensitic stainless steels 403 410S 410Cb 412 414 414 mod 415 416Se 420 424 425 425mod HP13Cr-1 NT-CRS Chapter 9: Martensitic Stainless Steels / 125 ...... Other 0.010 N 0.20-0.30 V 0.015 N 0.010 N ...... Other 0.5 Cu 0.15 min Se 1.2 V 5.5 V 9.0 V 1.5 Cu 4.0 V 0.90-1.25 W Ni ...... 5.5 1.25-2.50 1 5.8 2.4 0.75 0.75 1 1 1 1 3.50-4.50 0.50-1.00 Mo . . 2 0.75 0.75 0.6 4 4 2 0.4 1 0.5 0.5 2 0.75 0.40-0.60 4 4 0.15-1.00 0.5 0.40-1.00 0.90-1.25 Cr 11.5-14.0 11.5-14.0 11.5-14.0 11.5-14.0 11.50-14.00 11.00-12.50 11 12 15.0-17.0 16.0-18.0 16.0-18.0 14 14 14 17 13 10.50-12.50 12.3 16.0-18.0 16.0-18.0 16.0-18.0 14.5 14 Si ...... 1 1 0.3 0.3 1.5 1 1 1 1 0.3 0.35 0.65 1.5 1.5 1 1 1 S ...... 0.03 0.03 0.04 0.03 0.03 0.10-0.35 0.03 0.04 0.04 0.020-0.040 0.02 0.03 0.03 . . . . . Mn 1 2 1 0.5 0.45 0.4 1 1 1 1.25 1.25 0.4 1 1 1 0.5 1 0.50-1.00 C 0.01 0.20 max 0.02 0.60-0.75 1.05 2.15 2.2 0.15 max 0.20-0.28 0.01 0.75-0.95 0.95-1.20 0.95-1.20 0.95-1.20 1.15 1.05 1.05 1.45 0.15 max 0.20-0.40 0.20-0.40 0.06 max 0.06 max Form Wrought Wrought Wrought Wrought Cast Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought Wrought PM PM PM Cast Cast Cast Cast Cast Cast UNS J91150 J91153 J91151 J91154 J91650 J91540 J91422 S43100 S44002 S44003 S44004 S44020 S44023 Nippon JFE Nominal Nominal JFE Nominal Nominal Nominal Nominal Nominal Nominal Nominal Nominal (continued) NT-CRSS KL-HP 431 440A 440B 440C 154 CM CA-40F CA-28MVW Table 1Table Compositions (wt%) of martensitic stainless steels Alloy KL-12Cr 12Cr 440F 440FSe BG-42 ATS-34 14-4 CrMo CPM S30V CPM S60V CPM S90V CA-15 CA 15M CA-40 CB-6N CB-6MN 126 / Stainless Steels for Design Engineers

tungsten should also be considered precipitation- on temperature. The amount is given by the hardening alloys, but they customarily are not Koistinen and Marburger equation (Ref 1): and will not be in this work. 1−=VMsTexp{β ( − )} The ferritic alloys often have compositions α′ (Eq 1) that allow them to be partially martensitic under some conditions. 430 (UNS S43000) and The martensite is coherent with the parent 3CR12 (UNS S41003) can contain some austenite and resembles the passage of slip dis- martensite if their heat treatment is such that locations through the crystal. The sum of many austenite is allowed to form and is followed by such dislocations is shear, and this can be rapid cooling. Even 409 (UNS S409XX) can macroscopically visible as in Fig. 1. form some austenite if chromium is at the high The formation of martensite is essentially me- end of its possible range and nickel and man- chanical (i.e., via deformation, not diffusion). ganese residual levels are high. The martensitic The shear and volume expansion, about 4%, alloys themselves can be made to be partially which accompanies the transformation, in- ferritic by forcing their carbon contents to low volves a great deal of strain energy that must be levels as is customarily done with 410S (UNS taken into account. This is shown diagrammati- S41003). Not understanding these alloys can cally in Fig. 2 (Ref 2). lead to unexpected consequences in mechanical properties or corrosion performance.

Martensite Formation

Martensite as a phenomenon deserves a brief review. Martensite forms as result of the diffusionless transformation of austenite. The austenite may be supersaturated with carbon or nitrogen, but that is not necessary for the transformation. The driving force for the transformation is simply the much lower free energy of the ferrite phase over the austenite phase, which can be attributed largely to large mutual repulsion between iron atoms that possess unpaired outer electrons with the same quantum number and magnetic polarity. This free-energy differential increases with decreasing temperature. At Fig. 1 Martensite platelets emerging from the surface. a certain temperature, the martensite start tem- Source: Ref 2 perature Ms, the transformation occurs spontaneously via the coordinated movement of atoms in a shearing-type mode at very high speeds approaching the speed of sound in the material. The composition of the martensite is identical to that of the parent austenite. There is regularity to the relationship between the parent austenite and the martensite. Greninger and Troiano determined that the close-packed planes of the austenite {111} varied from the {011} of the martensite by only 0.2¡. Further, the direction of the <101 bar> of the austenite was only 2.7¡ from the <1 bar 11 bar> of the martensite. These relationships de- ﬁne the habit plane that constitutes the austenite martensite boundary. Martensite forms essentially independent of Fig. 2 The martensite reaction ab contrasted to the nucleation time and the fraction transformed depends only and growth-type transformation of austenite to ferrite, ac Chapter 9: Martensitic Stainless Steels / 127

This energy differential between ferrite and stitial sites and forms various carbides, leaving martensite is stored in the high-strain energy the parent martensite less strained, softer, and matrix. Applied strains affect the transforma- tougher. Figure 4 shows that the large strain en- tion. Indeed, metastable austenite can readily be ergy in martensite varies with the carbon con- transformed to martensite by deformation. tent, and Fig. 5 (Ref 4) shows how hardness However, the untransformed austenite is hin- varies with carbon content. dered from transforming by the compression it Nitrogen behaves similarly to carbon in both receives from the already-formed martensite. austenite and martensite, but its solubility is Thus, some residual austenite is commonly lower, and it is less signiﬁcant as an alloying el- found between lathes of martensite. ement accordingly. Hydrogen and boron, as in- At the Ms temperature, the body-centered terstitials, also raise hardness. cubic (bcc) phase becomes preferable energetically, but this temperature is too low for diffusion transformation, and a slight shear in the Phase Structure austenite lattice causes a rearrangement of the atoms from a face-centered cubic (fcc) to a dis- Figure 6(a) to (h) shows a series of photomi- torted bcc structure. The amount of distortion is crographs of various martensitic alloys (Ref 5). proportional to the amount of carbon in the in- A stainless martensitic alloy should have the terstices of the structure. These interstices are following characteristics: considerably smaller in the bcc structure even though it is expanded from the fcc. The octahe- ¥ It must have at least 10.5% chromium to dral sites change from 2.86 by 3.56 A to 2.86 by qualify as stainless and even more for better 2.86 A, as shown in Fig. 3 (Ref 3). The distor- corrosion resistance. tion is accommodated by accommodation from ¥ It should be fully austenitic at some temper- site to site at low carbon levels, but above about ature. 0.018% carbon this can no longer be accommo- ¥ The temperature at which austenite forms on dated and a tetragonal distortion occurs (Ref 3). heating should be sufficiently high to permit The carbon is in a state of supersaturation in the tempering above the temper embrittlement as-formed martensite. When the martensite is range. tempered, the carbon diffuses from these inter- These criteria are somewhat challenging. Fig- ure 7(a) shows that at low-carbon (0.05%) levels austenite is stable up to about 12% chromium, above which some δ-ferrite tends to be stable at all temperatures below the melting point. Increasing carbon slightly expands the chromium level at which full austenitization can occur (Fig. 7b) (Ref 3).

Fig. 3 Change in size of the octahedral interstitial site with the change from face-centered cubic (fcc) to body- Fig. 4 Strain energy of martensite dependence on carbon centered cubic (bcc). Source: Ref 2 content. Source: Ref 2 128 / Stainless Steels for Design Engineers

Fig. 5 Variation in martensite hardness with carbon content The interplay between chromium and carbon The variety of martensitic stainless steels is further explained in Fig. 8(a) and (b), in would be very limited if only chromium and which it becomes clear that for higher- carbon were available as alloying elements, but chromium alloys the range over which full fortunately nickel again can make an important austenitization can occur is further restricted. contribution. Nickel greatly expands the Chapter 9: Martensitic Stainless Steels / 129

Fig. 6 (a) Annealed 410 showing carbides within an equiaxed ferrite matrix. (b) 410 quenched and tempered. (c) 416 quenched and tempered: white ferrite and gray sulfides in a martensite matrix. (d) 420 quenched and tempered showing fine carbides in a martensite matrix. (e) 420 quenched and tempered showing surface decarburization. (f) 440A annealed displaying primary and smaller secondary carbides in a ferrite matrix. (g) 440B quenched and tempered displaying both primary and secondary carbides. (h) 440C quenched and tempered displaying significant primary carbides plus finer secondary carbides in a martensite matrix 130 / Stainless Steels for Design Engineers

Fig. 7 Iron-chromium phase diagrams at two low-carbon levels

1800 1800

1600 Liquid 1600 Liquid

L + α 1400 1400 σ α + γ γ γ 1200 1200 α + γ γ + carbide

Temperature, ° C Temperature, 1000 ° C Temperature, 1000 γ + carbide

800 800 α + carbide α + carbide 600 600 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Mass, %C Mass, %C

Fig. 8 (a) Iron-chromium phase diagram at 12% chromium; (b) iron-chromium diagram at 17% chromium

chromium levels and temperatures at which amount of total alloy that can be used and in the austenite is stable as is shown in Fig. 9 (Ref 3). end puts an upper limit on the ability of marten- Table 2 quantifies the influences of the vari- sitic stainless steels to achieve high corrosion re- ous possible alloying elements on the key prop- sistance. This is because as the main corrosion erties of martensitic stainless steels. fighters, chromium and molybdenum, which are It can be seen that the elements that promote ferritizers, are increased, so must austenitizers austenite, with the exception of cobalt, all de- such as nickel. The coordinated increase in these press the Ms temperature. This puts a limit on the elements lowers the Ms to such a degree that the Chapter 9: Martensitic Stainless Steels / 131

alloys become stably austenitic before much Austenitizing is complicated in martensitic higher corrosion resistance is obtained. stainless steels because many grades contain The ability to temper without austenite rever- carbon at levels intended to produce carbides sion is an important trait. Obviously, if transfor wear resistance purposes. Since carbon solu- forming martensite to austenite during temper- bility varies strongly with temperature at ing caused subsequent untempered martensite austenitizing temperatures (Fig. 7a and b), con- or other undesirable phases, this would limit trol of temperature is vital to have the correct one’s ability to temper at a high enough temper- balance of carbon in solution versus carbon as ature to achieve desired toughness. This limits carbide since carbon in solution has such a the use of nickel while encouraging the use of strong influence on ferrite content, Ms, and me- elements like molybdenum. chanical properties. Copper has become an important alloying el- Austenitizing temperature also determines ement in martensitic stainless steels because it austenite grain size. This affects Ms, but more greatly improves corrosion resistance in certain importantly it influences subsequent toughness. environments without diminishing an alloy’s Phosphorus precipitates at prior austenite grain ability to be tempered. boundaries during tempering with a maximum effect at 475 ¡C. This is the infamous temper embrittlement. Figure 10 (Ref 3) shows the sig- Thermal Processing nificant toughness change that occurs as increasing austenitizing temperature increases The main concerns with processing marten- austenite grain size and permits greater phos- sitic stainless steels are austenitizing, quench- phorus concentrations at grain boundaries. ing, tempering/stress relieving, and annealing. Refining phosphorus from any chromium- containing steel is quite challenging thermodynamically, so achieving low phosphorus levels depends mainly on restrictions on raw materials for melting. Because this is difficult or costly, grain size control is the main tool for controlling temper embrittlement. The higher-carbon grades, those above 0.20% carbon, should be heated gradually through stage heating to avoid cracking due to thermal stresses. Soaking at 800 ¡C until uniform temperature is achieved minimizes this risk. Another concern during austenitizing is superficial carbon loss, an example of which is shown in Fig. 6(e). Heating in air to 1050 ¡C can cause surface carbon to decrease by approximately 0.10% per hour, resulting in much lower surface hardness. This loss increases with base carbon level and austenitizing temperature. Carbon or nitrogen pickup could also occur if the atmosphere was rich in these elements. The carbon potential of the furnace atmosphere must be controlled to avoid potentially serious prob- Fig. 9 The expansion of the range of austenite stability with lems. If hydrogen atmospheres are used the nickel content

Table 2 Inﬂuence of alloying elements on ferrite, Ms, and austenite start

Element N C Ni Co Cu Mn Si Mo Cr V Al Lowering of % Ð220 Ð210 Ð20 Ð7 Ð7 Ð6 6 5 14 18 54 ferrite per % element Lowering of MS Ð475 Ð475 Ð17 0 to 10 Ð17 Ð30 Ð11 Ð21 Ð17 Ð46 . . . per % element Change of AC 0 to 280 0 to 250 Ð30 to Ð115 0 0 Ð25 to Ð66 25 to 73 25 to 60 0 to 35 50 to 290 30 to 750 per % element 132 / Stainless Steels for Design Engineers

+120 P: 0.047% [ ] P: 0.035%

+80 [ ] P: 0.021%

+40

P: 0.003-0.004% [ ] 0 [ ] Transition temperature, ° C temperature, Transition [ ]

−40 [ ]

−80 5 10 20 50 100 200

Austenite grain size, μm

Fig. 10 Inﬂuence of austenite grain size and phosphorus level on toughness

danger of embrittlement after quenching must as the transformation of M3C into M7C3 can re- be recognized. Stress relief without delay would sult in a secondary hardening, a true precipita- be mandatory. tion-hardening effect. In the presence of strong The high chromium content of these alloys carbide-forming alloying elements such as renders them very deep hardening. Air harden- molybdenum, vanadium, and tungsten, the M2X ing is generally sufficient. Oil quenching which carbide can become the more stable species and is faster may be slowed by heating the oil. be responsible for the secondary hardening. At ¡ Avoiding quench cracking and excessive 500 C, coarser M23C6 and M7C3 begin to grow warpage is almost always a greater concern than at grain boundaries. This is accompanied by a depth of hardening so air quenching is standard. pronounced softening. The hardening reduction Because the quenching and the transforma- with stress relief and tempering for a 12% Cr tion it causes are inevitably accompanied by alloy is shown in Fig. 11 (Ref 6). residual stresses in a brittle material, stress re- Separately at the 475 ¡C range, the previously lieving should be immediate to avoid cracking. mentioned phosphorus segregation to prior Higher-carbon grades should not even be al- austenitic grain boundaries occurs. This effect lowed below room temperature before stress begins to disappear above 550 ¡C. Thus true relief. Pickling should never be done on as- tempering is conducted above this temperature. quenched material because this could easily re- The microstructural changes at these tempera- sult in hydrogen uptake and delayed cracking tures are the above-mentioned loss of carbon by hydrogen embrittlement. from solid solution, carbide precipitation and Heating as-quenched material to between 150 coarsening, and, of course, stress relief. The re- and 400 ¡C produces stress relieving. Besides sult is a pronounced softening and toughening. the normal ﬂow on a microscopic scale, which If the material contains retained austenite, it we understand as stress relieving, there is a may decompose to ferrite and carbide with a slight growth in the number of ﬁne cementite negative effect on toughness. particles and a corresponding decrease in the The molybdenum, vanadium, and tungsten- amount of carbon in solid solution. This results alloyed grades will resist softening during tem- in a slight decrease in hardness. At 400 ¡C, a pering because of the strength of the secondary further precipitation of M2X and M7C3 as well hardening they undergo due to precipitation Chapter 9: Martensitic Stainless Steels / 133

500 Initial hardness

450 300°C 300°C 350°C 350°C 400°C 400°C ° 400 450 C 450°C 500°C 550°C 350 600°C 650°C 300 700°C Hardness 750°C 500°C 550°C 250 600°C

650°C 200 700°C 750°C

150 11 12 13 14 15 16 17 18 19 20 21 22 23

T (20 + LOG t) × 10−3

Fig. 11 Influence of tempering on hardness hardening of carbides and nitrides. Nickel most sophisticated martensitic stainless alloys. seems to amplify this action by its influence on While 420 is the common alloy and is quite diminishing the solubility of carbon in the ma- serviceable, much more wear- and corrosion-re- trix. Thus, the tempering of the higher-alloy sistant alloys exist. At one time, 440C was the martensitic stainless steels can truly be consid- maximum step up from 420; however, further ered a precipitation-hardening reaction. alloying with molybdenum for corrosion resist- The higher-carbon, higher-chromium grades ance and vanadium for hardness of the carbide are typically only stress relieved because the re- phase has led to improvements. The wear resist- moval of chromium from solution by carbide ance of a blade is largely determined by the formation at higher temperatures causes an un- hardness and amount of carbides while the acceptable loss of corrosion resistance. toughness is governed by the matrix properties. These alloys are used at very high hardness levels, so cleanliness is very important to tough- Applications ness, which measures the ability to withstand chipping in use. Electroslag remelting (ESR) or High-Temperature Use. The basic 12 % Cr vacuum induction melting-vacuum arc remelt- martensitic alloy has been the basis of alloying ing (VIM-VAR) provides the cleanliness re- improvements that were done to produce better quired, while powder metallurgy is optimal for high-temperature performance, especially for obtaining very fine carbide size and uniformity. turbines. The nominal analyses of some prominent The addition of vanadium and niobium, both grades are shown in Table 3. of which form much more stable carbides than The martensitic alloys have a tendency to- chromium, results in alloys that have vastly im- ward centerline segregation during solidifica- proved creep resistance in the 550 ¡C range, as tion as well as toward the formation of primary shown in Fig. 12 (Ref 3). carbides. This has produced limitations in the Tool and Cutlery Alloys. A high-profile use amount of highly wear-resistant constituents of martensitic stainless steels is in cutlery. Hunt- such as vanadium carbide (hardness Rc 75), ing knives, sport knives, and chefs’ cutting tools which can be introduced into the matrix in con- are highly valued items and contain some of the ventional production. Powder metal techniques 134 / Stainless Steels for Design Engineers

4 50 2 45

40 2 35 3 30

1 25 Applied stress, kg/mm Applied stress,

20 1 (0.2C-10.5Cr) 2 (0.2C-10.5Cr-0.1Nb) 3 (0.2C-10.5Cr-0.1V) 16 4 (0.2C-10.5Cr-0.1V-0.1Nb)

31030 100 300 1000 3000 10,000

Rupture life, h

Fig. 12 Inﬂuence of vanadium and niobium on high-temperature properties

Table 3 Tool and cutlery martensitic stainless steels alloy compositions

Alloy UNS Form C Mn S Si Cr Mo Ni Other Other 420 S42000 Wrought 0.15 min 1 0.03 1 12.0-14.0 ...... 4116 DIN 1.4116 Wrought 0.5 ...... 14.5 0.65 ...... 0.15 V Nominal 440A S44002 Wrought 0.60-0.75 1 0.03 1 16.0-18.0 0.75 ...... 440C S44004 Wrought 0.95-1.20 1 0.03 1 16.0-18.0 0.75 ...... BG-42 Nominal Wrought 1.15 ...... 0.3 14.5 4 . . . 1.2 V . . . ATS-34 Nominal Wrought 1.05 0.4 . . . 0.35 14 4 ...... 14-4 Nominal Wrought 1.05 0.5 . . . 0.3 14 4 ...... CrMo 154 CM Nominal Wrought 1.05 0.45 . . . 0.3 14 4 ...... CPM Nominal PM 1.45 ...... 14 2 . . . 4.0 V . . . S30V CPM Nominal PM 2.15 0.4 ...... 17 0.4 . . . 5.5 V . . . S60V CPM Nominal PM 2.2 ...... 13 1 . . . 9.0 V . . . S90V are not subject to the same limitations as contin- good toughness without tempering and mini- uous casters and have alloys the production of mize the loss of chromium to carbides, main- alloys with high volume content of VC. One taining it in solution for corrosion resistance. such alloy is Crucible CPM 90V with 14% Cr, The addition of nickel and molybdenum yields 9% V, 1% Mo, and 2.3% C. This alloy has equal full austenite and martensite transformation and or better toughness and corrosion resistance as improves corrosion resistance. Table 4 lists sev- 440C but has ten times the wear resistance at eral such alloys by JFE: the ﬁrst two can be the same macrohardness. made to meet L80 speciﬁcations and produced Oil Country Tubular Good and Line Pipe. as seamless. JFE reports production of over The need for corrosion resistance in oil produc- 100,000 tons per year of this product (Ref 7). tion tubulars has grown as the quality of petro- The third alloy is a near match for the precipita- leum deposits has become less optimal. Use of tion-hardening stainless Custom 450 (UNS stainless can eliminate for corrosion inhibitors S45000) (Ref 8). Like other precipitation-hard- in H2S and CO2 environments. This has led to enable steels, it shows excellent resistance to the use of low-carbon martensitic stainless SCC at high strength levels. Figures 13 and 14 steels. Low carbon and nitrogen levels give show the improvements in corrosion resistance Chapter 9: Martensitic Stainless Steels / 135

Table 4 JFE Steel/Nippon Steel oil country tubular goods and line pipe alloys

Alloy UNS Form C Mn S Si Cr Mo Ni Other Other HP13Cr-! JFE Wrought 0.025 0.45 ...... 13 1 4 ...... Nominal HP13Cr-2 JFE Wrought 0.025 0.45 ...... 13 2 5 ...... Nominal NT-CRS Nippon Wrought 0.03 1.45 ...... 12.7 1.4 4.5 1.5 Cu 0.040 N Nominal NT-CRSS Nippon Wrought 0.02 2 ...... 12.3 2 5.8 1.5 Cu 0.015 N Nominal KL-12Cr JFE Wrought 0.01 ...... 11 . . . 2.4 0.5 Cu 0.010 N Nominal KL-HP JFE Wrought 0.01 ...... 12 2 5.5 . . . 0.010 N 12Cr Nominal

ronments to about Rc 22 to avoid SCC by the hydrogen embrittlement mechanism. The stainless can resist this failure mode at higher strengths. As specifying bodies such as the American Petroleum Institute (API) approve the use of stainless tubulars at higher strength levels than carbon steel tubulars are safely capable of handling, then the strength improvement, coupled with the orders of magnitude improvement in corrosion resistance, will cause a great increase in their use. Lower carbon levels permit the use of field welds without tempering so that similar alloys can be used for line pipe. These are corrosion Fig. 13 Corrosion rates of stainless versus carbon steel resistant and yet meet X70 and X80 class specifications. These alloys are the last two in Table 2. These uses of martensitic stainless steels for oil production represent possibly the greatest growth area for any kind of stainless steel in the first decade of the 21st century.

REFERENCES 1. D.R. Koistinen, R.E. Marburger, “A General Equation Prescribing the Extent of the Austenite/Martensite Transformation in Pure Iron,” Acta Met, Vol 7, 1959, p 59 2. http://www.msm.cam.ac.uk/phase-trans/ 2002/martensite.html 3. Bletton, Aciers Inoxidables, Les Editions de Physique les Ulis, Paris, 1993, p 481 4. ASM Handbook Desk Edition, 1985, p 28Ð9 5. http://products.asminternational.org/mgo/ Fig. 14 Corrosion rates for stainless oil country tubular goods 6. K.J. Irvine et al., JISI, Vol 195, ISIJ Interna- (OCTG) alloys under severe operating conditions tional 1960, p 386Ð405 over carbon steel L80 oil country tubular goods 7. S. Deshimaru et al., “Steels for Production, (OCTG) under test conditions representative of Transportation and Storage of Energy, JFE difficult real-use environments (Ref 7). Technical Report (No. 2), March 2004 The improvements in martensitic steels for p 55Ð67 these applications are hardly more than a thor- 8. M. Kimura et al., “High CR Stainless ough revisiting of the developments of the OCTG with High Strength and Superior 1950s and 1960s. This does not diminish their Corrosion Resistance,” JFE Technical Re- importance. Carbon steels are limited in sour envi- port (No. 7), Jan 2006, p 7Ð13 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 137-146 All rights reserved. DOI: 10.1361/ssde2008p137 www.asminternational.org

CHAPTER 10

Precipitation-Hardening Stainless Steels

Summary Introduction

THE PRECIPITATION-HARDENABLE (PH) The PH stainless steels exploit the low austen- grades are a highly specialized family of stain- ite stability possible in the chromium/nickel less steels whose existence derives from the stainless steels by making the alloys so lean in need for very high-strength materials with good composition that they can be made to trans- corrosion resistance. The workhorse alloys are form nearly entirely to martensite by thermal the martensitic PH grades, which are used in or mechanical treatment. This martensite can many forms. Primarily used as forgings, bar, then be further hardened by the coherent pre- and other hot-worked forms, they can also be cipitation of intermetallic compounds, ele- obtained in cold-rolled sheet and strip, although mental copper, nitrides, or even phosphides. not with the flatness expected from non-PH This precipitation hardening can also be made stainless. The semiaustenitic alloys are more to occur in a fully austenitic matrix, and this amenable to production as sheet, strip, and wire also provides a commercial PH alloy. But, the and are designed for applications that require martensitic PH grades are by far the more extensive forming before hardening. The fully common. The border between the more highly austenitic PH alloys fill a small niche where alloyed martensitic stainless steels, which un- high mechanical properties are required at tem- dergo secondary hardening during tempering, peratures above or below which the other PH and the PH alloys is indeed vague. Some au- grades are found lacking, when a nonmagnetic thors have astutely treated them as a single material is required, or when the higher thermal group. Here, we treat them separately because expansive coefficient of an austenitic material is they are traditionally considered as separate desired. In no case is corrosion resistance better alloys. than that of normal 304 found in PH stainless The advantage of the PH alloys over the steels. If enhanced strength and very high corro- strictly martensitic stainless steels is that they sion resistance are required, then the designer attain great strength with higher toughness and should look to duplex stainless steels for the op- corrosion resistance than can be obtained timal material. If cost is a greater concern than through the hardening of martensite through corrosion resistance or toughness, then marten- carbon. In addition, they can be fabricated in a sitic stainless steels should be considered for relatively soft state and then hardened with very applications where strength and hardness over little dimensional change. that of annealed ferritic and austenitic stainless The PH grades were developed at the begin- is required. ning of World War II, with Stainless W (UNS The increased use of titanium alloys and ad- S17600) by U.S. Steel generally acknowledged vanced composite materials may occur at the as the first. The later-developed grades are dis- expense of the stainless PH alloys and at the tinguished from the first by their more uniform, same time may create some new niche applica- and therefore tougher, microstructure through tions for them. the elimination of residual δ-ferrite and retained 138 / Stainless Steels for Design Engineers

austenite and by more astute alloy design and movement. As time and temperature of precipi- chemistry control. tation increase, the zones can grow to sizes that The mechanism of precipitation hardening is cannot accommodate the small size differential; parallel to that used to strengthen aluminum al- coherency is lost, and with it the hardening ef- loys in which the precipitation of a coherent sec- fect diminishes. The precipitation has the dual ond phase from a supersaturated solid solution is function of stress relieving the martensite while produced by an aging heat treatment. The coher- further hardening the matrix through the precipi- ent precipitate strains the lattice and impedes the tation of the coherent precipitate. The mechani- motion of dislocations, producing strengthening. cal properties of the final microstructure depend Overaging causes the precipitates to lose co- on the initial strength of the matrix before aging, herency, and softening follows. The precipitate the amount of precipitate, and the coherency of that causes the hardening is normally nickel the precipitate. The ideal microstructure for the (aluminum/titanium) (Ref 1). Figure 1 shows the initial matrix is 100% martensite. To the extent compounds that can form from the precipitation there is δ-ferrite or retained austenite, properties, of supersaturated aluminum and titanium in an especially yield strength and toughness in the iron alloy matrix. transverse direction, are compromised. The It is also possible to produce a hardening reac- aging temperatures can also be high enough that tion by the precipitation of elemental copper reversion of martensite to austenite occurs, (Ref 2). In nitrogen-bearing alloys, a hardening which also lowers subsequent tensile properties. may be produced by the precipitation of Cr2N While the presence of persistent, large bands (Ref 3). The precipitation begins with the diffu- of either δ-ferrite or γ-austenite is undesirable, sion of the precipitating species to sites on the but both also have benefits. The presence of existing matrix. These enriched zones are called some fine bands of δ-ferrite promotes easier and Guinier-Preston (GP) zones. Close dimensional more reproducible precipitation of chrome car- matchup between the precipitating species and bides at the δ/γ interface during the “austenite the parent matrix is required. The differential conditioning” or “trigger anneal” heat treatment should be on the order of a percent. This allows step for semiaustenitic alloys (17-7, AM350, not only coherency but also strain. The coherent etc.). Although bands of stable austenite are precipitate is a effective barrier to dislocation undesirable, it is the presence of residual interlath

γ + + Ni(AlTi) Ni3AlTi Ni(AlTi)

Ni(AlTi) + Ni2AlTi

4 Ni2AlTi

Ni2(AlTi) + Ni3AlTi

Ni2AlTi γ + Ni2AlTi γ + Ni(AlTi) Ni3Ti 3 σ

Limit of austenite ferrite region in solution treated Aluminum, wt% Aluminum, 2 conditions

γ + + Ni2(AlTi) Ni3AlTi γ + Ni3(AlTi) 1 γ γ + + Ni3Ti Ni3(AlTi) Cellular γ + Ni Ti 3 precipitation

0 1 2 3 4

Titanium, wt%

Fig. 1 Possible aluminum/titanium precipitates Chapter 10: Precipitation-Hardening Stainless Steels / 139

austenite that provides the work-hardening abil- strength than that of which the martensitic or ity in many of these PH alloys. It is this work semiaustenitic alloys are capable. The austenitic hardening that gives the PH alloys, especially the PH strength is better above 750 ¡F. semiaustenitic ones, their unusual combination of high strength plus ductility and toughness in the fully hardened state. Martensitic Precipitation-Hardenable The complexity of PH steels comes from the Stainless Steels processing involved in producing the martensitic structure in which the precipitation will occur. The martensitic PH alloys are, as stated, fully The most straightforward alloys are the marten- martensitic at room temperature. Their martensitic, also called the martensitic PH alloys. These site is a relatively soft, low-carbon (less than steels are supplied in the fully martensitic condi- 0.05%) martensite as opposed to the higher car- tion with hardness in the low Rc 30s. This is con- bon found in the martensitic stainless steels. The fusingly called the annealed condition, or condi- early alloys of this type, 17Ð7 PH and 17-4 PH, tion A, even though the matrix is untempered contained up to 10% δ-ferrite stringers, which martensite. After the material is fabricated, it is caused poor through-thickness toughness. This subjected to an aging treatment designated by would be expected from the Schaeffler-Delong the aging temperature in Fahrenheit (e.g., diagram, but this is asking too much of the Scha- H-950). These aging temperatures range from effler-Delong diagram, which was developed for 950 ¡F (510 ¡C) to 1150 ¡F (620 ¡C). welds, to predict the phase composition of alloys A second major group of PH grades is the that have been homogenized by hot working. semiaustenitic. These grades in the normally The inaccuracy of the diagram for more com- furnished condition A are fully austenitic. This plex systems was overcome, and alloys were de- is accomplished by adding elements that lower signed that had minimal δ-ferrite and still trans- the martensite start temperature, such as more formed entirely to martensite, if not at room chromium, molybdenum, and nickel. The temperature, at least at a reasonably attainable austenite is more formable than martensite, and subzero temperature. This was done first by trial it has the possibility of superior corrosion resist- and error and more recently by using thermody- ance because of higher chromium content. This namic computer models, such as ThermoCalc, to is balanced by the need to use either cold work, predict equilibrium phase composition. This de- cryogenic treatment, or a destabilizing anneal to velopment was very significant for making the cause the matrix to become martensitic before alloy family useful as a high-strength/high- its precipitation aging treatment. toughness material for demanding applications Last, if the austenite is made very stable by requiring high mechanical properties and corro- further alloying additions, a precipitation reac- sion resistance. The most advanced PH alloys tion can still be made to occur by the same type are martensitic PH grades by Cartech, Custom of aging treatment without martensite ever form- 465 and 475. ing. The precipitation takes place in austenite Table 1 shows the more significant of these al- and therefore results in lower room temperature loys compared on a strength basis; Table 2 shows

Table 1 Mechanical properties of martensitic precipitation-hardenable alloys

Toughness, Alloy UNS Condition Yield, MPa Tensile, MPa Elongation, % HRC CVN ft-lb Stainless W S17600 H-950 ( 510) 1240 1340 14 42 . . . 17-4 PH S17400 H-925 (495) 1210 1310 14 41 40 15-5 PH S15500 H-925 (495) 1210 1300 15 41 20 H1100 (595) 930 1025 17.50 34 70 13-8 PH S13800 H-950 (510) 1450 1550 12 47 25 H1050 (565) 70 Custom 450 S45000 H-900 (480) 1270 1350 14 42 60 H1100 (595) 460 970 23 . . . 180 Custom 455 S45500 H-950 (510) 1515 1585 10 48 8 H1050 (565) 1205 1310 14 40 25 Custom 465 S46500 H-950 (510) 1650 1765 11 49 13 H1000 (535) 1500 1600 13 48 28 Custom 475 . . . H-975 (525) 1855 2005 5 54 . . . H 1100 (595) 1315 1572 13 48 Ferrium S53 ...... 1565 1985 14Ð16 54 18 140 / Stainless Steels for Design Engineers

Table 2 Composition of martensitic precipitation-hardenable alloys

Alloy Designation C Mn Si Cr Ni Mo Al Cu Ti Other Stainless W S17600 0.1 0.5 0.5 17 6.3 . . . 0.2 . . . 1 P 0.3 17-4 PH S17400 0 0.6 0.6 16 4.3 ...... 3.2 ...... 15-5 PH S15500 0 0.6 0.6 15 4.3 ...... 3.2 ...... 13-8 PH S13800 0 0.1 0.1 13 8.5 2 1.1 ...... Custom 450 S45000 0 0.3 0.3 15 6 0.8 . . . 1.5 . . . 0.3 Nb Custom 455 S45500 0 0.3 0.3 12 8.5 ...... 2.5 1 0.3 Nb Custom 465 S46500 0 0.2 0.2 12 11 1 ...... 2 . . . Custom 465 (275) . . . 0 0.2 0.2 12 11 1 ...... 2 0.2 Nb Custom 475 . . . 0 0.4 0.4 11 8 5 1.2 ...... 8.0 Co Ferrium S53 . . . 0.2 0.1 0.1 10 5.5 2 0 ...... 1 W . . . 0.3 V . . . 14 Co

Fig. 2 Typical microstructures of precipitation-hardenable (PH) stainless steels: (a) 15-5PH as-quenched martensite; (b) 13-8 PH solution treated and aged displaying ﬁne martensite; (c) 17-7 PH displaying ferrite stringers in a martensite matrix; (d) 17-7 PH showing residual ferrite stringers and inclusions their nominal compositions. Figure 2 shows a se- less W to the latest in Custom 475 besides the ries of photomicrographs of PH alloys. elimination of δ-ferrite is in the volume fraction The main advancement metallurgically in of the precipitating phase and the elimination of these alloys from the top, and earliest, in Stain- retained austenite. Alloy designers found that to Chapter 10: Precipitation-Hardening Stainless Steels / 141

reduce δ-ferrite they also tend to stabilize ¥ Molybdenum to offset loss of corrosion re- austenite, which also happens to reduce the sistance by minimization of chromium, to temperature at which martensite forms, MS. increase the temperature at which austenite Last, the higher levels of molybdenum reduce forms, and to form another hardening pre- the tendency to form secondary austenite during cipitate in the presence of cobalt aging. To minimize δ-ferrite requires reducing ¥ Cobalt to stabilize austenite while raising chromium or molybdenum, which also reduces MS corrosion resistance. As a result, as alloy ¥ Aluminum or titanium to form intermetallic strength increases, corrosion resistance is com- precipitates with nickel or copper to precipi- promised. The alloys with the greatest strength tate as elemental copper potential, Custom 465 and 475, barely qualify It is possible to quantify these various influ- as stainless with around 11% chromium. But, ences on phases. This is summarized in Table 3 from a utility point of view, these alloys are de- in terms of the influence of the element on dif- signed to have maximum mechanical properties ferent factors measured in degrees Centigrade with adequate corrosion resistance, so this is for a 12% chromium alloy. viewed as an acceptable compromise. Rapid quenching of these alloys is not re- The newest alloy is Ferrium S53, one of the quired. They are air hardenable. But, the cool- recent alloys designed by computer-assisted ing of these alloys must be completed expedi- thermodynamic calculations. It was designed to tiously through the final stages of martensite replace 300M, 4340, and AerMet 100 on an formation with minimal delay. During delays equal mechanical properties basis but also pro- after the start of martensite transformation has vide the corrosion resistance necessary to be occurred, the remaining austenite tends to stabi- used in aircraft components without cadmium lize, and full transformation to martensite does plating. Its composition superficially seems de- not occur. When this happens, the higher levels ficient in chromium to provide “stainlessness,” of austenite reduce subsequent mechanical but the cobalt level raises the thermodynamic properties after aging. activity of chromium sufficiently that the equiv- As with any alloy used at such high strength alent of 12% chromium in a non-cobalt-contain- levels, microstructural cleanliness is essential, ing alloy is achieved. The precipitation harden- but air melting and argon oxygen decarburiza- ing mechanism is the precipitation of Mo2C. It has been established that this hardening mecha- tion (AOD) refining are quite adequate. nism optimizes resistance to stress corrosion Corrosion Resistance. The martensitic PH cracking (SCC) for a given strength level. stainless steels obey the same rules as other The alloying characteristics of these grades stainless steels with regard to corrosion resist- are: ance. The martensite carries no nitrogen in solution, so the resistance to pitting is given by: ¥ Low carbon, nitrogen, silicon, and manganese because these elements lower MS PREN = %Cr + 3.3%Mo (Eq 1) without contributing to age hardening ¥ Low chromium to suppress δ-ferrite In the martensitic PH alloys, no chromium is ¥ δ Sufficient nickel to suppress -ferrite and rendered ineffective by the formation of Cr23C6 provide for precipitates without excessively since carbon is either held low or stabilized by depressing MS titanium or niobium. Thus, the corrosion

Table 3 Inﬂuence of alloying elements on key transformations

Element N C Ni Co Cu Mn Si Mo Cr V Al Lowering Ð220 Ð210 Ð20 Ð7 Ð7 Ð6 6 5 14 18 54 of % ferrite per % element

Lowering of MS Ð475 Ð475 Ð17 0 to 10 Ð17 Ð30 Ð11 Ð21 Ð17 Ð46 per % element

Change of AC 0 to 280 0 to 250 Ð30 to Ð115 0 0 Ð25 to Ð66 25 to 73 25 to 60 0 to 35 50 to 290 30 to 750 per % element 142 / Stainless Steels for Design Engineers

resistance will equal that of stabilized ferritic al- ments, the austenite could be made stable at loys of the same pitting resistance equivalent room temperature. This would make the alloy number (PREN) for which voluminous data are softer and more fabricable and, most impor- available. tantly, permit them to be manufactured as cold- A greater concern is the risk of SCC in these rolled sheet and strip. If the austenite could then alloys. While the mechanism of SCC in be transformed to martensite by cryogenic treat- austenitic alloys is still debatable, it has long ment, cold work, or special heat treatment, then been clear that, for martensitic alloys, SCC is it could be age hardened just like the martensitic simply a manifestation of hydrogen embrittle- PH grades. This has been accomplished for a ment in which the hydrogen is provided by local group of alloys called the semiaustenitic PH corrosion. The existence of pitting is a suffi- grades. The “semi” signifies that the austenite cient, if not necessary, condition for SCC to in these alloys is metastable rather than stable at occur if the temperature is within the range of ambient temperatures. Also, it should be noted susceptibility and the material is inherently sus- that these semiaustenitic alloys usually contain ceptible. The material susceptibility is largely a some δ-ferrite in their predominantly austenitic function of resistance to crack propagation in microstructure after annealing. any given alloy that is measured by fracture These alloys are complex metallurgically be- toughness. The martensitic PH grades have ex- cause of the technique used to make the austenite cellent toughness and low rates of crack propa- stable at room temperature after a full solution gation, but none should be considered immune anneal. The austenite is rendered stable by fairly to SCC since their hardness is never below the high levels of carbon, a powerful austenite stabi- Rc 22 level, which is considered to be the lizer, in solution. The amount of carbon that can threshold hardness for susceptibility to SCC in be held in solution is a function of annealing tem- body-centered cubic (bcc) ferrous alloys. perature. The 1050 ¡C anneal of the condition A The suitability of high-strength alloys for use mill anneal puts all the carbon in solution, giving in potential SCC-provoking environments con- the austenite the stability of a normal 301-type taining H2S is regulated in many locales by Na- alloy. This permits extensive forming. The key is tional Association of Corrosion Engineers to apply a subsequent lower-temperature anneal (NACE) International Standard MR01-75. In it, so that less carbon goes into solution. Some of it the use of S17400 is permitted if it is double will thus form M23C6. This is in a sense deliber- tempered at 620 ¡C and its hardness is 33 HRC ately sensitizing the alloy, but the sensitization or less, while S45000 can be used if it has been takes place at such a high temperature that aged at 620 ¡C for 4 h and its hardness is 31 chromium deficits around precipitated carbides HRC or less. These permissible hardness levels are minimized by diffusion. This causes a higher are significantly higher than allowed in non-PH Ms temperature because of the lower amount of martensitic alloys, 22 HRC, which reflects the carbon, and chromium, in solution in the austen- fact that the martensitic matrix has the tough- ite. Depending on the temperature at which the ness of a lower-hardness martensite. anneal is done, the Ms temperature can be con- In marine environments, the PH alloys are trolled so that a transformation to martensite can susceptible to SCC if used at a high strength be raised to either room temperature or some at- level. S17400 aged at 480 ¡C with a yield tainable cryogenic temperature. Figure 3 shows a strength of 1240 MPa is susceptible to SCC, chart of these heat treatment options. while higher aging temperatures (above 540 ¡C) The lower strength levels achieved by the producing lower strengths renders the materials condition T route in Fig. 3 reflect the lower car- immune at stresses near the yield strength, ap- bon content of the martensite, while the highest proximately 1170 MPa. The threshold strength strength of the condition C route reflects the for ordinary martensitic stainless steels would compound influence of cold work and marten- be 1030 MPa (Ref 4). site hardness with similar subsequent contribution from age hardening. The main alloys of this group are listed in Semiaustenitic Precipitation- Table 4. Examination of the chemistries in this Hardenable Stainless Steels table shows that the first two alloys rely on the precipitation of Ni3Al for the hardening, while If a martensitic stainless steel were alloyed the last two have no apparent precipitating more strongly with austenite-stabilizing ele- components. Their hardening is a more subtle Chapter 10: Precipitation-Hardening Stainless Steels / 143

Fig. 3 Processing routes for S15700 Source: Ref 5

Table 4 Compositions of semiaustenitic precipitation-hardenable alloys

Alloy Designation C Mn Si Cr Ni Mo Al N 17-7 PH S17700 0.1 0.5 0.3 17 7.1 . . . 1 0 15-7 PH S15700 0.1 0.5 0.3 15 7.1 2.2 1 0 AM-350 S35000 0.1 0.8 0.4 17 4.3 2.8 . . . 0.1 AM-355 S35500 0.1 0.9 0.4 16 4.3 2.8 . . . 0.1 secondary hardening from the tempering of titanium would preferentially deplete the alloy martensite rather than the classic precipitation of carbon and nitrogen, precluding the action of hardening via precipitation of intermetallic the conditioning heat treatment, which relies on compounds. In 17-7 and 15-7, aluminum rather manipulating the amount of carbon in solution. than titanium is the precipitating agent because In AM350 and AM355, it is the precipitation of 144 / Stainless Steels for Design Engineers

Table 5 Mechanical properties of semiaustenitic precipitation-hardenable alloys

Yield, Tensile, Alloy UNS Condition MPa MPa Elongation, % HRC 17Ð7 PH . . . TH 1050 (565) 1100 1310 10 42 RH 950 (510) 1380 1520 9 46 CH 900 (480) 1585 1655 2 49 15Ð7 PH . . . TH 1050 (565) 1380 1450 7 45 RH 950 (510) 1550 1650 6 48 CH 900 (510) 1720 1790 2 50 AM-350 . . . SCT 850 (450) 1210 1420 12 46 SCT1000 (540) 1020 1165 15 40 AM-355 . . . SCT 850 (450) 1250 1510 13 48 SCT 1000(540) 1035 1124 22 38

(Cr,Fe)2N within the martensite phase that is that all carbon is present as chromium carbide, responsible for age hardening. In addition, and that the chromium content is diminished by molybdenum produces a secondary hardening that amount before applying Eq 1. In addition, in carbon-bearing martensite. These two early carbon can remove some molybdenum in the alloys did not possess the hardening potential form of carbides, and nitrogen can remove that alloys employing copper-, titanium-, or chromium as a nitride. The effective corrosion aluminum-based precipitates enjoy. resistance of these alloys thus is similar to fer- The mechanical properties of these alloys ritic 430. The designer is thus advised to consult are not greatly different from the martensitic al- with producers about corrosion resistance de- loys, as can be seen in Table 5. Their separate pending on the thermal processing that will be existence is due to the need for alloys that are used, especially if double aging is performed, more fabricable at room temperature than are the which can cause some degree of chromium de- alloys that are martensitic at room temperature. pletion at grain boundaries. This beneﬁt is offset by the necessity to condi- While the general and pitting corrosion re- tion anneal before age hardening. If corrosion sistance of the semiaustenitic PH alloys are resistance is a high concern, then the alloy and never quite as good as most austenitic stainless, heat treatment that yields the greatest amount of they have very good resistance to SCC com- chromium in solution should be chosen. Thus pared to ordinary martensitic stainless steels. condition CH is better than RH, which is better The δ-ferrite and the generally well-tempered than TH, the order of ascending solution anneal- martensitic matrix provide a crack-arresting ing temperature and ascending Ms. feature and good inherent toughness that resist The martensitic PH grades have somewhat SCC at higher strength levels than in straight better strength because they have a more uni- martensitic stainless steels. AM-355 in the SCC formly martensitic structure, and they employ a (850 ¡F) condition can withstand stresses of tougher, lower-carbon martensite. The fact that 75% of 0.2% offset yield strength in salt spray the semiaustenitic alloys are typically sheet without SCC failure. products generally makes their service toughness requirements less onerous, so that their retained δ-ferrite is not a crippling drawback because of its detrimental effect on through Austenitic Precipitation-Hardenable thickness toughness. Stainless Steels Corrosion Resistance. The semiaustenitic PH alloys tend to higher values of PREN than The austinitic PH class consists of just one the martensitic alloys inherently since they are important commercial alloy, A-286. The impor- alloyed to have lower Ms temperatures. The tance of the alloy is that it is entirely stable thermal processing of these alloys causes a sig- austenite in both the solution-annealed and the niﬁcant portion of the chromium to be removed age-hardened condition. This means it is very from solution as chromium carbide. This lowers formable and nonmagnetic. And, because the the corrosion resistance from what would be ex- precipitation takes place in an austenite matrix, pected based on the bulk composition. From an the precipitation takes place at a higher temper- engineering point of view, it is best to assume ature, around 700 ¡C. This gives the alloy the Chapter 10: Precipitation-Hardening Stainless Steels / 145

Table 6 Austenitic precipitation-hardenable composition

Alloy UNS C Mn Si Cr Ni Mo Al V Ti A-286 S66286 0.05 1.5 0.5 15 25.5 1.3 0.15 0.3 2.15 Discalloy S66220 0.04 1.6 0.5 14 26 3 ...... 1.7

Fig. 4 A-286 properties as a function of test temperature. Source: Ref 5

potential to be used to near 700 ¡C without fear PH alloys, makes it attractive for a variety of of overaging. Thus, austenitic PH stainless rep- aerospace and nonaerospace uses. resents a way to strengthen the austenite matrix, The hardening mechanism of A-286 is the which has the following advantages: precipitation of Ni3 (aluminum, titanium). Dif- fusion, even at the higher temperatures, is ¥ High ductility and therefore high formability slower in austenite, so aging treatments are typ- in the soft, unaged condition ically 16 h. Table 6 gives the typical composi- ¥ High toughness at all temperatures and tion of A-286, and Fig. 4 shows some typical strength levels properties as a function of temperature for a ¥ Excellent creep and stress rupture properties standard 980 ¡C solution treatment followed by ¥ Excellent oxidation, corrosion, and SCC re- a 720 ¡C, 16 h aging. sistance The mechanical properties can be greatly en- This alloy is rightly considered an iron- hanced by cold working prior to aging, as is based superalloy and is the root of the group shown in Fig. 5. that succeeds it in properties, the nickel- and The toughness of the austenitic matrix is cobalt-based superalloys. These last alloys can abundant and quite temperature insensitive. attain greater strength and creep resistance Charpy V-notch values of over 60 J are typical than A-286, effectively ending further devel- from Ð200 to 800 ¡C. opment of austenitic PH alloys. The lower cost The corrosion resistance of A-286 is compa- of the A-286 alloy, compared to nickel-base rable to that of 304 and 316. It has slightly 146 / Stainless Steels for Design Engineers

Fig. 5 The inﬂuence of cold work on aging response in A-286. DPH, diamond pyramid hardness. Source: Ref 5 better resistance to SCC despite its higher allurgical and Materials Transactions A, strength level. Vol 30A, Feb 1999. 3. G. Aggen, Ph.D. thesis, Carnegie Mellon REFERENCES University 4. E.E. Denhard, “Stress Corrosion Cracking 1. F.B. Pickering, Physical Metallurgical De- of High Strength Stainless Steels in Atmos- velopments of Stainless Steel, Stainless ’84, pheric Environments”, paper presented at Goteborg, Sept 3Ð4, 1984, p 2Ð28. the Twenty-fourth Meeting of the AGARD 2. M. Murayama, Y. Katayama, and K. Hono, Structures and Materials Panel (Turin, Microstructural Evolution in a 17-4 PH Italy), April 17Ð20, 1967. Stainless Steel After Aging at 400 ¡C, Met- 5. Allegheny Technology Blue Sheets Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 147-154 All rights reserved. DOI: 10.1361/ssde2008p147 www.asminternational.org

CHAPTER 11

Casting Alloys

Summary essential element. Thus, the casting alloys listed in Tables 1 and 2 (Ref 1) are recognizable as ap- WITH TYPICAL ALLOY SYSTEMS, cast- proximate counterparts of the co-listed wrought ing is often the most convenient method by alloys (AISI grade). This cross reference to which to produce components. This is true for wrought equivalents is helpful when looking for stainless steels—both for corrosion-resisting data about an alloy that may be more easily and for heat-resisting applications. This chapter found for wrought alloys than for cast. discusses primarily the alloys used for stainless The High Alloy Product Group of the Steel steel castings and their metallurgy. Foundry Founder’s Society of America employs a nam- methods are discussed to the degree they are ing system (ACI, the Alloy Casting Institute) specific to the stainless alloys. for cast alloys that is significant; these designations are currently assigned by ASTM as grades and are added to ASTM specifications. The first Stainless Steel Casting Alloys letter, “C” or “H,” indicates corrosion resisting. The second letter indicates the relative amount Essentially any wrought stainless alloy compo- of nickel, from a minimum of 0 to 1% for “A” sition can be modified to be made as a cast alloy. up to 30% nickel for “N” alloys. The number The systemic difference between cast alloys and following the hyphen for “C” alloys designates their wrought equivalents is that cast alloys gen- the maximum carbon in hundredths of a percent. erally contain between 1.0 and 2.5% silicon. As The suffix letters designate additional alloying with other ferrous alloys, this is done to increase elements, such as Cu for copper, M for molyb- the fluidity of the melt to make it cast more effec- denum, N for nickel or nitrogen, F for free ma- tively. Silicon has strong metallurgical effects, chining, and C for columbium (niobium). The both beneficial and detrimental, which should be heat-resisting, “H,” alloys have generally only a understood by the user of cast stainless steels. second letter designating relative nickel level on These are explained. A second general observa- the same scale as “C” alloys but going past stain- tion is that the stabilized ferritic stainless steel al- less steels all the way to nickel-based alloys. The loys, which constitute almost half the tonnage of inclusion of a number after the first two letters all stainless steel used, are notably absent from indicates the center of the carbon range ex- the cast alloys. This is because these alloys are pressed in hundredths of a percent by weight. single phase at all temperatures in the solid state To learn more about the influence of alloying and because they have large as-cast grain sizes elements, refer to the chapters on the individual that can only be refined by heavy cold work fol- alloy families; see Section 3. Here, we briefly lowed by annealing. This makes them quite lack- summarize: ing in toughness as cast. Since heavy cold work ¥ Pitting and crevice corrosion resistance, as defeats the purpose of casting to achieve a near- well as general corrosion resistance, are en- net shape, stabilized ferritic stainless steels are hanced by chromium, molybdenum, tung- seldom used as castings. Also, the standard stabi- sten, and nitrogen and carbon in solution. lizing alloy, titanium, is too readily oxidized for ¥ Localized corrosion is caused by chromium normal foundry practice to avoid the loss of this depletion, which occurs when precipitates 148 / Stainless Steels for Design Engineers

Table 1 Compositions of cast stainless corrosion resisting alloys Composition(a), wt%—maximum or range Nearest ACI designation AISI grade UNS %C %Mn %Si %Cr %Ni %Mo %Other Chromium alloys CA-15 410 J91150 0.15 1.00 1.50 11.5Ð14.0 1.0 0.50 CA-15M J91151 0.15 1.00 0.65 11.5Ð14.0 1.0 0.15Ð1.00 CA-40 420 J91153 0.40 1.00 1.50 11.5Ð14.0 1.0 0.50 CA-40F 420F J91154 0.2Ð0.4 1.00 1.50 11.5Ð14.0 1.0 0.20Ð0.40 Ss CB-30 431,442 J91803 0.30 1.00 1.50 18.0Ð22.0 2.0 CC-50 446 J92613 0.30 1.00 1.50 26.0Ð30.0 4.0 Chromium-nickel alloys CA-6N J91650 0.06 0.50 1.00 10.5Ð12.5 6.0Ð8.0 CA-6NM S41500 J91540 0.06 1.00 1.00 11.5Ð14.0 3.5Ð4.5 0.4Ð1.0 CA-28MWV 422 J91422 0.20Ð0.28 0.50Ð1.00 1.00 11.0Ð12.5 0.5Ð1.0 0.9Ð1.25 0.9Ð1.25 W, 0.2Ð0.3 V CB-7Cu-1 17-4PH (AISI 630) J92180 0.07 0.70 1.00 15.5Ð17.7 3.6Ð4.6 2.5Ð3.2 Cu, 0.2Ð0.35 Nb, 0.05 N max CB-7Cu-2 15-5 PH (XM-12) J92110 0.07 0.70 1.00 14.0Ð15.5 4.5Ð5.5 2.5Ð3.2 Cu, 0.2Ð0.35 Nb, 0.05 N max CD-3MN 2205 (S32205) J92205 0.03 1.50 1.00 21.0Ð23.5 4.5Ð6.5 2.5Ð3.5 1.0 max Cu, 0.10Ð0.30 N CD-3MCuN 255 (S32550) J93373 0.03 1.20 1.10 24.0-26.7 5.6Ð6.7 2.9Ð3.8 1.4Ð1.9 Cu, 0.22Ð0.33 N CD-3MWCuN (S32760) J93380 0.03 1.00 1.00 24.0Ð26.0 6.5Ð8.5 3.0Ð4.0 0.5Ð1.0 Cu, 0.5Ð1.0 W, 0.20Ð0.30 N CD-4MCu J93370 0.04 1.00 1.00 24.5Ð26.5 4.75Ð6.0 1.75Ð2.25 2.75Ð3.25 Cu CD-4MCuN J93372 0.04 1.00 1.00 24.5Ð26.5 4.7Ð6.0 1.75Ð2.25 2.75Ð3.25 Cu, 0.10Ð0.25 N CD-6MN J93371 0.06 1.00 1.00 24.0Ð27.0 4.0Ð6.0 1.75Ð2.25 1.75Ð2.5 Cu, 0.15Ð0.25 N CE-3MN 2507 (S32750) J93404 0.03 1.50 1.00 24.0Ð26.0 6.0Ð8.0 4.0Ð5.0 0.10Ð0.30 N CE-8MN J93345 0.08 1.00 1.50 22.5Ð25.5 8.0Ð11.0 3.0Ð4.5 0.10Ð0.30 N CE-30 312 J93423 0.30 1.50 2.00 26.0Ð30.0 8.0Ð11.0 CF-3 304L J92500 0.03 1.50 2.00 17.0Ð21.0 8.0Ð12.0 CF-3M 316L J92800 0.03 1.50 2.00 17.0Ð21.0 8.0Ð12.0 2.0Ð3.0 CF-3MN 316LN J92700 0.03 1.50 1.50 17.0Ð21.0 9.0Ð13.0 2.0Ð3.0 0.10Ð0.20 N CF-8 304 J92600 0.08 1.50 2.00 18.0Ð21.0 8.0Ð11.0 CF-8C 347 J92710 0.08 1.50 2.00 18.0Ð21.0 9.0Ð12.0 Nb CF-8M 316 J92900 0.08 1.50 2.00 18.0Ð21.0 9.0Ð12.0 2.0Ð3.0 CF-10 304H J92590 0.04Ð0.10 1.50 2.00 18.0Ð21.0 8.0Ð11.0 CF-10M 316H J92901 0.04Ð0.10 1.50 1.50 18.0Ð21.0 9.0Ð12.0 2.0Ð3.0 CF-10MC 316H J92971 0.10 1.50 1.50 15.0Ð18.0 13.0Ð16.0 1.75Ð2.25 (10xC)Ð1.2 Nb CF-10SMnN NITRONICª 60 J92972 0.10 7.00Ð9.00 3.50Ð4.50 16.0Ð18.0 8.0Ð9.0 0.08Ð0.18 N CF-12M 316 0.12 1.50 2.00 18.0Ð21.0 9.0Ð12.0 2.0Ð3.0 CF-16F 303 J92701 0.16 1.50 2.00 18.0Ð21.0 9.0Ð12.0 1.5 max 0.2Ð0.35 Se CF-20 302 J92602 0.20 1.50 2.00 18.0Ð21.0 8.0Ð11.0 CG-6MMN NITRONICª 50 J93790 0.06 4.00Ð6.00 1.00 20.5Ð23.5 11.5Ð13.5 1.5Ð3.0 0.1Ð0.3 Nb, 0.1Ð0.3 V, 0.2Ð0.40 N CG-8M 317 J93000 0.08 1.50 1.50 18.0Ð21.0 9.0Ð13.0 CG-12 308 J93001 0.12 1.50 2.00 20.0Ð23.0 10.0Ð13.0 CH-8 309S J93400 0.08 1.50 1.50 22.0Ð26.0 12.0Ð15.0 CH-10 309H J93401 0.04Ð0.10 1.50 2.00 22.0Ð26.0 12.0Ð15.0 CH-20 309 J93402 0.20 1.50 2.00 22.0Ð26.0 12.0Ð15.0 CK-3MCuN 254SMOª J94653 0.025 1.20 1.00 19.5Ð20.5 17.5Ð19.5 6.0Ð7.0 0.5Ð1.0 Cu, 0.18Ð0.24 N CK-20 310 J94202 0.20 2.00 2.00 23.0-27.0 19.0Ð22.0 CN-3M 904L J94652 0.03 2.00 1.00 20.0Ð22.0 23.0Ð27.0 4.5Ð5.5 CN-3MN AL-6XN¨ J94651 0.03 2.00 1.00 20.0Ð22.0 23.0Ð27.0 6.0Ð7.0 0.18Ð0.24 N CN-7M 320 N08007 0.07 1.50 1.50 19.0-22.0 27.5Ð30.0 2.0Ð3.0 3.0Ð4.0 Cu CN-7MS J94650 0.07 1.50 3.50 18.0Ð20.0 22.0Ð25.0 2.5Ð3.0 1.5Ð2.0 Cu CT-15C N08151 0.05Ð0.15 0.15Ð1.50 0.50Ð1.50 19.0Ð21.0 31.0Ð34.0 0.5Ð1.5 Nb (a) Balance Fe for all compositions. Source: Ref 1 Chapter 11: Casting Alloys / 149

Table 2 Compositions of cast heat-resistant stainless and nickel base alloys

Composition(a), wt%—maximum or range ACI designation Nearest AISI grade UNS %C %Cr %Ni %Si max HA 504 J82090 0.20 max 8Ð10 1.00 HC 446 J92605 0.50 max 26Ð30 4Ðmax 2.00 HD 327 J93005 0.50 max 26Ð30 4Ð7 2.00 HE 312 J93403 0.20Ð0.50 26Ð30 8Ð11 2.00 HF 302B J92603 0.20Ð0.40 19Ð23 9Ð12 2.00 HH 309 J93505 0.20Ð0.50 24Ð28 11Ð14 2.00 HI J94003 0.20Ð0.50 26Ð30 14Ð18 2.00 HK 310 J94224 0.20Ð0.60 24Ð38 18Ð22 2.00 HK-30 J94203 0.25Ð0.35 23Ð27 19Ð22 1.75 HK-40 J94204 0.35Ð0.45 23Ð27 19Ð22 1.75 HL N08604 0.20Ð0.60 28Ð32 18Ð22 2.00 HN J94213 0.20Ð0.50 19Ð23 23Ð27 2.00 HP N08705 0.35Ð0.75 24Ð28 33Ð37 2.00 HP-50WZ 0.45Ð0.55 24Ð28 33Ð37 2.50 HT 330 N08605 0.35Ð0.75 13Ð17 33Ð37 2.50 HT-30 N08603 0.25Ð0.35 13Ð17 33Ð37 2.50 HU N08005 0.35Ð0.75 17Ð21 37Ð41 2.50 HW N08006 0.35Ð0.75 10Ð14 58Ð62 2.50 HX N06050 0.35Ð0.75 15Ð19 64Ð68 2.50

(a) Balance Fe for all compositions. Manganese content: 0.35Ð0.65% for HA, 1% for HC, 1.5% for HD, 2% for the other alloys. Phosphorus and sulfur contents: 0.04 (max) for all but HP-50WZ. Molybdenum is intentionally added only to HA: 0.90Ð1.2%. Maximum molybdenum for other alloys is 0.5%. HH contains 0.2% N (max). HP-50WZ also contains 4Ð6% W, 0.1Ð1.0% Zr, and 0.035% S (max) and P (max). Source: Ref 1

form in the solid state. These precipitates are minimize solidification hot cracking or to allow carbides, oxides, and sulfides as well as inter- weld repair of cracks that do form. It has been metallic phases richer in chromium, molyb- shown that the existence of ferrite can increase denum, or nitrogen than the matrix. the resistance to stress corrosion cracking. ¥ General corrosion resistance follows the above guidelines but is also helped by copper and nickel, which do not assist in pitting resistance. Metallurgy of “C” Alloys ¥ High-temperature oxidation resistance is enhanced by increasing chromium and silicon. The corrosion-resistant “C” series have Wrought alloys employ aluminum and rare wrought counterparts from which they differ es- earths to help oxidation resistance, but the sentially only in silicon content. This silicon has difficulty of keeping these elements from no significant influence on corrosion resistance being oxidized requires special techniques or mechanical or physical properties, so an un- such as vacuum induction melting and inert derstanding of these alloys by approximating refractories for molds. them to their wrought counterparts is justified. ¥ Iron-chromium (ferritic) alloys have better The main difference between the cast and thermal fatigue resistance but poorer creep wrought product forms of these alloys is the resistance than iron-chromium-nickel grain structure. In wrought grades, the grain (austenitic) alloys. structure can be manipulated by deformation and heat treatment. The use of deformation is The alloy designation system largely ignores not an option in cast alloys; consequently, the the wrought alloy distinctions by microstructure opportunities for grain refinement in cast alloys (i.e., ferritic, austenitic, duplex, PH [precipita- are limited. To counteract the problems of lower tion hardening], and martensitic). One reason is corrosion resistance of cast grades, a homoge- that the most widely used wrought-stabilized nizing solution heat treatment is necessary to ferritics (e.g., 409, 439) do not exist as common counteract the chromium depletion that occurs casting alloys, and nominally austenitic alloys due to solidification segregation and precipitating in the cast form contain enough ferrite to be sig- phases. Representative mechanical properties nificantly magnetic. Thus, the distinctions based for “C” alloys are listed in Table 3 (Ref 1). on phase are not as well defined for casting al- Martensitic. CA alloys are martensitic. The loys. The high ferrite content in the nominally metallurgy is straightforward and equivalent to austenitic casting alloys is to avoid or at least their wrought counterparts. The mechanical 150 / Stainless Steels for Design Engineers . Specimen V-notch V-notch V-notch V-notch V-notch Keyhole notch Keyhole notch V-notch Keyhole notch V-notch Keyhole notch V-notch V-notch Keyhole notch Keyhole notch Keyhole notch V-notch V-notch Keyhole notch Keyhole notch Keyhole notch Keyhole notch Izod V-notch V-notch Keyhole notch 7 2 2 70 20 25 55 26 70 70 30 80 30 50 74 60 75 70 . ft-lb 110 110 100 120 100 140 Charpy toughness Charpy J 2.7 2.7 9.5 . 94.9 27.1 33.9 74.6 35.3 94.9 81.4 94.9 67.8 40.7 40.7 94.9 162.7 149.2 135.6 100.3 135.6 101.7 190 108.5 HB 269 225 310 400 195 210 253 305 190 140 160 140 156 163 150 170 170 149 150 176 190 144 185 130 Hardness, 60 55 30 54 ...... Reduction in area, % in area, 50 24 10 15 55 39 45 48 22 14 18 25 20 18 60 50 50 50 55 45 50 52 38 37 Elongation % in 50 mm, 2 in. 60 36 31 65 81 92 63 36 42 37 45 38 45 42 38 40 44 50 38 53 ksi 125 100 100 170 862 448 255 262 310 345 262 214 689 689 414 558 634 434 248 290 310 248 290 262 276 303 365 MPa 1172 1172 Yield strength, 0.2% offset strength, Yield 77 90 80 87 77 85 77 80 77 77 82 88 69 95 97 97 76 ksi 115 115 112 130 120 150 190 108 Tensile strength Tensile MPa 669 669 531 745 896 600 531 586 827 793 655 531 552 621 552 531 531 565 607 524 770 476 1034 1310 F), WQ F), A o o F), AC, T F), FC to F), WQ F), AC, T F), AC, T A F), OQ, F), AC F), AC F), FC to F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ F), WQ o o o o o o o o o o o o o o o o o o o o o o o o C (1900 C (1900 o o C (1750 C (2050 C (1900 C (1950 C (2100 C (1800 C (1800 C (1900 C (1450 C (1900 C (2050 C (2000 C (1900 C (1950 C (2000 C (1900 C (2000 C (2100 C (2050 C (1900 C (1900 C (2000 C (1900 C (1900 o o o o o o o o o o o o o o o o o o o o o o o o Heat treatment(a) 1040 1040 980 980 790 1120 1120 1120 1120 1150 1150 1150 1120 1040 1040 1095 >955 >1040 >1040 >1040 >1040 >1095 >1040 >1040 >1065 >1065 >1095 >1040 >1095 CA-6NM CD-4MCu CF-3A CF-8 CF-8A CF-20 CF-3M CF-3MA CF-8M CF-8C CF-16F CG-8M CH-20 CK-20 CN-3MN CN-7M A, age. Source: Ref 1 temper; T, WQ, water quench; AC, air cool; FC, furnace OQ, oil quench; (a) Table 3Table cast stainless alloys resisting of corrosion properties mechanical Room temperature Alloy CA-15 CA-40 CB-7Cu CB-30 CC-50 CE-30 CF-3 Chapter 11: Casting Alloys / 151

properties are governed by the thermal pro- high level of nickel negates much of the potential cessing, and strength, hardness, and toughness cost savings duplex alloys offer. All castings are can be varied over a wide range. The CB 30 solution annealed and quenched to eliminate em- and CC 50 alloys are ferritic and, as such, have brittling intermetallic phases. negligible toughness but effectively deliver The cast duplex alloys may offer a better engi- corrosion resistance. The toughness of CB 30 neering approach than the equivalent austenitic can be improved by balancing the chromium cast alloys because they have greater strength and silicon to a lower part of the range and the and lower alloy cost for the same level of corro- carbon and nickel to the higher end to render sion resistance. They do not have the same prob- the microstructure partly martensitic. lems of hot cracking that make casting austenitic Precipitation Hardening. The cast PH alloys steels difficult. The poor hot workability of include CB-7Cu-1, which behaves in a similar duplex steel is not an issue for castings. It is im- way to 17-4PH, which has an overlapping com- portant for designers to understand that cast du- position range. Note that most other major plex steels are totally compatible galvanically wrought PH grades rely on titanium and alu- with wrought or cast austenitic alloys of the minum to form coherent strengthening precipi- same corrosion resistance. Mixing components tates and so do not have cast counterparts. Cop- with different microstructures does not create a per, which can harden ferrite but not austenite, galvanic differential when corrosion resistance is thus the only strengthener available. There is levels are similar. Reluctance to mix alloys for one cast PH alloy that has no wrought counter- galvanic reasons can be an expensive error when part. It is CD-4MCu; however, it is rarely used their similar corrosion resistances makes them in the precipitation-hardened condition and is compatible, even if they are quite different most commonly classified as a duplex stainless microstructurally. steel in which the nitrogen level is closely con- Austenitic-Ferritic. The typical CF alloys, trolled. This is a highly alloyed duplex grade which make up about two-thirds of U.S. stain- that contains copper to precipitation harden the less steel castings, are nominally austenitic but ferrite phase. Oil field CO2 corrosion is resisted always contain ferrite. This is not detrimental by alloys that resemble the martensitic PH and improves resistance to stress corrosion grades. These alloys are discussed in Chapter cracking and sensitization. Homogenization an- 22, “Petroleum Industry Applications” and can nealing can reduce the amount of ferrite and re- be considered castable alloys. sult in lower yield and tensile strength and Duplex. The cast equivalents of alloys 2205 higher elongation and toughness. The composi- and 2507, J 92205, and J 93380 have similar tion balance is the main determinant of ferrite properties and corrosion resistance. Modern level. Increasing the nickel, nitrogen, manganese, wrought duplex alloys rely on nitrogen to parti- or carbon content decreases ferrite. Increasing tion the alloy with uniform corrosion resistance chromium, silicon, or molybdenum content in- in each phase and to suppress intermetallic phase creases ferrite. Increasing the solidification rate formation. Cast alloys are effectively limited to will increase the ratio of austenite to ferrite in 0.25% nitrogen before gas porosity becomes ex- duplex or austenitic-ferritic alloys. The predom- cessive. Porosity can be reduced by replacing inantly austenitic matrix has a very high tough- some nickel with manganese, which increases ni- ness even at cryogenic temperatures. Ferrite, if trogen solubility. Doing so would expand the continuous, decreases toughness. Fortunately, it most promising area of stainless steel develop- is seldom present as a continuous phase. The ment, lean duplex alloys such as 2101 and 2003, loss of toughness associated with high ferrite to the cast grades. Alloy 2101 with 4 to 6% man- content can be aggravated by heating the ferrite ganese provides the corrosion resistance of CF- above 475 oC (885 oF) for a sufficient time for 8M or 316L with total nickel plus molybdenum the ferrite to decompose to the brittle α and α'. of only 2% versus the 12% required for the At higher temperatures, development of the σ austenitic alloy. The duplex alloys also have phase would have a similar embrittling effect. greater strength and are nearly immune to stress These phases thus formed are quickly redis- corrosion cracking. These alloys represent signif- solved and removed by annealing. Note that icant cost-savings potential for the foundry and sometimes copper is added to austenitic alloys for its customers. CE-30 is duplex steel, which is to improve corrosion resistance in sulfuric acid fairly simple metallurgically and uses only environments. It has no precipitation hardening chromium for corrosion resistance. However, its effect in austenite, as it does in ferrite. When 152 / Stainless Steels for Design Engineers

used as a precipitating hardening agent, copper has higher thermal expansion and lower thermal does not increase corrosion resistance. conductivity than ferrite, which aggravates ther- Virtually any non-titanium-bearing, corro- mal fatigue and oxide spalling. Nevertheless, sion-resistant, austenitic, wrought alloy can the better high-temperature strength of austenite have a cast counterpart. Curiously, the 2xx low- generally is the predominant consideration, and nickel alloys are not found in most cast alloys most “H” alloys are austenitic. Tables 4 and 5 lists. If a specific wrought alloy cannot be found list properties of “H” alloys (Ref 1 to 3). to have a published cast counterpart, the de- Ferritic HA, HC, HD. Of the ferritic alloys signer should not avoid requesting a producer HA, HC, and HD, HA with less than 10% to supply a version that the foundry is confident chromium is not quite stainless but is useful to of making. The designer must thoroughly un- 650 oC (1200 oF) for petroleum refinery applica- derstand the design of the alloy desired so that tions. HC and HD are very high chromium any alterations to its composition necessary to ferritics that have very low toughness and creep allow castability will not compromise expected resistance but are quite oxidation and sulfida- performance. tion resistant. They can be cost-effective materials when high-temperature strength is not an overriding concern. Metallurgy of “H” Alloys Austenitic HE-HP. The predominant high- temperature grades are the austenitic HE The heat-resisting “H” alloys are principally through HP, after which come the nickel alloys, austenitic. Alloying elements and impurities which are not generally classified as stainless diffuse more slowly through the face-centered steels because they contain less than 50% iron. cubic (fcc) austenitic structure than the bcc The high material cost of the nickel base alloys ferrite structure, making the austenite more re- restricts their use to those specific environments sistant to diffusion-controlled creep. Austenite where maximum carburization or nitriding

Table 4 Mechanical properties of heat-resistant cast stainless alloys at room temperature

Tensile strength Yield strength

Alloy Condition MPa ksi MPa ksi Elongation, % Hardness, HB Standard grades HA N + T(a) 738 107 558 81 21 220 HC As-cast 760 110 515 75 19 223 Aged(b) 790 115 550 80 18 . . . HD As-cast 585 85 330 48 16 90 HE As-cast 655 95 310 45 20 200 Aged(b) 620 90 380 55 10 270 HF As-cast 635 92 310 45 38 165 Aged(b) 690 100 345 50 25 190 HH, type 1 As-cast 585 85 345 50 25 185 Aged(b) 595 86 380 55 11 200 HH, type 2 As-cast 550 80 275 40 15 180 Aged(b) 635 92 310 45 8 200 HI As-cast 550 80 310 45 12 180 Aged(b) 620 90 450 65 6 200 HK As-cast 515 75 345 50 17 170 Aged(c) 585 85 345 50 10 190 HL As-cast 565 82 360 52 19 192 HN As-cast 470 68 260 38 13 160 HP As-cast 490 71 275 40 11 170 HPNb(d) As-cast 450 220 8 . . . HPNbTi(e) As-cast 450 220 8 . . . HT As-cast 485 70 275 40 10 180 Aged(c) 515 75 310 45 5 200 HU As-cast 485 70 275 40 9 170 Aged(f) 505 73 295 43 5 190 HW As-cast 470 68 250 36 4 185 Aged(g) 580 84 360 52 4 205 HX As-cast 450 65 250 36 9 176 Aged(f) 505 73 305 44 9 185 (a) Normalized and tempered at 675 ¡C (1250 ¡F). (b) Aging treatment: 24 h at 760 ¡C (1400 ¡F), furnace cool. (c) Aging treatment: 24 h at 760 ¡C (1400 ¡F), air cool. (d) ISO 13583-2 speciﬁcation minima. (e) ISO 13583-2 speciﬁcation limits for microalloyed grade. (f) Aging treatment: 48 h at 980 ¡C (1800 ¡F), air cool. (g) Aging treatment: 48 h at 980 ¡C (1800 ¡F), furnace cool. Source: Ref 1 Chapter 11: Casting Alloys / 153

Table 5 High-temperature mechanical properties of “H” alloys

Tensile Creep rate Stress to Stress to Yield strength strength 0.0001%/h 1% in 100,000 rupture rupture Alloy Temp ksi, MPa ksi, MPa psi, MPa h psi, MPa in 1000 h in 10,000 h HA 1400 oF 16 27 1800 oF HC 1400 oF 8.7 10.5 1.3 1.3 1800 oF 2.1 2.5 3.6 0.6 HD 1400 oF 36 3.5 7.0 1800 oF 15 1.0 2.5 HE 1400 oF 3.5 11.0 1800 oF 1.0 2.5 HF 1400 oF 35 6.0 4.4 9.1 1800 oF 1.6 (est) HH 1400 oF 17 33 3.0 TYPE 1 1800 oF 6.3 9 1.1 HH 1400 oF 18 35 7.0 2.0 8.0 TYPE 2 1800 oF 7 11 2.1 1.6 0.9 HI 1400 oF 6.6 8.5 1800 oF 1.9 2.6 HK 1400 oF 6.8 6.3 12.0 1800 oF 2.7 0.9 2.8 1.7 HL 1400 oF 50 7.0 15 1800 oF 18.7 2.8 (est) 5.2 HN 1400 oF 1800 oF 2.4 2.1 HP 1400 oF 1800 oF 2.1 HPNb(a) 800 oC ...... 51 MPa . . . 55 MPa HPNbTi(a) 800 oC ...... 54 MPa . . . 64 MPa HT 1400 oF 26 35 8.0 12 1800 oF 8 11 2.0 2.7 1.7 HU 1400 oF 40 8.5 1800 oF 6.2 10 2.2 2.9 1.8 HW 1400 oF 23 32 6.0 7.8 1800 oF 8 10 1.4 2.6 HX 1400 oF 19.5 42 6.4 1800 oF 6.9 10.7 1.6 2.2

(a) Data from ISO 13583-2. Source: Ref 2, 3

resistance is mandatory. For oxidation and sulfi- alloying also are discussed in the oxidation sec- dation resistance, the iron base alloys are pre- tion of Chapter 6. ferred. These cast stainless steels derive their The major problem that all producers of stain- oxidation resistance from their chromium level. less steels face is that of transferring molten The chromium near the surface acts as a reser- metal from the furnace to the mold cavity. This voir to replenish the protective iron/chromium problem is heightened when the foundry makes oxide scale as explained in Chapter 6 in the sec- complex shapes. Methods developed to protect tion on oxidation. Silicon, another stable oxide the molten stream from exposure to air to prevent former, assists in forming this protective scale reoxidation have shown great promise and have and resistance to carburization. Other typical been demonstrated by the wrought alloy produc- alloying elements do not aid in oxidation resist- ers who tend to produce much simpler shapes ance. If it were possible to cast these alloys with than the foundry. Protection of the molten stream aluminum or rare earth additions without them could result in castings with much better high- being lost to oxidation before solidification, temperature performance that could be used in- there could be some impressive benefits. Such stead of some use of higher nickel alloys. alloys exist in wrought form, for example, High-temperature strength is modestly im- 153MA and 253MA. The metallurgical basis of proved by higher levels of chromium and the benefits from aluminum and rare earth nickel. Molybdenum improves high-temperature 154 / Stainless Steels for Design Engineers

strength, but its detrimental effect on oxidation higher-performance grades. It is also possible to resistance and its promotion of intermetallic use AOD-refined master melt stock to achieve precipitation limits its use. Carbon is very effec- the same benefits as AOD refining while using tive for promoting high-temperature strength induction melting. and suppression of intermetallic phase forma- Welding of cast stainless alloys is a common tion. All “H” alloys, therefore, employ much practice and does not present problems when higher carbon levels than the “C” alloys. This using approved weld procedures and qualified does, however, directly imply that the corrosion welders. Chapter 17 describes joining methods resistance of “H” alloys, should it be an issue, is in detail. The same precautions about sensitiza- significantly degraded over otherwise similar tion apply to castings. Welding of non-niobium- “C” alloys. stabilized “C” alloys with carbon levels above The HP grades have undergone significant 0.03% will require postweld annealing to redis- development over the last 30 years. This devel- solve chromium carbides, which will otherwise opment has come about through the addition of make the alloy susceptible to corrosive attack in niobium to increase creep and rupture proper- the chromium-depleted regions of the heat- ties. The use of microalloying additions has de- affected zone. livered creep and rupture properties some 30% Iron and nickel base “H” alloys that are fully higher than the HP grade without niobium mi- austenitic can suffer from hot shortness due to croalloying. It is unfortunate that these HP sulfide films that precipitate along grain bound- grades have not at this time found their way into aries even at low bulk sulfur levels. This makes ASTM standards; however, work is under way them susceptible to hot cracking of welds. Al- to remedy this omission. Currently, the most up- loys with some ferrite are less susceptible to hot to-date collection of these grades can be found cracking, so most “C” alloys are highly resistant in ISO 13583-2. to this problem.

Foundry Practice REFERENCES Metals Handbook While the scope of this book does not extend 1. Cast Stainless Steels, , to the production of castings, certain aspects are desk ed., J.R. Davis, Ed., ASM Interna- important to the user of castings. For the last 50 tional, 1988, p 386Ð390 2. International Organization for Standardiza- years, virtually all stainless has been refined in tion, www.iso.org, ISO 13583-2 argon oxidation decarburization (AOD) vessels 3. Steel Founders Society of America, online or versions thereof. For “C” alloys, this refining documents: http://www.sfsa.org/sfsa/pubs/ method should be considered a basic require- index.html ment for good quality where the carbon levels are restricted to low levels (e.g., CF3M). “H” alloys are less refined inherently and can be SELECTED REFERENCE simply arc or induction melted; however it may be necessary to use refining techniques for the ¥ http://www.sfsa.org Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 155-160 All rights reserved. DOI: 10.1361/ssde2008p155 www.asminternational.org

CHAPTER 12

Melting, Casting, and Hot Processing

Summary austenitic alloys, impossible to produce. The advent of AOD, continuous casting, ladle metal- THE PRIMARY PRODUCTION PROC- lurgy, and powerful hot rolling mills has led to ESSES of melting, casting, and hot processing stainless steels of much higher quality produced are invisible to the end user. The vast majority of at lower cost. Ironically, the low processing cost stainless steel is made by arc furnace melting fol- of stainless steel has spurred demand and made lowed by argon oxygen decarburization (AOD) some of its ingredients, such as molybdenum refining and continuous casting. It is not normal, and nickel, which are relatively scarce and ex- and it is seldom beneficial for the end user to pensive commodities, even more costly, forcing specify processing paths. The end user should, the cost of many alloys to spike even higher however, be knowledgeable and require the pro- than in earlier years. ducer to document the process and the producer’s control of it. Melting and Refining

Introduction The arc furnace is nearly universally used for the first step in the production of stainless steel. The manner in which stainless steel is made The arc furnace is quite flexible in the types of at the producing mill can have a great impact on charge materials it can accept. Since the charge its final properties. These production methods materials for stainless steel are typically carbon have undergone a major evolution over the last steel and stainless steel scrap, this flexibility al- 50 years and are mainly responsible for stain- lows scrap of all types to be used. The necessary less steels becoming the practical, widespread chromium is added as ferrochromium, whose engineering materials they are today. Traditional cost is inversely related to its carbon content. carbon and alloy steel-making methods are not The carbon content of the heat of steel is roughly suitable for stainless steels. The fundamental 1.5 to 2.5% when it is melted and ready to difference is that the basic decarburization step, charge into the separate refining vessel. which is common to all steel making, is thermo- It is this carbon whose removal is the primary dynamically very difficult in stainless steel be- focus of refining. In the 1960s, Union Carbide cause the essential element, chromium, reacts engineers perfected a method, the previously more strongly with the purifying agent, oxygen, mentioned AOD process, of removing nearly all than does carbon. Thus, early stainless steel mak- the carbon from molten stainless steel without ing, done in an arc furnace, was a lengthy process significant loss of chromium. This process is that necessarily involved high chromium losses based on the following chemical reaction: to the slag as carbon was removed. This process Cr O (Solid) + yC = yCO (gas) + Cr (Eq 1) was not only very expensive, the carbon levels 3 4 that could be achieved were not much below The equilibrium for this reaction is: 0.10%, making most of today’s stainless steels, −ΔG whose carbon levels range from 0.010% in sta- Ln (K ) = (Eq 2) bilized ferritic alloys to about 0.07% in normal 4575T 156 / Stainless Steels for Design Engineers

where K is the equilibrium constant, and G is steel combine to reduce the oxygen content of the Gibbs free energy. the steel to around 100 ppm. This could be fur- Working through the thermodynamics yields ther reduced by aluminum, but aluminum-based the relationship that summarizes the important inclusions are generally undesirable. The ther- relationship among carbon, chromium, and CO modynamic activity of aluminum is consider- (Ref 1): ably reduced in iron as chromium levels increase, so its role as a deoxidizer is less valuable in stainless steels. Titanium, on the other hand, is %Cr 13,800 Log ==−+−8..log 76 0 925 p CO enhanced as a deoxidizer in chromium-iron al- %C T loys, and consequently small amounts of it are (Eq 3) sometimes used as a supplementary deoxidant in alloys even though an alloy specification may Thus, increasing the temperature works to in- not call for any. Titanium is believed to reduce crease the elimination of carbon as CO, which hot working defects. More active deoxidants, evolves from the melt. This is similar in principle such as calcium and magnesium, can be used to the basic oxygen furnace (BOF) process for when required. Also note that even if no inten- carbon steel in which oxygen is injected into tional addition of metallic calcium is made, molten steel to remove carbon by oxidizing it. strong deoxidation with aluminum or titanium The key to the AOD process, though, is the in- can reduce small amounts of calcium from the jection of oxygen and argon into the bath to keep CaO in the slag, producing measurable calcium the partial pressure of CO (pCO) very low. This is content in the metal. done at a temperature consistent with economic Besides carbon and oxygen, other impurities refractory life. The injection is done through can be removed from the molten stainless. Once tubes called tuyeres in the bottom of the barrel- the steel has been deoxidized, sulfur can be shaped vessel. The injection and the reaction readily removed by contact with a basic slag. cause extremely thorough mixing, which would Sulfur can be reduced to less than 0.001% in never happen in the flat, stagnant, arc furnace the AOD, and this excellent purity level is com- bath. This mixing not only allows the CO-pro- mercially furnished without additional price ducing reaction to reach equilibrium, but also premium. Sulfur, although a harmful impurity the mixing of the slag and metal also permits from a corrosion standpoint, is often deliber- desulfurization. By increasing the ratio of argon ately kept at moderate levels (0.008 to 0.015%) to oxygen in the injected gas as the refining pro- for tungsten inert gas (TIG) welding penetration ceeds, the carbon is selectively oxidized with- (see Chapter 17) and at high levels (0.15%+) for out concurrent chromium oxidation. A typical machinability (see Chapter 15). These trade-offs, starting ratio is 3 to 1 oxygen to argon/nitrogen which are beneficial to processors, should be by volume. The ending ratio can be as low as 1 viewed with skepticism by end users, whose to 9, oxygen to argon/nitrogen. The choice of product integrity is compromised. There are which inert gas to use, argon or nitrogen, is based processing methods for which higher levels of on cost and final nitrogen content desired. Stabi- sulfur are not necessary that are preferable to lized stainless steels require low carbon and ni- the end user while not compromising welding trogen levels, for instance, so the more expensive or machining costs. For example, machinability argon must be used. can be improved by calcium additions that pro- It is possible to use a vacuum system to keep duce malleable oxides to replace the deleterious the partial pressure of CO low when refining sulfides (see Chapter 15), and welding methods, with injected oxygen. This is the vacuum oxy- such as laser welding, can be used in many gen decarburization (VOD) process. The VOD cases to eliminate the need for the weld penetra- process can achieve slightly lower carbon levels tion enhancement of sulfur while increasing but does not achieve cleaner steel as some welding speeds. believe. Phosphorus is an impurity for which no prac- In both processes, after final carbon content tical removal technology exists in stainless has been achieved ferrosilicon is added to reduce steel. Any known process to remove it first re- the chromium in the slag and have it return to the moves chromium. Thus, it exists in almost all molten steel. The excellent mixing of the slag stainless steel at levels close to its normal speci- and metal in the AOD permits this to be done fication limit, about 0.030% in austenitic alloys efficiently. The silicon plus the manganese in the and 0.020% or less in ferritic alloys, which are Chapter 12: Melting, Casting, and Hot Processing / 157

made from a higher percentage of low-phospho- on interstitial solubility. The higher solubility rus carbon steel scrap. The deleterious effects of of carbon, nitrogen, and oxygen in stainless phosphorus on corrosion are not avoided unless steels is significant. A manganese/silicon deox- much lower levels are achieved. Consequently, idized stainless steel will still have about 100 its presence is tolerated since it has no differen- ppm of dissolved oxygen at the freezing tem- tial effect over the range in which it is found. perature as opposed to the less than 10 ppm of Heavy metals are eliminated by high-temper- oxygen found in aluminum-killed carbon steel. ature AOD blowing, as is hydrogen. Care must This oxygen precipitates as oxides in the solid be taken not to reintroduce such impurities after state. refining, which is a risk when using damp or Vacuum induction melting (VIM) is another contaminated scrap for coolant. method of melting stainless steels. This is a Alloy adjustment can be done in the AOD or nearly slag-free process, and little refining is preferably in a treatment-and-transfer ladle. The possible. Melt purity is largely controlled by the tapped molten steel generally has excess heat purity of the starting material, and use of AOD from the highly exothermic refining process. master melt stock for VIM remelting is com- This allows the composition to be measured and mon. Limited decarburization is possible via in- adjusted before it must be cast. This can be done jection of oxides such as Fe3O4 or SiO2 to create very precisely by wire feeding of alloying ele- CO evolution inside the vessel. Using this tech- ments through the slag into the heat, which can nique, very low carbon levels (less than 50 be stirred by argon bubbling via porous plugs. ppm) are achievable commercially. Use of VIM This technique is very effective for the fine- is generally limited to high-value, high-purity, tuning of reactive elements such as titanium. or low-tonnage melts. The refining treatments used for carbon steel and stainless steel are very similar, but there are subtle differences because of the difference in Remelting the thermodynamics of dilute solutions like carbon steel and highly alloyed, nondilute solutions Some stainless steels and related alloys are like stainless steel. Table 1 shows the factors by remelted to refine composition or ingot struc- which additions of various elements to stainless ture. There are two principal remelt processes: steel (j) alter the thermodynamic activity of vacuum arc remelting (VAR) and electroslag other alloying elements (i). remelting (ESR). Equation 4 is used to calculate the activity of In VAR, the material to be remelted is cast elements in steel. The activity coefficient γ into a cylindrical electrode and placed inside a varies with the concentration of alloying ele- cylindrical water-cooled vacuum chamber. A ment x by: high-current direct current (dc) arc is established between the electrode and a starter plate n δγln at the bottom of the chamber. The end of the RTlnγγ=+ RT ln 0 ∑ RT i ii (Eq 4) electrode is melted, and the molten drops fall = δx jn1... j through the intervening vacuum. Volatile constituents escape from the molten drops, and the This calculation is best left to computer pro- purified drops collect to form a molten pool on grams such as Thermo-Calc that have been per- top of the starter plate. VAR parameters are ad- fected for these lengthy procedures. It should justed to maintain a shallow pool, which solidi- be noted that chromium, which is always pres- fies in a bottom-up fashion. The shallowness of ent in nondilute quantities, has a powerful effect the molten pool produces a refined grain

Table 1 Inﬂuence of alloying elements on the thermodynamic activity of carbon, nitrogen, sulfur, and oxygen J i Al C Cr Mn Mo N Ni O S Si Ti W O .04 .14 –.02 –.01 –.01 .11 .01 –.34 .05 .08 . . . –.005 N –.03 .13 –.05 –.02 –.01 0.0 .01 .05 .01 .05 –.53 –.001 S . . . .11 –.01 –.03 .003 .01 0.0 –.27 –.03 .06 –.07 .01 0 –.39 –.45 –.04 –.02 .003 .06 .006 –.20 –.13 –.13 –.6 –.01 158 / Stainless Steels for Design Engineers

structure with less solidification segregation ceramic tube, called the submerged entry than found in typical cast product. nozzle, into the mold, which is covered with a In ESR, the material to be remelted is cast consumable protective and lubricating slag into an electrode of similar shape, but slightly cover, called a mold powder. The mold powder, smaller than the water-cooled mold. A gap be- which melts in the mold as it is added, contains tween the electrode and a starter plate at the bot- ceramics, fluxes, and carbon. The level of the tom of the mold is filled with a prepared slag. molten metal should be carefully controlled by Typically, this slag is calcium fluoride-based ultrasonic measurement, or other methods, to with high lime (CaO) content. Additional ingre- prevent fluctuations in level that may entrap dients control the basicity, fluidity, oxidizing slag in the slab surface. The entire water-cooled, potential, and other properties of the slag. A copper alloy mold oscillates in a precise pattern high current is used to melt the slag, which in as the solidifying strand of steel is withdrawn turn melts the end of the electrode, and the from the mold bottom by pinch rolls and molten drops fall through the slag. Reaction of sprayed with water to cool it. The pinch rolls the molten drops with the slag removes sulfur apply enough pressure to slightly deform the and some other impurities, and the purified slab. This deformation has a crucial, seldom-rec- drops collect to form a molten pool on top of the ognized effect. It causes a beneficial recrystal- starter plate. ESR melting typically is done at a lization that improves hot working characteris- higher rate than VAR, and the molten pool is tics of austenitic and duplex alloys. In ferritic deeper. This deeper pool produces a grain struc- alloys, it can cause excessive grain growth, ture between that of VAR and typical cast prod- which detracts from hot workability. The initial uct, with commensurate intermediate segrega- portion of slab cast in a sequence is seldom tion patterns. of adequate quality to be used because of exogenous inclusions, entrapped mold powder, and non-steady-state solidification structure. Casting The defective portion must be identified and scrapped or diverted to low-quality requirement Continuous slab, billet, and bloom casting end uses. have become the standard methods of making The strand is bent from an initial slightly stainless steel primary products, replacing the curved shape to flat and cut into slabs. More obsolete ingot method. There are some alloys than one heat of steel may be cast sequentially that cannot be continuously cast, but these repre- without restarting the process. This is ideal eco- sent a miniscule percentage of stainless produc- nomically and for quality reasons since initial tion. Continuous casting produces slabs directly, and final segments of a casting can contain thus removing the costly soaking and slab- more inclusions and aberrant structure. Some rolling processes. In a well-executed continuous end users stipulate that no first slabs be applied casting operation, slabs are of sufficient quality to their orders. Producers generally apply first that they require no surface conditioning before slabs to less-critical uses or discard suspect sec- being hot rolled. Slabs range in thickness from tions of them. If casting conditions are not opti- 13 to 63 cm (5 to 15 in.). The segregation in con- mal, the result can be slabs with poor surface tinuous casters is less than in ingots because of quality that must be surface ground. the smaller section size. It is not eliminated, Slabs are sometimes quenched to avoid pre- however, and certain alloying elements concen- cipitation of phases; however, they may be held trate at the centerline, where they defy homoge- at high enough temperatures prior to hot rolling nization. Carbon and molybdenum are examples to stay above the temperature range in which of alloying elements with this tendency. embrittlement can occur or to stay above the In properly executed continuous casting, the temperature at which an embrittled slab can ladle feeds by a slide gate, or preferably a stop- fracture. Ferritic and martensitic alloys are es- per rod gate, into a ceramic tube into the large pecially prone to these problems. tundish situated over the caster mold. The metal There has been great interest for decades in in the tundish is covered with a protective slag producing stainless steel coils directly from cover, and flow patterns within the tundish are the melt in so-called strip casters. Elimination designed to minimize dead spots and encourage of hot rolling could be quite valuable in stain- removal of inclusions by impingement with the less steel, whose hot rolling from slab can be slag cover. The metal feeds through another both expensive and problematic. There are a Chapter 12: Melting, Casting, and Hot Processing / 159

number of such machines in pilot or limited mills that routinely roll carbon steel to 1.5 mm production. They have not had sufficient com- (0.06 in.) can struggle to attain 4.5-mm (0.18-in.) mercial or technical success to have become a thickness for 316 stainless. factor in the industry. Since their development The high separating forces on the hot rolling is only being undertaken by those large stain- mill stands also cause greater roll deflection and less steel producers who already have the hot compression, which if not countered by roll rolling assets that strip casting would replace, bending or roll shifting schemes can lead to sig- it seems unlikely that strip casting will soon nificant variation in thickness across the sheet, become a major factor even if it is perfected as much as 0.25 mm (0.01 in.). This variation as technically. a percentage of thickness is not reduced by cold Another method of shortcutting the casting/ rolling and is a major cause of tolerance loss in ingot step has been perfected: the powder met- sheet and strip. Hot-rolled bands vary in thick- allurgy approach. In powder metallurgy, the re- ness along the length of the coil because the tail fined molten metal is atomized by gas or liquid end of the slab is colder and harder to roll. Coil and made to freeze into small particles. These boxes (on reversing mills) address this problem particles, having been quenched extremely to a degree by permitting the semirolled coil to rapidly, are quite homogeneous. Powder tech- equalize in temperature. nology methods allow for the design of alloys Hot strip tandem mills powerful enough to that would otherwise freeze with too much successfully roll high-quality stainless steel hot- segregation and too coarse a structure with rolled bands are massively expensive and are conventional production methods. Traditional seldom justified for the tonnage of stainless steel powder metallurgy production methods are rolling a given melt shop produces, although used to make small near-net shape compo- rolling stainless on hot tandem mills used pri- nents, avoiding most of the costly machining marily for carbon steel can be an excellent pro- steps. More impressively, powder technology duction method. is also used to produce massive components. Hot Steckel mills have become the favored For example, very high carbon/vanadium method of hot rolling stainless steel because stainless tool steel components can be made by their throughput better matches stainless steel encapsulating powder in an evacuated canister melt shop production outputs. This permits the in which it can be sintered and hot worked to melt shop and caster to be adjacent to the hot 100% density and virtually complete homo- mill, which permits energy-saving hot charging geneity. Chapter 9 on martensitic alloys dis- of slabs. In hot Steckel mills, typically a four- cusses these materials. high reversing rougher rolls slabs to about 3-cm (1.2-in.) thick. Then, the transfer band is rolled to final gauge on a separate reversing four-high Hot Rolling finishing mill with coil boxes to preserve temperature. The economy of having only two mill Hot rolling remains an essential process for stands makes these mills ideal for typical stain- the vast majority of stainless steel used. Hot less production quantities and permits the cost rolling characteristics of stainless steels vary of sophisticated mill capabilities, such as roll greatly. Ferritic stainless steels are extremely shifting, roll crossing, or roll bending, not to easy to hot roll since they have a soft, single- have to be duplicated among many stands. This phase structure at hot rolling temperatures. is the same justification for using Sendzimir Martensitic stainless steels roll like their carbon mills to cold roll stainless. In both cases, the and alloy steel counterparts since their mi- logic applies more to austenitic alloys than to crostructure during hot rolling is a moderately the easily rolled ferritic stainless alloys. alloyed austenite similar to alloy steels. The mi- In either case, the hot-rolled band carries a crostructure during hot rolling is the crucial fac- heavy, embedded scale that must be removed tor. Austenitic stainless steels have high strength from the surface before further processing in at hot rolling temperatures. Furthermore, the most cases. Some alloys can be cold rolled in the low diffusion rates in austenite slow recrystal- “black band” state at a cost of coarser surface lization so that the steel does not always soften finish and greater rolling mill roll wear. If nor- between stands in tandem mills. This increases mal cold rolling or use as hot-rolled coil is fore- mill loads, and lower reductions must be taken seen, the hot-rolled band must then be annealed than for alloy steels. Powerful hot strip tandem and pickled since the as-rolled hot-rolled band 160 / Stainless Steels for Design Engineers

has poor corrosion resistance, poor mechanical rily as edge cracks and slivers. Edge cracks are properties, residual cold work and hardness vari- simply a lack of ductility at the colder strip edge. ations, as well as a heavy oxide layer. Stainless hot ductility often has a narrow temperature window, and many factors can affect the size of that window depending on alloy type. Defects The most inherently challenging alloys for hot working are the duplex alloys and the alloys that Stainless steel hot-rolled bands can contain solidify in the fully austenitic state. The former many types of defects. These are seldom seen has a mixed-phase structure, and the phases can by the end user because they are removed exhibit mechanical incompatibility at certain when they are not prevented. They do have temperatures. The latter alloys reject sulfur and repercussions on delivery. The major cate- oxygen during solidification and slab reheating gories are: to the grain boundaries, where they form very • Hot mill defects weak films. But, even alloys such as 304 and • Inclusion-related defects 316 can have very poor hot ductility if they con- • Hot ductility-related defects tain much sulfur and oxygen or if they are reheated for long times or at temperatures above Stainless steel is less forgiving of hot mill 1250 °C (2280 °F), which facilitates diffusion of faults because its surface is not removed by oxi- sulfur and oxygen to the grain boundaries and dation to the degree carbon steel’s surface is. also encourages very large grains. This poor hot Thus, a skid mark from a slab-heating furnace ductility manifests itself as “slivers,” which can will remain through the hot rolling, annealing, require grinding of the entire hot band surface. and cold rolling processes. This is true of all hot These tendencies are fought by low oxygen and mill scratches, gouges, digs, etc. Rolling stain- sulfur levels and minimal slab-reheating temper- less requires a different mindset than rolling atures and times, as well as slab surface working carbon steel, which argues against the benefits in the caster pinch rolls. Sometimes, very poor of rolling stainless on a mill built and primarily hot working alloys are given a single hot reduc- used for carbon steel. tion pass on a hot mill to produce a full recrystal- Inclusion-related defects are all essentially lization that disperses grain boundary-weakening avoidable by using state-of-the-art technology. elements on subsequent reheat. Protecting metal from reoxidation and keeping precise mold-level control in the continuous caster prevents all inclusions of a size that can REFERENCE produce a defect. Hot ductility defects are more subtle. They 1. D. Peckner and I.M. Bernstein, Handbook of arise from many causes and are manifest prima- Stainless Steels, McGraw-Hill, 1977, p 3–13 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 161-171 All rights reserved. DOI: 10.1361/ssde2008p161 www.asminternational.org

CHAPTER 13

Thermal Processing

Summary desirable, and some are potentially very detrimental. Readers are encouraged to review the THE THERMAL PROCESSING of stainless earlier chapters on phases in stainless steel steel is a topic the end user should approach (Chap. 6Ð10) to familiarize themselves with with great respect. It is not simple in concept or these phases. in practice. Before attempting to carry out any Each of the alloy groups of stainless steels has thermal processing on stainless steel, the practi- radically different thermal processing objectives tioner must understand the alloy design, compo- and requirements; therefore, each is discussed sition, and processing history of the material in separately. question. The thermal processing then must be designed and executed in a planned, controlled manner. The consequences of failure in thermal Austenitic Stainless Steels processing can become catastrophic to mechanical properties and corrosion resistance. Thermal processes applied to austenitic stainless steels include: ¥ Soaking for homogenization and preparation Introduction for hot working ¥ Annealing to remove the effects of cold The thermal processing of stainless steels can work and to put alloying elements into solid have many purposes. Normally, the objectives solution (solution annealing) are simple: heating for hot working, annealing to ¥ Stress relieving soften after cold working, solution annealing The temperatures at which these processes to homogenize, heating to temper martensite, or are carried out are shown in Table 1 for typical to stress relieve. However, even if the objective austenite compositions. is simple, the processes that occur are anything but simple. Variations in temperature, times at temperature, heating and cooling rates, and at- Soaking mosphere can have complex and easily unin- Because virtually all stainless steel is continu- tended consequences. There is no substitute for ously cast, the older soaking process of holding understanding the processes that are occurring ingots in soaking pits for many hours is rarely when stainless steels are heated for successful used. The soaking had two functions. The obvi- heat treating to be achieved. ous one was to equilibrate at the right tempera- Stainless steels have many alloying elements ture for hot working. The second, less-obvious, in large amounts. Many of these elements are one was to achieve greater chemical homogene- highly reactive thermodynamically. The practi- ity. The lack of homogeneity comes from the cal consequence of this is that many phases are solute segregation that occurs as solute ele- thermodynamically possible at different temper- ments are rejected from the material that was atures. Stainless also reacts with its environment ﬁrst to freeze. Solute segregation was exagger- at high temperatures, causing changes in surface ated by the slow solidiﬁcation of ingots, and alloy content. Some of the resulting phases are continuous casting helped make the stainless 162 / Stainless Steels for Design Engineers

Table 1 Recommended thermal processing temperatures for austenitic alloys Annealing Annealing ASTM A480(a) Stress Stress Alloy temperature, ¡C temperature, ¡F 2006, ¡F relieving, ¡F relieving, oC Standard alloys ...... 1500Ð1600 non-L, 815Ð870 non-L, 1000Ð1600 L grades 540Ð870 L grades 201, 202, 201LN 1010Ð1120 1850Ð2050 1900 min ...... 301, 301LN, all versions 1010Ð1120 1850Ð2050 1900 min ...... 304, 304L, 305, all versions 1010Ð1120 1850Ð2050 1900 min ...... 316, 316L, 316N, 317, 317L 1040Ð1175 1900Ð2150 1900 min ...... 308, 309, 309S, 310, 310S 1040Ð1175 1900Ð2150 1900 min ...... Stabilized alloys ...... 1000Ð1600 540Ð870 321 955Ð1065 1750Ð1950 1900 min ...... 347, 348 980Ð1025 1800Ð1950 1900 min ...... 20Cb-3 925Ð955 1700Ð1750 . . . 925Ð1010 . . .

Moderately alloyed, Creq<26 1120Ð1175 2050Ð2150 Various 1500Ð1600 815Ð870 S31725, N08028, JS700 Highly alloyed, lower sigma 1120Ð1175 2050Ð2250 Various Not recommended . . .

alloys, Creq>30, high N AL6-XN, 4565, 654SMO, 254SMO Highly alloyed, sigma-prone 1205Ð1230 2200Ð2250 Various Not recommended . . .

alloys, Creq>30, low N AL6-X

(a) Standard specification for general requirements for flat-rolled stainless and heat-resisting steel plate, sheet, and strip. steel more homogeneous. Nevertheless, cast rounding matrix by sufficient soaking. Welds that slabs and blooms must be soaked to eliminate are unannealed have such precipitated inclusions as-cast segregations. This process, to the extent in an unequilibrated state, and the result is dimin- it is done, occurs as they are reheated to the ap- ished chromium concentration and poorer pitting propriate temperature for hot working. Soaking resistance. dissolves the few percent of residual delta ferrite that are present on slab solidification. It is important to soak at the highest temperature at Annealing which delta ferrite is not a stable phase so that it Annealing serves two main functions in will dissolve, about 1250 ¡C (2280 ¡F) for most stainless steel: It removes the effect of cold austenitic stainless steels. work by replacing strained microstructure with Soaking at higher temperatures causes ferrite new strain-free grains, that is, recrystallization. levels to increase, negating the homogenization New grains nucleate and grow. If stored strain and causing very poor hot workability. Longer energy is insufficient, as happens often with times at temperature than the minimum required ferritic stainless steels, true recrystallization is for thermal uniformity also cause problems as difficult to achieve, and the annealing process any sulfur and oxygen impurities are rejected may only produce recovery without recrystal- from austenite and can diffuse to grain bound- lization, leaving the same grains relieved of aries, where they form weak, plastic films that stress. This leaves the surviving grains with the also degrade hot workability. Grain growth, by same crystallographic orientation that deforma- reducing grain boundary area, exacerbates this tion produced and may or may not be the de- effect. Thus, soaking times are best minimized sired outcome. Second, annealing returns into and closely controlled. Alloys are therefore de- solution solute that has been precipitated as un- signed to have only a slight amount of delta fer- wanted phases, principally carbides, but also rite to be redissolved during soaking. Ferrite is intermetallic phases. Annealing also may help useful because it has a high solubility for oxygen to reduce solute segregation remaining from and sulfur. Having none would result in impurity the casting process, making the composition rejection of these elements to grain boundaries more homogeneous. The homogenization during solidification, which results in the worst- process is accelerated by the reduction in di- possible hot working characteristics. The oxygen mensions from hot and cold working. A reduc- and sulfur trapped in the ferrite during solidification in dimension by a factor of two reduces the tion precipitates in the solid state as inclusions, time to achieve a given degree of homogeniza- which also must be equilibrated with the sur- tion by a factor of four. Chapter 13: Thermal Processing / 163

Annealing to recrystallize is fairly rapid. To a carbides can be redissolved and carbon dif- first approximation, it is instantaneous, and the fused away from the carbide, but this does not results are merely a function of the maximum mean that all composition gradients have been temperature attained. This may not be the case reduced to zero. This means that precipitates for continuous annealing lines, in which transit may re-form more rapidly in such a material time can be short enough, less than a minute at than they would in a completely homogeneous temperature, to limit the grain size attained. The alloy. driving force for recrystallization is the strain The annealing temperature for a given alloy is energy stored in the lattice from deformation. chosen based on the temperature required to put The strain energy in a given material is propor- all alloying elements into solution. Higher car- tional to the square of the material’s flow stress. bon levels, for instance, require higher tempera- As the material is heated, recovery occurs first. tures to dissolve all the carbon. This is one of This is the change in physical and mechanical the principle values of accurate phase diagrams. properties associated with dislocation annihila- If it were simply a consideration of recrystal- tion and polygonalization that occurs before the lization, all alloys could be annealed at similar nucleation and growth of new grains. temperatures at the low end of the recom- The nucleation of new grains occurs at high- mended range. Within the recommended range, angle grain boundaries and proceeds by the the temperature selected should be determined movement of roughly hemispherical growth by the desired grain size. End use determines fronts into strained areas. The percentage re- whether a fine or coarse grain size is preferable. crystallized, once a sufficient temperature is Table 1 lists recommended annealing tempera- reached, grows sigmoidally. It is normal for tures for austenitic stainless steels. the time to fully recrystallize to be rather less The overall interplay between prior cold work than the time to attain that temperature. Even and annealing temperature on mechanical prop- at the lower range of annealing temperatures, erties of annealed material can be summarized times are generally less than 1 min. Recrystal- as (Ref 3): lization will not occur if the lattice contains insufficient strain energy. Thirty percent cold ¥ Grain size of a given alloy is the most im- work should be used as a rough threshold for portant parameter in characterizing mechan- the required amount. Annealing after lower ical properties. amounts of cold work is characterized by ¥ Yield and tensile strength are essentially scarce nucleation sites and can result in very constant for a given grain size regardless of large and nonuniform grain size. Hot-worked the amount of prior cold work; however, the material often has a composite structure that elongation depends on the prior reduction. may have already had some recrystallization ¥ Yield strength, tensile strength, and hardness depending on the final reduction temperature. are essentially linear functions of grain size. Annealing may not produce a clear recrystal- ¥ Elongation decreases with finer grain size lized structure in this case. and at an increasing rate as grain size be- The relative rapidity of recrystallization an- comes finer as long as cross-section size of nealing is due to the fact that it is rate con- the test specimen is not extremely small. trolled by short-range diffusion. Solution an- This is not true for very coarse-grained ma- nealing requires longer-range diffusion and terial. thus can require much longer times. Some stud- ¥ Maximizing elongation comes from maxi- ies have shown that welds, for instance, do not mum prior cold work and medium annealing recover completely from their loss of corrosion temperatures properties that arise from local alloy depletion ¥ Anisotropy coefficients, or plastic strain ra- unless they have been annealed for times on the tios r are constant up to about 40% reduction order of 1 h (Ref 1). Others have seen homoge- after, which r45 and rn increase sharply, while nization in as little as 10 min (Ref 2). Wrought rt decreases. This leads to earing during materials can require much shorter times be- drawing. cause reductions during hot working have re- ¥ The increase in properties for a one ASTM duced diffusion distances. It should be noted grain size increment is: that precipitates can be redissolved and not ap- a. 13 MPa (2 ksi) for tensile strength parent in the annealed microstructure without b. 20 MPa (3 ksi) for yield strength full homogeneity being achieved. For example, c. 2 HRB for hardness 164 / Stainless Steels for Design Engineers

Atmospheres for annealing are important. metallic phases. Use of high-chromium and- Austenitic stainless steels heated in air, of molybdenum alloys without enhanced nitrogen course, form oxide scales. Beneath this oxide, is no longer recommended, and the use of lower- the metal matrix becomes significantly depleted nitrogen alloys should be reexamined and ques- of chromium (Ref 4), often more than 5% lower tioned if specified. in chromium and to a depth of as much as 10 µ Last, stainless surfaces should be scrupu- (395 µin.). So, not only must any oxide be re- lously clean before annealing. Even hard water moved, so must the chromium-depleted layer. deposits can cause differential oxide growth, This requires aggressive pickling, which while which can cause etched spots on the surface, done commonly, may not be practical for many where the postanneal pickling attacks the differ- stainless users. The chromium-depleted zone, ent oxide more strongly. Carbonaceous materi- however, does pickle rapidly precisely because als left on the surface are even more objection- it does have less chromium. To avoid oxide able because they can cause carburization and scale formation, vacuum, hydrogen, or inert gas subsequent loss of corrosion resistance. atmospheres may be used. Stabilizing anneals are sometimes conducted If vacuum is used, it should be less than on stabilized alloys such as 321 and 347. This is 2 × 10Ð3 torr (0.3 Pa). If an inert gas or hydrogen useful when carbon levels are sufficiently high is used, the key consideration is moisture con- that significant dissociation of carbides occurs tent. The dew point must be Ð40 ¡C (Ð40 ¡F) or at annealing temperatures. A second anneal at lower. More stringent levels may be required if lower temperature, about 900 ¡C (1650 ¡F), mirror finishes are desired. Cool down must be then is done to permit the carbon to combine rapid as oxidation potential increases as tempera- with the stabilizing element rather than leaving ture decreases. Vacuum or inert gas is preferable it available to form chromium carbides. Current to hydrogen for alloys containing stable oxide preferred practice for these alloys is to maintain formers such as aluminum or titanium or for al- carbon and nitrogen below 0.03% for corrosion- loys containing boron. resistant service, which renders this stabilizing Austenitic alloys that are subject to sensitiza- unnecessary. Alloys used for high-temperature tion must be cooled rapidly enough from anneal- service benefit from the creep-resisting contri- ing temperatures to avoid carbide precipitation butions of higher carbon levels. during cooling. If forced air or water quenching are impractical or if section size prohibits rapid cooling, then using stabilized or low-carbon Stress Relieving grades is indicated. Austenitic stainless steel weldments often con- Superaustenitic stainless steels, and even al- tain residual stresses, which can cause distortion loys like 317, present a special problem because or lead to stress corrosion cracking in service. these alloys have significant sigma-forming ten- They are commonly stress relieved at tempera- dencies. Sigma forms initially because solidifi- tures slightly below the annealing temperature, cation segregation causes local enrichment of so that residual stresses may be relieved by sigma-promoting elements, such as molybde- creep. One hour at 900 ¡C (1650 ¡F) reduces num. It can also form from slow cooling of slabs residual stress by about 85%. Lower tempera- or hot bands. This latter sigma forms at grain tures require exponentially longer times for the boundaries and will cause embrittlement and re- same stress relief, with times doubling for each duced corrosion resistance, so it must not only 100 ¡C (180 ¡F) decrement as decreasing diffu- be redissolved, but also the alloy must be ho- sion rates, which govern creep, are encountered. mogenized to remove the residual concentration Cold-worked austenitic stainless steels have a gradients from the sigma. If this is not done, markedly diminished proportional limit, partic- chromium- and molybdenum-depleted regions ularly in compression. This Bauschinger effect, will still exist, and sigma will re-form much which arises from the easy mobility of disloca- more rapidly during subsequent exposure to tions, can be eliminated by stress relieving at high temperatures. For this reason, the higher around 350 ¡C (660 ¡F) for 2 h, which provides ends of the annealing ranges are recommended, the thermal energy for dislocation interactions and annealing times should be generous. Newer to lock into place. This produces a sharp yield alloys have higher nitrogen contents to suppress point without premature nonproportional elastic formation of sigma and other deleterious inter- deformation. Chapter 13: Thermal Processing / 165

Ferritic Stainless Steels phases that require dissolution. Alloys with high chromium and molybdenum contents can form Ferritic stainless steels, from an annealing σ and/or α', the brittle, ordered body-centered point of view, must be discussed in two cate- cubic (bcc) phase, at temperatures below an- gories. First are the modern, stabilized alloys, nealing temperatures, so rapid cooling is pru- which are ferritic at all temperatures. These al- dent when chromium plus molybdenum ex- loys behave as interstitial-free (IF) alloys be- ceeds 20%. cause the interstitial carbon and nitrogen are re- The driving force for recrystallization in these moved from solution as a stable precipitate. In alloys is limited by the lower stored energy the second category are the older ferritic steels, from deformation inherent in the bcc structure. which have enough austenitizing elements, usu- In addition, the pronounced deformation texture ally carbon, in solution to cause them to form leads to annealing responses that are more accu- austenite at what would otherwise be a good an- rately characterized as recovery and grain nealing temperature. This makes them truly growth with diminished recrystallization. These quasi-martensitic alloys, and they must be alloys retain this texture after annealing, and treated accordingly. Table 2, which lists heat- this characteristic anisotropy is exploited for treating temperatures for ferritic stainless alloys, good drawability. The major concern is to avoid also shows which grades ﬁt into which category. excessive annealed grain size, which greatly reduces toughness. Anneal at the higher end of the Soaking range only if the loss of toughness associated Heating of ferritic stainless for hot working is with large grain size is not a concern. Stabiliz- straightforward. Whether stabilized or not, ing anneals are normally unnecessary for stabi- these alloys are heated to the 1000 to 1100 ¡C lized ferritics as their high diffusion rates ensure (1830 to 2010 ¡F) range for hot working. The freedom from knife-edge attack due to sensiti- superferritics can be heated to up to 1300 ¡C zation from free unbonded carbon combining (2370 ¡F). At this temperature, no debilitating with chromium at grain boundaries. The stabi- phases occur, and ductility is good. The high lizing additions of titanium and/or niobium tie diffusion rate inherent to the ferritic structure up the carbon as stable TiC or NbC, which does makes homogenization easy. As long as hot not redissolve during annealing. working is completed at temperatures above The interstitial-bearing ferritic stainless that at which austenite forms, good hot ductility steels must be annealed subcritically, or the for- is expected. This is not a concern with IF alloys, mation of austenite at higher temperatures which do not form austenite. would make martensite formation on cooling virtually unavoidable. Thus, a typical primary Annealing anneal cycle for a typical alloy such as 430 would be nearly 24 h at 750 ¡C (1380 ¡F), the The IF ferritics do not undergo any phase majority of which is thermal equilibration of change with temperature during the course of the large coil mass. The actual time at tempera- properly executed heat treatment. The objective ture required is less than 1 h. Continuous an- of annealing is generally simply to remove the nealing is not practical because the diffusion of effects of cold work. This is because they do not carbon is too slow to occur in the dwell time at need to have carbon put into solution and, ex- temperature typical in continuous annealing cept in rare cases, do not have intermetallic lines. This cycle also precipitates essentially all the carbon and nitrogen as mixed Cr/Fe car- Table 2 Recommended annealing temperatures bides and nitrides and homogenizes chromium for ferritic alloys content. This necessarily slow process permits Annealing Annealing subsequent subcritical annealing for mechani- Alloy temperature, oC temperature, oF Stabilized, Cr+Mo<20 cal properties (to alleviate the effects of cold 409, 439,18 SR 870Ð925 1600Ð1700 work) to be done in a few minutes since carbon Unstabilized, Cr+Mo<20 has been eliminated from solution by the for- 405, 430, 434, 436 705Ð790 1300Ð1450 Stabilized, Cr+Mo>20 mation of fairly stable carbides. Since the ma- 29-4C, Monit, Seacure, 444 1010Ð1065 1850Ð1950 terial is generally purchased in the annealed Unstabilized, Cr+Mo>20, condition, the user need never be concerned 446 760Ð830 1400Ð1525 with such lengthy anneals. 166 / Stainless Steels for Design Engineers

Stress relieving is rarely a concern for any martensite cannot be avoided by furnace cool- type of ferritic stainless. Unstabilized grades ing from austenitic temperatures, then only sub- should not be welded, and if they are, full subcritical annealing is feasible. But, even for critical annealing is required. Stabilized grades nickel-free alloys the hardenability is so great have no need for postweld heat treatment. Low- that annealing by slow cooling is quite difficult. temperature heat treatment runs the risk of α' Martensitic alloys are put into the annealed con- formation and is best avoided. dition for processing before they are quenched and tempered for their ﬁnal use. Thus, the more economic subcritical anneal is the predominant Martensitic Stainless Steels annealing heat treatment. The nickel-bearing alloys have such high The martensitic stainless steels resemble the hardenability that annealing in the critical range unstabilized ferritic stainless steels described. cannot produce softening by any practical cool- The martensitic stainless steels form essentially ing rate, so subcritical annealing is always rec- 100% austenite on heating and have very high ommended for these alloys. Nickel reduces the hardenability, so their ability to be softened by temperature at which austenite is stable as annealing is limited. The traditional martensitic shown in Chapter 9, Fig. 9. Other additions like stainless steels are iron/chromium/carbon al- vanadium, molybdenum, and tungsten promote loys, sometimes with a small amount of nickel secondary hardening and tempering resistance, and/or molybdenum. More recently, alloys have and subcritical annealing of these alloys be- been developed for petroleum applications that comes a slow, difficult process. This is a charac- contain high copper, nickel, and/or molybde- teristic of the so-called super 12Cr alloys. num and low carbon. The principles of heat Martensitic alloys have lower corrosion re- treatment of the two alloy categories are the sistance in the annealed condition than in the same. The more highly alloyed newer alloys hardened condition because in this state they are, in fact, simpler to heat treat because their have the maximum amount of chromium tied up low carbon and nitrogen levels alleviate the as chromium carbide. need to temper. Austenitizing Soaking Table 3 lists the austenitizing and tempering Hot working should be carried out in the ranges for martensitic stainless steels. Full austenitic range. Temperatures for this are austenitizing is crucial to producing a micro- listed in Table 3. Forging and hot working structure that is fully martensitic. Only austenite should always be followed by annealing to transforms to martensite. If other constituents, δ avoid stress cracking due to the deep hardening such as ferrite or carbides, exist during the of these alloys. austenitizing heat treatment before quenching, they will not transform to martensite. Some alloys, such as the 440 group, have enough car- Annealing bon that the austenitizing temperature deter- Martensitic stainless steels can be annealed mines how much carbon is put into solution. by subcritical anneal and sometimes by full an- The carbon in solution in the austenite will be- neal depending on alloy level. If the alloy level come the carbon level in the martensite, which is such, as in the nickel-containing grades, that directly determines strength and corrosion

Table 3 Recommended annealing, austenitizing, and tempering temperatures for martensitic alloys Alloy Subcritical anneal, Full anneal, Austenitizing, Tempering, low Tempering, high oC (oF) oC (oF) oC (oF) range, oC (oF) range, oC (oF) Straight 650Ð760 830Ð885 925Ð1010 205Ð370 400Ð700 Cr, C<0.20, 410, (1200Ð1400) (1525Ð1625) (1700Ð1850) (565Ð605) (1050Ð1125) 416,403 Ni/Mo, C<0.20, 620Ð705 Not recommended 980Ð1065 205Ð370 400Ð700 414, 431, 415, (1150Ð1300) (1800Ð1950) (565Ð605) (1050Ð1125) 425, C>0.20 440A,B,C F, 420 675Ð760 845Ð900 1010Ð1065 150Ð300 Not recommended (1245Ð1400) (1555Ð1650) (1850Ð1950) (300Ð700) Chapter 13: Thermal Processing / 167

resistance because undissolved carbides contain precipitate with the abundant iron atoms first. At chromium, which diminishes that available for higher temperatures and longer times, more ther- corrosion resistance. Austenitizing temperature modynamically stable carbides, such as Cr23C6, and holding time become most significant when form. Carbide formation is a complex function carbon exceeds 0.20%, where its solubility is a of temperature, time, and composition. The steep function of temperature. growth of carbides reduces strain and hardness. The ␦ ferrite is a generally undesirable phase There are exceptions, such as the precipitation of that can be produced by temperature excursions Mo2C, whose morphology produces a precipita- or composition variations. Excessive austenitiz- tion hardening (PH), called secondary hardening temperatures can cause its formation, as can ing. Niobium and vanadium also form carbides low levels of carbon, which may be originally that result in higher hardness at all tempering present in the alloy or arise from decarburiza- temperatures. Had this been understood early tion. It will cause lower hardness and toughness on, these steels could have been correctly in- if present. cluded in the PH stainless group. The PH steels, Heating rates should be such that a uniform AM-350 and AM-355, both derive their PH temperature is attained before the allotropic from the precipitation of Mo2C and Mo2N. In all transformation from bcc to face-centered cubic other cases, higher tempering temperatures lead (Fcc), which involves a more than 1% linear di- to lower hardness. mension change and can cause distortion or The nickel-bearing alloys have a restricted cracking. Oxidation during austenitizing can upper tempering temperature because of the cause serious carbon loss on the surface, which danger of re-forming austenite, which would will result in serious loss of surface hardness. then transform to untempered martensite during Heating 410 in air for 10 min at 1100 ¡C (2010 cooling, requiring a second tempering operation. ¡F) can cause surface carbon to decrease by Intermediate temperatures can lead to the one-half, lowering hardness from HRC 45 to phenomenon of temper embrittlement, which is under 20. caused by the precipitation of phosphorus and Quenching rate is not a significant issue for other species, such as, but not limited to, car- the martensitic stainless steels since they have bides, at prior austenitic grain boundaries. This such high hardenability, but some, especially phenomenon is distinct from the precipitation of those with higher carbon levels, may have re- α', which causes the so-called 475 ¡C embrittle- tained austenite, which can lower hardness and ment, which occurs more severely in alloys cause problems with dimensional stability. The with higher chromium levels. Because the for- quenching rate must be sufficient, however, to mation of martensite is diffusionless, the avoid precipitation of carbides in the austenite austenite boundaries maintain the microcompo- during quenching since the sensitization would sitions they have at high temperatures. Auste- persist in the final microstructure. If this oc- nite has low solubility for impurities such as curs, a subzero treatment at below Ð75 ¡C phosphorous, so phosphorous is more highly (Ð100 ¡F) should be undertaken immediately concentrated in the grain boundary regions. after quenching. This compositional inhomogeneity can be eliminated by higher tempering temperatures or by the addition of molybdenum, which combines Tempering with the phosphorus and prevents the embrittle- Untempered martensite has insufficient tough- ment. The existence of temper embrittlement is ness to be a useful engineering material. During the reason that Table 3 recommends avoiding tempering, carbon is precipitated from the super- certain temperature ranges for tempering. saturated state it is in when it is quenched into Lower tempering temperatures and higher the bcc martensite structure during the diffusion- austenitizing temperatures are best for corrosion less transformation. The strain energy associated resistance because both minimize the amount of with the lattice strain of the bcc martensite chromium tied up as carbide. Quenching from caused by the poor fit of the carbon in the tetrag- higher temperatures also enhances toughness. onal interstices is very large. Heating to even low temperatures allows carbon enough mobility to diffuse and precipitate as carbide. Since Stress Relieving carbon diffuses 106 times as fast as iron, If quenched martensitic steels are not imme- chromium, or other carbide formers, it tends to diately tempered, then they should be promptly 168 / Stainless Steels for Design Engineers

stress relieved. Otherwise, the residual stress eliminating retained austenite, enhances dimen- from quenching could result in stress corrosion sional stability but diminishes toughness. Table 4 cracking even in seemingly benign environ- lists the solution treatments for all PH grades. ments. Stress relieving is simply a low-tempera- The as-quenched state is called condition A. ture tempering operation, but some elimination This is the normal condition in which the mate- of residual stress does occur. Higher tempera- rial is supplied from the mill and is intended to tures and longer times produce greater stress be soft enough for machining and some forming. relief and maximize elastic properties, but opti- If softer material is required, the H-1150M con- mal toughness is obtained at intermediate times dition exists in which the material is first highly and temperatures. overaged at 760 ¡C (1400 ¡F), allowing some austenite to re-form. The subsequent aging then overages that martensite while retaining some Precipitation Hardening Stainless Steels stable austenite. The result is a very tough microstructure. There are three classes of PH steels, and each Aging. The time and temperatures required requires totally different heat treatment. The to produce this precipitation are also given in classes are martensitic, semiaustenitic, and Table 4. The condition code itself tells the tem- austenitic. The most straightforward alloys are perature at which the aging is conducted in the martensitic PH grades. Like the plain marten- that the code numbers are based on thermal sitic alloys, the martensitic PH alloys are care- processing temperatures expressed in degrees o fully designed to produce a nearly fully austenitic Fahrenheit ( F), for example, TH 900 means structure at high temperature that quenches to a transformed to martensite (T) and aged at 900 nearly fully martensitic structure on cooling. The degrees Fahrenheit. The final properties are a martensite is low in carbon, so it is relatively soft function of both aging time and temperature. and not prone to brittleness. So, the hardness and Lower temperatures result in higher possible strength of these alloys is derived from a subse- hardness but lower toughness. The precipi- quent tempering-type heat treatment during tates, as mentioned, are optically invisible and which various constituent elements form ex- cause very little dimensional change. Contrac- tremely fine coherent precipitates that greatly tion on the order of 0.0005 in./in. from aging is strain and therefore strengthen the matrix. There typical, often permitting machining to final di- are numerous precipitates that can provoke this mensions in condition A. All aging treatments effect, and they are described in detail in the PH are above the temper embrittlement range to chapter (see Chapter 10). All require the short- which these alloys are susceptible. Service range diffusion of substitution elements to form temperatures in this range would result in em- these optically invisible precipitates. brittlement, so use above 350 ¡C (660 ¡F) should be avoided. Molybdenum-bearing grades should be selected to minimize this Martensitic PH Grades phenomenon if high-temperature use is con- Solution treatment of these alloys is con- templated. ducted to achieve a fully austenitic structure. Solution Heat Treatment and Conditioning. The constituent elements are easily dissolved, The semiaustenitic grades are more complicated so excessive temperature or time is unnecessary than the martensitic PH alloys. These alloys are and could be counterproductive if it were to re- designed so that they are austenitic when sult in ferrite formation or surface oxidation, quenched from the solution heat treatment tem- which would be detrimental to final mechanical perature. This is also called condition A, and it properties. The presence of retained ferrite is permits them to be highly formable. This stabi- mainly a function of alloy and composition lization of the austenite comes mainly from within the allowed range. Earlier grades such as higher chromium and carbon levels. These al- 17Ð4 and the obsolete stainless W intrinsically loys essentially resemble a lean 301 austenitic contained some ferrite. The subsequent alloys alloy, many with some molybdenum substitut- are substantially ferrite free. Most alloys may ing for part of the chromium. The key to these retain some austenite after quenching to room grades is making them behave as a martensitic temperature, in which case subzero treatment alloy. This is done by precipitating some of the should be done within 24 h to avoid further sta- carbon as chromium carbide at a temperature at bilizing the austenite. Subzero treatment, by the high end of what would normally be consid- Chapter 13: Thermal Processing / 169

Table 4 Recommended annealing and stress-relieving temperatures for martensitic PH grades Alloy Condition code Solution anneal Conditioning Aging Martensitics 13-8 A 925 ¡C 15 min, oil or air cool below 15 ¡C Ð75 ¡C 8 h RHxxx 925 ¡C 15 min, oil or air cool below 15 ¡C Ð75 ¡C 8 h xxx ¡F for 4 h Hxxx 925 ¡C 15 min, oil or air cool below 15 ¡C . . . xxx ¡F for 4 h 15-5 A 1035 ¡C 30 min, oil or air cool below 15 ¡C . . . Hxxx 1035 ¡C 30 min, oil or air cool below 15 ¡C . . . xxx ¡F for 4 h 17-4 A 1035 ¡C 30 min, oil or air cool below 15 ¡C . . . Hxxx 1035 ¡C 30 min, oil or air cool below 15 ¡C . . . xxx ¡F for 4 h 450 A 1035 ¡C for 1 h, water quench . . . Hxxx 1035 ¡C for 1 h, water quench . . . xxx ¡F for 4 h 455 A 830 ¡C for 1 h, water quench . . . Hxxx 830 ¡C for 1 h, water quench . . . xxx ¡F for 4 h 465, 275 A 980 ¡C for 1 h, cool rapidly Ð75 ¡C 8 h Hxxx 980 ¡C for 1 h, cool rapidly Ð75 ¡C 8 h xxx ¡F for 4 h 475 A 925 ¡C 1 h, air cool Ð75 ¡C 8 h Hxxx 925 ¡C 1 h, air cool Ð75 ¡C 8 h xxx ¡F for 4 h

Semiaustenitics 17-7, 15-7 A 1065 ¡C for 30 min, air cool ...... T 1065 ¡C for 30 min, air cool 760 ¡C (1400 ¡F), 90 min, . . . air cool to RT for 30 min. C 1065 ¡C for 30 min, air cool Cold reduce . . . R 1065 ¡C for 30 min, air cool 955 ¡C (1750 ¡F) 10 min, air . . . cool, chill to Ð75 ¡C for 8 h THxxx 1065 ¡C for 30 min, air cool 760 ¡C (1400 ¡F) xxx ¡F for 4 h 90 min, air cool to RT for 30 min CHxxx 1065 ¡C for 30 min, air cool Cold reduce xxx ¡F for 4 h RHxxx 1065 ¡C for 30 min, air cool 955 ¡C (1750 ¡F) 10 min, air xxx ¡F for 4 h cool, chill to Ð75 ¡C for 8 h AM-350 A 1010Ð1065 ¡C . . . L (equivalent to T) 1010Ð1065 ¡C 930 ¡C for 90 min, air cool . . . SC (equivalent to R) 1010Ð1065 ¡C 930 ¡C for 90 min, air cool, . . . 180 min at Ð75 ¡C SCTxxx 850 ¡F or 1010Ð1065 ¡C 930 ¡C for 90 min, air cool, xxx ¡F for 180 min 1000 ¡F 180 min at Ð75 ¡C

DA (double aged) . . . 930 ¡C for 90 min, air cool, 450Ð540 ¡C 180 min 730Ð760 ¡C 180 min Am-355 A 1025Ð1040 ¡C ...... L (equivalent to T) 930 ¡C for 90 min, air cool . . . SC (equivalent to R) 930 ¡C for 90 min, air cool. . . . 180 min at Ð75 ¡C SCTxxx 850 ¡F or 930 ¡C for 90 min, air cool, xxx ¡F for 180 min 1000 ¡F 180 min at Ð75 ¡C DA (double aged) 930 ¡C for 90 min, air cool. 440Ð470 ¡C 730Ð760 ¡C 180 min for 180 min Equalized and 930 ¡C for 90 min, 540Ð590 ¡C for overtempered air cool, 730Ð760 ¡C 180 min 180 min Austenitic A-286 ST1650 900 ¡C 120 min, oil/water quench ...... ST1650A 900 ¡C 120 min, oil/water quench 730 ¡C 16 h ST1650DA 900 ¡C 120 min, oil/water quench 730 ¡C 16 h, 650 ¡C 8 h ST1800 980 ¡C 120 min, oil/water quench ST1800A 980 ¡C 120 min, oil/water quench 730 ¡C 16 h

PH, precipitation-hardenable; RT, room temperature.

ered sensitization, 760 ¡C (1400 ¡F). The car- ﬁnishes (Mf) near room temperature. This bide precipitation occurs at the interfaces of the process is called austenite conditioning. The small amount of residual ferrite these alloys heat treatment scheme just described would be have and also at grain boundaries. The deletion condition A-1400. This material, after quench- of carbon and chromium from the matrix ing to room temperature, would be said to be changes the matrix composition sufficiently that condition T. the temperature for the start of the martensitic The higher the temperature of the condition- transformation (Ms) of the depleted austenite in- ing, the less carbon is precipitated and the lower creases from below zero to about 65 to 100 ¡C the resulting Ms. The highest conditioning tem- (18 to 212 ¡F). The martensitic transformation peratures of 955 ¡C (1750 ¡F) cause a sufficiently 170 / Stainless Steels for Design Engineers

low Ms that subzero treatment is required to for a normal solution anneal of non-PH austen- obtain the fully martensitic structure required for ite. The subsequent precipitation aging requires age hardening. This would be called A-1750. higher temperatures and longer times because After subzero treatment at Ð73 ¡C (Ð100 ¡F), it diffusion is much slower in austenite. would be called condition R-100. It is even possible to obtain the fully martensitic structure by cold work from full condition Duplex Stainless Steels A, which is quenched from 1065 ¡C (1950 ¡F). This is called condition C and requires heavy Duplex stainless steels are both ferritic and cold rolling to accomplish. Since there is no heat austenitic, so their heat treatment combines the treatment to precipitate carbon from the austen- same elements and principles as their principle ite matrix, the resulting martensite is the hardest. phases. The aging treatments of the semiaustenitic alloys are identical to those for the martensitic Soaking alloys because the treatments are standardized. The resulting mechanical properties vary in a Duplex alloys are multiphase at all useful complex fashion with the alloy composition and working temperatures, making their hot worka- thermomechanical treatment history before bility quite poor. It is extremely important to aging. The principles at work that determine the drive sulfur to the lowest possible levels, less mechanical properties, besides the strain in- than 0.001%, to achieve satisfactory hot ductil- duced by the precipitate phase, are: ity. Otherwise, soaking should proceed the same as for ferritic stainless alloys since ferrite con- ¥ Carbon level of the quenched martensite ¥ stitutes the continuous phase to be worked. Amount of cold work of austenite prior to Because these alloys always contain high aging chromium and generally high molybdenum, These factors do not come into play with the they should be cooled as rapidly as possible martensitic PH grades because all of the carbon from high temperatures to avoid formation of is in solution, and they are not signiﬁcantly cold sigma or other intermetallic phases. worked prior to aging. The semiaustenitic PH alloys, because of the variable amount of car- Annealing bide precipitation, have an equally variable chromium content, so corrosion resistance will The function of annealing in the duplex alloys vary accordingly. The highest-temperature solu- is generally to: tion and carbide precipitation treatments pro- ¥ Remove the effects of cold work vide the best corrosion resistance, as do the ¥ Restore the balance between the volume lowest aging temperatures. fraction of ferrite and austenite ¥ Achieve equilibrium composition within Austenitic PH Alloys both the austenite and ferrite ¥ Dissolve unwanted intermetallic phases In contrast to the martensitic and semiaustenitic PH alloys, the austenitic PH alloys The annealing range of duplex alloys is some- are austenitic under all conditions. Their what restricted, approximating the overlap be- strengthening reaction comes from precipitation tween what each of the two constituent phases of titanium- and aluminum-nickel intermetallic would be annealed at separately. Table 5 lists phases within the austenite matrix as occurs in the normal annealing temperatures for these al- superalloys. Solution treatment is very typical loys. The use of nitrogen as a key alloying

Table 5 Recommended annealing and stress-relieving temperatures for duplex alloys Annealing Annealing Alloy temperature, ¡C temperature, ¡F ASTM A480 2006 Stress relieving, ¡F Stress relieving, ¡C Lean duplex, Cr+Mo<23, 1010Ð1100 1850Ð2010 Various Not recommended 2003-2101, 19-D-2304 Medium alloy, Cr+Mo<26, 1040Ð1100 1900Ð2010 1040 min Not recommended 2205 Cr+Mo>26, 2507, 52N+, 1050Ð1150 1925Ð2100 Various Not recommended Zeron 100, 255 Chapter 13: Thermal Processing / 171

element has improved the annealing behavior of tions, either from ␣' beginning at 350 ¡C (660 these alloys since its diffusion is quite rapid, ¡F) or from ␣/␹, which takes over at 600 ¡C causing ferrite-austenite equilibrium to be at- (1110 ¡F), would occur and so is not indicated tained very rapidly. Nitrogen also hinders (for- for these alloys. mation and facilitates the dissolution of secondary austenite, which can form after quenching from welding temperatures and cause regions of REFERENCES poor corrosion resistance. These alloys are not very susceptible to car- 1. A. Garner, The Effects of Autogenous bide sensitization and normally have very low Welding on Pitting Corrosion in Austenitic carbon content. Thus, the guiding principle in Stainless Steel, Corrosion, Vol 35 (No. 3), annealing is simply to achieve phase balance 1979, p. 108Ð114 and avoid cooling so slowly that intermetallic 2. J.F. Grubb, personal communication, June phases may form. 4, 2006 The strengthening of duplex is normally 3. Data courtesy J&L Specialty Steel, October achieved by the strong grain reﬁnement and 2, 1998 solid solution hardening. No strengthening heat 4. J.F. Grubb, Proceedings of the International treatments are used. Stress relief would have to Conference on Stainless Steels, 1991, occur at temperatures at which embrittling reac- Chiba, ISIJ Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 173-180 All rights reserved. DOI: 10.1361/ssde2008p173 www.asminternational.org

CHAPTER 14

Forming

Summary shown that it is important to understand and exploit these characteristics to optimize forming STAINLESS STEELS RANGE in formability of stainless steels. from the extremely formable austenitic alloys to the hard-to-form martensitic alloys. For steels with equivalent corrosion resistances, formabil- Flat, Rolled Stainless Steel ity increases with the level of austenitizing alloying elements. The ferritic alloys are the least The vast majority of carbon steel and, espe- alloyed, least expensive, and least formable; cially, stainless products are flat products. These duplex steels are next, and austenitic steels are steels are formed by bending, roll forming, spin- the most formable but most expensive. How- ning, hydroforming, and deep drawing. Roll ever, if the specific structural anisotropy of forming is most commonly used to produce ferritic alloys, which gives particularly good welded pipe and tubing and is simply bending deep-drawing characteristics for a given level of done in a continuous manner. Bending is a sim- ductility, can be exploited, the best forming ple operation, and there is no meaningful change economies can be gained. in thickness of the sheet during the operation. The higher yield strength and work-hardening rates of most stainless steels will result in greater Introduction springback than would be experienced in carbon steel. Tooling must be adjusted to compensate The technology for forming stainless steel is for this. In neither of these forming methods is quite similar to that for forming carbon steels. there large motion of the formed material across The primary difference is the higher strength of the die surface, so lubricant is not normally all types of stainless steels compared to drawing- used. The reader is referred to the Forming and type carbon steels. This higher strength requires Forging in Volume 14 of the ASM Handbook for greater sophistication in tooling and lubricants, detailed charts on bend radii and springback re- and it requires more powerful forming equip- lated to bending. ment. The higher initial strength is also Other forming techniques employ more com- accompanied by a higher work hardening rate in plex deformation processes. Deep drawing is austenitic stainless steels, which further distin- the foremost of these. Figure 1 (Ref 1) shows guishes them from carbon steel. Galling also schematically what occurs during drawing. A must be recognized as a danger and prevented. round blank is held between dies over a cavity, Stainless steel has lower thermal conductivity and a punch pushes the material into the cavity than carbon steels, which can cause it to retain to produce the part. If the dies pinch the blank heat from deformation and friction, thereby to only a small degree, the process is normal, or decreasing lubricity. Last, stabilized stainless al- ordinary, drawing. If the dies significantly loys contain abrasive carbide microconstituents. restrain the periphery of the blank so it cannot The various types of stainless steel have very move, stretch forming occurs. Material proper- different deformation characteristics in terms of ties determine whether a material is most suc- strain hardening and anisotropy. It will be cessfully formed by stretch forming or drawing. 174 / Stainless Steels for Design Engineers

rate. If it work hardens faster than it becomes thinner, the strain is distributed, and local failure is prevented. Austenitic materials have the face-centered cubic (fcc) crystal structure with many slip systems and low stacking fault energies. This means that they can generate many complex arrays of tangled dislocations, which cause strain hardening. They can also transform during stretching to the much harder martensite, for which the deformation is greatest, again redistributing the deformation away from the potentially thinning area. This makes austenitic stainless steels particularly suitable for stretch forming. Deep Drawing. Deep drawing, referred to simply as drawing, without stretching requires a different material characteristic. For drawing, a low work-hardening rate is desirable so that the material can compress in the circumferential direction while elongating in the radial direction. Obviously, a high ability to elongate is always useful regardless of any other characteristic. But, ferritic material has one other advantage: Body-centered cubic (bcc) alloys have more slip systems than fcc alloys. When bcc alloys are rolled to become flat stock, they may retain a preferred crystallographic orientation, called texture, as a result of the deformation. This non- random crystal structure can cause the material to have higher strength in the through-thickness direction. This directional variation in properties is called anisotropy. When a material with desirable texture is stretched, it flows in the stretching direction and contracts laterally at lower stresses than are required to initiate plastic flow in the through- thickness direction. As long as the work-hardening rate keeps the flow stress below the through-thickness yield strength, there will be no thinning. The geometry of deep drawing with constraint fits such materials’ capabilities. This is why carbon steels and ferritic stainless steels deep draw well. If the material were constrained from contracting while being stretched, the tensile strength would be exceeded before through-thickness flow occurred and the material would fracture with little deformation. So, Fig. 1 Deep drawing schematic. Source: Ref 1 these materials cannot be stretch formed. Even without stretching, some hold-down pressure is Stretch Forming. In stretch forming, the required to prevent wrinkling of the blank material is constrained from moving wholly before it is pulled into the die. This is more pro- into the die. Thus, the section that enters the die nounced for thinner blanks and for material is stretched more and must become thinner. with higher work-hardening rates. In many Whether the material becomes so thin locally forming situations, adjustments to the drawing that it fails is governed by its work-hardening process (i.e., hold-down force adjustment, draw Chapter 14: Forming / 175

bead contour, blank size, lubrication, die radii, and 2.0 are more common. Figure 2 shows the etc.) may be more important than material prop- LDR as a function of the Lankford ratio. It erties in determining whether the desired part should noted that the very ductile 304 fares no can be made successfully. better than carbon steel because of the advantage The material properties that are important to of a beneficial anisotropy, which the ferrous bcc formability are ductility, as measured by tensile structure has. Indeed, the best-performing deep- test elongation; the work hardening rate, which drawing stainless steels are low interstitial ferritic is the instantaneous slope of the true stress, true steels with boron added. A number of stainless strain curve and is called n; and the anisotropy. steels are compared in Table 1 (Ref 4). The measure of anisotropy is the Lankford ratio It shows that when comparing different types (Ref 2), expressed as: of materials, some tests are not good predictors of deep-drawing performance. Optimizing the material/drawn component combination is far rr+2r + R = 04590 (Eq 1) from simple, especially when other considera- 4 tions, such as cost and material performance in service must be factored in. End users are where R is the average strain ratio, r0 is the strain encouraged to deal directly with the producing ratio in the longitudinal direction, r45 is the strain mill early in the design stage of any new high- ratio measured at 45¡ to the rolling direction (of production, deep-drawn component. The pro- the sheet metal-forming operation), and r90 is the ducing mills, while not necessarily exhaustive strain ratio in the transverse direction. R deter- sources of information, are certainly reservoirs mines the average depth (that is, the wall height) of knowledge of current practice. of the deepest draw possible. When this expres- The most widely used summary of a mater- sion equals 1, then a material may be considered ial’s formability is contained in its forming limit isotropic, that is, the material properties are the diagram (FLD), developed by Keeler and Back- same for all crystallographic orientations. As a ofen (Ref 5). This diagram shows the locus of first approximation, the Lankford ratio equals failure under varying strain states. Figure 3 the ratio of the lateral strain to the through-thick- shows a comparison of the FLDs for austenitic ness strain during the tensile test of a sheet spec- stainless steel and carbon steel. The FLD tells imen. As the value increases from 1, the drawability increases because the material tends to maintain a constant thickness while changing shape from a flat blank to a cup shape. The ability to be deep drawn is measured by the limiting drawing ratio (LDR), the ratio of the diameter of a disc to that of the deepest cylinder into which it can be drawn. The ferritic stainless steels in sheet form have LDRs of around 2.2 compared to 2.0 for 304. For austenitic steel, the ratio is about 1.0, while for flat-rolled carbon steel and ferritic stainless steel, Fig. 2 Limiting drawing ratio variations with Lankford ratio. it can be greater than 2.0, but values between 1.5 Source: Ref 3

Table 1 Deep-drawing materials comparison 0.2% proof Tensile Erichsen Conical cup stress, strength, Elongation, Hardness Lankford value value, value, 1 2 Steel N/mm N/mm % HV n value ro r45 r90 rømmmm YUS 190 343 497 33.8 173 0.20 1.60 1.47 2.10 1.66 9.5 26.7 YUS 436S 275 459 34.8 135 0.21 1.67 1.63 2.12 1.76 9.8 26.9 (B-added) YUS 436S 284 483 34.5 137 0.22 1.49 1.90 2.01 1.83 9.8 26.8 (B-free) YUS 4O9D 239 424 37.2 116 0.24 1.51 1.77 2.11 1.79 9.7 26.7 SUS 430 308 472 31.8 159 0.21 0.94 0.92 l.50 1.07 8.9 28.5 SUS304 281 705 64.0 172 0.44 0.91 1.19 0.83 1.03 12.5 27.0 Source: Ref 2 176 / Stainless Steels for Design Engineers

Fig. 4 Forming limit diagrams for categories of stainless steels. A, austenitic stainless steel; F, ferritic stainless Fig. 3 Forming limit diagram of carbon steel compared to steel; HAS, high-strength austenitic stainless steel; HSF, high- austenitic stainless steel. Source: Ref 3 strength ferritic stainless steel; FA(50), ferritic-austenitic stainless steel. Source: Ref 3 the point of failure for a given sheet material with a given thermomechanical history over a full range of combined strain states. These dia- total interstitial content (i.e. carbon plus nitro- grams are generated by examining circle grids gen) and by thermomechanical working to give printed on material that is deformed to failure in a fine-grained, fully recrystallized, yet benefi- various modes. The single most important value cially anisotropic, microstructure. Figure 5 on the curve is the intersection of the curve with shows how the FLD of an enhanced 409 ferritic the major strain axis at zero minor strain. This stainless steel, 409 Ultra Form, compares to the can be used as an index of formability and is the already highly evolved 409. elongation possible for plane strain conditions. For austenitic stainless steel, maximum This value within a given class of materials is drawability is obtained by low work-hardening proportional to the strain-hardening exponent rates coupled with maximum elongation, as because a higher work-hardening rate causes exemplified by 305 and high-nickel 304. This higher localized resistance to thinning, which is comes with a cost penalty as the easiest way to the precursor of failure. improve formability is to increase the nickel There is much more variety within the fami- level, although replacing the expensive nickel lies of stainless steel than within carbon steel. with copper or manganese, as in 204Cu, has Figure 4 shows generalized FLDs for austenitic, been shown to be effective. Many austenitic ferritic, high-strength ferritic, high-strength components are made by stretched deep draw- austenitic, and duplex stainless steels. ing. In this case, the preferable alloys are the Deep drawing of components is seen as a way leaner austenitics, 201 and 301. These alloys to obtain near-net shape. Since tooling is costly, it form martensite more rapidly than do 304 and is necessarily a high-volume application. Quite 305 during stretching. Martensite has a 4% often, designers push component design to the greater volume than the austenite from which it limit of a material’s ability to be formed. There forms and a much greater strength. This gives are various drivers that cause this. One is to elim- 201 and 301 the ability to redistribute deforma- inate extra operations or components by consoli- tion from thinning areas elsewhere and stretch dating them into one more complex deep-drawn extensively, making them an optimal material part. Another is to use the least-expensive alloy. for objects such as sinks. In some industries, such as the household appli- The specific alloy composition is often finely ance industry, as many components as possible tuned for a given part and tooling design, and are deep drawn from ferritic stainless alloys and small deviati ons can dramatically increase the more costly austenitic alloys are used only breakage rates. Even such minor process when the part cannot be made from a ferritic. changes as blank temperature variations due to There has been much research to develop fer- ambient temperature can alter work-hardening ritic stainless alloys with improved formability. rates enough to cause breakage problems. Cer- This has been accomplished by reducing the tainly, this can occur when designs push the Chapter 14: Forming / 177

Fig. 5 Optimized 409 for forming versus normal 409. Source: Ref 5 envelope of a material’s capability, but cost A last key variable in which practice and ma- pressures generally drive designers to this terial interact is strain rate. Ferritic steels ﬂow extreme. more easily at lower strain rates and are thus On very severe forming, intermediate more formable. Austenitic steels experience the annealing may be required to either enhance opposite effect if they are susceptible to marten- ductility or reduce required pressure. Stainless site formation. Adiabatic heating can retard the steel in the as-drawn condition will have resid- martensitic transformation and reduce the work- ual stress and may have sufficient hardness to hardening rate, changing their forming charac- be susceptible to delayed failure if placed in a teristics. When drawing, this is good, but for corrosive environment. Bright-annealed alloys stretching it may not be. with high martensite levels from forming can Hydroforming, a variation on deep drawing fail by hydrogen embrittlement with just the in which hydrostatic pressure forces a blank residual hydrogen from annealing. Therefore, into the die cavity, can improve the degree to the use of bright-annealed lean alloys such as which stainless steels can be deep drawn. The 301 is not recommended. hydroforming process avoids friction between 178 / Stainless Steels for Design Engineers

the blank and the tool. Deformation is spread forming difficulties and lack of compliance with more evenly across the blank, and the material mechanical property requirements. They also forms close to its theoretical best. Productivity have significant rolling anisotropy, which causes using this technique is relatively low, so its use the yield strength transverse to the rolling direc- is justified mainly when it is not otherwise pos- tion to be consistently higher than it is in the sible to make a certain design in one drawn rolling direction. This behavior is contrary to the component. An example of this may be the pro- general behavior of single-phase alloys. The dif- duction of a complex exhaust manifold that reference in yield strength is sufficient, reportedly quires a higher-alloyed ferritic stainless with up to 15% lower in the longitudinal direction relatively low formability. than in the transverse direction required for ten- Besides failure by breakage, there are other sile tests (Ref 7), that it is both serious design less-severe defects found on deeply drawn and forming considerations. parts. Austenitic steels can develop a surface Ferritics undergo a more specialized surface condition known as orange peel, the result of relief because their anisotropic grain structure slip planes within a grain disrupting the sur- can yield in a more concerted fashion and give face. Orange peel is prevented by keeping even greater surface relief, called ridging and grain size fine so that the surface relief is too roping. This is minimized by refining grain small to be seen. Austenitic stainless steels can size, achieving full recrystallization versus just also develop anisotropy, which while less recovery during annealing, and temper passing severe than ferritic steels, can cause “earing,” (i.e., elongation of about 1% on a cold-rolling in which round blanks deform nonuniformly in mill) to suppress the yield point phenomenon a four-, six-, or eight-fold symmetry, causing that is characteristic of ferrous bcc materials. excess ear-shaped material to extend beyond End users should always make their use of the the intended dimensions of the component. material known to the producing mill so that Material is wasted because larger blanks have the correct thermal processing path can be to be used. The steel producer can minimize employed for the manufacturing process the the phenomenon by keeping cold roll reduc- material will undergo. tions above about 60%. One measure of the One of the most important material consider- earing tendency is derived from the Lankford r ations for deep drawing is surface finish. Flat- measurements: rolled stainless should be fully annealed and pickled so that the surface holds lubricant well and yields as readily as possible. Temper pass- rr++2 r Δr = 09045 (Eq 2) ing will drastically reduce the drawability of 2 stainless by smoothing the surface (rather than increasing the yield strength). Temper passing The left side of Eq 2, ⌬r, is a measure of the with roughened rolls does not significantly variation of plastic strain ratio r with direction harm drawability. The surface finishes that are in the plane of a sheet. Values of ⌬r near zero produced by temper rolls with special finishes, generally indicate minimal tendency toward such as Koolline, retain lubricant well and can earing, while ⌬ values significantly above or be drawn with minimal distortion. below zero indicate increased tendency toward Tooling for stainless must be strong and wear earing. A combination of a high R value from resistant. Traditional tooling materials are D2 Eq 1 and a low ⌬r value provides optimal tool steel and high-strength aluminum bronze. drawability. D2 tool steel must be hardened to HRC 60 to It should be noted that deformation always 62 and must have smooth surfaces. The use of produces some surface relief, so highly reflec- powder metal techniques to produce tool steels tive surfaces become spectrally diffuse, or for dies has permitted much higher volume frac- cloudy, after plastic deformation. This has been tions of ultrahard microconstituents such as an issue for items such as automotive bright vanadium carbide to be introduced, thereby trim. Mechanical buffing can restore the luster, vastly improving wear resistance without harm- but the time and expense of buffing increase ing toughness or even raising overall hardness. dramatically if orange peel or roping (a similar The benefits of cast aluminum bronzes are low surface defect) is excessive. friction, high thermal conductivity, and low ten- Duplex stainless steel flat products exhibit dency to gall. They are preferred when finished significant in-plane anisotropy that can cause part surface appearance is more important than Chapter 14: Forming / 179

Table 2 Suitability of various lubricants for use in forming of stainless steel Blanking Press- Drop Contour and brake Press Multiple-slide Deep hammer roll Lubricant piercing forming forming forming drawing Spinning forming forming Embossing Fatty oils and CBC A C A C B B blends(a) Soap-fat NR NR C A B B C B C pastes(b) Wax-base B B B A BBCBA pastes(b) Heavy-duty B NR B A B B NR A B emulsions(c) Dry ﬁlm (wax BBB NR B A B NR A or soap plus borax) Pigmented B NR A B A C NR NR NR pastes(b)(d) Sulfurized or A A B+ A C NR A B A sulfochlorinated oils(e) Chlorinated oils or A(h) NR A NR A NR A(i) A NR waxes(f) high- viscosity types(g) Chlorinated oils or B+ A A A B NR A(i) A A waxes(f) low- viscosity types(j) Graphite or NR (l) (l) NR (l) NR (l) NR NR molybdenum disulﬁde(k)

A, excellent; B, good; C, acceptable; NR, not recommended; SUS, Say-bolt universal second. Ratings consider effectiveness, cleanliness, ease of removal, and other suitability factors. (a) Vegetable or animal types; mineral oil is used for blending. (b) May be diluted with water. (c) Water emulsions of soluble oils; contain a high concentration of extreme pressure (EP) sulfur or chlorine compounds. (d) Chalk (whiting) is most common pigment; others sometimes used. (e) EP types; may contain some mineral or fatty oil. (f) EP chlorinated mineral oils or waxes; may contain emulsifiers for ease of removal in water-base cleaners. (g) Viscosity of 4,000 to 20,000 SUS. (h) For heavy plate. (i) For cold forming only. (j) Viscosity (200 to 1000 SUS) is influenced by base oil or wax, degree of chlorination, and additions of mineral oil. (k) Solid lubricant applied from dispersions in oil, solvent, or water. (l) For hot forming applications only. absolute die life and forming pressures are medium-sized hardware items, such as screws, moderate. bolts, nuts, rivets, and specialized fasteners. Lubricants for stainless steel forming must be As with flat-rolled forming operations, the able to prevent metal-to-metal contact under primary difference between carbon/alloy steels higher pressures than those seen with carbon and stainless steels comes from the higher yield steel. The ASM Metals Handbook, Desk Edi- strength and higher work-hardening rates of tion, lists common lubricants as shown in Table stainless. Anisotropy is not a significant consid- 2 (Ref 1). Not listed in the table are the newer eration for long products. The ferritic stainless thermoplastic acrylic polymers that, when ap- steels are the most easily cold headed. The use plied to the surface at a density of around 1 g/m2 of the most formable stainless alloys for flat (0.004 oz/ft2), provide a dry film with lubricat- products, the stabilized ferritic alloys, is limited ing properties surpassing any of those listed in for long products because of the severe lack of Table 2. toughness these alloys show for cross sections greater than about 2 mm (0.08 in.). But, the nonstabilized ferritics, the martensitic, precipitation-hardenable (PH), austenitic, and duplex Stainless Long Products grades are all cold formable. In cold-heading terminology, the maximum Cold heading, one of the most important possible deformation an alloy can tolerate is ex- forming operations conducted on stainless long pressed in terms of the length of long product products, is a forming process that increases the exposed beyond the die that can be successfully cross-sectional area of a room temperature forged into the upset. This is measured in the blank at one or more points along its length. number of diameters of initial stock. So, an op- Cold heading is typically a high-speed process timal ferritic such as 430 can tolerate upsets up in which the blank is progressively moved to about 2.25 diameters, while a very low work- through a multistation machine. The process is hardening austenitic, such as 384, can tolerate widely used to produce a variety of small- and 3.0. The martensitic, PH, and richer duplex 180 / Stainless Steels for Design Engineers

their work hardening is not instantly removed by annealing. For example, the initial pressure requirements for a given deformation range for 304 can be three times what is required to deform carbon steel and up to five times for higher molybdenum alloys. For greater deformation, this multiple increases (see Fig. 6a and b). Austenitic stainless steel loses ductility when heated above 1280 ¡C (2335 (F) because of low melting phases in the grain boundaries. As much as possible, all alloys should be hot worked in a single phase field of the phase diagram to avoid mixtures of ferrite and austenite since the great difference in their great strengths can cause failures. Duplex stainless steels and other alloys (e.g., 17-7 PH) that have high levels of ferrite in austenite or austenite in ferrite at the hot-working temperature exhibit reduced hot ductility compared to either fully austenitic or fully ferritic stainless steels and are more difficult to hot work.

REFERENCES 1. Forming of Sheet, Strip, and Plate: Deep Drawing, in Metals Handbook, desk ed., J.R. Davis, Ed., ASM International, 1998, p 782Ð829 2. U.F. Kocks, C. Tomé, H.-R. Wenk, Ed., Tex- ture and Anisotropy, Cambridge University Press, Cambridge, UK, 1998 3. E. Schedin, “Forming Stainless Steel,” Fig. 6 Forces required for hot working. Source: Ref 1 ACOM Technical Paper, www.outokumpu .com 4. H. Sumitomo and T. Tanoue, Nippon Steel alloys are in the 1.5 to 2.0 range of formability. Technical Report 71, October 1996 The lean duplex, when they ﬁnd their way into 5. S.P. Keeler and W.A.A. Backofen, ASM wider use, should resemble 430 with 2.25 diam- Trans. Q, Vol. 56 (No. 163), 1963, p 25Ð48 eters maximum. 6. “409 Ultra Form Stainless Steel,” Product Hot forming of stainless steel is done as an in- Data Bulletin, www.AKSteel.com tegral part of their production; therefore, all 7. R. Cordewener et al., “Duplex Stainless stainless alloys can be forged. The main issue is Steels,” Paper 109, TWI conference, Glas- that the high hot strength of stainless requires gow, 1994 much more force than would be required for carbon steels. Martensitic stainless steels re- SELECTED REFERENCE quire 10 to 100% more force than 4340 alloy steel, while the austenitics require much more ¥ ASM Handbook, Vol. 14, Forming and because of their high hot hardness and because Forging, ASM International, 1988 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 181-191 All rights reserved. DOI: 10.1361/ssde2008p181 www.asminternational.org

CHAPTER 15

Machining

Summary From a more focused viewpoint, the machinability of a material is further described by: MACHINING STAINLESS STEELS is a complex operation. Not only does a shop need 1. Consistency: Does the material machinabil- the correct supporting equipment and supplies, ity stay the same when bundles are changed? a better understanding of the metal itself is ad- 2. Tool life/wear: How long does the tool last vantageous. Technology in the production of a in the machining operation? This could be more machinable stainless steel is advancing. minutes, hours, shifts, or days. The incorporation of complex oxides has led to 3. Productivity: How many parts were made the development of materials that allow higher in an hour, shift, or day? machining speeds and increased productivities, 4. Cost per part: What is the cost of the final both of which are reducing machining costs and geometry? keeping shops competitive. 5. Cycle time: How fast can a part be completed? 6. Surface finish: How smooth or shiny is the part? Introduction 7. Chip control: Are the chips manageable? 8. Maximum cutting speed: How fast can the Stainless steel forgings, castings, plate, and part be cut without affecting tool life? long products all are frequently machined. This 9. Maintaining tolerances: How long can the fundamentally involves the removal of a layer machining operation continue before ad- of material from the workpiece with a cutting justments are made? tool one or multiple times until a finished or 10. Minimal operator intervention: Does the semifinished part is produced. Machining, in it- operator need to constantly adjust setup? self, is a complex topic with many variables. Rather than attempt to understand all aspects of machining, it is helpful to consider a material’s machinability, that is, its ability to be machined and the factors that affect its ability to be machined. In Fig. 1, a macroview shows how the machinability of a material is influenced by the interaction of humans, machine, methods, material, and management. Some of the variables can affect the appearance of the material, while others affect the performance of the piece, making machining an art as well as a science. Opti- mum machinability is obtained when each of these sectors come together, providing the best possible conditions for efficient machining. Any change in one of these sectors can change the behavior or efficiency of a machining job. Fig. 1 The 5 M’s of machinability 182 / Stainless Steels for Design Engineers

This list is somewhat empirical or job re- ƒ(chemistry, cleanlinesss, lated, but it provides guidelines for defining a Machinability of stainless steels = structure, processing, job since cutting conditions can be very differ- ccross section) ent for each material and part. For example, if the surface finish of a part is very important, it may be necessary for chip control and tool life Each variable contributes uniquely to machin- to be sacrificed. Clearly, machining involves ability. Machine shops and users generally have much more than simply cutting a piece of very little influence on these material variables. metal. Because no two mills are exactly identical, there Machining is a very empirically mature sub- will be differences in machinability of a steel ject. The recommended feed rate, depth of cut, grade provided by different mill suppliers. How- tool material, and cutting fluid for a given mate- ever, having an understanding of how these vari- rial/material condition (thermomechanical his- ables contribute to machinability is invaluable. tory) can be found in readily available pub- Armed with an understanding of the material and lished tables. Books such as the ASM how it is made, one can determine the tooling, Handbooks; Machinery’s Handbook, published coolant, and setup of the machining job. by Industrial Press; Marks’ Standard Hand- Let us take a closer look at these variables. book for Mechanical Engineers, published by McGraw-Hill Book Company; or the Machin- Chemistry ing Data Handbook, 3rd edition, by the Machinability Data center at the Institute of Ad- The role of chemistry is to define not only vanced Manufacturing Services (IAMS; for- the different grades of stainless steel (ferritic, merly known as Metcut Research Associates martensitic, etc.), but also how the grade Inc.) in Cincinnati, OH; include much of the is chemically balanced within the specific data used in industry today. Material manufac- grade; for example, the amount of carbon in a turers are also a source of valuable machining martensitic stainless can change tool wear char- data. A typical guide from ASM is shown as acteristics, or a change in nickel content within Table 1 (Ref 1). specification limits can alter the stringiness of a Rather than simply reproducing data, the chip. Combined, both will be the basis the mate- focus of this chapter is the metallurgical factors rial’s machinability. governing the machinability of stainless steels. Each of the elements used to produce stain- Most of the information regards machining less steels will contribute some general machin- stainless bar products; however, many of the ing attributes. The effects of the elements as de- concepts could be applied to forgings as well as scribed next are general, and slight deviations castings. may be encountered depending on the stainless grade. However, for the more common stainless grades used today, these effects of these alloying elements are fairly accurate. Physical and Mechanical Properties Iron is the base element in a stainless steel. It is a soft, gummy material that has high work- The machinability of stainless steels is very hardening characteristics. Iron is characterized difficult to characterize in definitive terms be- by surface finishes that are difficult to obtain cause of the broad nature of these materials. A and chips that are stringy, and it has a high ten- ferritic stainless steel, such as type 430, will dency toward tool built-up edge (BUE). machine very differently from the martensitic. Chromium strengthens and reduces ductility In some sense, this is like comparing brass to of stainless steel. Machine and tool setup re- carbon steel. Both type 410 and type 430 are quire more rigidity. Chromium allows chips to stainless steels, but the chemistry and structural begin breaking. differences create diversity in machining char- Carbon content increases strengthen stainless acteristics. steels and promote carbide formation. Low car- The machinability of stainless steels can be bon levels, typical in ferritic stainless steels, do thought of as a function of the steel’s chemistry, not help machinability much. Increasing cleanliness, structure, processing history, and amounts of carbon to greater than 0.08% will aid the cross-section size of the stock, with no one in chip breakability and reduced BUE in these factor more important than another: grades. However, as carbon content increases, Chapter 15: Machining / 183

Table 1 Machining setup recommendations for turning wrought stainless steels

Uncoated Coated High-speed steel tool Tool Tool Hardness, Depth of Speed, Feed, Tool material Speed, fpm Feed, material Speed, Feed, material Material HB Condition cut(a), in. fpm ipr AISI Brazed Indexable ipr grade fpm ipr grade Ferritic steels 405, 409, 429, 135-185 Annealed 0.040 150(235) 0.007 M2, M3 575 650 0.007 C-7 850 0.007 CC-7 430, 434, 0.150 120(190) 0.015 M2, M3 450 500 0.015 C-6 650 0.015 CC-6 436, 442, 0.300 95(150) 0.020 M2, M3 350 400 0.030 C-6 525 0.020 CC-6 446(c) 0.625 75(115) 0.030 M2, M3 275 310 0.040 C-6 ...... Austenitic and duplex steels 201, 202, 301, 135-185 Annealed 0.040 95 0.007 M2, M3 325 375 0.007 C-3 500 0.007 CC-3 302, 302B, 0.150 75 0.015 M2, M3 300 325 0.015 C-3 425 0.015 CC-3 304, 304L, 0.300 60 0.020 M2, M3 225 250 0.020 C-2 325 0.020 CC-2 305, 308, 0.625 45 0.030 M2, M3 175 200 0.030 C-2 ...... 309, 309S, 225-275 Cold drawn 0.040 80 0.007 T15, M42(b) 300 325 0.007 C-3 425 0.007 CC-3 310, 310S, or 0.150 65 0.015 T15, M42(b) 250 275 0.015 C-3 350 0.015 CC-3 314, 316, duplex 0.300 50 0.020 T15, M42(b) 290 215 0.020 C-2 275 0.020 CC-2 316L, 317, annealed 0.625 40 0.030 T15, M42(b) 140 165 0.030 C-2 ...... 321, 330, 347, 348, 384, 385(c) 2205, 2507 295-310 Annealed

Martensitic and PH steels 403, 410, 420, 135-175 Annealed 0.040 155 0.007 M2, M3 475 620 0.007 C-7 800 0.007 CC-7 422, 501, 0.150 125 0.015 M2, M3 400 480 0.015 C-6 625 0.015 CC-6 502(c) 0.300 100 0.020 M2, M3 320 380 0.030 C-6 500 0.020 CC-6 0.625 80 0.030 M2, M3 240 300 0.040 C-6 ...... 175-225 Annealed 0.040 145 0.007 M2, M3 460 570 0.007 C-7 850 0.007 CC-7 0.150 115 0.015 M2, M3 385 450 0.015 C-6 550 0.015 CC-6 0.300 90 0.020 M2, M3 300 350 0.030 C-6 450 0.020 CC-6 0.625 70 0.030 M2, M3 235 265 0.040 C-6 ...... 275-325 Quenched 0.040 95 0.007 T15, M42(b) 360 465 0.007 C-7 700 0.007 CC-7 and 0.150 75 0.015 T15, M42(b) 280 360 0.015 C-6 450 0.015 CC-6 tempered 0.300 60 0.020 T15, M42(b) 225 280 0.020 C-6 375 0.020 CC-6 375-425 Quenched 0.040 65 0.007 T15, M42(b) 290 320 0.007 C-7 475 0.007 CC-7 and 0.150 50 0.015 T15, M42(b) 225 250 0.015 C-6 300 0.015 CC-6 tempered 0.300 40 0.020 T15, M42(b) 180 200 0.020 C-6 250 0.020 CC-6 PH, precipitation-hardenable. Source: Ref 1 (a) Caution: check horsepower requirements on heavier depths of cut. (b) Any premium high-speed steel (T15, M33, M41–M47). (c) Free machining versions.

the amount of carbide increases, the structure as duplex and 200 series alloys, manganese has changes to martensitic, and the wear on tools the same relative effect as nickel when used in increases. greater amounts, as, for instance, in the 200 se- Nickel increases the toughness and ductility ries stainless steels. of stainless and reduces the work hardening Molybdenum increases the strength and elerate. Nickel also increases elevated temperature vated temperature mechanical properties. This mechanical properties. This causes chips to be increase in hot hardness and strength means more difficult to break. Nickel will have a ten- more energy will be needed to cut the material, dency toward increased BUE; however, better thus creating hotter cutting conditions. While tool life will generally result. the molybdenum helps in chip breakability, it Sulfur reduces mechanical and corrosion will require more rigid setups and will reduce properties and can be a cause of hot cracking in tool life. the resulfurized grades. It is best known as a Copper improves ductility and reduces the free-machining contributor that promotes better strain-hardening or work-hardening rate (with tool life and greater machining speeds. the exception of participation-hardening alloys, Manganese is generally added to combine for which copper is used as the precipitant). with sulfur to form manganese sulﬁde (MnS), Chips can be more difficult to break, which in- which acts as a self-lubricant and improves creases the tendency of BUE and promotes bet- machinability. In high-manganese grades, such ter tool life. 184 / Stainless Steels for Design Engineers

Nitrogen strengthens stainless steels. It aids measure and control it to varying degrees. The in chip breakability and reduces BUE but in- material’s grain size results from the thermal creases tool wear. and mechanical history during manufacturing Titanium promotes carbide formation and in- and from the mill’s equipment capability and creases tool wear. practices. Niobium promotes carbide formation and in- The grain size of a particular product can dra- creases tool wear. matically change its machinability. It is entirely The production of stainless steels is identi- possible for the grain size difference between fied by industry specifications, such as AISI, two lots of material to be large enough to pre- UNS, EN, JIS, etc. These specifications are all vent both lots from being effectively machined defined with fairly broad elemental chemical with the same setup, requiring adjustments in compositions. For example, an AISI 304 has a the machining setup to remedy the situation. 2 wt% window for the nickel content; that is, Finer grain sizes strengthen the stainless steel, this grade can have a nickel level of 8 to 10%. cause hotter cutting conditions, and have a A type 304 with 8% nickel can have different higher tendency of BUE. On the brighter side, machinability characteristics from a type 304 finer grain sizes yield better surface finishes and with 10% nickel. This 2% difference alters chip smoother roll thread crests. morphology and surface finish capability. Since today’s mill technology can meet very tight ele- Process mental targets within the grade specification, how the mill balances the grade’s chemistry The type of equipment used by the stainless will provide the foundation of its machinability manufacturer, the manufacturing sequence, and characteristics. the practices employed by the mill can affect machinability as well as mechanical properties, but more important, processing affects how Cleanliness consistently the material can be machined. The The cleanliness of steel is determined by the melt type, hot rolling parameters, thermal treat- amount and type of inclusions it contains. Vac- ments, cold-finishing parameters, and sequence uum and argon oxygen decarburization (AOD) of these operations can affect how consistently melting and refining along with proper steel- a material machines. Many times, the culprit is making techniques can reduce the inclusions to equipment operational procedures or practices negligible levels. It is beneficial to machinabil- that can vary one day to the next. Equipment ity to avoid hard inclusions. However, certain types can also play a role in machinability. For inclusions are plastic and act as solid-state lu- example, machinability can vary when the bricants and chip breakers and prevent adhesion same-size material is drawn across two different of the material to the tool. The beneficial effect draw benches using different pulling mecha- of controlled inclusions is discussed in this nisms and two different straightening mecha- chapter. nisms. Whether the material is continuously annealed or batch annealed can cause different strain distributions across the material cross Structure section as well as material strength differences. Material structure consists of both the phases Various annealing lines vary in time/tempera- that are present and the microstructure of those ture profiles and therefore result in different phases. Each type of stainless steel belongs to a grain size and mechanical properties. larger family, which is characterized by a single With all this in mind, manufacturing consis- predominant phase or a combination of two. tency can be a great asset in machinability. A These are ferritic, austenitic, martensitic, pre- machine shop can adjust when a material is con- cipitation hardening, and duplex (see the chap- sistently bad, but it is very difficult when one lot ters on stainless steels, Chapters 6 to 10, in this is easy to machine followed by a lot that is Volume). Their machining characteristics are tough to machine, while a third bundle performs described in the next section. The microstruc- differently from the first two. Mills that promote ture of a given alloy is independent of the grade machining consistency pride themselves by type and composition and is mainly influenced practicing manufacturing consistency. Toler- by grain size. Grain size is not normally speci- ance variation will be tighter and machining fied or reported on certifications; however, mills costs will be lower with their products. Chapter 15: Machining / 185

Cross-Section Size will tend be stringy but can be broken through aggressive chip breaking, and surface finishes Mill processing equipment dictates a manu- will be somewhat of a challenge. Ferritics are facturing route based on size. Smaller diameters the easiest of the stainless steels to machine, do are cold drawn, while larger diameters are not require much horsepower, have a low work- straightened/cut/turned, yielding a softer prod- hardening rate and better tool wear, and will uct. This can have an impact on machining per- generally have higher speed and feed capabili- formance despite all other factors being the ties than other stainless families. same. Cold finishing of stainless steels can be accomplished via a couple of general manufacturing routes. Martensitic The first is by cold drawing to bar, and the sec- Martensitic stainless are also very basic ond is simply a straightening-turning operation. straight chromium stainless steels, 400 series The mechanical properties of the straightened- stainless grades, and are similar to the ferritic turned bars will be softer than the bars made by grades. The difference is that the martensitic cold drawing. Typically, sizes greater than 1 in. grades have much higher carbon levels, which (25 mm) are annealed/turned and straightened, further strengthen the materials and allow these with virtually no strain in the product. materials to be hardenable by heat treatment. These grades will have higher carbide levels, which will lead to higher tool wear. This is es- Machinability of the Stainless pecially true if the material is being machined in Steel Families the hardened condition. The higher strengths will require more horsepower to cut and will Comparing the machinability of stainless need more rigid setup than ferritic steels. The steels with other materials such as carbon steels, work-hardening rate of martensitic stainless is brass, or aluminum, there are some striking dif- lower than for ferritic stainless. Martensitic ferences. In general, stainless steels have: stainless also has a small yield-to-tensile ratio, making chips easier to break. 1. Low thermal conductivity 2. High work-hardening rates Austenitic 3. High tensile strengths 4. High toughness The austenitic grades, the 300 series stainless 5. High ductility grades, are more difficult to machine than the 6. Large spreads between the yield and tensile ferritic and martensitic families. Austenitic strengths stainless steels are more highly alloyed and are more prone to higher work-hardening rates. This Each stainless steel family (ferritic, marten- leads to the need for higher horsepower and sitic, etc.) brings its own general set of machin- more rigid setups. These grades are very prone ing rules. This is mainly due to the chemistry of to BUE and hence are prone to poorer surface these families and its resultant effect on the finishes and tend to tear. The yield-to-tensile ra- physical and mechanical properties. A general tios of austenitic stainless steel is very large, description of the machining behavior is pro- making chips hard to break. Chips in this family vided next. One must keep in mind that these are of alloys tend to be long and stringy. The higher general characteristics. Further alloying of these strength and higher ductility of these grades also families, such as with a sulfur addition, can re- tend to increase cutting temperatures, necessitat- sult in a radical change in machining behavior. ing tooling with higher heat resistance.

Ferritic Precipitation Hardening Ferritic stainless steels are the most basic Precipitation hardening stainless steels are stainless steels and are part of the 400 series characterized by higher strength and toughness. grades. Their basic chemical composition is The solution-annealed hardness of AISI 630, for iron and chromium. These grades generally ex- instance, is HRC 36 versus HRC 23 for a 304. hibit lower strengths, more ductility and soft- Higher horsepower requirements, high tendency ness, and a close yield-to-tensile ratio. These to BUE, higher tool wear, and difficulty break- grades will have a high tendency to BUE, chips ing chips are familiar scenarios for this class of 186 / Stainless Steels for Design Engineers

stainless. Except for alloy A-286, the precipita- machinability of a material. It was once a lead- tion-hardenable (PH) grades are all martensitic ing addition to carbon steels, but its reported en- alloys and can be treated as such for machining vironmental toxicity has diminished its role. purposes. Many carbon mills are looking at other machining agents to replace lead. Lead, however, has Duplex not been a large factor for stainless steels because of the extremely negative effect it has on The duplex is a unique class of stainless char- hot workability, always a serious consideration acterized by a dual-phase structure. Duplex al- in stainless steel design. loys have a structure that is roughly a 50% mix Selenium and tellurium have similar charac- of austenite and ferrite; thus, two hardness ma- teristics to lead as additives but also are non- terials with different hardnesses coexist side by competitive due to cost, toxicity, and incompati- side. The tool will alternate cutting between soft bility with stainless. and hard grains of the duplex structure, leading to an automatic tendency to initiate chatter in the cutting system. Strength levels of duplex al- Sulfur loys are quite a bit higher than austenitic grades. It became obvious very early to metallurgists Between the duplex structure and high-strength that higher sulfur levels correlated to better levels, high horsepower is necessary, and highly machinability, and sulfur remains the popular rigid setups are required. Some work at Ugitech additive choice. Sulfur is cost-effective as a found that to effectively machine these grades, machinability additive and can be easily re- highly alloyed carbide tooling with high hard- moved with modern refining methods. Sulfur is ness and high heat resistance, such as the a natural impurity and has negative effects on C7/C8-type carbides, should be used. mechanical and corrosion properties, discussed separately here. The role of sulfur as a machin- Super Stainless Steels ability agent in stainless steels is very complex Super stainless steels are today’s highly spe- and not necessarily straightforward, but in gen- cialized stainless grades. These grades, like the eral sulfur has been extremely beneficial in in- duplex alloys, are being developed to increase creasing the machinability of stainless steels. corrosion performance parameters to meet some Generally, as the molten metal cools to solid of today’s increasing performance require- form, sulfur combines with manganese to form ments. These alloys are more highly alloyed manganese sulfide inclusions. Manganese be- than the duplex materials. Strength levels are comes a very important variable during this re- higher and toughness is greater, driving machin- action. Two basic sulfide forms are found in ability downward. stainless steels—manganese sulfides and chromium sulfides—and the form the sulfide will take depends on the manganese content. Role of Inclusions When manganese levels are less than 0.4%, chromium sulfides and chromium-rich sulfides Metallurgists have long known that the pres- will be present. As manganese levels reach 0.4 ence of a soft second phase dispersed in the ma- to 1.8%, chromium-rich manganese sulfides are trix of a parent metal can improve its machin- present. For manganese levels beyond 1.8%, ability. These particles provide a solid-state comparatively pure manganese sulfide will be lubricant between the chip and tool or a discon- found. The manganese-to-sulfur ratio is also im- tinuity in the material to aid in chip breaking. portant. For highest machinability levels, a high The challenge to the alloy designer has been to manganese-to-sulfur ratio is desired. However, develop second phases that produce these bene- if corrosion resistance is desired, a low man- ficial effects with a minimal of degradation to ganese level is preferred to encourage the for- the material itself. mation of chromium sulfides or chromium-rich manganese sulfides since these sulfide forms have superior corrosion resistance. Lead, Selenium, Tellurium Sulfides form initially as spherical inclusions The range of additions possible to stainless within the cast structure. Hot working, as well as steel is the same as for carbon steel. Lead addi- cold working, elongates these inclusions as tions probably are the best source for improving shown in Fig. 2. This elongation increases the Chapter 15: Machining / 187

machining speeds can be increased, improving machining productivity. Much has been written about the benefits of sulfides in a machining operation. A metallurgical perspective shows the more complex nature of sulfides. And, there is more to this than just adding sulfur. The discussion of the combination of manganese and sulfur revealed that there is a particular balance of manganese to sulfur to achieve desired needs. In addition to chemistry (manganese-sulfur balance), the size and shape (relative elongation or globular nature) of sulfides contributes to the machinability of stainless steels. Sulfides are defined into four categories based on morphology: type I to IV sulfides. Type I sulfides form from the melt as large globular sulfides and are assumed to be best for machinability. Type II sulfides generally form as a eutectic-like distribution of finer rod-like sulfides in interdendritic regions. Type III sulfides form as angular- shaped particles. Type IV sulfides form from the melt as plate-like sulfides in a ribbon-shaped pattern. It has been shown that the larger, more Fig. 2 Typical AISI 303. Source: Ref 2 globular (less-elongated) type I, homogeneously dispersed sulfides are better than the other types for enhancing machinability in both carbon and stainless steels. However, it seems that sulfides that are too large or too small can be detrimental to machinability. The coarsest type I sulfides, once thought to be best for machinability, are difficult to attain with today’s stainless manufacturing equipment. Ingot casting, with the slow cooling and solidification rates, is beneficial to the creation of the coarser type I sulfides. With the transition to continuously cast blooms, solidification rates are much faster, effectively creating a finer type I sulfide. The effect of sulfur on the machinability of stainless steels is more effective in smaller Fig. 3 AISI 303 chip breaking at the sulfides. Courtesy of amounts than in larger amounts. Figure 4 shows Ugitech a graph of drill penetration results in varying amounts of sulfur in an 18-9 stainless. It is eas- surface area of the sulfides and will form weak ily seen that small additions of sulfur have the planes that mechanically weaken the steel, per- greatest effect at sulfur levels <0.10%. This also mitting chip breakage. The deformation caused can be proven with a 304 or 316 stainless steel, by the severe bending of the chip during cutting for which sulfur levels are inherently lower than causes the chip to break at the sulfide striations, 0.030%. Machinists readily can see differences shown in Fig. 3. An additional benefit of these in tool wear between a 316 with 0.023% and manganese sulfide stringers is that when these one with 0.028% sulfur. That small addition of stringers touch and pass the cutting tool, a small sulfur has a dramatic effect. On the opposite amount of manganese sulfide is deposited on the side of the scale, with sulfur levels greater than tool surface, providing a lubricating layer be- 0.20%, the curve flattens out. tween the chip and tool. This reduces friction, From a machining perspective, stainless thus reducing heat to the tool. Consequently, steels can be classified into three groups based 188 / Stainless Steels for Design Engineers

Sulfur hurts corrosion resistance by locally depleting chromium from the matrix to precipitate as a sulfide. These manganese sulfide inclusions become exposed on the surface of the bar and become initiation sites for pitting corrosion. Passi- vation of the components will help, but the hole left behind by the sulfide becomes a collection site for contaminants. Other sulfide inclusions that form on the grain boundaries contribute to intergranular corrosion. Hot workability is also hurt by sulfides. Sulfur increases the hot shortness of materials during hot-forming operations. Fig. 4 Effect of sulfur on stainless machinability. Source: Ref 2 Manganese sulfides form stress risers within the material, which lead to reduced mechanical on the sulfur additions. In the first group, sulfur properties in notch-sensitive alloys, especially in content is limited to 0.010%. For these stainless the transverse direction. As the amount of sul- steels, corrosion resistance, weldability, or me- fides increases in stainless steels, a susceptibility chanical properties are more important than to longitudinal cracking can become an issue. machinability. Grades in this group are With these stress risers in place, any cold de- 316LVM and other remelted stainless grades, formation can lead to cracking along the sul- many duplex and super stainless grades. The fide stringers. As a general rule, the smaller the second group has sulfur contents ranging from bar, the higher the sulfur, and the higher the 0.010 to 0.030%. This group of stainless steels strain produce a high probability for a crack to still holds corrosion resistance, welding, and initiate. mechanical properties as critical but has the ad- Despite these deficiencies, the sulfides found vantage of increased machinability. Grades in in stainless steels are very effective in improving this group include 316/316L, 304/304L, 321, machinability, especially in austenitics, where 347, 410, and 430, among others. The third contact forces are very high. The benefits of sul- group with sulfur levels у0.15% are considered fur to improve machinability outweigh the resulfurized grades. Sulfur levels are generally losses due to defective parts, at least from the 0.25 to 0.35%, with some grades reaching machinist’s viewpoint. The question of perform- 0.45%. The grades in the third group include ance of the finished component is another issue, 303, 420F, 430F, 430FR, 1.4570, and others. A which concerns the end user; for the end user, fourth group of grades can be included in this sulfur is a major negative factor for corrosion re- classification, but the amount of sulfur is not the sistance. This has led to the development of al- grading criterion. This is where mills will rebal- ternative methods of improving machinability. ance chemistries and processing to enhance machinability beyond what the first three groups offer to create enhanced machining grades. Oxides Product offerings include Ugitech’s Ugima and The basic machinability-enhancing agent dis- Ugima XL (Ugima 2) materials, Sandvik’s San- cussed thus far has been sulfur. The beneficial mac materials, Outokumpu’s Prodec materials, effects of sulfur are undeniable, but the detri- and Carpenter’s Project 70+ materials. mental effect is equally evident. This has led Sulfides in stainless steels have a dark side, steel producers to look at other inclusion sys- especially when sulfur levels exceed 0.15%. It tems for a viscoplastic inclusion without the is well documented that sulfides negatively af- negative effects of sulfur. Since the 1900s, fect corrosion resistance and mechanical prop- steelmakers have known that injecting calcium erties of stainless steels. Sulfur tends to form into the melt converts refractory inclusions into segregated films with low melting points during soft, malleable, complex oxides that act as free- solidification. And, since these films are low machining agents with high-temperature lubri- strength, they may induce the formation of mi- cating capabilities. The oxide inclusion chem- crocracks brought on by solidification shrinkage istry is based on the CaO-Al2O3-SiO2 system. stresses. Further processing of these materials Figure 5 shows that small amounts of calcium can induce cracking even further, leading to can greatly increase tool life. These calcium- poor processing yields. based oxide formulations have been commer- Chapter 15: Machining / 189

Fig. 6 Complex Ugima oxides populating the 303 matrix. Courtesy of Ugitech

Fig. 5 Effect of calcium on machinability of 303. Source: Ref 2

Fig. 7 X-ray examination showing manganese sulfides and Ugima oxides coating the tool surface. EDAX, energy dispersive analysis by x-ray. Courtesy of Ugitech cialized for stainless steels, but the machinabil- The Ugima oxide works similarly to sulfur by ity agent is, at this point, not standardized as coating the cutting tool and acting as a lubricant. this process is difficult to reproduce consis- Figure 7 shows EDAX (energy dispersive analy- tently from heat to heat. However, Ugitech SA sis by x-ray) spectra of the surface of a carbide (formerly Ugine Savoie and now part of tool, proving the existence of coatings of man- Schmolz and Bittenbach) developed a propri- ganese sulfide and Ugima oxide. The Ugima etary and patented process sold under the trade oxide performs synergistically with sulfur. Since name Ugima and Ugima XL (Ugima and Ugima the oxide alone has limited lubricating abilities 2 in Europe and Asia). Figure 6 shows the at slow machining speeds, the manganese sul- Ugima oxide coexisting with sulfur in AISI. fides in the stainless are the dominant machin- 190 / Stainless Steels for Design Engineers

ing agent. As machining speeds increase, manganese sulfide eventually loses its lubricity, and the oxide acts as a high-temperature lubricant that will allow faster machining speeds without the need to use specialized tooling. Data generated at Ugitech as well as data generated by other producers have shown that once sulfur content reaches 0.30%, the contribution toward machinability flattens and possibly decreases, which seems to contradict the experience of many machine shops. However, the combination of sulfur and the Ugima oxide extends the machinability range for sulfur levels beyond 0.30%, as seen in Fig. 8. Tool lubrication is only one function the com- Fig. 8 Comparison of machinability of AISI 303 at different plex oxide performs. Like sulfur, the Ugima sulfur levels with and without the Ugima oxide. oxide is a discontinuity and will aid in chip The vertical axis, VB30/0.3, represents 0.3 mm of tool wear in breakage. Figure 9 exhibits cross-sectional pho- 30 min. tomicrographs of 304L chips with a 0.025% sulfur level. As shown, the chip is shearing along Tooling and Coolants complex oxide stringers, helping the chip break. Ugitech has seen synergistic effects with The machining of stainless steels can be copper additions as well. Machining tests at much more complicated than for other materi- Ugitech and many field experiences have shown als. Machine and tool rigidity, machine power increases in machining performance of 30% requirements, sharp cutting tools, and optimum with the addition of 1.4 to 1.8% copper to a lubrication type and amounts are very impor- 303 stainless steel with the Ugima oxide. The tant. Feed rates need to be sufficiently more ag- grade chemistry meets EN 1.4570 and is sold gressive than with carbon steels to prevent work under the trade name 303 Ugima UX (4570 hardening and glazing of the material as well as Ugima in Europe). Further additions of copper to avoid reducing tool life. up to 4% have exhibited even better machining High-Speed Tool Steels. The cutting tools performance. themselves are the main variable other than the

Fig. 9 Comparison of 304L chips with and without the Ugima oxide. Courtesy of Ugitech Chapter 15: Machining / 191

material in determining the quality of the ma- hard layer on the tool surface that will provide chined part. The earliest cutting tools were additional lubricity between the tool and chip as those made of the high-speed steels (HSSs). The well as potentially providing heat resistance for wrought HSS tooling was very versatile, al- the tool. Like the carbide, there are certain coat- lowed resharpening of the tools many times, ings that are more beneficial than others. Tita- and is still a preferred choice in many shops. nium nitride (TiN) was a great coating when it With time, mills perfected their ability to pro- was first introduced, but further development in duce more highly alloyed tool steels to meet the this area has created other coatings that work increasing demands of the machining industry. even better. Grades like titanium-aluminum- Today, tools are being made with a powder met- nitride (TiAlN) and aluminum-titanium-nitride allurgy process, by which ingots of compacted (AlTiN) are great coatings for turning, form cut- high-speed tool steel powders have more struc- ting, cutoff, grooving, drills, reamers, and tural homogeneity and thus better wear and heat milling. These tools are usually subjected to resistance. Powder metallurgy techniques pro- higher temperatures from high speeds, deep duce even more highly alloyed high-speed tool cuts, and limited coolant. The aluminum in the steels with properties approaching those of car- coating breaks down and combines with the bide tooling, allowing better machining per- surrounding oxygen to form aluminum oxide formance. Although high-speed tool technology (Al2O3), a material with higher hardness and has improved, its limiting factors are heat and better thermal resistance. When tooling will not wear resistance. Use of HSS tooling still limits be subjected to much heat, such as with taps, machining speeds and performance. high wear resistance coatings such as titanium Carbides. Carbide tooling is the next genera- carbonitride (TiCN) coatings are good. tion of tool materials after HSS, although it Coolants. Stainless cutting is hot, has high tends to be hard and brittle. In the past, unless it frictional forces, and has tendencies of the was possible to feed hard, maintain fast speeds, metal to stick to the tools. Coolants need to re- and have uninterrupted cuts, carbide tooling move this heat and provide lubrication to reduce was a good choice. However, carbide tool tech- friction and minimize BUE. Coolants available nology has come a long way in grades and are petroleum based, semisynthetic, synthetic, technology. Micrograin and nanograin carbides water soluble, and the new type: vegetable are providing carbide tooling with increased based. All of these coolants need to be highly toughness. Single-point tooling is now able to fortified for use in stainless machining. For ex- withstand the punishment of an interrupted cut, ample, petroleum-based oils need high sulfur, and drills are able to withstand some flexing. chlorine, and fat levels for lubricity at higher Generally, the carbide grades to be used when temperatures. The other coolants need high- machining stainless steels are the C5- to C8- pressure additives, again to help lubricity. It is type carbides. These classes of carbides are always a good idea to discuss the options with harder and have more heat resistance, properties the coolant supplier. that are needed when cutting stainless. The discussion in this chapter has stated that the cutting of stainless is more difficult and generates more heat. Cutting materials that withstand these cir- REFERENCES cumstances are needed. The C2- to C4-type car- 1. Machining Data Recommendations, in Met- bides are not well suited for stainless steels be- als Handbook, desk ed., 2nd ed., J.R. Davis, cause they do not have the heat resistance Ed., ASM International, 1998, p 917–950 needed. There are always exceptions. For exam- 2. T. Kosa and R.P. Ney, Sr., Machining of ple, it is possible to use the C2- to C4-type car- Stainless Steels, in ASM Handbook, Vol bides when using the older cam-operated multi- 16, Machining, ASM International, 1989, spindle and Swiss machines, for which machine p 681–707 speeds are limited. Tooling manufacturers such as Kennametal, Sandvik, Iscar, etc. have their own proprietary grade designations, but many correspond to the C5 to C8 types. SELECTED REFERENCE Coatings. Tool coatings have contributed to machining improvements. Coatings add a very • www.ugitech.com Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 193-199 All rights reserved. DOI: 10.1361/ssde2008p193 www.asminternational.org

CHAPTER 16

Surface Finishing

Summary Beyond such considerations, one can also create surfaces on the stainless that enhance its SURFACE TREATMENTS are extremely beauty and performance. The surface can be important to the end user, and they are totally made reflective or matte, ground or mechanically within the end user’s control and specification. patterned, coated, painted, plated, or oxidized. They include the proper cleaning of stainless, It can be treated chemically or electrolytically. the various means of descaling after thermal The surface can be altered on an atomic basis, treatment, and the choice and application of sur- sometimes producing profoundly different me- face finishes. The cost ramifications of improper chanical and corrosion-resistant properties. All or suboptimal surface treatments are immense of these are discussed in this chapter. because of the possibility of them compromising corrosion performance, which is a characteristic for which the end user pays dearly. Function of Surface Treatments

Removal of Oxide Scale Introduction Oxide scales form on stainless steel during annealing, hot-forming, and joining operations. Surface finishing is usually very important for Removal of this scale is important to proper cor- stainless steel. The underlying economic justifi- rosion resistance. This is because the chromium cation for using stainless steel is that it does not in the steel oxidizes much more readily than corrode if properly specified for the environment other elements, so the surface of the steel under it faces. Thus, its surface appearance remains the oxide is chromium depleted and therefore normally intact throughout its life. This appear- has lost possibly a significant amount of corro- ance should therefore be aesthetically pleasing, sion resistance. An oxide scale is quite different even in an industrial setting, and the surface fin- from a passive film. Scales deplete chromium; ish should not detract from its performance. passive films enrich the surface in chromium by Raw stainless surfaces resulting from rolling selective loss of iron. and annealing operations are not considered at- Oxide scales are arguably best removed by tractive and are used only for functions in which pickling. Pickling is the chemical dissolution of aesthetics are a negligible consideration. Even the oxide scale. The acids commonly used are ni- so, there are surface treatments required for tric (HNO ), hydrofluoric (HF), and sulfuric stainless steel intended for such uses. The stain- 3 (H2SO4). HCL, which is commonly used to less surface must be freed of: pickle carbon steel, is not recommended for stain- ¥ Oxides resulting from annealing, joining, or less because it locally attacks (i.e., pits) the sur-

hot forming face. HNO3 is an oxidizing acid that by itself does ¥ Accumulated ambient foreign material not pickle stainless. It is used in combination with ¥ Applied process materials, such as forming HF to modulate the attack by the strongly reduc- lubricants, ﬂuxes, etc ing action of the HF. This combination allows ¥ Contamination from other materials, espe- good control of pickling rates by varying the ratio

cially iron. of the two acids. H2SO4 is used alone. It is often 194 / Stainless Steels for Design Engineers

used as a preliminary pickling because of its ag- The HNO3 provides: gressive, and less-controllable, action. ¥ a source of H+ ions Pickling involves both scale dissolution and ¥ stabilization of the passive film metal dissolution. Thus, pickling can be consid- ¥ elevation of the redox potential ered a deliberately imposed corrosion process. ¥ an oxidizing agent of the base metal The scale formed during hot rolling may be an ¥ a dissolving agent for the scale embedded scale with a minimal chromium- depleted layer beneath the scale. Annealing and The nitrates carried off from this pickling shotblasting are required before pickling so that process are an environmental problem. It has the acid can penetrate to the chromium-depleted been shown that the HNO3 can be eliminated region under the scale and selectively dissolve and replaced by hydrogen peroxide (H2O2). this layer. Nearly all the action of pickling de- This is the proprietary, patented UG3P process rives from undercutting of the oxide scale and developed by Ugine (Ref 1) or Henckel’s dissolution of the chromium-depleted layer. It is “CleanOx” process. also vital that this layer be removed as it dimin- Electrolytic pickling is commonly used in the ishes the corrosion resistance of the surface. production of cold-rolled stainless. This process Simply removing the scale mechanically may uses alternating positive and negative polariza- be insufficient if only the oxide is removed and tion in baths using sulfates or nitrates. The H+ the compromised (depleted) metal surface is ions are produced by the polarization instead of left. Metal removal rates are proportional to the by an acid, so neutral solutions can be used. chromium level. Thus, sensitized material will The low solubility of trivalent chromium undergo grain boundary attack if pickled. All (Cr+3), typically found in oxide scale, makes stainless has slightly greater oxidation in the such scale hard to remove. Use of strong acids grain boundaries when it is oxidized, and this and complexing agents is required to get pick- leads to “ditching” of the grain boundaries dur- ling to occur at an economically acceptable rate. ing pickling, resulting in the matte appearance Hexavalent chromium (Cr+6) is much more of a pickled surface, as seen on two-dimensional readily soluble in aqueous solutions, for exam- flat-rolled material. ple, oxidizing treatments such as Kolene,* a Typical concentrations for H2SO4 pickling are molten mixture of nitrate salt and hydroxide, or 8 to 15% by weight. Bath temperatures are gen- electrolytic conversion treatments such as elec- erally 150 ¡F (65 ¡C). The rate of pickling de- trolytic sodium sulfate. Both treatments convert creases rapidly as iron builds up in the bath, and the trivalent chromium to hexavalent form, al- the bath must be replaced for efficient pickling. lowing for easy scale dissolution. Their use The attack of H2SO4 on the base metal can be may, however, lead to environmental problems severe, and undissolved constituents can remain as hexavalent chromium compounds are toxic on the surface as “smut.” This smut must be and heavily regulated. physically removed or dissolved by subsequent Pickling may be assisted by prior treatment of HF/HNO3 pickling. the scale in molten 85% sodium hydroxide The HF/HNO3 pickling is carried out between (NaOH), 14% sodium nitrate, and 1% sodium 120 and 140 ¡F (50 and 60 ¡C). Higher tempera- chloride bath. The strong oxidizing action of tures cause excessive HF evaporation and can such a bath chemically alters the chromium also lead to visible emission of nitrogen oxides. oxide in the scale to a more soluble oxide with- The NO2 (nitrogen dioxide) produced during out attacking the metal, making it more easily the reduction of the HNO3 (oxidation of the removed by subsequent pickling. metal) is visibly brown-red. The rate of NO2 formation increases at higher temperatures. Cleaning HNO3 concentrations are normally from 10 to 25%. HF concentrations can vary from 1% for Stainless steel is very resistant to chemicals light scales to 8% for maximum aggressiveness of many kinds, permitting it to be cleaned by and difficult-to-pickle alloys. many aggressive agents. Because contamination The HF provides: is “on” the surface of stainless rather than “in” the surface as is the case with many materials, ¥ a complexing agent for iron and chromium simply using the cleaning agent appropriate for ¥ destabilization of the passive film ¥ stabilization of the redox potential * Kolene is a registered trademark of Kolene Corporation. Chapter 16: Surface Finishing / 195

the contaminant will permit very thorough carbonate or brush material that is sufficiently cleaning of the stainless surface. The stainless soft not to mar the stainless surface. The user is surface will not be harmed by cleaning as long encouraged to test whether a product meets this as strong halides, iron utensils, and abrasives requirement by testing it on a small, preferably that alter the surface finish are avoided. Table 1 unexposed, area. Blasting the surface with car- summarizes some cleaning recommendations bon dioxide pellets is a rapidly growing process based on the contaminant to be cleaned. for removing paint and other adherent, soft The greatest controversy regarding cleaning coatings and deposits without damaging the sur- stainless steels involves the use of sodium face of stainless steel. hypochlorite. Many highly qualified experts These cleaning recommendations apply to in- flatly disapprove of its use because it can easily dustrial, architectural, and domestic uses of leave a residue of chloride on the stainless sur- stainless. For special levels of cleanliness re- face, leading to localized corrosion. But, in real- quired for medical, pharmaceutical, or semicon- ity, health considerations make this position un- ductor applications, refer to the chapters dealing tenable. Sanitation concerns in the food service with those applications. industry take precedence over this prohibition. Passivation is a very commonly used surface In fact, every stainless steel, as explained in the treatment to remove surface contamination, no- “Corrosion and Oxidation” section of this Vol- tably iron, and to form a passive film. The film ume, is resistant to some level of chloride con- forms of its own accord when a clean surface centration at a given temperature and pH. Thus, encounters moist air, but film formation can be short-term, room temperature use of hypochlo- accelerated by controlling the environment. rite bleaches will generally not harm most types The surface to be passivated should first be of stainless steel if well rinsed after application. cleaned by one of the methods discussed. This Field experience has shown that stainless steels allows uniform passivation and avoids contami- with less than 16% chromium can be harmed by nation of the passivating solution, especially by hypochlorite bleaches. Damage occurs mainly chlorides, which can cause a rapid attack. Those on abrasively polished surfaces in alloys with stainless steels with more than about 17%

16% chromium, such as 430, but does not occur chromium can be pickled in 20% HNO3 at 50 to on roll-ﬁnished or bright-annealed surfaces. 60 ¡C (140 ¡F). Precipitation-hardenable (PH), Stainless steels with slightly higher chromium martensitic, and straight chromium grades with and nitrogen, such as 201, 301, and 304, are not less than 17% chromium should have 2.2 g/L attacked unless concentrated chloride solutions sodium dichromate added to that solution to are permitted to stay on the surface, particularly avoid attack. The free-machining grades are in crevices. the most easily attacked of all types. They The recommendation not to use abrasives should be immersed in 5% NaOH for 30 min at does not include soft abrasives such as calcium 75 ¡C (165 ¡F) followed by a rinse before the

Table 1 Recommended cleaning methods

Contaminant Cleaning recommendation Comments Exterior soiling Soap, detergent, or dilute ammonia Use a soft cloth or sponge, clean water; dry with forced air or a dry cloth. Fingerprints Detergent and warm water or a hydrocarbon Wax and oil polishes minimize fingerprinting. Glass solvent cleaner is appropriate for mirror finishes. Grease, oil Hydrocarbon solvent Alkaline cleaners may also be used in severe cases but may require cleaning the entire surface to maintain visual uniformity. More severe stains, discolorations, Nonscratching creams or polishes Do not use HCl-containing products. and rust stains Hypochlorite bleaches must be well rinsed to avoid pitting. Hard water scale, mortar 10Ð15% phosphoric acid, sulfamic acid- Neutralize with ammonia, rinse, and dry. containing, or oxalic-containing cleansers Do not use HCl-containing products. Oxides, heat tint If severe, treat by pickling; Scotchbrite©, If abrasives must be used, they should blend stainless scouring pad, or nonscratching with existing surface finish in size and direction. cream or polish Paint Alkaline, trisodium phosphate, or hydrocarbon Follow manufacturer’s directions solvent 196 / Stainless Steels for Design Engineers

nitric/dichromate passivation is performed. This Electrochemical methods permit more uni- should be followed by a rinse, another 30-min form and reproducible color and tougher films. NaOH treatment, and a final rinse. The most prominent is the International Nickel In recent years, citric acid has become a pop- Company (INCO) process. In this two-step ular replacement for HNO3 because it avoids process, the steel is first immersed in a mixture the problem of toxic nitrates and hexavalent of one part chromic acid, two parts H2SO4, and chromium. When citric acid is used, a 10% one part water at 30 ¡C (85 ¡F). This is followed solution is applied for the same time and tem- by an electrochemical treatment with the same perature as with HNO3. The use of NaOH be- bath or with H3PO4 substituted for H2SO4. fore and after is still recommended for low- The colors produced are correlated to thick- chromium and free-machining alloys. Refer to ness and treatment parameters (Ref 2) as shown ASTM documents A 967-01, A 380-99, B 912- in Table 2. Rocha-Fila et al. confirmed that 00, as well as Federal Specification QQ-P-35. these coloring treatments did not degrade the pitting corrosion resistance since their forma- Brightening tion mechanism more closely resembles passive film formation than oxidation. Stainless steels can be brightened by chemical or electrolytic action, which selectively dissolves the surface in such a way that it becomes microscopically smoother. Electropolishing is Aesthetic Surface Finishes the most effective means of accomplishing this. It causes the surface roughness to decrease by When steel is used in other than the as- annealed and pickled state, it is often for aes- approximately one-half in Ra (the arithmetical average of surface peaks and valleys as meas- thetic rather than functional reasons. The major ured over a straight line). Electropolishing, like exception to this is temper-rolled strip, which pickling, also selectively removes any exposed has a bright surface from cold rolling, but chromium-depleted regions, leaving only the whose normal use is strictly functional. The hot- bulk alloy with the intended corrosion resist- rolled or cold-rolled annealed and pickled sur- ance on the surface. Apparent pitting resistance face finish is nonuniformly dull and unattractive is thus increased. The resulting surface is bright to most observers. When stainless steel first ap- and cleanable and provides the optimal corro- peared on the market, a highly polished surface sion resistance that a given alloy can achieve. was the paragon, as showcased on the Chrysler The reflectivity of the surface is a function of Building in New York City (Fig. 1). the preelectropolished surface. This surface finish was very expensive to pro- The most commonly used electropolishing duce since it had to be done by polishing and solution is 40% H SO and 45% phosphoric buffing. Soon, the more economical abrasive 2 4 polishing became the standard, and it had the acid (H3PO4), balance water, used at 90 ¡C (194 ¡F) with a current density of between 1.0 benefit of removing the many cosmetic defects and 3.0 amp/m2. Other baths using perchloric that then were common to the manufacture of acid are technically good but carry the risk of stainless. Much higher levels of surface quality explosion. were made possible with the development of the Sendzimir mill to the bright anneal process per- Coloring mitted much higher levels of surface quality. Stainless steel passive films are so thin that they are quite invisible. Oxide films are thicker Table 2 Parameters for oxide film coloring of and through optical interference can cause differ- stainless steel ent colors. The films are formed thermally in a Oxide film process known as heat tint. The color depends on Heat treatment time Color thickness, nm oxide film thickness, which is a function of the 10 Gold brown 70 time at a given temperature, well metal composi- 15 Brown/red/blue 100 tion, and oxygen partial pressure. The colors typ- 20 Brown/blue 120 25 Green/blue 140 ically range from light yellow formed at 300 ¡C 30 Green/gold/blue 165 (570 ¡F) through violet formed at 420 ¡C (790 35 Gold/green 185 ¡F) and dark blue formed at 600 ¡C (1110 ¡F). 40 Gold/green/brown 210 Reproducibility is sometimes a problem. 50 Red/brown 250 Chapter 16: Surface Finishing / 197

Stainless steel surface finishing is governed by: ¥ High strength and high work-hardening rate require more power for metal removal. ¥ High surface hardness influences which abrasive will be effective. ¥ Low thermal conductivity can cause excessive temperatures during processing and distortion. ¥ Oxidation (heat tinting) that occurs over about 250 ¡C (480 ¡F) ¥ Residual stress due to surface working, especially in austenitic alloys Grinding is a relatively coarse procedure with usefulness that is largely confined to defect removal. Polishing for aesthetic purposes is generally done with abrasive coated belts and is done to both coils and sheets. There are no offi- cial roughness values for the various surface finishes. ASTM merely describes the finish by a grit with which it is typically made. Table 3 shows the varying surface roughnesses that are typical of polished stainless. Neither the producer nor the customer is protected by a clear specification as of this writing. Thus, the producer is encouraged to publish standards, and the customer should specify min- View of the Chrysler Building in New York. Fig. 1 imum and maximum Ra values for his purpose. ©iStockphoto.com/stevenallen The typical number 4 polish varies in roughness from the beginning of the coil to the end be- Mirror finishes could be achieved with cold cause of the wear of the abrasive particles on rolling followed by bright annealing. This has the belts. The difference can be more than 10 μ μ R not entirely replaced sheet buffing to obtain in. (0.25 m) in a, is quite visually apparent, mirror finishes, but sheet buffing has been rele- and will make adjacent pieces of stainless from gated to those few applications where a high the same coil look different. This may be objec- degree of perfection is demanded. The main tionable for certain products, especially appli- methods of producing aesthetic surface finishes ances and architectural panels. on stainless steel are abrasive finishes and rolled Polish appearance varies also with the pres- finishes. The latter are superior in uniformity ence or absence of lubricant. Lubrication during and corrosion resistance, but the polished fin- polishing (i.e., wet polishing) gives sharper cuts ishes are still more common. and less heat tint from frictional heating, resulting in more surface brightness. Polish appearance also depends on the length of the grit line Polished Finishes caused by an individual particle of abrasive. The Polishing is carried out with coarse- to length of the grit line varies with the speed of the medium-sized abrasives that are bonded to a material as it passes the rotating abrasive belt, flexible backing. This is distinct from grinding, in which abrasives are bonded rigidly to each other on a rigid backing. Buffing is done by Table 3 Polished finish designations based on very fine abrasives, which are loose and do their grit sizes to achieve target surface roughnesses work by being forced along the surface by a soft Finish number Grit number Ra max, µm Ra max, µin. material. The standard rule is that materials 3 60Ð120 1.0 40 used in finishing must not permit iron or iron 4 120Ð180 0.75 30 oxide particles to come in contact with the 7 240Ð320 0.30 8 stainless surface or passivation layer. 8 500 0.15 4 198 / Stainless Steels for Design Engineers

and the rotational speed and diameter of the and rolling by which the surface is made flaw- backup roll for the abrasive belt. This combina- less and very smooth. Final brightening occurs tion of variables, while seemingly controllable, with temper passing after bright annealing. usually varies enough so that no two polished sheets look identical to the trained eye. This Rolled Finishes variability also causes problems in field repair. Repair of a polished surface damaged, for The bright-annealed and temper-passed sur- example, by welding or scratching is quite face may be used in its mirror-like condition, challenging. A perfect match and blend to the or it can be used as a basis for rolled finishes, surrounding original surface is an art, and no which can take on any appearance and can be practitioner of that art can accomplish it per- engraved onto a temper mill roll. All polished fectly. If the repaired area is small, then it can finishes can be duplicated as rolled finishes. be acceptable, if not undetectable, but for large But, in addition, many other designs such as areas, repolishing of the entire surface is often replicating fabric or leather, geometric de- the best remedy. signs, or matte finishes can also be made. The A solution to the problem of matching pol- pattern on the roll is impressed into the stain- ished surfaces is to use very long grit lines for less surface, and elastic flattening of the roll R finishes of the same roughness. By eliminating and the stainless cause about 50% less a on the variable of grit length, only correct pressure the stainless than on the roll. This effect in- and grit size are required to achieve good visual creases with the yield strength of the stainless matching. This can be done with a belt sander. being processed. The benefits of rolled finishes These finishes are called “hairline” in Japan and are: “grainline” and other names in the United ¥ They can be made identically from coil to States. Lack of uniformity, difficulty of repair, coil since roll engraving is quite precise. and a decrease in corrosion resistance are the ¥ They retain the enhanced corrosion resist- chief drawbacks of polished finishes. ance of bright-annealed material. ¥ They are less expensive to produce, so the Bright Annealing final product is generally priced lower than Aesthetic finishes that do not depend on abra- polished material. sion are derived from bright annealing. Bright In theory, such finishes can be made from air- annealing is annealing in a very low oxygen annealed material, but finishes applied by roll to atmosphere, either dissociated ammonia* or two-dimensional surfaces that have been dulled hydrogen/argon. This process was originally by pickling are somewhat gray and indistinct. If developed as a means of producing bright fer- the pickling is kept mild enough to retain surface ritic trim for automobiles. That use has largely brightness, then it is possible insufficient prick- passed out of favor, but the process itself is in- ling has occurred to remove the chromium- trinsically superior to annealing in air because depleted layer caused by the air anneal. Some of no oxide is formed only to be later removed, the finishes produced by rolling are shown in the usually at significant expense. Because the orig- architectural chapter (see Chapter 18). These fin- inal product was strip, many older bright-an- ishes can be applied to all the normal sheet nealing facilities are narrower than normal sheet alloys, including the lean duplex alloys. This has width. The wider, more modern lines are high permitted rolled finishes to be used on UNS speed and wide enough to produce bright sheet. 32003 for building exteriors in Doha, Qatar, Bright-annealed sheet will only be as bright and where corrosive conditions require a PREN flawless as the cold-rolled sheet that is an- (percentage chromium equivalent) of 25. nealed. Therefore, a quality product, generally The most critical applications from a surface as mirror-like as possible, must be produced perfection point of view use rolled finishes with great care. The better producers have pro- rather than abrasively finished number 4 pol- prietary methods of prior pickling, annealing, ishes. These finishes also permit type 430 stainless steel to be used successfully in restaurant * At one time, “dissociated ammonia” referred to a 3:1 applications without corrosion issues where (mole ratio) mixture of hydrogen and nitrogen produced by the catalytic decomposition of anhydrous ammonia. Today, polished type 430 material previously had this composition is prepared by mixing gases. For many corroded. There are also instances for which stainless steels, the 3:1 H2:N2 can be varied without difficulty. polished UNS S31600 was corroding in coastal Chapter 16: Surface Finishing / 199

architectural applications in the United States. exposing an activated stainless surface to a When the abrasively polished surface was high carbon fugacity at 470 ¡C (880 ¡F) for replaced with rolled finish UNS S30400 around 200 h a 50 nm thick layer with 12 at.% (Koolline), there was no further corrosion. Be- carbon in supersaturation can be achieved with- cause alloying is such a high component of cost out carbide formation. The properties of this in stainless, it makes sense to employ rolled fin- layer are phenomenal; the hardness is 1000 ishes whenever possible. However, it must be Vickers 25 compared to 200 for the base alloy noted that when manual polishing is used to re- (Ref 3). In addition, the corrosion resistance in- pair scratches or other damage in rolled finish creases significantly. Only by such supersatura- material, its corrosion resistance may be reduced. tion with carbon could it be determined that carbon, like nitrogen, is a powerful antipitting Surface Alteration alloying element when kept in solid solution. These processes are at the very initial stages of In carbon steels, surface chemistry can be commercial use. changed to affect certain properties. Carburiz- ing and nitriding are examples of such processes. Simply using these processes on REFERENCES stainless cannot be done because these elements combine too strongly with chromium as 1. Stainless Steels, Les Editions de Physiques, carbides or nitrides, dramatically reducing the 1989 corrosion resistance. There have been modifi- 2. C. Rocha-Filo et al., J. Braz. Chem. Soc., cations to carburizing and nitriding that permit Vol. 15 (No. 4), 2004, p 472Ð480 austenitic stainless steels to have very high car- 3. Y. Cao, F. Ernst, and G. Michal, Acta bon levels implanted to a thin surface layer. By Mater., Vol 51, 2003, p 4171Ð4181 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 201-212 All rights reserved. DOI: 10.1361/ssde2008p201 www.asminternational.org

CHAPTER 17

Welding

NEARLY ALL WELDING of stainless steel characteristics of austenitic stainless steels that is done by end users or processors. Like thermal distinguish them from ordinary carbon steels in processing, it is complex in theory and practice. welding are: This chapter gives a basis for understanding the ¥ Austenitic stainless steels have lower thermal influence of alloy composition and metallurgy conductivity and higher thermal expansion on the welding process, which must be re- than carbon steels or ferritic stainless steels, spected as a process that combines melting, re- which can localize the heating, thus increas- fining, and thermal processing. Knowledge of ing the potential for residual stress and each aspect is required for the process to be de- therefore hot cracking. signed and executed properly. ¥ Stainless steels contain readily oxidized The welding and joining of stainless steels re- chromium, which must be protected. quires knowledge of both the technology of the ¥ Surface oxidation during welding depletes welding or joining process and the response of chromium in all types of stainless steel from the steel to the thermal and mechanical effects the underlying surface, resulting in reduced of the process. The welding process must, of corrosion resistance unless this layer is re- course, produce a sound joint, but it must also moved. result in the weld and its surrounding affected ¥ The possible formation of chromium car- metal having correct strength, toughness, corro- bides in the heat-affected zone (HAZ) can sion resistance, etc. for the intended service cause susceptibility to grain boundary corro- conditions. This chapter does not attempt to sion (sensitization). teach welding. The main objective is to show ¥ The possible precipitation of intermetallic how standard welding technology is correctly phases in the HAZ can lower toughness and applied to stainless steels. corrosion resistance. The foremost special consideration of weld- ¥ There is increased microsegregation in the ing stainless steel as opposed to carbon steel is fusion zone with increasing alloy content. that the chromium in stainless steel, which is ¥ There are thermodynamically metastable what makes it stainless, must be protected from conditions due to the low diffusion rates in oxidation, so that: the face-centered cubic (fcc) matrix. 1. It stays in solution as a corrosion-resisting The influence of carbon has been well ad- element. dressed using low-carbon versions of all grades 2. It does not form refractory oxides that would whenever welding involves significant time be- diminish weld soundness. tween 600 and 900 ¡C (1110 and 1650 ¡F). This prevents rapid precipitation by reducing the supersaturation of carbon. The older method of Welding Characteristics of preventing sensitization is to stabilize the alloys Stainless Steels with titanium, as in type 321, or with niobium, as in 347. This is foolproof only if carbon levels Austenitic stainless steels are readily are low, less than 0.04%, since TiC can dissoci- welded by nearly all welding techniques. The ate at elevated temperatures and not be able to 202 / Stainless Steels for Design Engineers

recombine successfully with titanium during some highly alloyed grades with compositions cooling, permitting a thin zone of sensitization that do not permit a ferritic solidification mode. called knife-line attack. Fortunately, most 321 In such alloys, sulfur and other contaminants, and 347 are produced with carbon levels below such as phosphorus, oxygen, zinc, and copper, 0.03%. The higher carbon-stabilized alloys and must be excluded from the weld zone. Welds of the high-carbon (>0.03%) unstabilized alloys less highly alloyed austenitics, generally those must be annealed after welding to redissolve with less than 20% chromium, which are bal- chromium carbides if the cooling was suffi- anced to freeze in a ferritic mode, retain some ciently slow for the carbides to have formed. ferrite at room temperature, normally between This is avoided only in thin-gauge (>1.5 mm, 3 and 10%. This is not harmful since the ferrite 0.06 in.) material or when the HAZ is drasti- is richer in chromium and in molybdenum, if cally reduced, as in laser welding. present. The high thermal expansion of austenitic The amount of ferrite expected can be meas- stainless steel can cause high residual stress ured by magnetic devices and estimated from around welds, which may require annealing to the Schaeffler diagram, a useful empirical map- eliminate. Another serious threat posed by ther- ping of weld metal phase composition shown in mal stresses is hot cracking. This can occur to Fig. 1. This diagram has an arbitrary cooling material that has just solidified when geometric rate resembling that of tungsten inert gas (TIG; constraints to contraction imposed by the sur- described in a separate section of this chapter) rounding material imposed act on weak grain welds. Faster or slower cooling will change the boundaries. This weakness occurs when the relative amounts of ferrite and austenite because steel solidifies in an austenitic mode. When of the need for diffusion to achieve the most sta- austenite freezes, it strongly rejects sulfur to the ble phase balance. Very rapid cooling, as with intergranular areas, where it forms weak films. laser welding, tends to make austenitic welds This is solved by balancing the composition so less ferritic and has the opposite effect in duplex that alloys solidify first as ferrite, which does alloys. not reject the sulfur, forcing it to precipitate as The Schaeffler diagram has been improved sulfide inclusions within the grains. This ap- by the Welding Research Council’s adoption of proach is highly effective but cannot be used for the modification shown in Fig. 2, which super-

Fig. 1 The Schaeffler diagram. Source: Ref 1 Chapter 17: Welding / 203

Fig. 2 Welding Research Council’s (WRC’s) 1992 constitution diagram imposes the solidification mode as a function of Restricting heat input to under 16 kJ/mm the composition. The crucial line on this dia- (400 kJ/in.) and interpass temperature to under gram is dotted-dashed line AF, which delineates 150 ¡C (300 ¡F) helps to minimize each of those compositions that solidify in a primary these risk factors inherent to the more highly ferrite mode, precluding the problem of inter- alloyed austenitic grades. Note that the influ- granular solidification cracking. ence of microsegregation of alloying elements Another problem particular to the more is separate from and in addition to the negative highly alloyed grades is the formation of inter- influence of sulfur on the corrosion resistance metallic phases from long cumulative exposure of welds. Austenitics at the alloy level of 316 to temperatures in the 600 and 900 ¡C (1110 and and above should not have sulfur above 1650 ¡F) range (coincidentally, the same as for 0.001% for these alloys to deliver the expected carbide precipitation). The slow diffusion of al- corrosion resistance. loying elements in austenitics makes this a The austenitic stainless steel weld metal com- lesser problem than in ferritics or duplex. This position can be altered by the gases to which the adverse precipitation is largely prevented in the molten base metal is exposed. Lack of shielding modern, nitrogen-alloyed grades, so these al- can lead to oxygen combining with chromium loys are recommended if extensive welding is and other elements, creating slag and depleting planned. The more highly alloyed grades also the alloy of needed elements. Thus, oxygen-free suffer from greater microsegregation during so- gas mixtures are used to exclude the ambient at- lidification. This causes austenitic dendritic mosphere from the molten pool during electric cores to have lower chromium and molybde- arc welding. Inert gases provide the barrier, num content and consequently lower corrosion while the addition of 3 to 5% by volume of ni- resistance. Thus, the welds have lower resist- trogen gives the necessary partial pressure to ance to localized corrosion. This is addressed by ensure that welds will not be depleted of vital using more highly alloyed filler metal or by so- nitrogen content. Figure 3 (Ref 1) shows the in- lution annealing the welds. fluence of nitrogen content of the shielding gas 204 / Stainless Steels for Design Engineers

¥ Duplex alloys are more sensitive to problems in the HAZ because their generally high chromium and molybdenum content plus their ferritic content make the precipitation of embrittling intermetallic phases more rapid than in austenitics, so minimizing the total time at high temperature is the overriding concern. ¥ While carbide sensitization is not an issue with the duplex alloys, the formation of intermetallic phases can cause loss of corrosion resistance. ¥ Duplex, like all stainless types, must be protected from oxidation by shielding gas, and Fig. 3 Effect of weld shielding gas composition on crevice corrosion resistance of autogenous welds in AL-6XN since nitrogen is a crucial alloying element, alloy tested per American Society for Testing and Materials especially in duplex alloys, it must be a (ASTM) G-48B at 35 °C (95 °F) component of the gas mixture. ¥ Cleaning before and after welding is equally on corrosion resistance of a highly alloyed important in duplex as in austenitics. austenitic grade. Excess nitrogen in the shield- Modern duplex alloys derive their impressive ing gas (e.g., more than 10%) can cause poros- strength, toughness, and corrosion resistance ity in the weld, and greater than 5% is detrimen- from their nearly equal percentage of ferrite and tal to the life of the tungsten electrode. austenite. The nitrogen content of the austenite The heat from welding can produce a surface brings its corrosion resistance up to that of the oxide composed mainly of iron and chromium. ferrite phase, which is richer in chromium and The underlying surface can be significantly de- molybdenum. Nitrogen additions partition to pleted of chromium because of the loss of the austenite and thus both strengthens it and in- chromium to this scale and therefore signifi- creases its corrosion resistance to close to that cantly lower in corrosion resistance. Pits can of the ferrite. The early duplex alloys had a ten- start in this thin layer and propagate into sound dency to form excessive ferrite when welded metal beneath. For heat-tinted surfaces, the and formed embrittling intermetallic phases darker the tint, the stronger will be the effect. To rather rapidly. The additions of larger amounts fully restore corrosion resistance, the area must of nitrogen stabilized the austenite to higher be ground to remove the oxide and any depleted temperatures, so welds did not become so fer- base metal. This should be followed by acid ritic. The nitrogen also decreased the speed at pickling, which completes the removal of the which intermetallic phases form, enlarging the oxide and depleted zone. time window for welding without their precipi- Duplex stainless steels differ from austenitic tation. And, by promoting greater austenite for- stainless steels in their metallurgical response to mation at high temperature, the addition of high welding mainly because their approximately (>0.12%) nitrogen actually reduces the ten- 50% ferrite causes greater thermal conductivity dency for chromium nitride precipitation. De- at lower temperatures, and ferrite has greater spite these advances, the key precaution in diffusion rates. These alloys solidify in a com- welding duplex alloys is to prevent the forma- pletely ferritic mode, and since ferrite rejects tion of embrittling phases while preserving as little sulfur on solidification, hot shortness is not close to a 50/50 austenite/ferrite structure as a problem. So, compared to austenitic stainless possible. Minimizing time at red heat tempera- steels, duplex stainless steels have the following tures (500 to 900 ¡C, 930 to 1650 ¡F) is the ob- distinguishing factors: jective. But, sufficient time must be spent above about 1000 ¡C (1830 ¡F) to promote the forma- ¥ The ferritic solidification mode of duplex tion of sufficient austenite. If the weld cannot be stainless steels provides very good hot annealed, increased nickel filler metal (e.g., cracking resistance. The rapid cooling of 2209 with 2205 base metal) should be used. welds produces welds and HAZ with more Thus, joint preparation must be done correctly ferrite than the parent metal by quenching in and not left to the welder to correct using time- the high-temperature ferrite. consuming remedial procedures. Chapter 17: Welding / 205

Duplex stainless steels, because of their mod- ¥ The stabilized group can lose toughness via erate thermal expansion and higher thermal excessive grain growth. conductivity, can tolerate relatively high heat ¥ The grades more highly alloyed with inputs since these factors determine the stress chromium and molybdenum can form α' and intensity that will be generated by thermal gra- σ, leading to embrittlement. dients. However, excessively low heat inputs can result in fusion zones that are predomi- The semiferritic alloys such as 430, 434, and nantly ferritic, with a resultant loss of toughness 436 are seldom welded and often called un- and corrosion resistance. At the other extreme, weldable. The reason is that the welds are in- heat inputs that are too high lead to the forma- variably partially martensitic and thus normally tion of embrittling intermetallic phases. This brittle. Only very specially controlled composi- issue concerns the HAZ, which must dwell in tions of 430 can be welded successfully, and σ-forming temperatures for some period of these are not generally available commercially. time. The key is to limit the time at those tem- While the technical remedy for this is simply peratures by not permitting interpass tempera- annealing, it is seldom economically viable. It is tures to exceed 150 ¡C (300 ¡F) because work- rare to see any welding more extensive than piece temperature has the greatest influence on spot welding of unexposed surfaces with these time at σ-forming temperatures. It is prudent to alloys. If for some reason they must be used and impose this limitation when qualifying the weld welded, then the techniques for welding procedure and then monitoring the production martensitic stainless steels should be employed. welding interpass temperature electronically to The stabilized ferritic stainless steels are ensure qualifying procedures are not more le- commonly welded. The levels of stabilizing ele- nient than are those of production. ments required to prevent austenite formation Postweld stress relief is not needed for duplex and sensitization are well known and are re- weldments and indeed could be harmful be- flected in the alloys’ chemistry specifications. cause of the danger of embrittlement. Full an- For 409, the required titanium level is Ti > 0.08 nealing can be done and can restore the original + 8(C + N), while the requirement for the higher phase balance and composition that gives the chromium 439 is 0.20 + 4(C + N). These are optimal toughness and corrosion resistance empirical relationships that take into account found in wrought material. that some titanium oxidizes before it can stabilize carbon and nitrogen. Niobium can replace Ferritic stainless steels can be split into two groups for purposes of welding: the older semi- some titanium. This is discussed in detail in ferritic group and the more prevalent stabilized Chapter 8 on ferritic stainless steels. Because of ferritic group. The first group, in which the low toughness these alloys have in large chromium is between 16 and 18% with carbon cross sections, these alloys are only rarely seen up to 0.08%, is exemplified by the alloy 430. with minimum section size of more than 3 mm These alloys form appreciable amounts of (0.11 in.) and normally have sections less than 2 austenite when heated above 800 ¡C (1470 ¡F). mm (0.08 in.). Thus, successful welding is sim- Unless they are cooled extremely slowly, more plified to making a sound, well-shielded weld slowly than can be done in welds, the austenite without producing excessive grain growth in the transforms to martensite, which is very brittle. HAZ. In practice, this can be achieved by limit- The stabilized grades commonly use titanium ing heat input to less than 6 kJ/cm. An empirical D or niobium to combine with the carbon and ni- relationship between grain diameter and heat E trogen, which otherwise would cause the forma- input (kJ/cm) has been reported (Ref 2). In tion of the high-temperature austenite, render- the fusion zone, the relationship is: ing the alloys ferritic at all temperatures. D = 206 × E Ð 585.6 (Eq 1) The salient metallurgical characteristics for welding of the two groups are: In the HAZ, it is:

¥ Both groups offer good thermal conductivity D = 29.6 × E Ð 50.6 for up to 6.6 kJ/cm (Eq 2) and low thermal expansion. ¥ Both groups require protection from oxida- and tion by shielding gases. The stabilized group D = 75 × E Ð 350 above 6.6 kJ/cm (Eq 3) should not be exposed to nitrogen. ¥ The semiferritic group will form martensite, The light gauges ensure sufficiently short which requires annealing to eliminate. times at high temperature that precipitation of 206 / Stainless Steels for Design Engineers

intermetallic phases should not be a concern, cleanliness and shielding. If mechanical re- even though they can form, especially in super- quirements permit, the use of austenitic (309L) ferritic alloys. weld filler metal should be considered. The soft The impact properties of ferritic stainless joint may deform to accommodate thermal steels are always a concern because their transi- strains and thus minimize weld cracking. tion temperature can become elevated to ambi- Precipitation-Hardening Stainless Steels. ent levels. It has been determined that there ex- Last, precipitation-hardening (PH) stainless ists an optimum level of titanium around 0.10%, steels, while very complex metallurgically, are which ensures this minimum transition temper- straightforward from a welding perspective. ature (Ref 3). Because it is difficult to have low Obviously, any heat treatment to achieve the enough carbon plus nitrogen to stabilize at this properties of which these alloys are capable titanium level, dual stabilization with titanium must be a final step. The considerations in weld- and niobium as well as not having excessive ing them are: heat input are the best way to ensure weld ¥ toughness. Shielding must be sufficient to prevent loss Especially in the superferritics, maintaining of oxidizable alloying elements such as titanium, aluminum, and, of course, chromium. the benefits of having the fairly precise balance ¥ of carbon plus nitrogen to the stabilizing ele- Filler metal must match the base metal if like properties are required. ments titanium and niobium requires that nei- ¥ ther carbon nor nitrogen come into contact with Postweld heat treatment solution annealing the weld pool. Likewise, oxygen must be rigor- must be adequate to homogenize weld solidification segregation. ously avoided because it will quickly deplete ¥ the essential titanium, which is even more read- Austenitic PH grades are fully austenitic and subject to hot short cracking. ily oxidized than chromium. Extraordinary sur- ¥ face cleaning at and near the weld will pay divi- The high aluminum or titanium contents of dends in final quality. many PH alloys cause their welds to be “slaggy,” and these slaggy welds have are ir- Martensitic stainless steels vary little in alloy content, ranging from 11 to 18% chromium with regular with objectionable recesses, small amounts of nickel and molybdenum. crevices, or prominences. Their carbon content ranges from 0.10 to over These alloys are easily welded and not prone 0.30%. Thus, the major challenge they present to cracking or developing embrittling phases. is avoiding the potential cracking, which is But, because these alloys are designed for ex- most likely to occur in the HAZ from stresses treme mechanical performance, it is essential to caused by the austenite-to-martensite transfor- preserve their correct chemistry by shielding mation on cooling. Since this transformation with a fully inert gas mixture. If mechanical cannot be avoided, the desired approach is to properties equal to that of the base metal are not start with a well-tempered or annealed material required in the weld, then austenitic filler, such and then preheat and maintain high interpass as 309L, can be used. temperatures. For low carbon levels, below Table 1 summarizes the major metallurgically 0.10%, preheat can be omitted, but between important parameters for the various types of 0.10 and 0.20% carbon, preheating to 250 ¡C stainless alloys. It is prudent to consult with the (480 ¡F) is advised and for higher carbon levels, manufacturer’s data sheets for specific recom- 300 ¡C (570 ¡F). The problem becomes more mendations on alloys that they produce as they severe with increasing carbon level because the are often privy to test data and user experience transformation takes place at lower tempera- that cannot be found elsewhere in the literature. tures in more brittle material. Even with preheating, distortion may be encountered. For all normal uses of martensitic stainless steels, a Material Selection and Performance final heat treatment is required to achieve the quenched and tempered properties for which Stainless alloys that are prone to precipitation these alloys are designed. of intermetallic phases require special preweld- Aside from the cracking consideration, ing consideration. Such alloys include duplex, martensitic welding considerations are similar superferritic, and superaustenitic alloys. Any to, but less stringent than, those of low-alloy amount of time for which these alloys have stabilized ferritic stainless steels with regard to been exposed to temperatures at which inter- Chapter 17: Welding / 207

Table 1 Welding parameters for various stainless steels

Heat input Shielding Interpass Postweld Alloy group Filler kJ/cm(max) gas Preheat max heat treat o o Austenitic . . . 20Ð40 Ar+2% O2, 150 C 150 C None or full anneal Ar/3% CO2/2% H2 He+7.5%Ar+2.5 CO2 301, 302, 304 308, 308L Same Same ...... 304L 308L Same Same ...... 309 309, 310 Same Same ...... 310 310 Same Same ...... 316L, 316Ti 316L, 317L Same Same ...... 321, 347 347, 308L Same Same ...... Superaustenitic 22, 675, 276 16 Argon/helium or 50 oC 100 oC None or full anneal argon + 3Ð5%

N2:no O2 PH grades Same as base alloy 20Ð40 Argon/helium no . . . Full solution anneal

Martensitic o o 410 410, 308, 309L 20Ð40 Ar+2% O2, 250 C 250 C min Slow cool He+7.5%Ar+2.5 CO2 o o 420 420, 308, 309L, 310 20Ð40 Ar+2% O2, 250 C 250 C min Anneal He+7.5%Ar+2.5 CO2 o 440 440, 308, 309L, 310 20Ð40 Ar+2% O2, 250 C ...... He+7.5%Ar+2.5 CO2 Supermartensiic Same as base metal 20Ð40 Argon/helium no . . . Full solution anneal

Ferritic

430 430, 309L 20Ð40 Ar+2% O2, no . . . Subcritical anneal He+7.5%Ar+2.5 CO2

434 309 Mo L 20Ð40 Ar+2% O2, no . . . Subcritical anneal He+7.5%Ar+2.5 CO2

409 410L, 308, 309L 6.0 Ar+2% O2, no n.a. none He+7.5%Ar+2.5 CO2

439 439L, 309L, 316L 6.0 Ar+2% O2, no n.a. none He+7.5%Ar+2.5 CO2 Superferritic 29-4C 6.0 Argon/helium no n.a. None or full anneal o 2003, 2101, 2304, 2209 5Ð25 Argon + 3% N2 no 150 C None or full anneal 19-D o 2205 2209 5Ð25 Argon + 3% N2 no 150 C None or full anneal

o 25 Cr duplex 25Cr-10Ni-4Mo-N 5Ð25 Argon + 3% N2 no 150 C None or full anneal o 2507 25Cr-10Ni-4Mo-N 5-25 Argon + 3% N2 no 150 C None or full anneal superduplex PH, precipitation hardenable metallic phases form without full subsequent This effect exploits the temperature-dependent homogenization anneal is time that the welder surface concentration of sulfur in the weld pool, cannot use to complete a satisfactory weld be- which causes a decreased surface tension to- fore precipitation occurs. Thus, accurate knowl- ward the hotter center of the pool, causing the edge of material history is vital. Likewise, vari- molten pool to flow toward the center on the ations within specification of nitrogen content surface and then flow downward, shooting the influence the time it takes intermetallic phases hottest metal to the bottom of the weld pool, as to form. Once a welding procedure is qualified shown in Fig. 4. This speeds welding and mini- for an alloy with given nitrogen content, use of mizes weld and HAZ width, which is a good lower nitrogen alloys would not be prudent en- thing. The effect on corrosion resistance is less gineering practice. desirable since the abundant MnS inclusions Austenitic stainless steels that are intended that result from the higher sulfur levels decrease for autogenous welding are often specified with pitting resistance. This decrease in corrosion re- elevated sulfur levels, on the order of 0.005 to sistance can only be eliminated by a long an- 0.015%. This is done to improve weld penetra- neal. Unfortunately, the pipe purchaser cannot tion through the so-called Marangoni effect. know if the pipe has had a sufficient anneal. 208 / Stainless Steels for Design Engineers

90 194 85 185 80 Unwelded 176 , ° C , ° F 3

75 Welded 167 3 70 158 65 149 60 140 55 131 50 122 Metal flow directions in a weld pool with (left) and Fig. 4 45 113 without (right) sulfur. Source: Adapted from Ref 4 40 104 35 95 30 86 In-line induction annealing is insufficient for 25 77 20 68 this purpose. Furnace anneals of about an hour 15 59 are required. For alloys like 304L and 316L, the 10 50 Critical in 6% FeCI pitting temperature user should always require material chemistry 5 41 Critical in 6% FeCI pitting temperature certifications and assume that any sulfur levels 0 32 −5 23 above 0.003% are going to result in decreased 123456 7 pitting resistance of 1 to 5 PREN (pitting resist- Molybdenum, wt% ance equivalent number), which means up to 10 ¡C (18 ¡F) decrease in critical pitting tempera- Fig. 5 The influence of molybdenum on critical pitting temperature. Source: Adapted from Ref 5 ture, roughly the difference between 304 and 316 in performance. This also applies to girth alter the composition of the welded joint, which welds done by the pipe user. in turn can alter corrosion and mechanical prop- Welds are essentially a casting in the midst of erties and compromise the entire structure. wrought material. In addition to inclusions de- Moisture, paint, dirt or grease, oil, and oxides creasing weld corrosion resistance as men- all can negate good material, good welding tioned, solidification segregation can also cause technique, and good procedural qualification. microscopic regions to be poorer in corrosion- Cutting fluids, especially sulfurized oils, are es- resisting alloying elements chromium, molyb- pecially detrimental and should be removed denum, and nitrogen. This effect is minimal for completely prior to welding. Preheating is never low-alloy material, but for highly alloyed strictly forbidden since it is required to elimi- austenitic grades, it is a major effect, as shown nate moisture. in Fig. 5. Eliminating this effect requires a thor- Joint design does not differ in principle from ough homogenization anneal. that of other steel weldments. There is, however The use of filler metal with higher corrosion an increased need for dimensional uniformity resistance does not totally offset the influence of for the alloys susceptible to intermetallic precip- welding on corrosion resistance because some itation since minimizing time at temperature is a of the base metal melts and is not altered in priority, and variations in joint geometry impede composition by the filler metal. This is called the swift completion of the weld. This is also the unmixed zone. It is essentially a zone with true for alloys that are susceptible to excessive properties equal to that which would occur in an grain growth, such as the stabilized ferritics, or autogenous weld, that is, the corrosion resist- to sensitization. ance is lower depending on total alloy level and Figure 6 shows some joint designs appropri- sulfur content. ate to stainless steels, including the more sensitive alloys. These, like all joint designs, aim to ensure full penetration without burn through. Welding Processes Gas tungsten arc welding (GTAW)/tungsten inert gas (TIG) is commonly used for the All stainless steels should be very clean prior automated production of stainless steel pipe and to welding. The chemistries of both base metals tube, as well as manual short runs. It is versatile and filler metals are carefully formulated to pro- and generally used when thicknesses are less duce the mechanical and corrosion properties than 6 mm (0.2 in.). It can produce very high- that these alloys have been designed to produce. quality welds. A constant-current power supply is Virtually any contaminant can either interfere preferred. It is best performed with the DCSP (di- with the welding procedure or detrimentally rect current straight polarity) electrode negative Chapter 17: Welding / 209

Thickness Gap d, Root K, Groove Process th, mm (in.) mm (in.) mm (in.) Bevel α(¡) GTAW 3Ð5 1Ð3 ...... t GMAW 3Ð5 1Ð3 ...... d SMAW 3Ð4 1Ð3 ......

a SMAW 4Ð15 1Ð3 1Ð2 55Ð65 GTAW 3Ð8 1Ð3 1Ð2 60Ð70 GMAW 5Ð12 1Ð3 1Ð2 60Ð70 d SAW 9Ð12 0 5 60

a SMAW >10 1.5Ð3 1Ð3 55Ð65

k GMAW >10 1.5Ð3 1Ð3 60Ð70

d SAW >10 0 3Ð5 80

a SMAW >25 1Ð3 1Ð3 10Ð15

k GMAW >25 1Ð3 1Ð3 10Ð15

d r = 6-8mm SAW >25 0 3Ð5 10Ð15

GTAW >3 0Ð2 ...... t GMAW >3 0Ð2 ...... d SMAW >3 0Ð2 ......

a SMAW 3Ð15 2Ð3 1Ð2 60Ð70 GTAW 25Ð8 2Ð3 1Ð2 60Ð70 GMAW 3Ð12 2Ð3 1Ð2 60Ð70 d SAW 4Ð12 2Ð3 1Ð2 70Ð80

a SMAW 12Ð50 1Ð2 2Ð3 10Ð15 GTAW >8 1Ð2 1Ð2 10Ð15 GMAW >12 1Ð2 2Ð3 10Ð15 d r = 6-8mm SAW >10 1Ð2 1Ð2 10Ð15 GMAW, gas metal arc welding; GTAW, gas tungsten arc welding; SAW, submerged arc welding; SMAW, shielded metal arc welding

Fig. 6 Joint designs. Courtesy Ugine S.A. technique. It is helpful to incorporate a high- The shielding gas must replicate the con- frequency circuit to aid in establishing the arc. trolled gas mixtures used to refine stainless steel Thoriated electrodes containing 1.7 to 2.2% tho- and establish the original composition. The ria are recommended because they have better weld pool exposes a great deal of surface area to emissive properties and provide better arc stabil- the atmosphere in a very turbulent manner. Gas ity at higher currents. If consumable electrodes flows, usually 12 to 18 L/min, must be adequate are used, the shielding gas precludes the need for to prevent air infiltration by aspiration or turbu- coatings. The weld metal alloys are not necessar- lence before arc contact, ideally until tempera- ily the same as the parent alloys but are chosen tures cool to below oxidation temperatures. based on their ability as weld metals to provide For manual GTAW using a filler wire, the the most acceptable corrosion and mechanical wire should be fed continuously into the weld properties. This sometimes means using pool. Intermittent wire addition can lead to cre- austenitic filler with a ferritic base or higher ation of zones of essentially autogenous weld, nickel content in an austenitic or duplex base to negating many of the benefits of filler metal ad- compensate for the solidification rate or inher- dition. Moving the tip of the wire in and out of ently lower corrosion resistance of the weld. the protection of the gas shield is especially 210 / Stainless Steels for Design Engineers

bad. The hot tip can carry oxides and nitrides for GMAW, but the consumable electrode, the into the weld, defeating the action of the shield FCW filler metal, has a flux core that supple- gas and impairing weld quality. ments the shielding gas. Because of the flux, the Gas metal arc welding (GMAW) is arc shielding requirements are reduced; gases can welding in which a consumable electrode pro- be argon/25% carbon dioxide for horizontal vides larger amounts of filler weld metal than welding with current and voltages from 150 to practical in GTAW. There are three GMAW 200 amp and 22 to 38 V, respectively. Vertical techniques: welds can use 100% carbon dioxide with am- perage of 60 to 110 amp and voltage of 20 to 24 ¥ Pulsed arc transfer V. Flow rates of gas are 20 to 25 L/min. It is ¥ Spray transfer possible to get high-carbon welds, which may ¥ Short-circuiting transfer not resist corrosion as well as desired, so as al- Pulsed arc transfer employs a power source ways, weld qualification, including corrosion that is switched rapidly to provide transfer of evaluation, is critical. weld metal droplets at regular intervals. Spray Oxyfuel gas welding (OFW), “torch” weld- transfer uses a high current to form a stream of ing, uses oxygen to accelerate fuel (typically fine drops from the end of the electrode. This is acetylene) combustion to produce temperatures done with high power, resulting in a large fluid that can melt steels. By controlling the fuel-air weld pool, and therefore limits the technique to mixture, the flame can be made nonoxidizing horizontal orientations and thick material. for low-alloy steels. However, these “neutral” Short-circuiting transfer uses arc contact with flames can simultaneously oxidize and carbur- the workpiece at low power to melt the elec- ize stainless steels. Thus, the OFW process is trode, after which the short circuit is broken, not suitable for use with stainless steels. and material transfer ceases. The technique cre- Laser welding has become a major produc- ates a minimal weld pool and is viable in many tion method when it can be automated, as for orientations. It is a low-heat process suitable for pipe and tube or high-production manufactured thin material but may cause lack of penetration items, such as air-bag canisters. Metallurgically, defects if used for thick-section welding. it resembles resistance welding in that both For all GMAW processes, excessive protru- have minimal HAZ and very high quenching sion of the wire should be avoided; otherwise, rates, both of which can have a pronounced ef- the full benefit of the inert gas shielding may be fect on some types of stainless steel. The effect lost. is to undercool the molten metal and suppress Submerged arc welding (SAW) employs a the transformation that would normally occur. consumable electrode immersed in a conductive So, an austenitic alloy that normally solidifies in flux that acts as a protective shield from the at- a ferritic mode before transforming to austenite mosphere. The arc is struck through the flux, freezes directly as austenite. The freezing is so and gravity deposits the molten metal to the rapid that the normal hot shortness of austenitic workpiece. The large weld pool has high heat solidification is avoided, so quality is not com- input and can deposit large amounts of metal promised. In fact, laser welds quench the mate- relatively quickly. Thus, SAW may be prefer- rial so rapidly that corrosion resistance is enable to multipass techniques for alloys such as hanced since inclusions cannot nucleate and duplex for which time at temperature is limited. grow. Duplex alloys, on the other hand, freeze It is restricted to horizontal orientations and rein their high-temperature ferrite structure be- quires postweld slag (flux) removal. cause the fast quench prevents the nucleation Shielded metal arc welding (SMAW) is and growth of austenite. Unless this ferrite is done manually with short lengths (“sticks”) of heated to permit austenite to form, lower-tough- coated electrodes. This method has great versa- ness welds will result. Ferritic, martensitic, and tility with some trade-off in cost and quality. PH alloys are not harmed by the rapid quench. This last aspect is arguable, but the lack of Resistance welding is readily done on most shielding gas may introduce oxygen to the weld types of stainless steel. Allowance must be made metal, which can be detrimental to toughness. for the lower thermal and electrical conductivity Flux cored wire (FCW) welding is a method of stainless steels compared to other common that is able to accommodate a large range of materials. Most resistance welds, including both thickness and orientations while providing high seam and spot welds, have deep, tight crevices deposition rates. The equipment is the same as adjacent to the welds. The possibility of crevice Chapter 17: Welding / 211

corrosion in these regions should be considered (Chapter 13) of this book. Brazes and solders when contemplating the use of spot welds in rarely match the corrosion resistance of stain- stainless materials. The possibility of entrap- less steels, and careful attention should be given ment of foreign material and the difficulty of re- to the potential for galvanic and other forms of moving it from such crevices should also be corrosion when considering the use of soldered considered, especially in equipment for food or brazed joints with stainless steels. handling, pharmaceutical production, etc. High-frequency induction welding of stainless steel is more difficult than for low- alloy steel because of the refractory nature of Welding Practices chromium oxide, which has a higher melting temperature than does the stainless base metal. Safety must always be considered when This is opposite from the situation in low-alloy welding. In addition to the normal hazards steels, for which the iron oxide melts at a lower (which are not discussed here) associated with temperature than does the iron base metal. The welding, welding of stainless steels presents a presence of this refractory oxide on the surfaces special hazard: hexavalent chromium. The fume to be joined makes it more difficult to obtain a created by welding stainless steel contains sig- defect-free weld. nificant concentrations of chromium trioxide Thermal cutting of stainless steels is rou- and other forms of hexavalent (Cr+6) chromium. tinely practiced, but the processes and parame- Hexavalent chromium is a carcinogen and regu- ters used are determined by the refractory na- lated by the Occupational Safety and Health ture of the chromium oxides that form on Administration (OSHA). Exposure to and in- stainless steels. The high temperatures attain- halation of stainless steel welding fumes must able with lasers or plasma arc torches provide be avoided. The product exposure limit for good cutting action, and these processes are fre- hexavalent chromium is 5 μg/m3 as of Decem- quently used. To expand the range of thick- ber 31, 2008. Refer to OSHA for further updates nesses that can be cut or to increase cutting on this limit. Use of fume extraction equipment speed, supplemental oxygen or nitrogen blast is generally the preferred method of minimizing jets may be used. Stainless steels may also be hexavalent chromium exposures. Positioning cut using oxyfuel equipment if supplemental and operation of the fume extraction device iron powder is used. Combustion of the iron in- must be done precisely to ensure effective fume creases the temperature, while the iron oxide removal while avoiding excess turbulence, helps flux the refractory chromium oxide. Ther- which can cause loss of effective inert gas mally cut edges of stainless steel usually require shielding of the weld pool. Thermal cutting of subsequent cleaning, typically by grinding or stainless steels also generates hexavalent milling. Chemical cleaning of all surfaces of cut chromium, and similar procedures are required pieces to remove heat tint, fume deposits, and to minimize exposure during such operations. other contaminants is advisable. Nondestructive Evaluation (NDE) is used Soldering and brazing are possible with all almost universally to ensure weld quality. All of stainless steels. Soldering is done below 450 ¡C the standard NDE techniques used with other (840 ¡F), while brazing is done above 450 ¡C materials are applicable to stainless steel weld- (840 ¡F). Solders are generally alloys of tin and ments. Allowance must be made for the differ- bismuth, lead, silver, or antimony or combina- ing physical properties of stainless steels, and tions of several of these. Brazes are normally ei- appropriate reference defect standards must be ther silver based or nickel based. The provided. However, one technique—magnetic chromium-rich oxide coating must be removed particle inspection—is problematic. The pres- by a suitable flux for bonding to occur. Fluxes ence of bands of persistent austenite in marten- are typically acid type with chlorides. Thus, sitic or PH stainless steels can lead to spurious after the soldering or brazing, the flux must be defect indications. For this reason, magnetic thoroughly removed to prevent subsequent pit- particle examination of stainless steel welds is ting corrosion. Brazing temperatures must be best avoided. chosen to avoid ranges at which unfavorable Recent developments in stainless steel have phases form. The best range can be determined been made with weldability as a major consid- from examining temperature ranges to be eration. Highly alloyed, low-carbon martensitic avoided in the thermal processing chapter alloys for line pipe have been developed with 212 / Stainless Steels for Design Engineers

the express purpose of use in the as-welded con- generation of weld fume (see the discussion of dition. The low carbon makes welds of this ma- safety). Other new welding processes, such as terial that are tough and do not require temper- multiple (GTA or GMA) torch welding, laser- ing, so girth welds in the ﬁeld are possible. assisted GMA or GTA welding, etc. promise Likewise, the lean duplex alloys have very greater productivity. delayed precipitation of intermetallic phases because of their higher nitrogen and lower chromium and molybdenum contents. This REFERENCES makes welding of these alloys much more foolproof than with the early duplex alloys, such as 1. D.J. Kotecki, Welding of Stainless Steels, S31803. The dual-stabilized ferritic alloys have Welding, Brazing, and Soldering, Vol 6, tougher welds than those stabilized with only ti- ASM Handbook, ASM International, 1993, tanium or niobium. p 677Ð707 New developments in welding also have an 2. B. Aziez and R. Feen, Sheet Metal Ind., 1, impact on stainless steels. The friction stir weld- 1983, p 28Ð34 ing (FSW) process offers the promise of reliable 3. S.D. Washko and J.F. Grubb, Proc. Int’l solid-state joining. By avoiding melting and Conf on Stainless Steel, 1991, Chiba, ISIJ resolidiﬁcation, issues associated with solute 4. Stainless Steels, Les Editions de Physiques, redistribution are eliminated. The relatively low 1992, p 786 temperatures involved essentially eliminate 5. A. Garner, Corrosion, 37, 1981, p 178 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 213-223 All rights reserved. DOI: 10.1361/ssde2008p213 www.asminternational.org

CHAPTER 18

Architecture and Construction

Summary can choose from a variety of stainless alloys with sufficient corrosion resistance to withstand STAINLESS STEEL IS OFTEN EMPLOYED any environment. as an architectural material; the material can it- The ability of stainless steel to resist corro- self be viewed as a metaphor for architecture, a sion resides in its chromium-rich superﬁcial discipline that must balance aesthetics, economy, passive layer. Stainless steel by deﬁnition must and structural integrity. Stainless steel’s unique contain slightly more than 10% Cr. The passive combination of beauty, strength, and economy layer forms spontaneously in air or water, and makes it a remarkably appropriate material for if it is removed, say by abrasion, it re-forms by uses as diverse as sculpture and concrete rein- itself. This is explained in greater technical forcing bar. However, stainless steels are complex; they come in many different grades (chemical analyses), and these grades have varying strengths, appearance, resistance to corrosion, availability, and costs. The success of a building project involves careful planning for the appropriate use of materials. This chapter deals with the technology of stainless steel as it pertains to its proper use in architecture, art, and construction.

Corrosion Resistance

Corrosion is the life-limiting factor for architectural metals. Steel, copper, aluminum, lead, bronze, and other alloys react with the environment and degrade over time, as does wood, stone, plastic, paint, and even glass. With stainless steel, it is possible to choose a material that can withstand attack from the environment indefinitely. One of the first major uses of stainless steel in architecture was in New York’s Chrysler building, completed in 1930. Despite the rather crude methods of early production, limited alloy options, and lack of application experience involved in its construction, the domed top of the Chrysler building still shines Fig. 1 The Chrysler building with its famous bright stainless undiminished by the harsh coastal and urban details. Copyright © iStockphoto.com/Steven Allen. climate (Fig. 1). With today’s technology, one Used with permission 214 / Stainless Steels for Design Engineers detail in this book in the chapters on corrosion, Balancing Corrosion Resistance, but the key aspects are that the chromium Processing Characteristics, and atoms on the surface of the metal react with oxygen in air and water to form with neighbor- Economy ing iron atoms into a tight, ionically noncon- The rule of thumb for grade selection is usu- ductive layer that prevents any further oxygen ally somewhat oversimplified to recommend the penetration. This layer is mere atoms thick and use of types 430 and 304 on interior applica- completely invisible. tions, type 304 on exteriors where salt is not a The strength of the layer in resisting corrosive problem, and 316 where road salts or seacoast attack is proportional to several key alloying effects make a more corrosion-resistant grade elements: chromium, molybdenum, and nitro- necessary. A leading architectural metals com- gen. They contribute according to the following pany (Ref 1) makes the following recommenda- formula: tions, which mirror these traditional views: PREN = %Cr + 3.3(%Mo) + 16(%N) (Eq 1) ¥ Type 304 should be used for most exterior applications. PREN stands for pitting resistance equivalent ¥ Type 316 should be used within ten miles of number. This number can be related to resist- saltwater bodies. However, if the building is ance to the mildest form of corrosive attack that subject to saltwater spray, a nobler grade of stainless steel undergoes, pitting corrosion. stainless steel, such as 2101 or 2003, should Pitting is a “weakest link” phenomenon in be specified. which corrosion begins in small, micron-size ¥ In close proximity to deicing salt use, even parts of the surface and then grows by virtue of on nearby roadways where vehicle traffic the more aggressive media that form within can create airborne particles, type 316 them because of the corrosion reaction’s prod- should be used. If periodic rinsing will not ucts. Pitting occurs in environments that contain occur on all exterior surfaces, these areas chlorides. Chloride ions compete with oxygen must be washed each spring. If dependable and disrupt the integrity of the protective pas- maintenance is not predicted, a nobler grade sive layer. As alloys become richer in the alloy- of stainless steel, such as 2101 or 2003, ing elements mentioned, their ability to maintain should be used. the passive layer can overcome the chlorides’ ¥ Specify types with low carbon, less than ability to destroy it. The key is to choose an 0.030%, if welding will be employed. alloy rich enough in chromium, molybdenum, ¥ Any grade, including type 430, may be used and nitrogen to withstand any environment the in interior applications. structure will experience. Pitting corrosion, and ¥ In the most severe environments—high heat another similar form of corrosion called crevice and humidity, low rainfall, and high salinity, corrosion, can be prevented by proper choice of such as are found in Middle Eastern coun- alloy, finish, and design. Crevice corrosion oc- tries—a grade with a PREN of 25 or above curs when recessed spaces are small enough to is recommended. act like a corrosion pit. The acidity within the crevice increases because of restricted diffusion These guidelines are based on the admittedly in and out of the crevice, just as happens within easy availability of these alloys and a lack of a pit. The buildup of iron and chloride ions concern for cost during times of peak raw mate- makes a very corrosive medium that disables rial prices. For projects where quick availability the passive film formation. A design that avoids is not more important than cost, 439 and 201 crevices is the best defense. should be considered as viable replacements for The decision criterion for material selection 304. Stainless steel 2003 (UNS S32003) or an with stainless steel should then be, Which grade equivalent lean-duplex grade can replace 316 at and finish will exclude the possibility of pitting a cost advantage during times of high alloy cost corrosion at the lowest cost? Then, proper design should be used to exclude the possibility of Table 1 Ranking of common stainless steels by crevice corrosion. Other factors involving pitting resistance equivalent number (PREN) strength and fabrication should also be consid- Alloy 430 439 201 304 316 2101 2003 2205 2507 ered. Table 1 ranks a number of stainless alloys PREN 15 17 17 19 24 26 28 35 38 by pitting resistance. Higher PREN values indicate greater pitting resistance. Chapter 18: Architecture and Construction / 215

as have been experienced on occasion, such as resistance in abraded stainless steel surfaces is during the period of 2004 to 2007. It should be not seen on finishes that are produced by pat- noted that the leaner alloys suggested (439, 201, terns imprinted by hard-rolling mill rolls that and lean-duplex alloys such as 2003) can be have been engraved with the desired pattern in somewhat more difficult to form. If panel de- reverse; for this reason alone, this method of signs call for 90¡ bends, this is not an issue. surface finishing is recommended. These archi- However, for applications requiring severe tecturally useful surface finishes are produced forming, as in the case of a double-lock seam on by the preferred rolled-on, or embossed, a standing seam roof, these grades can provide a method. In addition to their advantage in corro- challenge to the fabricator/installer. Further, sion resistance, they are extremely uniform these leaner grades can pose challenges with from batch to batch, unlike finishes produced by certain finishing methods, such as abrasive pol- abrasive-coated belts, which change in grit ishing and embossing. To avoid unwanted com- coarseness with use. plications related to grade selection, the speci- fier should consult a competent architectural metals supplier. This effort will ensure a viable Balancing Service Environment, specification is written that will balance cost Design Requirements, and with the necessary performance attributes to Maintenance Considerations make the part as well as resist corrosion once installed. An expert system has been developed that en- It is valuable to know, in times of high nickel ables designers and specifiers to analyze the prices, that both the low-nickel 201 and the no- trade-offs of climate, design requirements, and nickel 439 can be used in place of 304, while maintenance on grade selection (Ref 3). An- 2101 (UNS S32101) and 2003 (lean duplexes) swering the questions in Fig. 3 for a particular can replace 316. To obtain these grades usually application yields a score that can be used to involves working with a producing mill since identify an appropriate alloy according to the they are not typically stocked in service center scale shown Fig. 4. The Nickel Institute, for- inventories. However, any competent architec- merly the Nickel Development Institute, also tural metals supplier will not shy away from the offers excellent publications on topics related to use of specialty grades where appropriate. alloy selection for specific service environments and design requirements (Ref 4). Reviewing a map of the salinity of rainwater Surface Finish and Corrosion Resistance in the United States is instructive of the degree to which geography influences corrosion sever- Surface finish is usually an aesthetic choice, ity. The average atmospheric chloride levels but it has a significant influence on corrosion re- collected in rainwater are shown in Fig. 5 (Ref sistance and must be factored into grade selec- 5). The highest levels occur along the coastlines tion. Mill finishes such as 2B and 2D are incon- of the Atlantic and Pacific Oceans and the Gulf sistent because they are annealed and pickled to remove oxides. These unattractive surfaces, however, have correct corrosion resistance for their alloy content. Welding or abrading the surface degrades the corrosion resistance by a significant amount. An un-heat-treated weld has lower resistance to corrosion in proportion to the alloy content of the grade. Type 316 welds have the corrosion resistance of wrought 304. Abrasion has a similar effect. Type 316 with a No. 4 polish behaves like 304 with a 2B mill finish. Because welds are abraded, this compounds the effect. Very smooth abrasively polished finishes mitigate this reduced corrosion resistance, as Fig. 2 The decrease in corrosion resistance with increasing shown in Fig. 2. The effect of reduced corrosion surface roughness by abrasion. Source: Ref 2 216 / Stainless Steels for Design Engineers

Fig. 3 Stainless steel selection expert system. Source: International Molybdenum Association (Ref 3) Chapter 18: Architecture and Construction / 217

Fig. 4 Grades recommended based on the expert system. Source: International Molybdenum Association (Ref 3)

Fig. 5 Average chloride concentration (mg/L) in rainwater in the United States. Source: Ref 5 of Mexico. The maximum corrosion rate is re- erly maintained, will stay new looking indefi- lated to the maximum chloride in the atmos- nitely. phere. This will be related to the distance in- Surface finish aesthetics are arguably more land, the height above sea level, and the important architecturally than the influence sur- prevailing winds (Ref 6). face finishes have on corrosion. Numerous finishes have been developed to try to meet various objectives. Finishes vary in reflectivity, Aesthetic Considerations directionality, and subtlety. Figure 6 shows some of the finishes that go beyond the familiar A correctly chosen grade of stainless steel brushed look (Ref 7), while Fig. 7 shows special will have no degradation over time and, if prop- finishes created by one manufacturer. 218 / Stainless Steels for Design Engineers

Fig. 6 Various rolled-on stainless steel ﬁnishes. Source: Ref 7. Courtesy of Outokumpu

Fig. 7 Special finishes for 304/304L and 316/316L stainless steels available from one manufacturer. (a) Rolled-in low-glare finish (InvariMatte). (b) Rolled-in no. 4 finish (InvariBlend). (c) Rolled-in moderate-glare finish (InvariLux). Source: Contrarian Metal Resources (Ref 8)

The classic abrasively produced finishes are “Standard Test Method for Specular Gloss”). No. 3, 4, and 8. These are American Society for The standards for appearance do not exist Testing and Materials (ASTM) designations for within specifications, only the method of pro- abrasively produced finishes, which are tradi- ducing them. There is considerable difference tionally produced by abrading the surface with in appearance from sheet to sheet, coil to coil, different grit size abrasives. Finish No. 3 calls and manufacturer to manufacturer. The greatest for 80 to 100 grit abrasive; No. 4 calls for 120 to consistency of appearance comes from specify- 150 grit abrasive. Finish No. 8 is a mirror finish ing a brand of rolled-on finish from a given obtained by final polishing with 800 grit abra- manufacturer. Any of the traditional finishes sive. Finishes No. 3 and 4 are directional, with can be replicated by a rolled-on finish with grit lines typically 1 cm (0.4 in.) in length. Fin- greater uniformity, with the possible exception ish No. 3 has a surface roughness average (Ra) of bright annealed having a difficult time match- of 0.4 to 0.8 µm (15 to 30 µin.). Specular gloss ing the mirror quality of a No. 8 finish. This is at 85¡ is typically 40 to 60 (per ASTM D 523, crucial in architecture, where the discovery of Chapter 18: Architecture and Construction / 219

unacceptable visual nonuniformity on large Because stainless in sheet form is usually re- areas can be disastrous, especially when this ap- flective, small deviations from flatness can be pears late in the construction process, as is nor- very visible. A good standard for flatness that mally the case with exterior components. precludes visible distortion is five I units. Steel Reflectivity or gloss can be a major consider- producers have various means to produce this ation in the choice of a surface finish. Mirror level of flatness, the most extreme of which is finishes are often used for high impact, but actually stretching the steel sheet or coil until more diffusely reflecting surfaces are more all distortions are eliminated. Sometimes, common. Patterned surfaces provide consistent rather than aiming for high flatness a controlled reflectivity from a moderately reflective 40 to deviation from flatness is used, such as slightly 60 specular gloss at 85¡ to a dull matte of less concave panels or panels with a die-pressed than 20, the latter having been developed for design. Another option to ensure flatness is to airport roofing, such as at Reagan Airport in back light-gauge stainless steel with a stiff Washington, D.C., or the Pittsburgh Convention material. Center (Fig. 8). Deviations of sheets from squareness and Flatness is a special consideration for panels straightness (camber) are also objectionable be- where lack of flatness, such as by “oil-canning,” cause such deviations can cause gaps between can cause a very shoddy appearance. Flatness is panels. The degree to which this is objection- measured in I units. able is a function of design, and tolerances can be held tightly at a cost. Width tolerance is π × Flatness (Iunits) = 2 ( H / 2L) 105 (Eq 2) normally +1/16 in./Ð0 in 48 in., while length is where H is the height of the deviation from flat- held to +1/8 in./Ð0 in 10 ft or less. Maximum ness, and L is the distance between peaks of de- camber is 3/32 in. in 8 ft. Closer tolerances can viations, assuming a sinusoidal wave. be negotiated.

Fig. 8 The Pittsburgh Convention Center with low-gloss ﬁnish stainless steel roof 220 / Stainless Steels for Design Engineers

Maintenance and Repair polymer coatings are applied to stainless steel at some producing mills to permanently provide a Maintenance is a significant cost of any struc- film to which additional fingerprint oil cannot ture. One of the great values of stainless steel is add a noticeable discoloration. its low cost of ownership. Stainless steel can be The greatest ally of stainless on building exte- abused, however, and it does benefit from riors is the cleansing action of rain. Rain does proper maintenance. The main objective of the not completely clean the surface, but it does di- maintenance of stainless is keeping it clean. lute any harmful contaminants and forestall cor- There are two reasons for this. The obvious first rosion from accumulated chlorides. Without the reason is that whatever is soiling the surface is benefit of cleansing by rain, stainless exteriors probably not attractive. The second reason is should be washed during routine window wash- that it may harm the surface by allowing corro- ing operations. sion agents to concentrate. Table 2 provides rec- Given the importance of cleaning, the follow- ommended practices for removing various sub- ing design considerations are recommended: stances from stainless steel surfaces (Ref 9). One of the most common complaints about ¥ Designs that can collect dirt, such as hori- maintenance of stainless is the work involved in zontal surfaces and recesses, should be removing fingerprints. The oil from fingerprints avoided. makes an easily visible interference film on the ¥ Designs that create uneven flow or drainage reflective stainless surface. The typical remedy patterns producing uncleansed areas should is to clean stainless steel with a solution con- be avoided. taining light oil and a detergent. If the oil from a ¥ Sheltered areas and areas subject to splatter, hand contacts the uniformly thin film of clean- especially roadside spatter, should be de- ing oil, no visible mark is left. Alternatively, signed so that they are easily cleanable.

Table 2 Cleaning methods for uncoated stainless steel

Requirement Suggested method(a) Comments Routine cleaning of light soiling Soap, detergent, or dilute (1%) ammonia Satisfactory on most surfaces solution in warm clean water. Apply with a clean sponge, soft cloth, or soft-fiber brush, then rinse in clean water and dry. Fingerprints Detergent and warm water; alternatively, Proprietary spray-applied polishes available to hydrocarbon solvent clean and minimize re-marking Oil and grease marks Hydrocarbon solvent Alkaline formulations are also available with surfactant additions. Stubborn spots, stains, and light Mild, nonscratching creams and polishes. Avoid cleaning pastes with abrasive additions. discoloration; water marking; Apply with soft cloth or soft sponge; rinse Cream cleaners are available with soft light rust staining off residues with clean water and dry. calcium carbonate additions. Avoid chloride- containing solutions. Localized rust stains caused by carbon Proprietary gels or 10% phosphoric acid Small areas may be treated with a rubbing steel contamination solution (followed by ammonia and water block comprising fine abrasive in a hard rub- rinses) or oxalic acid solution (followed by ber or plastic filler. Carbon steel wool and water rinses) pads that have previously been used on carbon steel should not be used. A test should be carried out to ensure that the original surface finish is not damaged. Adherent hard water scales and mortar/ 10Ð15 vol% solution of phosphoric acid. Proprietary formulations available with cement splashes Use warm, neutralize with diluted ammonia surfactant additions. Avoid the use of solution, rinse with clean water and dry hydrochloric acid-based mortar removers. Heat tinting or heavy discoloration Nonscratching cream or polish. Apply Suitable for most finishes. with soft cloth or soft sponge. Rinse off residues with clear water and dry. Nylon-type pad Use on brushed and polished finishes along the grain. Badly neglected surfaces with hardened A fine abrasive paste as used for car body May brighten dull finishes. To avoid a patchy accumulated grime deposits refinishing. Rinse clean to remove all paste appearance, the whole surface may need to material and dry. be treated. Paint, graffiti Proprietary solutions or solvent paint stripper Apply as directed by manufacturer depending on paint type. Use soft, nylon or bristle brush on pretreated material. (a) Cleaning agents should be approved for use under the relevant national environmental regulations and should be prepared and used in accordance with the company’s or supplier’s health and safety instructions. Source: Adapted from Ref 9 Chapter 18: Architecture and Construction / 221

¥ Contamination by rust from carbon steel is ¥ Separation of tools and work areas between corrosive and must be removed. those used for stainless steel and carbon ¥ Designs should facilitate easy access for steel is prudent. Contamination of stainless cleaning. surfaces with carbon steel from welding, grinding, and cutting can stain the surface of While any structure would benefit from these stainless steel and result in corrosion. This guidelines, they are especially valuable in maxi- can be remedied by passivation, but it is mizing the benefits of stainless steel. much better to avoid it in the first place. Repair of more severe damage done to sur- ¥ Welding is better done in the shop than in faces, such as scratching, is difficult. If a deco- the field. Correct filler metals must be used, rative surface pattern has been damaged, the and proper weld finishing is essential. The challenge is in trying to replicate it in the field. ability of contactors to produce sound, at- Very few surface finishes are wholly repairable tractive welds is an indicator of their overall in the sense that they can be repaired in a spot competence with stainless steel. so that the repair is invisible. The reason for ¥ Fasteners used with stainless steel should that is mainly that abrasive finishes are applied also be stainless steel. Galvanized steel, car- by rotating belts, and it is virtually impossible bon steel, and aluminum will corrode more to match the pressure, grit size, and arc of con- readily than the stainless, and this corrosion tact that created the original surface. It can be is aggravated by galvanic contact. The re- more nearly done to a surface with a rolled-on sulting corrosion products are also harmful finish, which has a very consistent grit length as well as unsightly. Fasteners should not be and depth. The exception to this rule are abra- permitted to cause distortion of flat panels. sively applied or rolled-on long-grain finishes. These have very long grit lines, so grit length is easily duplicated with a belt sander, the Additional Service Considerations usual tool available for field repairs. Even welds can be removed and reblended to be in- Fire resistance is an important consideration distinguishable from the surrounding original in buildings. Stainless steel is the only com- surface. The ability to be repaired should be a mon building material that remains strong and top criterion in the choice of a surface finish tough at temperatures encountered in fires. Or- whose appearance is critical and that may be dinary carbon steel undergoes a phase change subject to damage. at about 760 ¡C (1400 ¡F). This change in atomic structure results in a sudden shrinkage of more than 1 linear percent. This can literally pull a building apart. When this occurs to a Fabrication Considerations structure already weakened by heat, as carbon steel is, catastrophic failure ensues. Austenitic Fabrication and joining of stainless steel em- stainless steel keeps the same atomic structure ploy the same techniques as for carbon steel and and remains much stronger than carbon steel at other metals. The specifics of cutting, forming, elevated temperatures. Thus, austenitic stain- joining, soldering, and welding are described in less steel has great value as a material for the processing section of this book and are not structures that must retain structural integrity repeated here. The main distinction in the use of in a fire. stainless steel in this regard is that its higher Tests have been conducted on glass-rein- strength and corrosion resistance permit the use forced plastic, aluminum, galvanized steel, and of lighter gauges. This in turn permits designs in austenitic stainless steel ladders under load and hollow or rolled-formed sections, which have exposed to flame temperatures of more than higher stiffness and low weight and potentially 1000 ¡C (1830 ¡F). The plastic and aluminum lower overall cost than using less-expensive ladders failed in less than a minute. The galva- metals. A second aspect of higher strength and nized carbon steel lasted 5 min, while the stain- lighter gauge is greater spring back in forming less steel remained intact (Ref 10). If the need operations, such as press braking. for fire resistance is serious, stainless steel Certain processing principles related to archi- becomes the material of choice. It is used on tectural and building applications of stainless offshore oil platforms for stairways, ladders, steels should be emphasized: walkways, handrails, gratings, floor systems, 222 / Stainless Steels for Design Engineers

Fig. 9 Graphic depicting low release of metal ions from two grades of stainless steel (304 and 316) to the environment, based on a 4-yr multidisciplinary research project involving both ﬁeld research and laboratory studies. Source: Ref 11

ﬁrewalls, blast walls, living modules, and so In bridges, parking garages, and other concrete forth. structures, saltwater can penetrate the cement Ecological considerations are never trivial over time. If the internal rebar corrodes, the when considering a construction material. expansion of the corrosion products spalls Many materials used in buildings degrade envi- the concrete, leading to the failure of the ronmentally, usually by corrosion, and enter the structure. general environment. Asbestos, lead-based This can be delayed by treating the concrete paints, lead coatings, and others are once-ac- to repel the incursion of water or by coating cepted materials whose long-term effects have the carbon steel rebar with epoxy. These are been dangerous and costly. less than 100% effective. A more certain ap- Stainless steel, because it does not corrode proach is to use stainless steel rebar. The stain- when properly used, does not enter the environ- less steel for this duty need not resist pitting ment. While this seems obvious, it has been the corrosion, which affects a tiny percentage of object of interdisciplinary studies that have the steel volume. Therefore, an inexpensive, demonstrated its innocuousness even under low-nickel grade, such as 409, 430, or 201, can conditions of heavy acid rain on freshly abraded be used. Most of the work to date has been surfaces (Fig. 9) (Ref 11). with more expensive grades, such as 316 and Stainless is a material that will never come 2205. The lean duplexes are ideal for this ap- back to haunt an architect years later. Its intrin- plication because of their high strength, resist- sic raw material content value ensures that even ance to corrosion and SCC, and moderate cost. with the end of the life of a structure, the stain- The use of even these alloys reduces the long- less in the structure will be recycled. term cost of these structures, so the future adoption of less-expensive stainless steels holds great promise. The more enlightened Concrete Reinforcing Bar transportation departments in the United King- dom; Ontario, Canada; and Michigan, New Concrete reinforcing bar is one of the least Jersey, and Oregon in the United States have glamorous uses of stainless steel in structures. led this development. Chapter 18: Architecture and Construction / 223

REFERENCES dustries, Vol 13C, ASM Handbook, ASM International, 2006, p 42Ð60 1. Stainless Steel Selection Criteria, Contrar- 7. Guide to Stainless Steel Finishes, 3rd ed., ian Metal Resources, www.metalresources. Euro Inox, 2005, www.euro-inox.org, ac- net, accessed June 2008. cessed June 2008 2. Bulletin of the National Dairy Federation 8. Contrarian Metal Resources, www.metalre- 189, 1985, p 3Ð12 sources.net, accessed June 2008 3. C. Houska, “Which Stainless Steel Should 9. Care and Maintenance of Stainless Steels, Be Speciﬁed for Exterior Applications?” In- Leda-Vannaclip, www.l-v.com.au, accessed ternational Molybdenum Association, www. June 2008 imoa.info, accessed June 2008 10. The Nickel Institute, www.nickelinstitute. 4. The Nickel Institute, www.nickelinstitute. org, accessed June 2008 org, accessed June 2008 11. D. Berggren et al., Release of Chromium, 5. H. Guttman, Atmospheric and Weathering Nickel and Iron from Stainless Steel Ex- Factors in Corrosion Testing, Atmospheric posed Under Atmospheric Conditions and Corrosion, W.H. Ailor, Ed., John Wiley and the Environmental Interaction of These Sons, 1982, p 51 Metals, European Confederation of Iron and 6. R.B. Griffin, Corrosion in Marine Atmos- Steel Industries, Oct 2004, www.eurofer. pheres, Corrosion: Environments and In- org, accessed June 2008 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 225-232 All rights reserved. DOI: 10.1361/ssde2008p225 www.asminternational.org

CHAPTER 19

Automotive and Transportation Applications

Summary the auto manufacturers have become the largest users of stainless steel. THE ADEQUATE DURABILITY and life Exhaust systems constitute the largest use of span of cars, trucks, or any transport system stainless steel in the automotive market, but requires freedom from corrosion. This has re- there are other important applications that can- quired subsystems, such as those for exhaust not be ignored: valves and gaskets, hose and fuel, to resist more corrosive environments clamps, seat belt and air bag components, tub- for longer periods of time. The main result has ing, hardware, and ﬁlters. And, there will be been a strong growth in the use of the leaner new applications that respond to new socioeco- ferritic stainless steels in many components. nomic needs, such as for greater crash worthi- As more exotic propulsion systems and fuel, ness, lighter weight, or resistance to the corro- such as fuel cells and ethanol, emerge, stain- sion of new fuels. But, since exhaust systems less steels may be required to endure the corro- currently predominate, they are covered ﬁrst. sive environments.

Exhaust Systems Introduction Laws enacted in the United States in the The use of stainless steel in automobiles 1970s mandated automotive emission stan- used to be mainly a story of decorative appli- dards that could be met only with catalytic cations: wheel covers and trim with a minor converters. The only practical materials that amount used for valves and hose clamps. How- could withstand the temperatures of the hot ever, as automobiles became more sophisti- end of an exhaust system using a catalytic con- cated technically and as durability and envi- verter were stainless steels. From Allegheny- ronmental demands grew, the role of stainless Ludlum’s MF-1 evolved a succession of fer- became increasingly functional and less orna- ritic alloys that grew in sophistication to meet mental. Stainless alloys in common automo- the increasing needs of corrosion resistance, tive use now are generally highly engineered oxidation resistance, creep, thermal fatigue re- for their speciﬁc application and represent sistance, and formability. Soon, entire exhaust some of the most highly evolved applications systems were made of stainless; often, they engineering in any use of stainless. An exami- would last the life of the vehicle, rendering nation of the preferred practices in materials obsolete an entire muffler replacement indus- selection in automotive systems is an excellent try. The compelling need of the automotive in- example of the rule of using the simplest and dustry for economy drove the widespread lowest alloy content grade that can do the job. adoption of argon oxygen decarburization Because automotive and steel-producing engi- (AOD), the continuous caster, and other high- neers have collaborated so well, both parties volume methods of the carbon steel industry. have beneﬁted greatly, as have consumers, and So, even while the traditional automotive uses 226 / Stainless Steels for Design Engineers

of stainless steels—wheel covers and other or- The metallurgy of ferritic stainless is dis- namental trim—faded to nearly nothing, the cussed in depth in Chapter 8. Rather than reex- use of stainless in automobiles grew to about plain these concepts here, we revisit only the 30 kg (65 lb) per vehicle by the turn of the cen- main points that are relevant to alloy selection tury. The stainless steel industry was changed for exhaust systems. from a boutique industry to a mass production ¥ industry by its embracing the needs of the au- As chromium level increases, so does re- tomotive market. sistance to oxidation and corrosion, but An exhaust system normally consists of a yield strength also increases, and ductility manifold to collect exhaust gases, a catalytic decreases. ¥ Alloying with silicon, aluminum, and converter to reduce NOx and CO emissions, and a muffler; each of these are connected by pip- molybdenum also increases oxidation resisting. Each component of the system has different ance, but these elements have the same requirements for formability, resistance to oxi- detrimental effect on ductility while increasing hot strength. dation, resistance to external corrosion, resist- ¥ ance to internal corrosion, and mechanical Niobium above that needed for stabilization properties. At the outset, it should be noted that is a powerful solid solution hardener and is effective at high temperatures. ferritic stainless steels, as opposed to austenitic, ¥ are optimal for oxidation resistance, especially Ferritic stainless steels have very anisotropic cyclic oxidation. It is not a difference in the forming properties. They resemble high- oxide scale. formability carbon steels in that they tend The reason is that the thermal expansion of not to thin when stretched, which greatly as- ferritic stainless more closely matches that of sists in formability. ¥ Ferritic alloys can form a hard, brittle phase the oxide scale than does that of austenite. This α prevents the fracturing and spalling of the scale. called ’ in a process commonly called 885 The intact scale of ferritic stainless is thus pro- ¡F (or 475 ¡C) embrittlement. This is only a tective up to the temperature at which oxygen factor in alloys with chromium of 18% or diffusion through the scale becomes great more, especially those containing molybde- enough so that “breakaway” oxidation occurs. num and aluminum. Cold work accelerates the formation of this phase. At the breakaway temperature level, scale ¥ σ growth is no longer parabolic with time but be- The phase does not readily occur in alloys comes linear and therefore no longer protective. containing less than 20% Cr, so it is not a The temperature of this breakaway increases as consideration for exhaust systems unless silicon or molybdenum are also elevated. chromium content increases. We will see that ¥ other alloying elements can also improve this Coating ferritic steel with aluminum is effec- performance. tive in preventing oxidation and corrosion. Not only are the steels in exhaust systems al- All these factors come into play in the design most exclusively ferritic, they are also stabi- of exhaust systems. Because the alloys have lized by titanium or niobium. This prevents sen- evolved so well to ﬁt the individual require- sitization and makes all the chromium content ments for each component, we discuss them seg- useful as alloy. Titanium stabilization greatly ment by segment through the exhaust system. improves corrosion resistance by removing not The exhaust manifold collects the hot, com- just the carbon and nitrogen, but also the oxy- busted gases from the engine and delivers gen and sulfur from solution. This sharply im- them to the front pipe. The exhaust manifold proves resistance to pitting corrosion. Niobium must possess good high-temperature strength is used to costabilize and ﬁght creep. These al- and resistance to thermal fatigue. It must also loys are therefore essentially interstitial free and be able to resist oxidation at the exhaust tem- have excellent formability, which the designs of perature, which can reach 950 (C (1740 (F). exhaust system components require. Exhaust manifolds had previously been heavy Their formability is further enhanced, when castings but are generally now formed from necessary, by low additions of tramp substitu- stamped sheet stainless steel or formed from tional alloying elements such as manganese, welded tubing that may have a double-wall nickel, and copper. Special thermomechanical structure to insulate the gases from heat loss, processing is also used to optimize texture and which could preclude successful catalytic con- grain size. version downstream. Chapter 19: Automotive and Transportation Applications / 227

As the highest-temperature component of the loys for spalling resistance. All approaches in- exhaust system, the exhaust manifold must pos- volve raising chromium content but use differ- sess the greatest resistance to high-temperature ent techniques to enhance the effect of oxidation damage. Risk of such damage is due chromium. Table 1 lists a number of the grades to the intermittent use of vehicles, which causes of stainless steel commonly used in exhaust sys- cyclic oxidation and the ensuing spalling of the tems and where they are used. The alloys are oxide scale. There are numerous alloying ap- listed in order of increasing severity of the re- proaches for optimizing the ferritic stainless al- quirements for each major system component.

Table 1 Alloys normally used for the major elements of automotive exhaust systems

Service temperature Alloys currently used, common name Component ¼C ¼F Requirements (related designation) Exhaust manifold 750Ð950 1380Ð1740 High-temperature ¥ T439HP (UNS S43035, dual-stabilized 439) strength, thermal ¥ 18CrCb (DIN 1.4509, 18CrCb) fatigue strength, ¥ 441 (DIN 1.4509) oxidation resistance, ¥ 304/304L/304H (UNS S30400, S30403, formability S30409) ¥ 321 (UNS S32100) ¥ 309S (UNS S30908) ¥ 310S (UNS S31008) ¥ 332Mo (S35125) ¥ 600 (N06600) ¥ 601 (N06601) ¥ 625 (N06625) Front pipe 600Ð800 1110Ð1470 High-temperature ¥ 409 ALMZ (aluminized 409) strength, thermal ¥ T439HP (UNS S43035, dual-stabilized 439) fatigue strength, ¥ 18CrCb (DIN 1.4509, 18CrCb) oxidation resistance, ¥ 441 (DIN 1.4509) formability ¥ 436S (type 436S) ¥ 444 (UNS S44400, T441) ¥ 433 (T443) Flexible pipe 600Ð800 1110Ð1470 High-temperature ¥ 304/304L (UNS S30400, S30403) strength, thermal ¥ T321 (S32100) fatigue strength, ¥ 316/316L (S31600/S31603) oxidation resistance, ¥ 316Ti (S31635) formability, salt ¥ 332Mo (S35125) attack resistance ¥ 625 (N06625) Catalytic converter shell 600Ð800 1110Ð1470 High-temperature ¥ 409HP (UNS S40930, dual-stabilized 409) strength, salt ¥ T439HP (UNS S43035, dual-stabilized 439) attack resistance, ¥ 441 (DIN 1.4509) formability ¥ 18CrCb (DIN 1.4509, 18CrCb) ¥ 444 (UNS S44400, T441) ¥ 433 (T443) Catalytic converter 1000Ð1200 1830Ð2190 Oxidation ¥ ALFA-IV (FeCrAl) substrate resistance, thermal shock resistance Center pipe 400Ð600 750Ð1110 Salt damage ¥ 409HP (UNS S40930, dual-stabilized T409) resistance ¥ 409 ALMZ (aluminized 409) ¥ T439HP (UNS S43035, dual-stabilized T439) ¥ 441 (DIN 1.4509) ¥ 18CrCb (DIN 1.4509, 18CrCb) ¥ 444 (UNS S44400, T441) ¥ 433 (T443) Muffler 100Ð400 210Ð750 Corrosion ¥ 409HP (UNS S40930, dual-stabilized T409) resistance, ¥ 409 ALMZ (aluminized 409) from inner and ¥ T439HP (UNS S43035, dual-stabilized T439) outer surface ¥ 436S (T436S) ¥ 441 (DIN 1.4509) ¥ 18CrCb (DIN 1.4509, 18CrCb) ¥ Type 304/304L (UNS S30400, S30403) Tailpipe 100Ð400 210Ð750 Corrosion ¥ 409HP (UNS S40930, dual-stabilized T409) resistance, ¥ 409 ALMZ (aluminized 409) from inner and outer surface Source: Adapted from Allegheny Technologies Inc. 228 / Stainless Steels for Design Engineers

The best choice for a given design is not obvi- some regions, so it must resist hot salt corro- ous. We attempt to simplify the choices. sion. This may force the use of 316L versus the Thus, the basic alternatives for exhaust sys- normal choice of 304L. tem alloys are: The catalytic converter is the next component of the exhaust system. It exposes the ex- ¥ Straight chromium alloying at 11 to 12% haust gases to noble metal catalysts, which with stabilization by titanium or niobium, complete the combustion of the gases to form the basic type 409 (UNS S40920) less-noxious compounds. This is an exothermic ¥ Straight chromium alloying at 17 to 18% reaction at temperatures equal to those in the with stabilization by titanium or niobium, exhaust manifold. Thus, the housing, while not the basic type 439 (UNS S43036) requiring great hot strength, must resist oxida- Either of these basic alloys can enjoy en- tion. The catalyst itself is supported by a ce- hanced oxidation resistance by additional alloy- ramic and ferritic stainless steel carrier that ing with molybdenum, aluminum, or silicon. In must resist thermal shock and possess low heat addition, they can be coated with hot-dipped capacity for rapid heating. Exotic alloys of aluminum-silicon alloy to increase oxidation re- 20% Cr with 5% Al are used for the carriers. sistance. The housing is generally made of a 17% Cr fer- Chromium or molybdenum alloy additions ritic stainless. The converter is usually directly increase corrosion resistance, whereas alu- beneath the passenger compartment, so a heat minum or silicon additions do not improve that shield of type 409 is used to separate it from trait. Aluminum coating is a powerful corrosion the floor. fighter, and it has the aesthetic benefit of not The center pipe conveys the converted gases showing red rust. to the muffler. The cooling exhaust gases no Use of molybdenum or niobium enhances longer present a major oxidation threat, but the high-temperature strength. Alloys with these condensing water vapor creates an internal cor- additions are thus useful for manifolds with a rosion risk, and road salt presents an external design that constrains expansion and contrac- one. However, a simple grade such as 409 tion, making thermal fatigue a problem. should generally provide sufficient resistance to All alloying additions detract from formabil- this environment. ity and toughness, as well adding to basic mate- The muffler, next in line, presents only a cor- rial costs. Thus, the objective must be to use rosion issue. The muffler must withstand corro- only those alloying elements that are indispen- sion from the outside, the worst of which comes sable to performance. from road salt or coastal salt sources. Internal The front pipe connects the exhaust manifold corrosion is also a major consideration because to the flexible joint and experiences nearly the condensing exhaust gases create a hostile, same temperatures as the exhaust manifold, but acidic environment. After startup, the heating of not the same risk of thermal fatigue. To reduce the muffler to temperatures above 100 ¡C (212 exhaust noise, a double-wall pipe is sometimes ¡C) evaporates these condensates, and internal used for this component. corrosion ebbs. On short runs, this may not The flexible joint is the one segment of the occur. This represents a worst case for internal exhaust system for which austenitic stainless corrosion. The dual internal and external corro- steels are preferred. The function of the flexi- sive attacks require the use of aluminized stain- ble joint is to prevent vibration from the en- less for best performance. gine from being transmitted to the rest of the The tailpipe is exposed to view in most vehi- exhaust system. It consists of a double-wall cles, and its appearance is therefore important. pipe in a bellows configuration with an outer For this reason, an austenitic such as 304 can be covering of braided stainless steel wire. It must used, as can chromium plating or aluminizing. have very good high-temperature fatigue The object here is to avoid visible corrosion. strength to withstand the cyclic stress of the vi- Truck exhaust systems are beginning to re- bration it absorbs. The material used must quire similar technical sophistication as their have exceptional formability to be formed into emissions come under increased regulation. a bellows. The greater hot strength and forma- However, they do not present any challenges bility of austenitic steels thus prevails. The not already confronted and solved for passenger flexible joint is also exposed to road salt in vehicles. Chapter 19: Automotive and Transportation Applications / 229

Structural Components but extraordinary resistance to localized thinning, necking, and therefore fracture. When The driving forces of durability, safety, and crash worthiness becomes a prime considera- weight reduction have spawned other, more var- tion, then this characteristic makes 301, or its ied applications for stainless in automotive en- low-nickel counterpart 201, an ideal material for gineering. Across the board, the main distin- structural, energy-absorbing components since guishing trait of stainless that qualiﬁes it as the austenitic stainless can be rivaled for such appli- optimal material is its corrosion resistance, but cations only by heat-treated alloy steel, titanium, this characteristic would be insufficient in many or aircraft aluminum alloys, all of which are cases without considering mechanical proper- more expensive, less durable, or less formable. ties. Indeed, even if stainless steel were not cor- Tables 2 and 3 show the properties of speciﬁc rosion resistant, its superior strength and tough- variations on basic 301 developed by Out- ness would qualify it for many automotive okumpu and how they stack up against the applications. most competitive carbon steels, dual-phase Austenitic stainless steels are the toughest and steels, and TRIP steels (Ref 1). The value of a stiffest practical materials available to the auto- material as an energy-absorbing structure (i.e., motive engineers. Common 301 can be cold one that enhances crash worthiness) is meas- worked to yield strengths anywhere from its an- ured by the energy it can absorb per unit of nealed level of about 300 MPa (44 ksi) up to mass. The kinetic energy of a collision that a 2000 MPa (290 ksi). In this higher-strength con- structure can absorb in deformation is propor- dition, it has become the standard material for tional to its strength multiplied by the amount it seat belt anchors and hose clamps. Type 301 in can deform before fracturing. The superiority the annealed condition is actually the original of metastable stainless steels (i.e., 201 and 301, transformation induced plasticity (TRIP) steel as those that most easily transform to martensite it can be tailored to have a controlled level of during deformation) is shown in Fig. 1. Even austenite stability. This allows it to transform at with its lower density, aluminum falls far short a known rate to martensite during deformation, of austenitic stainless in energy absorption per giving not only a very high work-hardening rate unit weight.

Table 2 Comparison of tensile properties of carbon steels and stainless steels for automobile structural components

True stress at 0.2% proof Ultimate tensile ultimate tensile Uniform Type Thickness, mm strength, MPa strength, MPa strength, MPa elongation, % Total elongation, % Carbon steels TRIP 700 1.58 473 703 818 16.4 17 DP 750 1.48 513 811 920 13.4 18.8 DP 800 1.44 573 896 976 8.9 9.9 Austenitic stainless steels HyTens X 1.16 306 937 1429 52.5 59.3 HT 1000 1.55 639 1068 1377 28.9 38.6 Source: Ref 1

Table 3 Comparison of resilience and toughness of carbon steels and stainless steels for automobile structural components

Type Resilience, J/m3 Toughness, j/m3 Carbon steels TRIP 700 0.996 105 DP 750 1.131 101 DP 800 1.32 74 Austenitic stainless steels HyTens X 0.536 364 HyTens 1000 1.726 269 Source: Ref 1 230 / Stainless Steels for Design Engineers

Fig. 1 True stress-true strain curves for 301 variants (HyTens X and HyTens 1000) versus two duplex carbon steels (DP750 and DP800) and a transformation induced plasticity (TRIP) steel (TRIP700). Source: Ref 1

These exceptional strength-to-weight and en- on individual body components is thus essen- ergy absorption-to-weight characteristics permit tially nil. However, if the entire system is stain- automotive engineers to reduce weight and in- less and the investment is avoided, then the ini- crease crash worthiness while designing vehi- tial cost of a stainless body actually can be lower cles with greater life span—because corrosion than one in coated carbon steel. This is the expe- resistance “comes along for the ride,” as it were. rience of Italian bus manufacturers, who began Some components in which these virtues are in the 1980s using 304 stainless steel in buses. most readily exploited are bumper systems (Ref Now, buses are 80% stainless. 2). Porsche uses austenitic stainless steel for Designers began the conversion to gain the front and rear side members, internal push rods normal advantages the stronger stainless gives: on front and rear axles, and lower rear wish- over 10% lighter weight and over 10% im- bones in its Carrera GT. provement in crash worthiness of the passenger Another manufacturer, Audi, engineered vari- compartment, the accompanying savings in fuel ous components of austenitic stainless steel into consumption, and the virtual elimination of its otherwise aluminum-intensive A6 series. The body maintenance. With essentially the entire use of stainless steel in strategic components body now in stainless, coating and painting enables greater weight reduction than that could be eliminated. A stainless bus body is which the vehicle would have in all aluminum. shown in Fig. 2. Volvo and Saab have designed austenitic stain- This swung the economic pendulum to stainless steel bumper systems that also serve to re- less in a major way. Now, not only was the duce overall vehicle weight. long-term cost of operating the bus lower, but While it is probably apparent to the reader that the initial cost of the bus was lower. The eco- essentially any body component can be made in nomic analysis is shown in Table 4 (Ref 2). stainless and be made better in stainless, the The design key was to use rectangular 304 question of when doing so is a better engineer- stainless structural tubing, which allowed strong, ing decision involves economic considerations. stiff sections to be welded into space frames. It is Large automotive companies generally have only a matter of time until this is improved on by large ﬁxed investments in painting and coating the use of 201 (with 3 to 4% Ni instead of 8 to systems to protect entire bodies from corrosion. 9% Ni) to lower cost and cold working of the The incremental savings of eliminating coatings tubing to achieve higher strength levels. Chapter 19: Automotive and Transportation Applications / 231

Microcars are now a familiar sight in Europe. few manufacturers, such as Volkswagen, have These vehicles are prized for their ability to be installed stainless steel fuel tanks in their vehi- driven and parked in very small or congested lo- cles (Ref 2). cations. Their economy of operation is also a major attraction. These considerations combine to make stainless the best material for many of Trucks their components. Figure 3 shows a stainless steel microcar frame. Over-the-road trailers are an excellent exam- The design by the famous design house Pinin- ple of stainless steel being used for utilitarian farina employs a stainless frame to give maxi- purposes. Trailers used for hauling foodstuffs or mum torsional stiffness and crashworthiness corrosive materials are now constructed almost while eliminating painting entirely. entirely of stainless steel; lined carbon steel tanks are now largely obsolete. The engineering basis for this is the same as for buses: high Other Automotive Components strength, no coating costs, and a product with long life and low maintenance costs. Structural members in trailers are typically 304, while Stringent emissions controls regulations, led tanks may also be 316L for corrosion resistance in the United States by the state of California, when the transported material requires it. Tank have made manufacturers reexamine the suit- wrappers are often made of bright annealed and ability of polymeric fuel tanks. These tanks con- buffed 304. Manufacturers of trailers would be tribute more to the required maximum 2 g/day well advised to consider upgrading to duplex of hydrocarbon emissions than is tolerable, so a grades such as 2003 or 2205 or to cold-worked austenitic stainless, which would permit major weight reduction. This weight reduction would directly translate into greater load-carrying capacity because the payload of liquid-carrying trailers is limited by total gross weight. The ability to add a few thousand more pounds of payload would quickly pay back a small premium in material cost. Normal cargo-carrying trailers also use some stainless where corrosion is problematic, such as in doors and door frames. Weight reduction is less important in these trailers, which reach maximum load at a volume limit rather than a Fig. 2 Stainless steel bus bodies. Source: Ref 2 weight maximum.

Table 4 Life-cycle cost calculation (LCC) for stainless steel versus carbon steel for a bus application

Cost of capital 10.00% Inﬂation rate 5.00% Real interest rate 4.76% Desired LCC duration 20.0 years Downtime per maintenance/replacement event 1.0 day Monetary unit U.S. $ Value of lost production 101 U.S. $/day Stainless steel Carbon steel Material costs 3.331 1.391 Fabrication costs 25.322 28.582 Other installation costs 2.185 4.050 Total initial costs 30.838 32.023 Maintenance costs 0 1.448 Replacement costs 0 2.897 Lost Production 0 57 Material-related costs 0 0 Total operating cost 0 4.402 Total LCC cost 30.838 36.425 Source: Ref 2 232 / Stainless Steels for Design Engineers

Fig. 3 Microcar frame fabricated from stainless steel. Source: Ref 2

Rail Transport been successful for decades, has not been carried over into trucks and buses even though it is tech- Passenger trains have exploited the high nically feasible to economically produce struc- strength-to-weight and toughness qualities of tural sections in very high-strength stainless. the 301 family of stainless steels for many years. The corrosion resistance of these alloys makes them corrosion free in long use, obviat- REFERENCES ing the need for painting and lowering maintenance costs. As with any other major use of type 1. R. Andersson, E. Schedin, C. Magnusson, 301, a 5 to 10% increase in economy could be J. Ocklund, and A. Persson, The Applicabil- achieved if type 201 were used instead of 301. ity of Stainless Steels for Crash Absorbing No loss in performance would occur. The transi- Components, ACOM, No. 3Ð4, AvestaPo- tion to 201 has not occurred simply because of larit AB, 2002 inertia and resistance to change on the part of 2. F. Capelli, V. Boneschi, and P. Viganò, designers and producers. “Stainless Steel: A New Structural Automo- Hopper cars made of 12% Cr martensitic tive Material, Vehicle Architectures: Evolu- stainless steels, typically 409Ni and 3Crl2, have tion Towards Improved Safety, Low-Weight, excellent abrasion and corrosion resistance as Ergonomics, and Flexibility,” paper well as high strength and therefore greater load- presented at Florence ATA 2005, 9th Interna- carrying capacity. Curiously, the use of cold- tional Conference (Florence), May 2005, worked austenitic stainless in railcars, which has www.centroinox.it, accessed June 2008 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 233-242 All rights reserved. DOI: 10.1361/ssde2008p233 www.asminternational.org

CHAPTER 20

Commercial and Residential Applications

Summary Stainless is without rival for ruggedness and durability. Steel and aluminum corrode. Glass, STAINLESS STEEL HAS BECOME the es- stone, and ceramics break. Plastic is weak. The sential material for products related to food, second, even more important, reason is that health care, and laundry because it combines stainless steel is essentially benign from a hy- strength and durability with an unexcelled abil- gienic viewpoint. Stainless steel itself is inert, ity to be cleaned, disinfected, and sterilized. both chemically and biologically, with respect These qualities have long been apparent to com- to food. Further, it provides minimal harbor for mercial food, laundry, and health care profes- unwanted biologic growth as do more porous sionals and have increasingly carried over into materials. Stainless also competes quite well es- equivalent domestic areas as consumers have thetically with other materials, offering the de- become more aware of the benefits of stainless. signer numerous surface finishes. Last, stainless is very amenable to nearly all manufacturing techniques. Its lack of need of coatings often Introduction makes components made from stainless less expensive to produce than equivalent designs that The last 20 years have seen the long-standing must be coated with paint, porcelain, or metal. pervasive commercial use of stainless steel for food preparation and serving; laundry; heating, ventilation, and air conditioning (HVAC); and Food Contact Qualifications other appliances penetrate the domestic market for the same types of goods. Whether this is a Setting aside cost, esthetics, and manufactur- fad of an increasingly affluent consumer or a re- ing considerations, a food contact material must flection of more design engineers and con- first meet three criteria: It must be chemically sumers being more interested in lasting value inert, biologically inert, and cleanable and able than they were in the “throwaway” society that to be disinfected. preceded that period remains to be seen. Stain- Chemical neutrality is achieved by a mate- less has been increasingly identified with high- rial when the material does not enter into the quality, high-end products. But, the case for food with which it comes in contact. This has value rather than fad seems to be stronger if the become an increasing concern as the effects of lessons of the harshly pragmatic automotive in- ions or chemicals released from food prepara- dustry, in which decorative use of stainless has tion materials have been viewed as potential virtually disappeared while utilitarian uses have toxic or disease agents. Medical knowledge is mushroomed, are any indication. not sufficiently advanced to convince con- The case for using stainless in appliances of sumers of the harmlessness of such contami- all types, whether they are commercial or resi- nants, so it is preferable to demonstrate the ab- dential, relates to stainless being able to provide sence of contamination if one is to win the the best value over the intended service life. public confidence in a food contact material. 234 / Stainless Steels for Design Engineers

Stainless steel contains many constituent ele- rally occurring level of up to 0.3 μg/g in cereals ments. Were they to enter the food with which to 1.1 μg/g in meat and fish. the stainless came in contact, then stainless These negligible levels of leaching simply in- would be a poor food contact material. The dis- dicate that foodstuffs are a benign chemical to tinguishing characteristic of stainless, however, stainless steel. Nevertheless, it is necessary to is the spontaneous passive film, which is so sta- apply the correct assessment of the corrosivity ble chemically. This film acts as a barrier to cor- of the foodstuff in question. In food production, rosion, which would result in metal release. as opposed to preparation for serving, more ex- Stainless therefore is effectively inert. Tests treme levels of acidity and salinity can be en- have been made of the rates at which metal ions countered. Nippon Steel reported (Ref 3) that can enter foodstuffs (Ref 1). Table 1 shows the materials used in the manufacture of soy sauce, vastly lower rates of metal ion release from which can have 15% salt, must withstand pro- stainless than from aluminum and carbon steel, longed contact at 45 °C (115 °F). Under such both of which are permissible, if not optimal, conditions, 316 stainless pits in about 1 day, food contact materials. Aluminum releases alu- while higher alloy grades, the 6Mo alloys, of minum ions into solution of both cooking oil which their YUS 270 is one (equivalent to UNS and 3% acetic acid at nearly equal rates of S31254), are projected to last 20 years before 15 mg/cm3 in 30 days. Carbon steel releases pitting. iron at over 100 mg/cm3 in the same period. This is significant because only pitting corro- Stainless, however, releases less than 0.010 sion is likely to release metals ions into a food mg/cm3 of iron. This is complemented by other substance. So, while guidelines exist for the studies showing that the transfer of ions from minimum alloy content permissible for normal food contact vessel to food is diminishingly and food contact, such as those promulgated by the negligibly small. Tests have been conducted on National Science Foundation (NSF), one must stainless steels, types 304, 439, and 444, that still verify the corrosion due to a particularly had both industrial finishes (2B and BA) as well aggressive food ingredient. Choosing the proper as freshly abraded and air-aged finishes. These grade of stainless, based on pH, salinity, and sample steels were subjected to boiling solu- temperature, is the responsibility of the design tions of oils, alcohols, water, and 3% acetic engineer. Referred to chapters in this book on acid. None caused the transfer of either both corrosion and individual alloy families for chromium or nickel to exceed the statutory 0.1 guidance in choosing an alloy based on resisting ppm level (Ref 2). Nickel levels of various pitting in a given environment. This having foods before and after cooking have been scien- been said, no alloy greater in pitting resistance tifically measured to assess the possibility of than 304 is required in residential or commer- leaching of that ion from 304 stainless steel cial cooking food contact. The higher alloy (Ref 2). No increase was noted from the natu- requirements come from the more aggressive

Table 1 Net metal migration into acetic solution (3%)

Metal migration during indicated time, µg/cm3

Material Time Iron Chromium Aluminum Nickel Austenitic stainless 30 min 2.4 0.12 ≤ 0.19 ≤ 0.12 10 days 4.2 0.22 0.22 ≤ 0.12 20 days 2.7 0.22 0.19 ≤ 0.12 30 days 2.3 0.28 ≤ 0.19 0.31 Ferritic stainless 30 min 3.0 0.43 ≤ 0.19 ≤ 0.12 10 days 7.3 0.40 0.19 ≤ 0.12 20 days 8.6 0.71 ≤ 0.19 ≤ 0.12 30 days 6.6 0.87 ≤ 0.19 ≤ 0.12 Aluminum 30 min 4.9 0.93 930 ≤ 0.12 10 days 18.2 3.42 5,300 ≤ 0.12 20 days 17.9 5.58 7,160 ≤ 0.12 30 days 31.3 12.40 15,350 0.22 Carbon steel 30 min 8,430 0.62 2.7 ≤ 0.12 10 days 57,700 7.40 26.7 ≤ 0.12 20 days 62,900 6.82 24.0 ≤ 0.12 30 days 112,000 14.00 36.9 ≤ 0.12 Source: Ref 1 Chapter 20: Commercial and Residential Applications / 235

concentrations and exposure periods that can be mercialized (Ref 5). Silver ions, like copper found in food-processing plants. ions, are powerful antimicrobial agents. The Certain abuses can even damage stainless combination of such a coating with stainless as cookware. Very high temperatures, such as can a corrosion-proof substrate may represent the occur when unattended pans have their liquids maximum in hygienic and chemical protection boiled away, can damage stainless but are more and is already being used in medical applica- harmful to less-rugged alloys such as copper tions where such concerns exceed those in ordi- and aluminum. nary food contact situations. Stainless steel is primarily composed of iron, Cleanliness. A necessary quality in any mate- chromium, and nickel along with small amounts rial considered for food contact is the ability to of manganese, silicon, and molybdenum. It con- be cleaned. This includes the removal of both tains trace amounts of copper, aluminum phos- organic and inorganic substances. The most im- phorus, and sulfur. Each of these elements is portant objective of cleaning is to remove the naturally occurring in food. Each can be found visible and invisible materials that can provide a in a typical multivitamin/multimineral supple- growth medium for microorganisms. This ment. Stainless is essentially devoid of heavy process is distinguished from disinfection, metals, such as lead and mercury, which are va- which is the reduction of the microbial popula- porized at the temperatures at which stainless is tion to a satisfactory level. What this level is de- refined. Even if toxic metals were somehow to pends on the standards of hygiene in force. And, be made to contaminate stainless, the passive although cleaning can and does reduce the pop- film would prevent their release. All these fac- ulation of microorganisms, true bacteriological tors combine to make stainless the most chemi- cleanliness is obtained only after disinfection. cally neutral metal found in food contact. Alter- The combination of cleaning and disinfecting is native alloys, such as copper and aluminum, important. Studies have shown that the efficacy actively leach into foods. Copper and aluminum of disinfectants is weaker on bacteria that have have been linked to but not demonstrated to been established in a surface biofilm than on cause Alzheimer’s disease. bacteria in suspension. The most complete form Biological Neutrality. Microorganisms ad- of disinfecting is sterilization, whose objective here to solid surfaces. When a clean surface is the complete removal of all microbial life and comes in contact with food, a surface deposit is viruses. formed from the food. The film may also con- The purpose of cleaning stainless steel is to tain molecules left from previous cleaning and rid it of contamination. Various stainless manu- disinfecting. The formation of this film is pre- facturers and associations have identified a sumably influenced by material characteristics number of effective of cleaning products (Ref 2, such as roughness, although there are no spe- 6, 7): cific studies on this. However, microorganisms ¥ Alkalines, which dissolve fats and oils adhere to this film and, as colonies of them ¥ grow, form a biofilm. This film consists of lay- Chelating or sequestering agents, which ag- ers of microorganisms that can produce an exo- glomerate contaminants. These are often or- cellular polymeric matrix, which protects the ganic acids such as citric acid or oxalic acid colony from cleaning and disinfecting. Geomet- and amine acids such as sulfamic acid and ric factors also can protect these colonies. ethylene diamine tetraacetic acid (EDTA) or Rough surfaces are intuitively more difficult to salts of these compounds. ¥ Hydrocarbon solvents clean. The ability to maintain a microscopically ¥ smooth surface is an asset in stainless that poly- Water with soap, detergent, trisodium phos- meric, enamel, and mineral surfaces lack. Stain- phate, or other surface active agents, which less steel is much less roughened by abrasion, emulsify ¥ Dilute oxidizing acids like nitric acid keeping the surface smooth (Ref 4).This will be ¥ seen to influence its ability to be cleaned and Mild acids such as phosphoric acid disinfected. The effectiveness of a cleaner relates mainly There is some technology to go beyond bio- to the contaminant to be removed. Some trial logical neutrality in the use of coatings that ac- and error may be required for a given contami- tively discourage or eliminate growth of mi- nant. Some precautions are worth mentioning. croorganisms. Polymeric coatings impregnated Abrasive cleaners should be used with caution. with silver ions have been developed and com- The abrasive size and hardness must be chosen 236 / Stainless Steels for Design Engineers

so that the stainless surface finish is not affected abrasive polishing, permits bacterial colonies to in an unwanted manner. If the abrasive is harder be removed. The greater roughness of the other than the stainless or coarser than the stainless materials may serve to protect the bacterial surface roughness, the underlying finish can be colonies from shear forces and provide greater disturbed. Care should also be taken to clean specific surface area on which the colonies can with the polish grain if a polished surface is bond. being cleaned. Also, cleaners containing chlo- Disinfection. The ability of a surface to be rides are common. Their use is not recom- disinfected is measured by the concentration of mended on stainless. Use of hydrochloric a given disinfectant required for a specific re- (muriatic) acid is especially detrimental. If duction in bacterial population. Numerous chloride-containing cleaners are used, then thor- studies have been published (e.g., Ref 8Ð10) ough rinsing should be conducted to avoid showing that glass and stainless steel have equal chloride concentration through evaporation, es- aptitude for disinfection, and that polyesters, pecially in crevices. Steel wool or steel brushes polyurethanes, rubber, and aluminum all re- should not be used on stainless under any cir- quired about one to two orders of magnitude cumstances as iron residue interferes with the greater concentrations of disinfectant for the integrity of the passive film. same result. These results indicate why stainless The ability of stainless to be cleaned is best is so essential to the food industry. Stainless can measured by the actual removal of bacteria be disinfected quite readily, which allows the colonies. This has been done to compare great invisible liability of food-borne diseases unabraded and abraded (to simulate new vs. to be minimized. used) stainless steel, enameled steel, mineral The effectiveness of sodium hypochlorite as a resin, and polycarbonate materials, which can disinfectant is inarguable, also. So, despite its be used for sinks, counters, food prep tables, potential corrosivity, it will be commonly used. etc. (Ref 4). Figure 1 shows that the reduction in Taking this into account requires that commer- bacteria count by the same cleaning technique is cial and residential food equipment be able to ten times more effective on stainless than on the withstand some chloride level greater than oth- other material types. Abrasion did not degrade erwise projected. Industry practice in the United the ability of stainless to be cleaned as it did States has shown that corrosion problems occur softer materials. The surface of stainless, even at an unsatisfactory level with mechanically with the seemingly protected recesses due to polished 430 but not with bright-annealed 430.

Fig. 1 Bacterial retention as a function of material and cleaning time. Source: Ref 4 Chapter 20: Commercial and Residential Applications / 237

Thus, alloys with less than 16% Cr should not thermal conductivity in a cooking utensil mini- be used unless corrosion can be accepted. Al- mizes differences in temperature across the sur- loys containing 16% Cr can be used with opti- face in contact with the food, permitting better mal surface finish. Alloys as low in carbon as control of the cooking process. The solution to 12% are used for cutlery applications where the problem of thermal conductivity is to make slight corrosive attack can be accepted. This is a composite materials. Stainless can be bonded to necessary trade-off required to achieve high copper and aluminum, which allows the stain- hardness for good cutting edge retention. less to be on both the food contact surface as Higher chromium grades such as 304 can be well as the exterior, with an inner layer of cop- used even with mechanically polished surface per or aluminum effectively spreading the heat. finishes. From a cost-effectiveness point of Aluminum and copper are nearly equally effec- view, there is no reason to use more expensive tive as inner conductive layers. Premium cook- alloys than 430 or 201 in the vast majority of ware features them both. The “sandwich” is the commercial and residential kitchen and laundry optimal design because it optimizes heating uni- applications from a corrosion standpoint as long formity even more than using aluminum or cop- as surface finishes that have not been produced per alone would since the high conductivity by abrasive polishing are specified. Many such inner core functions as an isotherm. The unifor- finishes are widely used. In North America, the mity is the more important consideration than rolled-on replicas of No. 4 finish, Koolline, the absolute thermal conductivity or even the Lustrite, etc., are quite common, while in Eu- thermal diffusivity. rope the bright-annealed finish has been pre- In a triple layer, the choice of the non-food- ferred. Both of these are preferable to mechani- contacting stainless is less stringent. Some- cally polished finishes. times, the exterior is made of a ferritic stainless The food industry, an immense consumer of steel. The ferromagnetism of ferritic stainless stainless steel, could do more than any other in- steel makes it ideal for induction heating. Al- dustry to help conserve nickel by specifying alloys such as 436 have been used for this appli- loys such as 430, 439, and 201 as their standard cation, while 304 is the pervasive choice for the alloys as well as by specifying nonabrasive fin- food contact surface. This is despite the fact that ishes. This can be done with no loss of function- 201 or 301 are quite adequate for this applica- ality or change of appearance and could save tion. It is also possible to produce a magnetic 23% to 50% in material cost. carbon steel core with stainless bonded to both sheet surfaces. The exposed edges are rolled to shield them from corrosion. Applications Nonstick coatings, such as polytetrafluoreth- ylene (PTFE), are very popular because of their Cookware. Any interaction between a food nonstick qualities. Above 350 ¡C (660 ¡F) these contact material and the food is most likely to coatings give off toxic fumes. This is a danger occur during the cooking process when temper- for certain types of cooking, such as wok cook- atures are greatest. Only glass and stainless are ing or blackening, but more likely to be encoun- excellent food contact materials. And, since tered by accidentally high temperatures above cookware must be flexible enough to handle any those intended. Since they can be scratched and potential food, the choice of material for cook- are not impermeable, their use does not alter the ware must be the most conservative. For this choice of the material to which they are applied. reason and because of the brittleness of glass, Kitchen Appliances. Every type of commer- stainless is the material of choice. The qualities cial kitchen appliance can be, and usually is, discussed make aluminum and copper less de- made of stainless steel, as are premium domes- sirable. Both leach into food. Copper can be tic kitchen appliances. This choice is based on tinned to combat this. The tin also corrodes over durability and ease of cleaning and disinfecting. time but has very low toxicity. A larger draw- And, because many commercial appliances are back is the expense of retinning copper utensils. visible to the customer, aesthetics are also a Aluminum is known as a toxic metal, with its driving force. Choice of alloy for a given appli- toxicity causing symptoms similar to those of ance is a crucial cost factor. As was noted that Alzheimer’s and osteoporosis (Ref 11). These 430 is marginal for kitchen use, because of the two metals do have one advantage over stain- prevalence of chloride-containing cleaners, unless, however: their thermal conductivity. High less it has been bright annealed. All austenitics 238 / Stainless Steels for Design Engineers

are satisfactory under normal use. Designers seem to generally neglect the possibility that their equipment may be used in coastal climates. In the high ambient salinity of coastal climates, corrosion will occur unless 304 with a bright-annealed finish or a brushed finish rolled onto a bright-annealed 304 is used. Mechani- cally polished 304 stainless is inadequate for coastal environments. These are the same guidelines used for architectural applications. Figure 2 shows how different alloys withstand coastal conditions. The corrosion on 430 would be considered excessive, while that on the 201 and 304 is acceptable given that some routine cleaning would have prevented the corrosion that is present on these samples, which were exposed to coastal salt and humidity for 10 years in North Carolina (Ref 12). In the vast majority of ambient conditions, coastal salinity is not a problem. This applies to inland conditions or coastal conditions where interior environments are protected by air conditioning or adequate cleaning of the stainless is practiced. This is normally the case for commercial equipment. Under these conditions, 201 is quite adequate, and the use of 304 represents wasteful overengineering. This choice is supported by decades of use by the major manufacturers of commercial appliances. Many who are large enough to specify their desired grade on bills of materials rather than simply buying from service center inventories have routinely used 201 and realized an approximately 8% lower cost before surcharges. Use of 201 versus 304 reduces surcharges by almost 50%, which can be a much larger savings than the base price savings. Smaller manufacturers are often precluded from these savings because of the general, if inexplicable, practice of service centers not stocking 201 despite its being the most cost- effective general-purpose stainless grade. The extended nickel price elevation from 2004 on- ward has a good chance of changing that situation as end users rebel against surcharges, which cannot be passed on to their customers. It has been pointed out that there is an array of 201-type grades, and that this is a drawback to their wider adoption. I recommend following American Society for Testing and Materials (ASTM) A240 and specifying UNS S20100 when substituting for 304 as this has very similar performance in forming, welding, and appearance to 304 and can be most easily interchanged without complications Fig. 2 Stainless steel samples exposed on a North Carolina in manufacturing and field performance. For beach for 10 yr. Source: Ref 12 Chapter 20: Commercial and Residential Applications / 239

parts made by deep drawing, substitution is still concentrations are in weight percent. This is dif- very possible, but deep-drawn grades are more ficult to achieve for 17% Cr alloys if stabiliza- finely tuned to specific process paths and must tion is by titanium alone. be more tightly specified than general-purpose Interior or working parts of appliances, to the grades. The more commonly used alloys for ap- degree they require high cleanability or contact pliances are listed in Table 2. food, are also often made of stainless. This is The greatest savings comes, of course, from especially true of dispensing machines, such as using ferritic grades, and they should be used for beverages, ice cream, and ice. Stainless inte- whenever forming requirements permit, which riors are often found in refrigerators and dish- is the majority of the time, since most appliance washers. In the case of dishwashers, forming re- components experience little more than cutting, quirements are often severe enough to require bending, and welding. There are important pre- the use of austenitic stainless. Rolled-on fin- cautions, however. Mechanical polishing results ishes are generally preferred. Not a small reason in unacceptable corrosion resistance, and the for this is that this finish requires only a single low work hardening rate of ferritics causes the temper pass to both flatten and provide the fin- mechanical polish to take on a different color ish. This yields very consistent forming charac- shade. This subtle difference can be magnified teristics, meaning much lower breakage during to objectionable levels when a mechanically press-forming operations. Rolled-on finishes polished ferritic stainless, such as 430, is put also have very high visual consistency, which is side by side with an austenitic such as 201 or usually a very important quality criterion for ap- 304. This can be solved by specifying rolled-on pliance manufacturers. finishes, which look the same on ferritics and Canisters, chafing dishes, serving pans, etc. austenitics. These finishes also supply the added are generally made from austenitic stainless corrosion resistance that makes alloys such as steel, which lends itself to the typical deep- 430 acceptable. It is still preferable to use a forming operations used in their manufacture. dual-stabilized grade such as 468, which can be Coatings are rarely used. If antimicrobial coat- welded without adverse corrosion effects and ing were to be used in food contact, this would has high formability and corrosion resistance at be an ideal application since already cooked as little as half the cost of 304 when alloy sur- food is most often in the intermediate tempera- charges are factored in. Use of dual stabilization ture danger zone at which bacteria can multi- permits keeping titanium levels to a minimum, ply. Food preparation tables also fit into this making it possible to avoid TiN-caused surface category. defects, which occur if significant TiN precipi- Appliance facades are increasingly using tation occurs before solidification in the original stainless. These include refrigerators, stoves, steel production. This occurrence is strictly a microwaves, drawers, etc. Shelves and exhaust thermodynamic phenomenon related primarily hoods also benefit from being made of stainless. to the titanium and nitrogen levels, which The drivers here are cleanability, durability, should be minimized so that the product of tita- and esthetics. There are important visual con- nium times nitrogen is less than 0.0025 when siderations in these applications. Consistent,

Table 2 Stainless steels commonly used for appliances

Composition, %

Alloy UNS No. C N Cr Ni Mn Si Mo Ti/Nb 201 S20100 0.15 0.25 16.0Ð18.0 3.5Ð5.5 5.5Ð7.5 1.00 ...... 301 S30100 0.15 . . . 16.0Ð18.0 6.0Ð8.0 2.00 1.00 ...... 304 S30400 0.08 0.10 18.0Ð20.0 8.0Ð10.5 2.00 1.00 ...... 316 S31600 0.08 0.10 16.0Ð18.0 10.0Ð14.0 2.00 1.00 2.0-3.0 . . . 430 S43000 0.12 . . . 16.0Ð18.0 0.75 1.00 1.00 ...... 439 S43035 0.07 0.04 17.0Ð19.9 0.50 1.0 1.0 . . . 0.20 + 4 × (C + N), to 1.10 468 S46800 0.030 0.030 18.0Ð20.0 0.50 1.00 1.00 . . . Ti + Nb: 0.20 + 4 × (C + N), to 0.80 436 S43600 0.12 . . . 16.0Ð18.0 . . . 1.00 1.00 . . . Nb + Ta: 5 × C, to 0.70 444, S44400 0.025 0.035 17.5Ð19.5 1.0 1.0 1.0 1.75Ð2.50 Ti + Nb: 0.20 + 4 × YUS190 (C + N), to 0.80 29-4C S44735 0.025 0.025 28.0Ð30.0 0.5 1.00 0.75 3.5Ð4.5 Ti + Nb: 0.20 + 4 × (C + N), to 0.80 240 / Stainless Steels for Design Engineers

defect-free surface finishes are paramount. This characteristic, so most cutlery is tempered at again can really only be achieved by rolled-on low temperatures. The vast majority of require- finishes since abrasively polished finishes vary ments for high-quality cutlery are satisfied by excessively in roughness, reflectivity, and color. 420 stainless. If greater cutting edge retention is Panel flatness is often very important and an- desired, then more or harder carbides are engi- other benefit from rolled finishes. If visible neered into the martensitic matrix. This is done welds are required, as is often the case for prod- by adding more carbon and chromium, as is ucts such as hoods and counters, then special found in 440A and to a greater extent in 440C. finishes with very long polish grains have a The wear resistance added by carbides is pro- major advantage in that the weld can be ground portional to their hardness and amount. The and polished with a belt sander of the appropri- chromium carbides of these straight-chromium ate grit size so that the weld blends impercepti- martensitic stainless steels are very hard, 1800 bly with the adjoining original surface. This is a HV, versus the 1100 HV hardness of iron car- practical impossibility with abrasively polished bides. The addition of higher levels of carbon finishes and very difficult with rolled finishes. ties up chromium so that it cannot add to corro- Freedom from fingerprinting can be another sion resistance, however, so that it can become valuable attribute for faÁade applications. This barely rust resistant. Furthermore, at high car- can be obtained on stainless by the mill applica- bon levels, carbides precipitate in the liquid and tion of a thin, bonded polymer film. All bare are much coarser. These large carbides can pull stainless finishes show fingerprints. With un- out during edge honing, making a ragged rather coated stainless, it is best avoided by using min- than a fine, smooth cutting edge. eral oil-based cleaners. However, vanadium and tungsten have even Although very high alloy stainless steels are harder carbides, 2800 and 2100 HV, respec- used for high-temperature kitchen applications, tively. Through conventional casting and hot such as heating element sheathing (American working, only a small amount of these carbides Iron and Steel Institute [AISI] type 334), it is can be introduced into the matrix. The problem seldom used for oven interiors because it does is that if primary carbides form during solidifica- take on a heat tint when exposed to tempera- tion, they tend to be coarse and to embrittle the tures above 300 ¡C (570 ¡F). Range tops, which alloy. Hard particles are much more useful for see lower temperatures, are normally stainless. wear resistance if they are small and widely dis- Outdoor cooking grills, because they must en- persed. To a degree, this refinement of the pri- dure exterior environments without corrosion, mary carbides can be achieved by raising nitro- are almost always 304 or a similar grade. Heat gen levels. These problems can be circumvented tint does not occur with these to a problematic by the use of powder metallurgy, which permits degree. Gas burner manifolds are also stainless. the solidification step on a macroscale to be In this case, ferritics are required because of the skipped. Larger volume fractions of hard car- need for extreme high-temperature oxidation re- bides such as vanadium carbide and tungsten sistance and the desirability of a low coefficient carbide can be added and dispersed. Table 3 lists of thermal expansion. The preferred alloys are the martensitic alloys used for cutlery. It is rea- those developed for automotive exhaust sys- sonable to say that most of these alloys far ex- tems, variations on 409 and 439. ceed the requirements of food preparation. Flatware and cutlery were among the origi- Improved corrosion resistance of these alloys nal uses of stainless. Stainless filled the gap be- is achieved by adding molybdenum at the ex- tween carbon steel, which was hard but whose pense of chromium, which would cause exces- rusting was an obvious problem, and silver, sive δ-ferrite retention if it were raised. This can which was soft and whose cost prohibited its be seen in alloys above the basic 420. use to all but the wealthy few. Cutlery is the do- Flatware has no hardness requirement, so main of martensitic stainless steel. The corro- grade selection is based on the need for per- sion resistance of martensitic grades cannot be ceived quality. At the high end is 304, which has improved above modest levels, never reaching all the corrosion resistance that could be needed that of 304, but this criterion is secondary to for flatware. However, type 301 is commonly hardness because of the need to keep a sharp used also, as are ferritic steels, such as 430, for cutting edge. Maximum corrosion resistance is low-cost flatware. Depending on the shape of achieved in the as-quenched condition. But, the final utensil, material is stamped or forged some toughness is a valuable but not crucial and then finished. Chapter 20: Commercial and Residential Applications / 241

Table 3 Stainless steels commonly used for cutlery

Composition, %

Alloy Designation Form C Mn S Si Cr Mo Ni Other 420 UNS S42000 Wrought 0.15 min 1.00 0.030 1.00 12.0Ð14.0 ...... 4116 DIN 1.4116, nominal Wrought 0.50 ...... 14.5 0.65 . . . 0.15 V 440A UNS S44002 Wrought 0.60Ð0.75 1.00 0.030 1.00 16.0Ð18.0 0.75 ...... 440C UNS S44004 Wrought 0.95Ð1.20 1.00 0.030 1.00 16.0Ð18.0 0.75 ...... BG-42 Nominal composition Wrought 1.15 ...... 0.3 14.5 4.0 . . . 1.2 V ATS-34 Nominal composition Wrought 1.05 0.4 . . . 0.35 14.0 4.0 ...... 14-4 CrMo Nominal composition Wrought 1.05 0.5 . . . 0.3 14.0 4.0 ...... 154 CM Nominal composition Wrought 1.05 0.45 . . . 0.3 14.0 4.0 ...... CPM S30V Nominal composition PM 1.45 ...... 14.0 2.0 . . . 4.0 V CPM S60V Nominal composition PM 2.15 0.40 ...... 17.0 0.40 . . . 5.5 V CPM S90V Nominal composition PM 2.20 ...... 13.0 1.0 . . . 9.0 V

Many kitchen utensils are also made entirely clothing. Those, along with the implied quality of or in part with stainless. Type 304 is the alloy stainless, are the main drivers for its use. most commonly used, but again any of the Heating and Water Heating. With the de- stainless steels with at least 16% Cr are ade- velopment of high-efficiency, natural-gas-fired, quate, and grade selection depends on forming forced-air furnaces, stainless has come into do- and joining requirements. mestic use as a heat exchanger material. These Laundry appliances have converted signifi- furnaces gain their extra efficiency by condens- cantly to stainless. This trend began in Europe ing water from combustion gas exhaust. This with the development of high spin speed, hori- condensate can, depending on incoming air, zontal axis washing machines. These washers contain corrosion elements, which has led to the use far less water and energy to achieve higher use of very highly alloyed ferritic stainless steel levels of cleaning with less damage to clothing. in their construction. Alloy 29-4C (UNS These features have eroded the share enjoyed by S44735) was the original alloy used nearly uni- vertical axis, agitator-type washers, whose low versally in the United States. The worst conse- speed allows them to be constructed of low- quence of perforation by pitting could be the re- strength materials such as plastic or porcelain- lease into the home of toxic gas, so pitting coated steel. The stresses induced by the high corrosion must not be allowed. spin speeds, which are necessary in horizontal The intermediate efficiency furnaces (80 to axis machines to take water removal from 80% 90%) require the use of corrosion-resistant vent to 95%, require the strength of stainless steel. pipe to prevent corrosion from condensation in Porcelain-coated carbon steel obviously can be the flue. High-temperature plastics were tried, strong, but the coating is cracked by strains that but failed joints in them caused their recall after the steel itself easily tolerates. An additional ben- several fatalities were reported. High-efficiency efit to stainless over porcelain is that stainless (90% or higher) furnaces can use low-tempera- starts smooth and becomes even smoother with ture plastic pipe, but these units require the use use, while porcelain becomes quite abrasive over of a corrosion-resistant secondary heat ex- time as wear opens voids with edges that can be changer to recover the latent heat of vaporiza- quite sharp and cause significant damage to tion of the water from combustion. Alloy 29-4C clothing. Washer tubs and drums are made of was the original choice for most secondary heat both ferritic and austenitic stainless. The selec- exchangers, but at least one used alloy 6XN tion is based on forming requirements rather than (UNS N08367) alloy for formability. Some corrosion or strength. If components can be made manufacturers have always used lower-alloyed by bending rather than stretching, then the lower- stainless steels. cost ferritics can be used. Ferritics should be a The proper handling of combustion products 17% stabilized grade, such as 439 or 468, and is an interesting problem in materials selection. austenitics can be 201, 301, or 304. Unstabilized The variability of the use environments leads to ferritic alloys, such as 430, should never be used a huge spread in corrosion conditions and mate- in welded applications. Dryers are less challeng- rials performance. In the end, one has to balance ing, and it is difficult to make a strong case for materials selection between cost (fortunately, the functional value of stainless. Those designs 29-4C alloy is nickel free) and probability of that use stainless will last longer and be gentler to failure. Given the number of units produced and 242 / Stainless Steels for Design Engineers

the potentially serious consequences of failure, 3. “The Application of High Corrosion failure rates must be less than 10 Ð4, while fail- Resistance Stainless Steel YUS270 in ure rates much less than 10 Ð6 are impossible to Food Processing Facilities and Equipment,” verify and hard to justify. In any case, the com- Nippon Steel Technical Report 87, Jan petition is always between various stainless 2003 steels. Galvanized steel will not work. The issue 4. J.T. Holah and R.H. Thorpe, Bacteria Re- is difficult enough for natural-gas-ﬁred fur- tention on Cleaned Surfaces, J. Appl. Bacte- naces. Oil ﬁred is a developing situation for riol., Vol 69, 1990, p 599Ð608 which there is no good consensus. Wood burn- 5. Agion Technologies, www.agion-tech.com, ers and other unconventional furnaces (such as accessed June 2008 corn burners) present additional challenges, and 6. Removal of Stains and Discolourations, Out- answers are even less obvious. okumpu, www.outokumpu.com, accessed Water heaters are sometimes made of stain- June 2008 less steel. It is not uncommon for water to have 7. “The Care and Cleaning of Stainless Steel,” a sufficient level of chlorides to lead to stress Specialty Steel Industry of North America, corrosion cracking if an austenitic stainless is www.ssina.com, accessed June 2008 used. Therefore, the recommended alloy for this 8. E.P. Kysinski et al., J. Food Processing, Vol application is UNS S44400. More recently, lean 55, 1992, p 246Ð251 duplex alloys have been developed, such as 9. A.A. Mafu et al., J. Dairy Sci., Vol 73, 2101 and 2003, which can perform quite well 1990, p 3428Ð4332 without corrosion or stress corrosion cracking. 10. P. Gelinas and J. Goulet, Can. J. Microbiol., More highly alloyed duplex alloys such as 2205 Vol 29, 1983, p 1715Ð1730 are more expensive but would work well. 11. R.A. Goyer, Toxicity of Metals, Properties and Selection: Nonferrous Alloys and Spe- cial-Purpose Materials, Vol 2, ASM Hand- REFERENCES book, ASM International, 1990, p 1233Ð 1269 1. M.J. Julio, M.L. Martin, and J.M. Baena, 12. Allegheny Ludlum research, as presented in Cation Migration Tests in Metal Containers, D.S. Bergstrom and C.A. Botti, AL Innovation Stainless Steel (Florence), Oct 201HPTM (UNS S20100) Alloy: A High- 1993, p 1.221Ð1.226 Performance, Lower-Nickel Alternative to 2. “Stainless Steel in Contact with Food,” 300 Series Alloys, Stainless Steel World, Document Ugine, June 1996 KCI Publishing, 2005 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 243-246 All rights reserved. DOI: 10.1361/ssde2008p243 www.asminternational.org

CHAPTER 21

Marine Systems Applications

Summary restricts their use to items of rather thin gauge, less than about 1.0 to 2.0 mm, depending on AS RECENTLY AS THE 1960s AND 1970s, alloy. Thus, their use is limited to tubing. Super- handbooks on stainless steel were stating that austenitic alloys can be used at any thickness, “stainless steels are not stainless in seawater,” although they are a costly material. The success and “successful prolonged corrosion-free serv- story for stainless steel and seawater and there- ice of stainless steel in seawater requires sophis- fore desalination is that of duplex stainless steel. ticated corrosion engineering, or enormous good With the same corrosion resistance as any super fortune” (Ref 1). The advances in stainless steel austenitic or superferritic alloy, it has nearly made since then have thankfully made these double the strength plus resistance to stress cor- statements obsolete. Not only have basic corro- rosion cracking. And while duplex stainless steel sion problems been solved, stress corrosion is not a cheap material, it does contain much less cracking also can be avoided. More impres- nickel than an equivalently corrosion resistant sively, this can be done with alloys with austenitic stainless steels, which is a major cost strengths much higher than those of the alloys, saving factor. such as 316, that they replace and that have been Desalination technology is relatively new if only marginally successful in marine environ- one ignores the fact that distillation has been ments. The inertia in changing from the weaker, around for a very long time. Desalination in com- less-corrosion-resistant, more expensive aus- mercially viable quantities began with multi- tenitic stainless steels is large because of less stage flash technology in the 1950s. The underly- availability of the newer, better alloys, and lack ing principle of this process is the evaporation of of familiarity with their benefits. Those who water vapor from salt water with the subsequent understand and use these newer duplex alloys condensation of the salt-free water vapor. In the will be rewarded. This chapter reviews the major multi-stage flash (MSF) approach feedwater is marine applications of stainless steels, including heated and the pressure is lowered so that the desalination equipment, shipping containers, water “flashes” into steam. A variation on this and heat exchangers that handle seawater. technology is multiple-effect distillation (MED), another low-temperature distillation process. The differences in all distillation-based systems Desalination reduce to the efficiency of the design in minimizing energy consumed per unit of pure water out- At one time not long ago stainless steel was put. All distillation processes require heating of thought to be an inadequate to marginal material the input water and some process power. for use in seawater. Its use in heated seawater The other basic engineering approach to de- was therefore all the more suspect. This was first salination is reverse osmosis (RO). The inven- changed with the development of superferritic tion of polymer membranes that could separate and superaustenitic alloys. The superferritic al- the salt ions from the water made this technol- loys such as Seacure (UNS S44660) and 29-4C ogy possible. No thermal energy is required. The (UNS S44735) are quite resistant to seawater, water is pumped at high pressure through these even at high temperature. Their low toughness permeable membranes physically separating the 244 / Stainless Steels for Design Engineers

salt from the water. The change in salt concen- for their processes. Type 316 stainless steel has tration across the membrane is a function of the passed from consideration as a material for han- pressure and the membrane itself. A second dling brackish water or seawater. treatment may be used to improve water quality. In distillation systems, the rule of thumb is The distillation methods require about 5 that 2205 alloy (UNS S32205), with its pitting kWh/metric ton of water output, while the RO resistance equivalent number (PREN) of 35, is methods require twice that. The distillation sufficient for seawater up to 20 ¡C (70 ¡F); al- methods require another 20 kWh of thermal en- loys 2507 (UNS S32750) or Zeron 100 (UNS ergy from some source for feedwater heating, S32760) should be used for seawater at elevated while the RO method requires none. Thus the temperatures or high salinity. For the output of ability to find energy from cogeneration or a fresh water, lesser alloying is required. Stainless source such as solar, etc may determine which steel types 304 (UNS S30400), 316 (UNS process is preferred. S31600), 2101 (UNS S32101), 2003 (UNS S32003), or even 439 (UNS S43035) may be used depending on the combination of salinity Materials Selection for Desalination and temperature of the output water. Materials used for distillation processes have Besides their high corrosion resistance for evolved from use of type 316 (UNS S31600) lower total alloy cost, the duplex stainless steels stainless steel, first as lining and then as have higher strength, which is a significant fac- cladding. The superaustenitic alloys, the 6Mo tor since distillation plants are large. The use of variations, came next because they truly solved duplex allows wall thickness reductions that the corrosion problem, but at a price. Then, sep- bring about larger savings than those based arately the duplex alloys were developed, with solely on their cost per unit weight. Figure 1 the first market the petroleum industry, whose shows the difference among the candidate stain- demands and research made these alloys possi- less steels in corrosion resistance (Ref 2). The ble. It was not a stretch to see that high-strength viable materials for seawater are those that can alloys that could withstand seawater in offshore withstand roughly 20,000 ppm ClÐ level at the applications could do well on land as well. To appropriate temperature. give full credit, the pulp-and-paper industry was The strengths of the various candidate materi- also beginning to employ duplex stainless steels als are given in Table 1. These are typical values.

Fig. 1 Corrosion resistance (pitting) as a function of salinity and temperature. 1. 304L (UNS S30403); 2. 316L (UNS S31603); 3. 2205 (UNS S32205); 4. 904L (UNS N08904); 5. 254SMO (UNS S31254). Source: Ref 2 Chapter 21: Marine Systems Applications / 245

Table 1 Typical analyses and properties of major marine alloys

Composition, % Yield strength Tensile strength Elongation, Alloy UNS Cr Mo N Ni PREN(a) MPa ksi MPa ksi % 2101 S32101 21.5 0.3 0.22 1.5 26 515 75 650 94 40 2003 S32003 20.5 1.5 0.18 3 29 515 75 725 105 40 2205 S32205 22 3 0.17 5 35 515 75 760 110 35 2507 S32750 25 4 0.27 7 42 550 80 800 116 35 304L S30403 18 0 0.05 8 18 220 32 520 75 50 316L S31603 16 2 0.05 10 24 220 32 520 75 50 317L S31703 18 3 0.05 14 29 230 33 540 78 45 6XN N08367 21 6 0.22 24 45 380 55 760 110 45 254SMO S31254 20 6 0.20 18 43 380 55 750 109 45 Zeron 100 S32760 25 3.5 0.27 7 42 550 80 750 109 35 (+0.75 W)

(a) PREN, pitting resistance equivalent number.

Refer to the appropriate design code for your may be justified based on strength alone; their particular application to find minimum proper- exceptional corrosion resistance would be simply ties. The reader is cautioned that duplex longitu- an excellent side benefit. Corrugated stainless dinal properties are slightly lower than the bulkheads are positioned within the carbon steel transverse properties that testing requires. hull. The stiff corrugated bulkheads are them- Pumps for seawater follow the same guide- selves structural strengtheners for the entire ship. lines as piping, tanks, and all other components. The vertical corrugations also facilitate tank Cast or wrought duplex are the alloys of choice. cleaning as internal stiffeners are eliminated. Cryogenic containers are still the bastion of austenitic stainless steels. As leaner austenitic Shipping alloys have become less expensive than 9% Ni alloy steel, a martensitic grade, they have be- The major uses of stainless steel in shipping come the material of choice. In this case, the are in bulk storage containment. Cargos range 201 types are preferred to 304 because 201 has from food and beverages to chemicals and liquid greater strength at the cryogenic operation tem- natural gas (LNG). Practice in the past has been perature and is, of course, less expensive. The to use austenitic grades of stainless with cathodic expanding market for LNG has made ocean protection when necessary to address inadequate transport increasingly important because large corrosion resistance. However, since 2000 ma- disparities in prices often are due to the diffi- rine chemical tankers have become the largest culty in transporting it. The two best materials consumer of duplex stainless steel. The reason are UNS S20153 and S20400, which perform for this is that cargo tanks ideally have the widest equally well. If higher strength is valuable to a potential range of cargos possible. This range is design for cryogenic uses, then UNS S21904 defined by corrosion resistance. This factor alone (21-6-9 or Nitronic 40) could be used. This is reason to choose duplex over austenitic alloys alloy has yield strengths of 460 MPa (65 ksi) at such as 316L (UNS S31603) or 317L (UNS room temperature and 1200 MPa (175 ksi) at S31703). An equally decisive factor is strength. Ð196 ¡C (Ð320 ¡F). It is completely resistant to With codes permitting the tank’s design to be martensite formation. based on yield strength, the use of duplex al- Other shipboard systems benefit equally from loys—with strengths about double those of the use of duplex stainless steel. This extends to austenitic steels—permits significant weight re- piping, hardware, propellers, shafts, etc. duction. This is a major economic factor for ship owners in that dead weight can be replaced by fee-paying cargo at the same operating cost. Heat Exchangers These incremental revenues, over the life of the vessel, are many times the original cost of the Coolers for captive water systems such as for material. Based on the high value for strength in power plants often need to resist corrosion by ship economics, it would seem that the highest- brackish water or seawater. To the extent that strength alloys, such as 2507 (UNS S32750), these are thin-wall tubing, ferritic alloys such as 246 / Stainless Steels for Design Engineers

Seacure (UNS S44660) or 29-4C (UNS S44735) sion cracking to very high temperature and have been used quite successfully. If thicker salinity. tubes are required, then the equivalent duplex or austenitic alloys can be used. This would include types 2003 (UNS S32003), 2205 (UNS S32205), REFERENCES or 2507 (UNS S32750) duplex stainless steels, depending on salinity and temperature; for 1. Peckner and I. Bernstein, Stainless Steel austenitics, the 6Mo alloys such as 254SMO Handbook, McGraw-Hill, 1966, p 37-1 (UNS S31254) and AL6XN (UNS N08367) may 2. Stainless Steel for Desalination Processes, be used. The duplex alloys have the advantage Feb 2006, Outokumpu, www.outokumpu. of lower cost. Both are resistant to stress corro- com, accessed June 2008 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 247-255 All rights reserved. DOI: 10.1361/ssde2008p247 www.asminternational.org

CHAPTER 22

Petroleum Industry Applications

Summary ginnings. Demand for steel for drill pipe, casing, and tubing has led to many developments, THE PETROLEUM INDUSTRY has had to such as the technology for producing high- deal with increasingly hostile environments in quality seamless and welded pipe and tubing. its search for new supplies of oil. And that pe- Pipeline needs have fueled the market for high- troleum, when found, often contains harmful in- strength, low-alloy plate. Offshore production gredients. The result is increasing demand for in often-hostile environments has presented se- steels with greater strength and corrosion resist- vere material challenges. And, as the light sweet ance. Martensitic and duplex stainless steels crude that was easily found and produced on have provided the corrosion resistance and land is exhausted, future supplies of hydrocar- strength to deal with higher levels of hydrogen bons are increasingly likely to contain sulfides, sulfide, carbon dioxide, chlorides, and acidity. carbon dioxide, and saltwater in sufficient This chapter reviews the selection of stainless amounts to make corrosion a top priority in se- steels for petroleum applications, including oil lecting materials. For reference in this chapter, country tubular goods (OCTGs), line pipe, off- Tables 1 through 5 list the relevant alloys for shore platforms, and refinery equipment. petroleum industry applications. Many, but not all, of these alloys are listed in the National As- sociation of Corrosion Engineers (NACE) Introduction MR0175, Sulfide Stress Corrosion Cracking Re- sistant Metallic Materials for Oil Field Equip- The petroleum industry has driven large segment; the tables in this chapter also include ments of the steel industry since both their be- some newer alloys not in the NACE document.

Table 1 Ferritic stainless steels for petroleum Table 2 Martensitic stainless steels for industry applications petroleum industry applications UNS Common name Hardness, S40500 405 UNS Common name HRC, max(a) S40900 409 J91150 CA15 . . . S43000 430 J91151 CA15M . . . S43035 439 J91540 CA6 NM . . . S43400 434 K90941 9Cr 1Mo . . . S43600 436 S14125 S/W 13Cr 28 S44200 442 S41000 410 22 S44400 444 (18-2) S41426 13CRS . . . S44500 … S41427 … 29 S44600 446 S42000 420 22 S44626 26-1 Ti, E-Brite S42400 F6NM 23 S44627 26-1 Cb S42500 15Cr 22 S44635 26-4-4, Monit JFE KL-12G . . . S44660 Seacure, SC-1 JFE KNHP12Cr . . . S44700 29-4 Nippon NT-CRS . . . S44735 29-4C Nippon NT-CRSS . . . S44800 29-4-2 420M . . . S46800 468 L80 13 Cr . . . Note: See Appendix 1 for alloy compositions. Source: Adapted from NACE Note: See Appendix 1 for alloy compositions. (a) As specified in NACE MR0175, “Sulfide Stress Corrosion Cracking Resistant Metallic Materials for MR0175. Source: Adapted from NACE MR0175, “Sulfide Stress Corrosion Oil Field Equipment” Cracking Resistant Metallic Materials for Oil Field Equipment” 248 / Stainless Steels for Design Engineers

Table 3 Precipitation hardening stainless steels Table 4 Duplex stainless steels for petroleum for petroleum industry applications industry applications

Hardness, UNS Common name PREN(a) UNS Common name HRC, max(a) J93345 Escoloy 31-47 S13800 13-8 PH . . . J93370 CD4MCu 30-34 S15500 15-5 PH 33 J93380 Z100 38-46 S15700 15-7 PH 32 J93404 958 39-47 S17400 17-4 PH 33 S31200 44LN 30-36 S17700 17-7 PH . . . S31260 DP3 34-43 S35000 AM-350 . . . S31500 3RE60 27-31 S35500 AM-355 . . . S31803 2205 (old) 30-36 S45000 Custom 450 31 S32001 19D 20-24 S45500 Custom 455 . . . S32003 2003 27-31 S46500 Custom 465 . . . S32101 2101 25-29 S66286 A-286 35 S32205 2205 (new) 34-38 Custom 465 (275) . . . S32304 2304 23-27 Custom 475 . . . S32404 U50 27-32 S32520 52N+ 37-48 Note: See Appendix 1 for alloy compositions. (a) As specified in NACE MR0175. Source: Adapted from NACE MR0175, “Sulfide Stress Corrosion S32550 255 32-44 Cracking Resistant Metallic Materials for Oil Field Equipment” S32750 2507 38-44 S32760 Zeron 100 40-46 S32803 2803Mo 33-41 The modern dilemma that makes stainless nec- S32900 329 26-35 S32906 2906 36-45 essary is the addition presence of wet carbon S32950 7-Mo Plus 32-43 dioxide, which is extremely corrosive to carbon S32977 AF 918 39-46 and alloy steel. As if this is not a sufficient ma- S39274 DP3W 39-47 terial problem, sometimes the wetness is from Note: See Appendix 1 for alloy compositions. (a) PREN, pitting resistance equivalent number. Source: Adapted from NACE MR0175, “Sulfide Stress Cor- saltwater, which further aggravates corrosivity. rosion Cracking Resistant Metallic Materials for Oil Field Equipment” This corrosion problem is compounded by the accelerating influence of high temperature in The chapters on martensitic and precipitation deeper formations. What is the answer to the hardening stainless steels discuss this in detail. corrosion problem? Inhibitors, coatings, ca- The martensitic stainless steels used for these thodic protection, or more corrosion-resistant applications are resistant to carbon dioxide- materials are the main responses. The first three enhanced corrosion up to partial pressures of responses are not always practical. They also 100 atm, after which further alloying is neces- represent an ongoing cost rather than a one-time sary. This cannot be achieved with a martensitic cost. Each situation must be evaluated regard- structure, but the duplex alloys have the corro- ing which is the optimal solution. sion resistance and strength to work in this regime. They have high annealed strength and can also be cold worked to higher strength levels. Combating Corrosion in Alloys for If hydrogen sulfide is present, the selection Petroleum Applications process can become more difficult. High- strength martensitic steels are susceptible to Alloying steel with chromium, copper, brittle delayed failure in the presence of hydro- molybdenum, and nickel can lower the corro- gen sulfide. Being stainless does not by itself sion rate of steel by a factor of 10,000. Figure 1 provide immunity. If localized corrosion occurs, shows the influence of chromium alone, which hydrogen uptake ensues, and delayed failure produces a 100-fold reduction in corrosion of follows. Only keeping hardness below well- steel in seawater and carbon dioxide (Ref 1). established levels can render a martensitic alloy Molybdenum is the most powerful alloying immune. If the localized corrosion can be pre- addition to magnify the benefit of chromium. vented, however, then the stress corrosion The effects of copper and nickel are also very cracking (SCC) cannot be initiated. Molybde- significant, as Fig. 2 (Ref 2) indicates. These ad- num alloying expands the pH and chloride ditions must be made in a very balanced way if range from which an alloy can be free of the pit- a tough, fully martensitic structure is to be ting corrosion that initiates SCC, as shown in maintained. Carbon must be kept low to avoid Fig. 3 (Ref 2). Martensitic steels of all types the formation of chromium carbides during have a maximum in susceptibility to SCC via tempering, which would counteract the benefit hydrogen embrittlement near room temperature. of the chromium. Nickel is necessary to prevent Duplex alloys and austenitic alloys become σ-ferrite formation, which reduces toughness. susceptible at higher temperatures and do not Chapter 22: Petroleum Industry Applications / 249

Table 5 Austenitic stainless steels for petroleum industry applications

UNS Common name PREN(a) UNS Common name PREN(a) J92500 CF-3 . . . S30100 301 . . . J92600 CF-8 . . . S30153 301LN . . . J92602 CF-20 . . . S30200 302 . . . J92701 CF-16F . . . S30215 302B . . . J92710 CF-8C . . . S30300 303 . . . J92800 CF-3M . . . S30400 304 . . . CF-12M . . . S30403 304L . . . J93000 CG-8M . . . S30409 304H . . . J93254 CK3MCuN . . . S30415 153MA . . . J93402 CH-20 . . . S30453 304LN . . . J94652 CN-3MN . . . S30500 305 . . . N06022 AL 22 . . . S30800 308 . . . N08007 CN-7M . . . S30815 253MA . . . N08020 20Cb-3 28 S30900 309 . . . N08020 AL 20 28 S31000 310 . . . N08024 20Mo-4 38 S31008 310S . . . N08026 20Mo-6 . . . S31254 254SMO 46 N08028 Sanicro 28 39 S31266 B66 59 N08031 Nicrofer 3127 hMo 54 S31600 316 . . . N08320 20Mod 38 S31603 316L 25 N08366 AL-6X . . . S31609 316H . . . N08367 AL-6XN 49 S31635 316Ti . . . N08700 JS-700 36 S31700 317 . . . N08800 332 . . . S31703 317L . . . N08925 25-6Mo 46 S31725 317LM . . . N08926 Cronifer 1925 hMo 47 S31726 317LMN . . . N08932 URSB-8 49 S31753 317LN . . . N80904 904L 39 S32100 321 . . . S20100 201 . . . S32109 321H . . . S20153 201LN . . . S32200 NIC 25 . . . S20200 202 . . . S32654 654SMO 64 S20400 Nitronic 30 (204L) . . . S33000 330 . . . S20430 204 . . . S33400 334 . . . S20500 205 . . . S34565 4565 54 S20910 Nitronic 50 . . . S34700 347 . . . S21800 Nitronic 60 . . . S34709 347H . . . S21900 Nitronic 40(219) . . . S35125 332Mo . . . S21904 21-6-9 LC . . . S35315 353MA . . . S24000 Nitronic 33 . . . Cronifer 2328 . . .

Note: See Appendix 1 for alloy compositions. (a) PREN, pitting resistance equivalent number. Source: Adapted from NACE MR0175, “Sulﬁde Stress Corrosion Cracking Resistant Metallic Materials for Oil Field Equipment”

Fig. 1 Influence of chromium on the corrosion rate of steel in environments experienced by oil country tubular goods. Test conditions: synthetic sea water; CO2 partial pressure, 0.1 MPa; test temperature, 60 (C °140 °F); test duration, Fig. 2 Influence of copper and nickel on the corrosion rate of 150 h; flow velocity, 2.5 m/s; specific volume, 800 mL/cm2. martensitic stainless alloys used for oil country tubular SSC, stress corrosion cracking. Source: Ref 1 goods. Source: Ref 2 250 / Stainless Steels for Design Engineers

The NACE recommendations of suitable materials are deﬁned by MR0175. Table 6 summarizes these recommendations. The reader is encouraged to refer to the latest version of this document for further details.

Oil Country Tubular Goods

Oil country tubular goods (OCTG) include the drill pipe, casing, and tubing and associated hardware used to construct oil and gas wells. Drill pipe is used to twist the drill bit and convey drilling fluids to the point of contact and flush away debris. Casing is put in place to stabilize the well walls, while tubing is placed within the casing to carry oil and gas to the surface. Each of these components sees significant stresses, and high strength-to-weight materials are needed. Drill pipe is in tension, torsion, and compression alternately throughout its life. Casing hangs from the wellhead under its own weight for distances from hundreds of meters to 7000 or 8000 m and must withstand very high collapse as well as burst pressures. Fig. 3 Influence of molybdenum on susceptibility to stress Well strings, the exact sequence of size and corrosion cracking in solutions containing (a) 3.5% strength pipe for each level of the well, are opti- NaCl and (b) 0% NaCl. Source: Ref 2 mized for the conditions of each well. The variety of strengths and sizes are standardized by the American Petroleum Institute. The use of exhibit the same increasing susceptibility with the highest strengths has always been limited by strength. So, when hydrogen sulfide, which en- hydrogen embrittlement accelerated by hydro- hances hydrogen uptake, levels exceed about gen sulfide, so that the maximum hardness for a 10Ð2 atm, the martensitic alloys should no longer given material must be strictly adhered to when be used, and the duplex alloys are preferred. As hydrogen sulfide is present. temperatures and hydrogen sulfide partial pres- The terms 13Cr, 22Cr, and 25Cr are com- sures increase, alloying must also, until at 1 atm monly used in the industry even though this of hydrogen sulfide nickel base alloys are re- greatly oversimplifies the alloying, and there- quired. Figure 4 shows this progression with the fore performance, options that exist. The 13Cr alloy recommendations of Sumitomo. The re- alloys are a family of martensitic stainless quirements behind this diagram are generic. steels. The 22Cr and 25Cr alloys are duplex Any producer’s alloys must comply with this grades. The former are used in the quenched diagram’s regions, which have been defined by and tempered condition, while the duplex alloys NACE. Stainless steels are required above a are used as annealed or cold worked. certain carbon dioxide level for all levels of hy- The 13Cr grades began as simply variations drogen sulfide. Martensitic alloys, commonly on 420, which is a straight-chromium marten- called “13Cr,” are the first step up from alloy sitic often used for cutlery. This alloy, while far steels. At higher levels of carbon dioxide and better (about 100 times) than alloy steel in cor- hydrogen sulfide, duplex alloys are required, rosion resistance, has nearly the least corrosion with the 22CR alloys such as UNS S32205 used resistance of all stainless steels. To achieve at temperatures up to 200 ¡C (390 ¡F) and the higher corrosion resistance molybdenum is 25CR alloys such as UNS S32507 at tempera- added. Molybdenum at 1% increases resistance tures up to 250 ¡C (480 ¡F). Nickel-base alloys to general corrosion in a sodium chloride/ are required at hydrogen sulfide levels above 1 hydrogen sulfide/carbon dioxide environment by atm partial pressure. about tenfold. Another 1% increases it another Chapter 22: Petroleum Industry Applications / 251

Fig. 4 Alloy suitability as a function of H2S and CO2 partial pressure. Source: Ref 1 tenfold. The 2% level of molybdenum also Martensitic alloys are susceptible to SCC by a greatly reduces pitting, which in turn eliminates hydrogen embrittlement mechanism. This sus- the initiation point of SCC. Simply adding ceptibility is strongly temperature dependent. It molybdenum would cause the alloy to have ex- decreases with temperature from a maximum at cessive δ-ferrite, which cannot transform to ambient to none at around 100 ¡C (210 ¡F). If the martensite and would therefore reduce mechan- hydrogen sulfide level exceeds 0.03 atm, then ical properties. Thus, nickel must be added to 22Cr alloys should be used rather than 13Cr counter the ferrite stabilizing effect, unfortu- because of this risk. Hydrogen sulfide may be nately, but necessarily increasing the cost. The contained in the petroleum, or it may come from nickel does help lower the general corrosion sulfate-reducing bacteria, introduced by flooding, rate. Carbon and nitrogen in these alloys are for example. This can cause a sulfide-free system kept at low concentrations. These alloys are oth- to become sulfide rich after the fact and make ini- erwise almost identical to precipitation harden- tial materials choice wrong after the fact. ing martensitic stainless steels without the pre- The 22Cr and 25Cr alloys have significantly cipitating phase. higher resistance to chlorides and wet hydrogen 252 / Stainless Steels for Design Engineers

Table 6 Restrictions in use recommended by NACE MR0175 for selected stainless steels used for petroleum industry applications

Common Temperature, Hardness, ClÐ, UNS name PREN(a) ¼C (¼F) pH, min H2S, kPa HRC mg/L J91150 CA15 ...... 3.5 10 22 . . . J91151 CA15M ...... 3.5 10 22 . . . J91540 CA6NM ...... 3.5 10 23 . . . J93254 ...... 100 HRB . . . J95370 ...... 150 (300) . . . 700 94 HRB 90,000 N08926 ...... 121 (250) 3.5 700 . . . 60,700 S15500 15-5 ...... 3.4 33 . . . S15700 15-7 ...... 32 . . . S17400 17-4 ...... 3.4 33 . . . S20910 ...... 66 (150) . . . 100 . . . 35 S41000 410 ...... 3.5 10 22 . . . S41425 ...... 10 28 . . . S41426 ...... 3.5 10 27 . . . S41427 ...... 3.5 10 29 6,000 S41429 ...... 4.5 10 27 . . . S41500 F6NM ...... 3.5 10 23 . . . S42000 420 ...... 3.5 10 22 . . . S42400 ...... 3.5 10 23 . . . S42500 ...... 3.5 10 22 . . . S45000 450 ...... 1 31 . . . S66286 A286 . . . 65 (150) . . . 100 35 . . . Austenitic A-2 ...... 22 . . . Duplex <40 232 (450) . . . 10 ...... >40 <45 232 (450) . . . 20 ...... Superaustenitic, (Ni+2Mo) >30 ...... type 3a Superaustenitic, >40 121 (250) . . . 700 . . . 5,000 type 3b 149 (300) . . . 310 . . . 5,000 171 (340) . . . 100 . . . 5,000

Notes: See NACE MR0175 for further use and processing restrictions. See Appendix 1 for alloy compositions. (a) PREN, pitting resistance equivalent number. Source: Adapted from NACE MR0175, “Sulfide Stress Corrosion Cracking Resistant Metallic Materials for Oil Field Equipment” sulfide and can resist SCC at ten times higher gen. The more common alloys are UNS S32654 concentrations than the 13Cr alloys. Besides the and N08367. inherently greater corrosion resistance that de- The recent development of lean duplex alloys rives from the chromium, molybdenum, and ni- has not yet made its way into OCTGs. These al- trogen levels of the 22Cr and 25 Cr alloys (see loys offer an inherent alloy savings over the the chapters on corrosion in this book), the du- 13Cr grades in nickel and molybdenum content plex alloys have very fine grain size and a while offering better corrosion and SCC resist- roughly 50/50 mixture of ferrite and austenite. ance. Their strength levels in the annealed con- This acts as a crack arrestor should one phase be dition, 450 MPa (65 ksi), are lower than those susceptible to cracking while the other is not. of the martensitic alloys, 600 MPa (87 ksi), so There have been no reported downhole fail- for most downhole applications they will re- ures of annealed or cold-worked duplex alloys. quire cold working. It is likely, however, that There was one instance of very high-strength these alloys will see their first service as line tubing cracking after cathodic contact with car- pipe, where they will not need to be cold worked bon steel casing. This was after removal from to higher strength levels to be widely used. the well and after handling damage had occurred. The affected microstructure was found to be high (70%) in ferrite (Ref 3). Line Pipe and Flow Lines For corrosion resistance above that furnished by superduplex materials such as the 25Cr al- With the awesome cost of corrosion, the loys, super austenitic alloys fill a gap before case for stainless line pipe is easily made. nickel base alloys are needed. These alloys Whether to use stainless depends on vulnera- achieve a tenfold increase in hydrogen sulfide bility of carbon steel. This evaluation is made resistance and very elevated SCC resistance. based on the carbon dioxide, hydrogen sulfide, These are the so-called 6Mo grades. The more water, salinity, temperatures, pressures, flow advanced of them contain high levels of nitro- conditions, and so forth. The normal basis for Chapter 22: Petroleum Industry Applications / 253

these calculations follows that published by C. and nitrogen, both relatively inexpensive alloy- de Waard of Shell (Ref 3). The competing tech- ing elements. Alloys S32001 and S32101 are nology when corrosion dangers arise is the use well formulated for medium and high levels of of corrosion inhibitors, cathodic protection, or corrosion resistance required for wet carbon to line carbon steel with a protective coating. dioxide, hydrogen sulfide, and trace chlorides. The use of inhibitors is subject to the risk of ve- The main precaution for duplex alloys is main- locity limitations, temperature limitations, and taining a nominally 50/50 mixture of ferrite and simply of the inhibitor working appropriately, austenite with no embrittling intermetallic not to mention cost. Cathodic protection is phases. The modern alloys have high (greater costly and complex. Coatings can be damaged than 0.14%) nitrogen, which helps to preserve by numerous occurrences in acidity, mechani- austenite levels after welding and suppress in- cal damage, or fluctuations in temperature or termetallic formation. Nevertheless, minimiza- pressure. Use of a stainless corrosion-resistant tion of time above 350 ¡C (660 ¡F) is important. alloy can have well-defined and controlled This tendency increases with chromium and costs and performance over the life of an instal- molybdenum content, which is another reason lation. Line pipe differs from downhole in hav- why the lean duplex alloys are so attractive. ing strength requirements more in line with that For subsea use, 22Cr duplex generally re- of annealed duplex alloys. These requirements quires cathodic protection because of the risk of gave birth to modern duplex alloys, starting crevice corrosion. 25Cr duplexes are used with- with UNS S31803 and evolving to UNS out cathodic protection. Duplex pipelines have S32205 as the value of higher nitrogen became been in service in the North Sea since the 1970s. understood. Nitrogen not only enhances corrosion resistance, but also suppresses the formation of unde- Umbilical Tubing and Risers sirable and embrittling intermetallic phases that might otherwise form at welding temperatures. Increasingly, wells are located undersea. It is It also keeps the desirable austenite/ferrite ratios standard practice to control and monitor these in weld metal. wells via bundled umbilical tubing. The tubing Since the development of the first widely ac- can provide hydraulic and electrical power, con- cepted duplex alloys, more alloys have trol and adjust pressure, carry communications, emerged. Superduplexes, such as UNS S32750, and even introduce chemical to the well. The have become accepted alloys. Then, the need to depth of wellheads can increase collapse pres- improve costs led in the 1990s to the use of sures to levels beyond the capability of thermo- martensitic alloys with high levels of nickel and plastics, which has led to the use of duplex molybdenum, which at the time were lower stainless steel because of its strength and resist- cost. The emergence in the early 2000s of lean ance to corrosion and SCC. When resistance to duplex alloys provided strength and more corro- seawater is the main concern, the rule of thumb sion resistance with lower nickel levels, giving is that a pitting resistance equivalent number them a cost advantage during periods of high (PREN) of 35 or greater is required, whereas re- nickel cost. sistance to crevice corrosion requires a PREN The main attribute required by line pipe that of at least 40. This has made the superduplex is not as important in OCTGs is weldability. UNS S32750 the standard. Such a critical item This is not an overwhelming challenge for du- as an umbilical may seem like a poor applica- plex alloys, but for martensitic alloys, it re- tion on which to economize, but again the lean quires a very low interstitial level so that the duplexes offer possibilities to do so. By zinc martensite is self-tempering and ductile in the coating lean duplexes such as alloy 19D (UNS as-welded condition. This can be achieved by S32001) and 2101 (UNS S32101), very long stabilizing the alloy with small amounts of tita- service lives can be safely extrapolated. These nium. It would appear that under current condi- alloys are being promoted on their lower sus- tions that alloy 2101 (UNS 32101) has a ceptibility to σ formation during welding, and if cost/performance edge over the martensitic welding thermal cycles cannot be controlled competition and should for the long term. The that may be an issue, but superduplex seems to main ingredients required in a duplex for have become a pervasive choice because it is strength and corrosion resistance are chromium superbly reliable. 254 / Stainless Steels for Design Engineers

Risers are now produced in coiled tubing of Liquefied Natural Gas Vessels over 100 mm (4 in.) diameter, so that very economical long lengths are feasible. Liquefied natural gas (LNG) is becoming an increasingly important commodity as the value of stranded gas makes it economically desirable Platforms to convert it to a transportable state. Converting natural gas to a cryogenic liquid presents a ma- Platforms present a special case in which the terial problem. Vessels to contain it must have costs of maintenance are high, the corrosion en- strength and toughness at temperatures below vironment is severe, and the penalty for excess Ð150 ¡C (Ð240 ¡F). The traditional material, 9% weight is also high. A savings of 1 ton in weight Ni martensitic steel, has become expensive topside can save over $100,000 in steel in the compared to the lower-nickel austenitic stain- subsea jacket. This leads to a rapid payback for less steels, such as 201LN (UNS S20153), the use of materials that are sufficiently resistant which have no transition temperature and to corrosion such that corrosion loss allowance strengthen with decreasing temperature. Alloy can be eliminated. Both titanium and stainless 201LN is cheaper, easier to weld and fabricate, alloy UNS S32750 are equal candidates for this and of course is stainless, which 9% Ni steel is service, depending on availability and current not. The extreme ductility of 201LN compared alloy prices. Except in rare cases, stainless steel to martensitic steel gives it a decided advantage wins the cost battle between these alloy systems. in terms of rupture resistance, which is a major Almost any structure is a candidate for stain- design and political concern with this poten- less topside processing: piping, pumps, flanges, tially explosive commodity. fittings, etc. Hardware of any type and construction materials benefit from being stainless. Sea- water systems often employ 22Cr duplex with Refinery Equipment cathodic protection or unprotected 25Cr duplex. A wise preventive action is to paint stainless Corrosion resistance is a major factor in the that is covered by insulation or similar material, choice of materials in refinery operations. As we which otherwise can result in concentration discussed, crude oil itself is sometimes a very cells and consequent pitting. corrosive fluid, but in refining the by-products,

Table 7 Stainless steels used in various reﬁnery processes

Process Corrosive agents Applications Alloys Notes Crude distillation Sulfur-containing acids Preheaters, distillation 405, 409, 410 … (SCAs) tower Vacuum fractionalization SCA, chlorides Towers 405, 410, 316 Depending on crude corrosivity Condensers S44735, 2205 Depending on chloride level

Coker SCA, H2S Coke drums 409 Depending on crude corrosivity Gas plants Ð Compressor coolers, AL-6XN, 2205, 2507 … H2S, water, Cl , ammonia reboiler tubes Trays 410S, 316L … Amine plant Ammonia, MEA, DEA Reboilers, trays, ﬁlters, 304L, 316L … condenser tubing Sulfuric acid alkylation Sulfuric acid Contactor, mixer 20Cb3 Low pH excursions possible Dilute sulfuric acid Effluent piping 316L …

Hydrotreating H2S, ammonia, PTA(a) Hot sections 321, 347 Long exposure at high temperature General 410S, 304 … Catalytic re-forming High-temperature strength Reactor internals 304 HCl catalyst regeneration needed HCl residue Heat exchangers 2205 … Fluid catalytic cracking High temperature Trays 410 Condensers may need 6Mo Cyclones, vapor lines 304 … Hydrogen plant Tubing, heat exchangers 304 … Hydrocracking Sulﬁdes, chlorides Heat exchangers 409, 321, 347, 2205, Depending on temperature, 6Mo risk of chlorides Sour water stripping Sulfuric acid, ammonium Stripper 304, 20Cb3, 2205 Severity depends on pres- bisulﬁde, chlorides ence of sulfuric acid

(a) PTA, polythionic acid. Source: Ref 4 Chapter 22: Petroleum Industry Applications / 255

chemicals used in refining and the temperatures using austenitics, which are prone to grain used may further aggravate that corrosivity. The boundary chromium depletion by sensitization, aggressive chemical agents that refinery materi- and instead use low-carbon grades and stabi- als must withstand include wet hydrogen sulfide lized grades. and carbon dioxide, napthenic acids, polythionic acids, chlorides, sulfuric acid, and alkalines as well as simple oxidation. Sometimes, tempera- REFERENCES tures of use are such that embrittling or sensitizing phase transformations may occur. Table 7 1. Sumitomo Products for the Oil and Gas In- lists some major refinery processes and the ma- dustries, www.sumitomometals.co.jp, ac- terials used in them (Ref 4). cessed June 2008 Most of these situations are discussed else- 2. H. Asahi et al., “Development of High where in this book in detail. One that is quite Chromium Stainless Line Pipe,” Nippon specific to refinery applications is polythionic Steel Technical Report 72, January 1997 acid (PTA) attack. These acids usually form ac- 3. C. de Waard and U. Lotz, “Prediction of cidentally when sulfide corrosion products react CO2 Corrosion of Carbon Steel,” Paper 69, with moisture and air. The attack is intergranu- presented at Corrosion/93, National Associ- lar, and materials respond to it much as they do ation of Corrosion Engineers, 1993 to the Strauss test. The remedies are to prevent 4. C.P. Dillon, Corrosion Resistance of Stain- the inadvertent formation of PTA and to avoid less Steels, Marcel Dekker, 1995 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 257-263 All rights reserved. DOI: 10.1361/ssde2008p257 www.asminternational.org

CHAPTER 23

Chemical and Process Industry Applications

Summary Single- and Dual-Environment Systems

ENGINEERS IN THE PROCESS industries Under ideal conditions, a material may need must have materials that can contain a huge vari- to resist one single major corrosion threat. If the ety of chemical species at many temperatures, most potentially damaging species can be pressures, and flow rates. This is applied corro- clearly identified, then candidate materials can sion engineering combined with physics and be found by searching published data. These structural design. It is obvious that this task de- data are available freely online from Web sites pends on the availability of corrosion data, more (such as Ref 1 and 2) or for a charge from than can be presented here. This chapter covers sources such as the National Association of what data are necessary and how they can be Corrosion Engineers (NACE; Ref 3) and ASM found. International (Ref 4). It is difficult for any published data to keep up with the latest developments. The testing alone of new materials can Introduction take a long time, and then it must wait for publication. All materials are not covered, especially The need to work with hostile chemicals be- when a manufacturer publishes data on propri- gins with the manufacture of those chemicals. It etary alloys and excludes competitive materials. was in the production of nitric acid that stainless That having been said, any improvements over had its first industrial application. These are in- standard alloys will first be reported by the de- dustries with purely need-driven material chal- velopers of the alloy, and they will logically lenges. New processes are constantly in devel- tout its strongest points. For this reason, dia- opment, and they present new environments in logue with the primary steel producers is en- which materials must perform. The choices are couraged. No one has more exposure to the lat- highly pragmatic. In an industrial environment, est trends in applications. the costs of a poorly performing material can be A single-environment system is typically one well known by its effect on downtime, mainte- in which the aggressive chemical species is the nance, liability, etc. The essential knowledge is only consideration. This is normally the case for which materials will work. piping, tanks, or reaction vessels holding the The selection of materials for the chemical species or materials immersed in the aggressive and power industries is first a study of corrosion species. resistance, including resistance to stress corro- A dual-environment system is typically en- sion cracking (SCC). Strength plays a secondary countered in heat exchangers, but it must also role but can be an important cost factor. These be extended to single-environment systems in considerations may occur at very high or very which the exposure of the nonreactant side of low temperatures, in which case corrosion resist- the material to the ambient environment cannot ance may become oxidation resistance and be neglected, as in the case of marine ambient strength may mean creep strength. environments. 258 / Stainless Steels for Design Engineers

The challenges that must be met are primarily ensuring adequate corrosion resistance and sec- ondarily having acceptable mechanical properties. The corrosion issues run the full gamut of potential forms of corrosion: ¥ General corrosion ¥ Pitting corrosion and crevice corrosion ¥ Intergranular corrosion ¥ Stress corrosion cracking ¥ Erosion corrosion In addition to these forms of corrosion associated with liquids, there are considerations of gas phase attack, which may be oxidation, sulfida- tion, or attack by other gases. Mechanical design considerations are normally limited to static stress allowances. Previ- Fig. 1 Isocorrosion chart for sulfuric acid. Source: Ref 1 ously, handbooks dealt very lightly with this topic because all the normally recommended of carbon steel and titanium gives a valuable steels had similar strength. The proliferation of frame of reference for the engineer. duplex stainless steels has changed that. Now, If the forms of localized corrosion discussed high-strength alloys of high corrosion resistance next can be avoided, the corrosion tables are and SCC resistance are available and are mak- sufficient to guide the designer to a reasonable ing traditionally chosen stainless steels less than selection of candidate materials for any process optimal. in which the chemical species involved have been identified. If the data have not been developed for a certain environment, then the tables Corrosion Types give a first approximation of which materials may be resistant from examination of similar A designer wants to deal with general corro- environments, and a final decision can only be sion. Its rate can be predicted, and thickness reasonably made though direct corrosion testing can be chosen to allow for its occurrence. Cor- of candidate materials. Refer to the chapters on rosion data for general corrosion are normally corrosion for a more thorough discussion of presented in isocorrosion charts. These present uniform corrosion. the temperatures and concentrations for a given environment at which various materials Pitting and Crevice Corrosion will exhibit the same corrosion rate. This rate is most often 0.1 mm/yr, an amount that can be Stainless steel is unique among metals and al- thought of as a tolerable level for many uses. loys in that it derives its corrosion resistance Figure 1 shows an isocorrosion chart for stain- from constituent alloying elements working to- less steels in sulfuric acid (Ref 1). The data are gether to form a thin passive layer that, when clear when presented in this fashion. It can fur- intact, is highly resistant to corrosion. The ther be appreciated that in general reducing strength of the passive layer in resisting attack alloy performance to a mathematical formula, by halide ions, which are the most disruptive such as the pitting resistance equivalent num- ions to the layer, is proportional principally to ber (PREN) equation, would not be reasonable the chromium, nitrogen, and molybdenum con- since the relative performance of alloys tents of the alloy. This relationship follows the changes considerably with concentration. formula: Thus, the design engineer must rely on experi- PREN = %Cr + 3.3%Mo + 30%N (Eq 1) mentally developed data. Since these data are available both online and in print, no attempt This formula is one of the commonly used will be made to reproduce them fully here. Ex- versions, none of which is universally correct. amples are given in Tables 1 and 2 (Ref 1). Both tungsten and carbon can increase pitting Such tables are very useful, although the pre- resistance, while sulfur diminishes it. This is sentation is not visually compact. The inclusion discussed in the corrosion section of this book. Chapter 23: Chemical and Process Industry Applications / 259

The important consideration is that this formula resistance, which is what causes pitting. assumes that the key alloying elements are ho- Chromium is very reactive: Its affinity for oxy- mogeneously distributed in solution. This will gen makes the passive ﬁlm strong. Pitting has only be true if correct thermomechanical pro- nearly always been associated with manganese cessing occurs because, thermodynamically, sulﬁde inclusions, and although there is still de- these alloys are not used in an equilibrium con- bate over the precise mechanism, it appears that dition. Were they to attain equilibrium, say by chromium depletion at the metal-inclusion in- overheating, alloy segregation by precipitation terface is to blame. Eliminating inclusions by could occur, causing localized loss of corrosion eliminating either manganese or sulfur improves

Table 1 Corrosion table for sulfuric acid (H2SO4)

Concentration, % 0.1 0.5 0.5 0.5 1 1 1 1 1 2 2 2 3 3 3 Temperature, ¡C 100 = BP 20 50 100 = BP 20 50 70 85 100 = BP 20 50 60 20 35 50 Carbon steel 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 13% Cr steel 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 18-2 (UNS S44400) 2 0 2 2 0 2 2 2 2 0 2 2 0 2 2 3R12 (UNS S30400) 2 0 1 2 0 1 1 2 2 0 1 1 0 1 1 3R60 (UNS S31600) 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 18-13-3 1 0 0 1 0 0 0 1 1 0 0 0 0 0 0 17-14-4 1 0 0 1 0 0 0 0 1 0 0 0 0 0 0 2RK65 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 (UNS N08904) Sanicro 28 . . . 0 0 0 0 0 0 0 0 0 0 0 0 0 0 (UNS N08028) 254SMO . . . 0 0 . . . 0 0 0 0 1 0 0 0 0 0 0 (UNS S31254) 654 SMO . . . 0 0 . . . 0 0 0 0 0 0 0 0 0 0 0 (UNS S32654) SAF 2304 1 0 0 . . . 0 0 0 0 1 0 0 0 0 0 0 (UNS S32304) SAF 2205 . . . 0 0 1 0 0 0 0 . . . 0 0 0 0 0 0 (UNS S31803) SAF 2507 . . . 0 0 . . . 0 0 0 0 0 0 0 0 0 0 0 (UNS S32750) Titanium 1 0 0 1 0 0 1 1 1 0 0 1 0 0 1

Concentration, % 3 3 5 5 5 5 5 5 10 10 10 10 10 20 20 Temperature, ¡C 85 100 = BP 20 35 60 75 85 101 = BP 20 50 60 80 102 = BP 20 40 Carbon steel 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 13% Cr steel 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 18-2 (UNS S44400) 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3R12 (UNS S30400) 2 2 1 1 2 2 2 2 2 2 2 2 2 2 2 3R60 (UNS S31600) 1 2 0 0 1 1 2 2 0 1 1 2 2 0 1 18-13-3 1 2 0 0 0 1 2 2 0 1 1 2 2 0 1 17-14-4 1 2 0 0 0 1 2 2 0 0 1 2 2 0 1 2RK65 (UNS 0 1 0 0 0 0 1 2 0 0 0 1 2 0 0 N08904) Sanicro 28 . . . 1 0 0 0 0 0 2 0 0 0 0 2 0 0 (UNS N08028) 254SMO . . . 1 0 0 0 0 1 2 0 0 0 0 2 0 0 (UNS S31254) 654 SMO 0 0 0 0 0 0 0 2 ...... 0 0 . . . 0 0 (UNS S32654) SAF 2304 . . . 1 0 0 0 0 0 2 0 0 0 2 2 1 2 (UNS S32304) SAF 2205 . . . 1 0 0 0 0 0 2 0 0 0 1 2 0 0 (UNS S31803) SAF 2507 . . . 1 0 0 0 0 ...... 0 0 0 0 2 0 0 (UNS S32750) Titanium 1 2 0 1 1 2 2 2 1 2 2 2 2 2 2 (continued)

Notes: 0, corrosion rate of less than 0.1 mm/yr. The material is corrosion proof. 1, corrosion rate of 0.1Ð1.0 mm/yr. The material is not corrosion proof but useful in certain cases. 2, corrosion rate of more than 1.0 mm/yr. Serious corrosion. The material is not usable. BP, boiling solution. Source: Adapted from Ref 1 260 / Stainless Steels for Design Engineers

Table 1 (continued)

Concentration, % 20 20 20 20 30 30 30 30 40 40 40 40 50 50 50 Temperature, ¡C 50 60 80 100 20 40 60 80 20 40 60 90 20 40 70 Carbon steel 2 2 . . . 2 2 2 2 . . . 2 2 2 2 2 2 2 13% Cr steel 2 2 . . . 2 2 2 2 . . . 2 2 2 2 2 2 2 18-2 (UNS S44400) 2 2 . . . 2 2 2 2 . . . 2 2 2 2 2 2 2 3R12 (UNS S30400) 2 2 . . . 2 2 2 2 . . . 2 2 2 2 2 2 2 3R60 (UNS S31600) 1 2 . . . 2 1 2 2 . . . 2 2 2 2 2 2 2 18-13-3 1 1 . . . 2 1 1 2 . . . 2 2 2 2 2 2 2 17-14-4 1 1 . . . 2 1 1 2 . . . 2 2 2 2 2 2 2 2RK65 0 0 1 2 0 0 1 . . . 0 0 1 2 0 0 2 (UNS N08904) Sanicro 28 0 0 . . . 2 0 0 1 . . . 0 0 1 2 0 0 1 (UNS N08028) 254SMO 0 0 . . . 2 0 0 1 2 . . . 1 ...... 0 1 . . . (UNS S31254) 654 SMO 0 0 0 2 ...... 0 0 0 . . . 0 0 . . . (UNS S32654) SAF 2304 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 (UNS S32304) SAF 2205 0 1 2 2 0 1 2 2 2 2 2 2 2 2 2 (UNS S31803) SAF 2507 0 0 1 2 . . . 0 1 2 0 1 2 2 1 1 2 (UNS S32750) Titanium 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

Concentration, % 60 60 60 70 70 70 80 80 80 85 85 85 85 90 90 Temperature, ¡C 20 40 70 20 40 70 20 40 60 20 30 40 50 20 30 Carbon steel 2 2 2 2 2 2 2 2 2 0 1 2 2 0 1 13% Cr steel 2 2 2 2 2 2 2 2 2 1 1 2 2 0 1 18-2 (UNS S44400) 2 2 2 2 2 2 2 2 2 1 1 1 2 0 1 3R12 (UNS S30400) 2 2 2 2 2 2 2 2 2 1 1 1 2 0 0 3R60 (UNS S31600) 2 2 2 2 2 2 1 2 2 1 1 1 2 0 0 18-13-3 2 2 2 2 2 2 1 2 2 1 1 1 2 0 1 17-14-4 2 2 2 2 2 2 1 2 2 1 1 1 2 0 1 2RK65 0 1 1 0 1 1 0 1 2 0 0 1 1 0 0 (UNS N08904) Sanicro 28 0 0 1 0 0 1 . . . 1 1 0 0 0 0 0 0 (UNS N08028) 254SMO 0 1 . . . 0 1 . . . 0 1 2 0 ...... 1 . . . (UNS S31254) 654 SMO 0 1 . . . 0 1 ...... 1 . . . (UNS S32654) SAF 2304 2 ...... 1 1 ...... 1 (UNS S32304) SAF 2205 2 2 2 1 ...... 2 2 2 1 ...... 1 1 (UNS S31803) SAF 2507 ...... 2 2 . . . 2 2 1 1 ...... 0 0 (UNS S32750) Titanium 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Concentration, % 90 90 94 94 94 94 96 96 96 96 98 98 98 98 Temperature, °C 40 70 20 30 40 50 20 30 40 50 30 40 50 80 Carbon steel 2 2 0 2 2 2 0 1 2 2 1 1 2 2 13% Cr steel 2 2 0 1 2 2 0 1 2 2 1 1 2 2 18-2 (UNS S44400) 2 2 0 0 2 2 0 0 1 2 0 1 2 2 3R12 (UNS S30400) 2 2 0 0 1 1 0 0 0 1 0 0 2 2 3R60 (UNS S31600) 1 2 0 0 0 1 0 0 0 1 0 0 0 2 18-13-3 1 2 0 0 1 1 0 0 1 1 0 0 1 2 17-14-4 1 2 0 0 1 1 0 0 1 1 0 0 1 2 2RK65 1 2 0 0 1 1 0 0 1 1 0 1 1 2 (UNS N08904) Sanicro 28 0 1 0 0 0 0 0 0 0 1 0 0 0 1 (UNS N08028) (continued)

Table 1 (continued)

Concentration, % 90 90 94 94 94 94 96 96 96 96 98 98 98 98 Temperature, °C 40 70 20 30 40 50 20 30 40 50 30 40 50 80 254SMO 1 ...... 1 ...... 0 2 (UNS S31254) 654 SMO 2 2 ...... 2 0 1 . . . 2 ...... 1 1 (UNS S32654) SAF 2304 1 ...... 1 ...... 0 ...... 0 1 (UNS S32304) SAF 2205 1 . . . 0 ...... 0 0 1 . . . 0 0 1 1 (UNS S31803) SAF 2507 0 . . . 0 0 0 1 0 0 0 1 0 0 0 1 (UNS S32750) Titanium 2 2 2 2 2 2 2 2 2 2 2 2 2 2

Table 2 Corrosion table for fuming sulfuric When a crevice is permitted to exist, it mim- acid (oleum), H2SO4 + SO3 ics the pH-altering action found within pits in which transport restriction leads to a buildup of Conc. H SO , % 100 100 100 100 100 100 2 4 metal and hydrogen ions and oxygen depletion. Conc. SO3, % 7 11 11 60 60 60 Temperature, °C 60 60 100 20 70 80 All alloys undergo crevice corrosion under Carbon steel 0 0 2 ...... less-aggressive conditions than those required 13% Cr steel 0 0 2 ...... 2 18-2 0 0 ...... to induce pitting, so care must be taken to avoid (UNS S44400) crevices. 3R12 0 0 1 0 0 0 (UNS S30400) 3R60 0 0 0 0 0 0 Intergranular Corrosion (UNS S31600) 18-13-3 0 0 ...... 0 . . . Intergranular corrosion is a problem that can 17-14-4 0 0 . . . 0 0 . . . 2RK65 0 0 . . . 0 0 . . . be avoided entirely by correct alloy selection (UNS N08904) and proper thermal processing. The principle Sanicro 28 (UNS ...... cause of grain boundary attack is alloy deple- N08028) 254SMO ...... tion at the grain boundaries. The most familiar (UNS S31254) form of this problem occurs when austenitic 654 SMO ...... (UNS S32654) alloys having carbon levels above 0.03% are SAF 2304 ...... welded. The region near the weld where tem- (UNS S32304) peratures reach 600 to 900 ¡C (1100 to 1650 ¡F) SAF 2205 ...... (UNS S31803) may have carbon migrate to and along grain SAF 2507 ...... boundaries, the fast diffusion paths, where it (UNS S32750) Titanium 2 2 2 2 2 2 combines with less-mobile chromium atoms and precipitates as chromium carbide. This Notes: 0, corrosion rate of less than 0.1 mm/yr. The material is corrosion proof. 1, corrosion rate of 0.1Ð1.0 mm/yr. The material is not corrosion proof but useful in lowers the chromium level in solution, result- certain cases. 2, corrosion rate of more than 1.0 mm/yr. Serious corrosion. The ing in poor corrosion resistance only at the material is not usable. Conc., concentration. Source: Adapted from Ref 1 grain boundaries. This is easily prevented by selecting alloys with low carbon levels. Duplex the potential at which passive film breakdown alloys, curiously, undergo chromium carbide occurs. This is especially important for welds, precipitation under the same conditions but do which, if not annealed, can have maximum not undergo significant chromium depletion deleterious segregation by both inclusions and because the neighboring ferrite grains, in which solidification segregation. All austenitic and chromium diffuses more rapidly, contribute duplex stainless alloys have best corrosion re- chromium, mitigating the depletion. Precipita- sistance when quenched from the solution an- tion segregation of all types, not just by carbides, nealing temperature. The precipitation harden- must be guarded against. Sigma phase, nitrides, ing, martensitic, and ferritic alloys are more secondary austenite, and others can cause local complicated but are less relevant to this topic. If breakdown of corrosion resistance if alloys are information on them is needed, they are dis- heated to a dangerous temperature for sufficient cussed in detail in their respective chapters. time. It is important to learn these potential 262 / Stainless Steels for Design Engineers

vulnerabilities by reviewing the metallurgy of austenitics. Designing within this limit is sensi- any alloy selected for service. ble practice. And, if alloy selection uses a rule of avoiding situations in which pitting can occur, SCC will also be avoided even if stress Stress Corrosion Cracking excursions occur since in general pitting is a necessary precondition for SCC. The theory of SCC is still under debate. The reader will find the arguments confusing as the debate generates more heat than light. We will Erosion skip the theory; it can be found in the corrosion chapters. SCC, like excessive general corrosion Flow velocities can reach levels at which ero- or pitting, is avoided by referring to published sion becomes problematic, especially if hard test data from the corrosion tables. If a material particles are suspended in a fluid. Assuming that must be used where a risk of SCC occurs, then the material has sufficient corrosion resistance stress levels must be managed to stay below the to survive well in the static environment, the threshold stress for SCC. Figure 2 shows how best performance under erosive conditions is various alloys resist SCC as a function of chlo- obtained by materials with higher surface hard- ride concentration and temperature, the two ness. Accordingly, the duplex perform better most important aggravating factors. Material than austenitic alloys of the same corrosion re- comparisons are made difficult because tests are sistance level. normally run at a given fraction of a material’s yield strength. Thus, the data in Fig. 2 (Ref 1) must be interpreted. Higher-strength duplex alloys, while having better SCC performance than Specific Environments austenitics of equal corrosion resistance (e.g., 316 vs. 2304), have much better SCC resis- The list of specific environments against tance. Furthermore, the stress at which failure which stainless steels are sufficiently resistant will occur is much higher since the yield to select for use in the chemical process indus- strength at which the testing takes place is about tries is too long to provide here. Some of the twice as high for duplex alloys. SCC also ex- most important specific corrosives, such as hibits a threshold stress below which failure nitric, sulfuric, phosphoric, hydrochloric, and does not occur. This is about 60% of tensile organic acids and others, are covered in the strength for duplex and about 30% for chapter on corrosion. The main caution to the designer is to make sure that the source material from which design guidance is sought is current. Many otherwise excellent handbooks are somewhat obsolete in that they do not include the very importance duplex stainless steel family or only include the oldest alloys in the group, such as 2205 (UNS S32205). Many new alloys now exist that range in corrosion performance from that of 316 to that of the 6Mo-plus-N austenitics. These alloys are usable in all gauges, have high strength and toughness, resist SCC, and can achieve the corrosion resistance levels of any ferritic or austenitic alloy. They can also provide significant savings in alloy cost at the same corrosion level because they have lower nickel levels.

REFERENCES Fig. 2 Stress corrosion cracking (SCC) resistance in neutral chloride solutions containing 8 ppm oxygen. Testing time, 1000 h. Applied stress equal to proof strength at testing 1. Sandvik Materials Technology, www.smt. temperature sandvik.com, accessed June 2008 Chapter 23: Chemical and Process Industry Applications / 263

2. Outokumpu Corrosion Handbook for Stain- 4. D.B. Anderson and B.D. Craig, Handbook less Steels, www.outokumpu.com, accessed of Corrosion Data, 2nd ed., ASM Interna- June 2008 tional, 1995 3. P.A. Schweitzer, Corrosion Resistance Tables, 5th ed., National Association of Cor- rosion Engineers, NACE 37755, 2004 \aq2\ Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 265-267 All rights reserved. DOI: 10.1361/ssde2008p265 www.asminternational.org

CHAPTER 24

Pulp-and-Paper Industry Applications

Summary exclusive domain of duplex stainless steels because of their lower cost per unit of corrosion THE PULP-AND-PAPER INDUSTRY has resistance, high strength, and near immunity to seen more benefits from developments in stain- stress corrosion cracking. Pricing changes less steel than any other industry. The harsh mainly with alloying element costs, principally chemicals used in this industry called for better those of nickel and molybdenum. At prices be- materials than the normal austenitic stainless tween the highs and lows of the first decade of steels without the expense of the 6Mo grades. the 2000s, duplex costs have been roughly one- This need has been met through the use of the third less than that of an equivalent corrosion- duplex alloys, which have become the new resisting austenitic stainless. This factored in standard. with strength nearly double that of the equivalent austenitic make them an overwhelmingly superior choice for pulp-and-paper equivalent Introduction except if very special corrosion requirements differ from the norm, such as in bleaching. The proximity of the Scandinavian paper industry to that region’s specialty steel industry has been symbiotic. As a result of the strong Paper-Making Processes interaction between engineers having well- specified needs for improved materials and met- The kraft (German for “strong”) process was allurgists capable of providing them, the ad- introduced in 1937, replacing the sulfite process. vances in materials in the pulp-and-paper In the kraft process, the lignin-connecting wood industry have been a model of rapid technology fibers are dissolved under conditions of elevated transfer and innovation. Beginning in 1988, du- temperature and pressure in acidic conditions of plex stainless steels first went into production in pH 2.0 to 4.0. This leaves a long fiber, which en- kraft digesters, and there has been no turning ables paper of high strength, hence the name back in the replacement of austenitic stainless kraft. Over the years the materials used for the steels by duplex. So, a discussion of the materi- vessels, called digesters, in which this process is als selection for the pulp-and-paper industry has carried out have been sequentially carbon steel, changed from a fairly complicated analysis of stainless steel, and stainless steel clad onto car- which austenitic steel to use while guarding bon steel. In the previous sulfite process, acid- against stress corrosion cracking and pitting resistant brick vessels were used. Now, the di- corrosion and when to use clad materials for gesters, essentially large vertical tanks, are cost savings, to a fairly simple discussion of constructed of 2205 (UNS S32205) as a rule (see which duplex stainless steel is most economical Fig. 1). for a given piece of equipment. Since this revo- The digestion is typically carried out at 150 lution occurred in the 1990s before the latest and 180 ¡C (300 and 360 ¡C) and 10 to 12 bar. surge in nickel prices, it is safe to say that future The pH of the sulfate is around 2.0 to 4.0. In pulp-and-paper projects will be essentially the this environment, 316L can survive, but it 266 / Stainless Steels for Design Engineers

pulp suspension is injected at high velocity. The environment is a mixture of alkaline liquid, while the vapor phase can contain organic acids. The hardness of the duplex helps mitigate erosion, while the alloy level is beneficial against corrosion. 2205 is the alloy of choice here, but 2003 (UNS S32003) would suffice. The next step, washing and screening, has seen increasingly severe environments as closed systems required for pollution control have become more common. This has rendered the previous choice of carbon steel untenable. This stage also sees erosion potential from hard particles, such as sand, in the pulp. The optimum solution is a lean duplex such as 2101, 2304, or 2003. The delignification of the pulp comes next. This oxygen process dates from the 1970s. At first, highly alloyed austenitic alloys were used. Subsequently, it was found again that duplex performed better in that they were suffi- Fig. 1 The first kraft digester fabricated from alloy 2205. ciently corrosion resistant, but also offered Courtesy of Outokumpu freedom from stress corrosion cracking as well as materials savings because of their higher strength. The bleaching of the pulp is important for many types of paper, and this can be done by requires maintenance and has a finite life. The chlorine bleaches or ozone/peroxide bleaches. 2205 is twice as resistant to corrosion, 0.005 The chlorine bleaching now must generally be mm/yr versus 0.011 mm/yr (Ref 1). In the done in closed systems, which results in a nonchloride environment, molybdenum is not buildup of chloride levels to a point at which an essential alloying element, so the introduc- corrosion levels are unacceptable unless very tion of the use of 2101 (UNS S32101) or 2304 highly alloyed materials are used. The 6Mo (UNS S32304) is a logical cost-saving move grades have been successful, but now they can without strength or corrosion compromises. The be replaced by duplex alloys such as 2507 reduction in wall thickness allowed by the (UNS S32750), which again save cost by virtue higher-strength duplex depends on the engi- of their higher strength. neering code required. The American Society of Bleaching can be accomplished without chlo- Mechanical Engineers (ASME) code require- rine in the so-called TCF, totally chlorine free, ment is based on tensile strength and permits process. This reduces the corrosivity of the en- only a 24% reduction in wall thickness, while vironment as the ozone and hydrogen peroxide the total kjeldahl nitrogen (TKN) code, based used in the process are relatively harmless to on yield strength, would allow a 46% reduction. stainless steel. Alloys such as 316 are adequate This large a difference in strength levels re- for this environment, but lean duplex, 2101 or quired by codes is unfortunate and reflects an 2304, offer cost reductions through their greater orientation to materials in which the yield/ strength. tensile ratio is closer to unity, unlike either du- In plants that use recycled paper and mechan- plex or austenitic stainless steel. In the more un- ical wood chip processing, the materials selec- usual case of digesters using the sulfite process, tion criteria remain the same. Duplex stainless the materials selected would be the same. has become the clear choice. As one proceeds downstream in the process, Further downstream, containers and process environments change greatly, but the optimal equipment benefit equally from duplex down to materials remain duplex for various reasons. the handrails and walkways. This wholesale use The subsequent stage is blow tanks in which the of duplex can make plants nearly maintenance Chapter 24: Pulp-and-Paper Industry Applications / 267

free from a corrosion point of view, a dramatic REFERENCES change in an industry in which the thousand-fold greater corrosion rates of carbon steel presented 1. A. Tuomi et al., Duplex America 2000 Con- operators with endless equipment downtime ference, Houston, KCI Publishing, 2000 problems. 2. H. Dykstra et al, Corrosion in the Pulp and Additional detail about corrosion challenges Paper Industry, Corrosion: Environments and the use of stainless steels in the pulp-and- and Industries, Vol 13C, ASM Handbook, paper industry can be found in Ref 2. ASM International, 2006, p 762Ð802 Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 269-278 All rights reserved. DOI: 10.1361/ssde2008p269 www.asminternational.org

APPENDIX 1

Compositions 270 / Stainless Steels for Design Engineers ...... Other V 0.1Ð0.3 V 0.15Ð0.40 0.03Ð0.08 Ce ...... Other Cu 1.75Ð2.25 Nb 0.1Ð0.3 Nb 0.75Ð1.25 Cu 3.0Ð4.0 Se 0.15 min B 1.00Ð1.20 Ce 0.04 1.0 Al Cu 0.50 Al 0.8Ð1.5 S . . . . . 0.040 0.030 0.030 0.030 0.030 0.030 0.015 0.015 0.030 0.18Ð0.35 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.15 min 0.06 min 0.25 min 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 P . . . . . 0.045 0.060 0.045 0.045 0.045 0.060 0.045 0.045 0.040 0.060 0.040 0.060 0.040 0.060 0.060 0.060 0.040 0.040 0.060 0.060 0.040 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 Si 0.75 1.00 1.00 1.00Ð2.00 1.00 0.75 3.00Ð4.00 1.00 1.00 1.00 1.00 1.00 3.50Ð4.50 0.30Ð1.00 1.20 1.00 1.00 1.00 1.00 1.00 3.2Ð4.0 1.4Ð2.0 1.00 0.75 1.00 1.0 1.00 1.00 1.00 1.00 1.00 1.00 2.00Ð3.00 1.00 1.00 1.00 1.00 0.75 3.75Ð4.25 1.00 0.75 Mo ...... 0.5 0.2 2.0Ð3.0 1.5Ð3.0 2.0Ð3.0 2.0 Composition, % 0.60 optional 0.60 optional 0.60 optional Mn . . 5.0Ð6.5 7.0Ð9.0 2.0 2.0 4.0Ð6.0 8.0Ð10.0 5.5Ð7.0 7.5Ð9.0 2.0 2.0 2.0 0.8 2.0 0.8 5.5Ð7.5 5.5Ð7.5 5.5Ð7.5 7.5Ð10.0 7.0Ð9.0 8.0Ð10.0 4.0Ð6.0 7.5Ð9.0 2.0 2.0 2.0 2.5Ð4.5 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 11.5Ð14.5 11.0Ð14.0 14.0Ð15.5 14.5Ð16.0 14.0Ð16.0 (continued) Ni 5.0Ð6.5 1.0Ð1.75 5.5Ð7.5 9.0Ð11.0 8.0Ð10.5 8.0Ð10.5 3.5Ð5.5 3.5Ð5.5 3.5Ð5.5 4.0Ð6.0 4.0Ð6.0 1.5Ð3.0 0.5Ð2.5 2.25Ð3.75 5.5Ð7.5 8.0Ð9.0 0.75 5.0Ð6.0 5.0Ð7.0 7.5Ð9.0 6.0Ð8.0 6.0Ð8.0 8.0Ð10.0 8.0Ð10.0 8.0Ð10.0 8.0Ð10.0 8.0Ð10.0 7.0Ð10.0 8.0Ð10.5 8.0Ð10.5 8.0Ð10.5 8.0Ð10.5 9.0Ð10.0 11.5Ð13.5 12.0Ð15.0 10.0Ð12.0 10.0Ð12.0 12.0Ð15.0 10.5Ð13.0 13.5Ð16.0 12.0Ð15.0 14.0Ð15.5 Cr 22.0Ð24.0 17.0Ð19.0 16.0Ð18.0 16.5Ð18.0 17.0Ð19.0 19.0Ð21.5 17.0Ð19.0 16.0Ð18.0 16.0Ð18.0 16.0Ð18.0 15.0Ð18.0 16.5Ð19.0 20.5Ð23.5 16.0Ð18.0 17.0Ð18.5 14.0Ð16.0 17.5Ð22.0 17.5Ð22.0 16.0Ð18.0 16.5Ð18.0 17.0Ð19.0 17.0Ð19.0 17.0Ð19.0 17.0Ð19.0 17.0Ð19.0 18.0Ð20.0 18.0Ð20.0 18.0Ð20.0 18.0Ð20.0 18.0Ð20.0 18.0Ð20.0 18.0Ð19.0 17.0Ð19.0 17.0Ð19.5 20.0Ð22.0 22.0Ð24.0 15.0Ð17.0 19.0Ð21.5 17.0Ð19.0 18.0Ð20.0 17.0Ð18.5 19.9Ð21.0 N ...... 0.08Ð0.20 0.25 0.32Ð0.40 0.15Ð0.40 0.25Ð0.50 0.10Ð0.16 0.12Ð0.18 0.25 0.20Ð0.40 0.20Ð0.40 0.20Ð0.40 0.08Ð0.18 0.35 0.10Ð0.20 0.10 0.10 0.10Ð0.16 0.14Ð0.20 0.25 0.25 0.15Ð0.30 0.15Ð0.40 0.25Ð0.50 0.25Ð0.50 0.16Ð0.30 0.10 C 0.08 0.15 0.08 0.04Ð0.06 0.05Ð0.10 0.03 0.03 0.15 0.08 0.0Ð3 0.08 0.04 0.10 0.03 0.03 0.15 0.03 0.08 0.03 0.12 0.15 0.15 0.12Ð0.25 0.08 0.06 0.12 0.12 0.08 0.15 0.15 0.08 0.15 0.15 0.15 0.04Ð0.10 0.08 0.08 0.018 0.16Ð0.24 0.08 0.20 0.08 S21800 S30400 S20100 S20103 S20153 S20161 S20300 S20500 S20400 S24100 S24300 S21900 S21904 S21400 S21600 S21603 S30153 S30200 S30215 S30300 S30223 S30403 S30451 S30424 S30415 S30500 S30600 S30800 S20200 S20910 S21460 S21500 S30100 S30430 S30310 S30409 S30452 S30453 S30615 S30815 S30900 S30908 Designation(a) Tenelon 230 EZ 304HN 305 RA 85 H 253MA 309S All compositions include Fe as balance. Single values are maximum unless otherwise noted. (a) Unified Number System, UNS numbers are S or N followed by 5 digits. Notes: Table A1.1 Table Composition of austenitic stainless steels Name 201 201L 201LN Gall-Tough 202 205 Nitronic 30 Nitronic 32 Nitronic 33 Nitronic 40 (219) 21-6-9 LC Nitronic 50 Nitronic 60 Tenelon Cryogenic Esshete1250 216 216L 301 301LN 302 302Cu 302B 303 303Se 303 Plus X 304 304L 304H 304N 304LN 304BI 153MA Cronifer 1815 308 309 Appendix 1: Compositions / 271 Other . Co 0.05 ...... Al 0.15Ð0.60 Al 0.15Ð0.60 ...... Co 0.2 . Co 0.2 Other to 0.70 to 0.70 to 1.10 to 1.10 to 1.10 to 1.0 to 1.10 Nb+Ta 0xC Nb+Ta Nb 10xC to 1.10 . . Nb 10xC to 1.10 . 0.15Ð0.60 Ti Ce 0.03Ð0.10 Nb 10xC to 1.10 Nb 10xC to 1.10 . . . . Nb 10xC to 1.10 ...... Ti 5xC to 0.70 . . 0.030 max P . 5x(C+N) Ti 4x(C+N) Ti . 0.15Ð0.60 Ti Nb 0.25Ð0.60 10xC Nb+Ta 10xC Nb+Ta 8xC Nb+Ta 0.10Ð0.40 Ti 10xC Nb+Ta S . . . . . 0.030 0.030 0.015 0.030 0.030 0.030 0.030 0.030 0.015 0.030 0.030 0.030 0.10 min 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.010 0.030 0.030 P . . . . . 0.040 0.045 0.045 0.045 0.20 0.045 0.045 0.045 0.045 0.045 0.030 0.045 0.045 0.045 0.040 0.040 0.045 0.045 0.045 0.045 0.020 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 0.045 Si 1.00 1.00 0.75 1.00 1.00 1.50Ð2.50 1.00 1.00 1.00 1.00 1.00 0.5 1.50Ð2.50 1.50Ð3.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.75 1.00 1.00 0.75Ð1.50 1.00 1.00 1.00 1.00 0.5Ð1.0 0.6Ð1.0 1.00 1.00 1.00 1.00 0.75 1.00 Mo ...... 1.5Ð2.5 2.0Ð3.0 2.0Ð3.0 2.0Ð3.0 2.0Ð3.0 2.0 2.0 2.0 1.6Ð2.6 2.0Ð3.0 1.75Ð2.5 2.0Ð3.0 2.0 2.0 2.0Ð3.0 Composition, % Mn 3.0Ð4.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 4.0Ð5.0 2.0 2.0 1.0Ð1.5 1.0 2.0 2.0 1.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 3.0Ð4.0 3.0Ð4.0 4.0Ð5.0 2.0 1.5 2.0 2.0 1.65Ð2.35 2.0 (continued) Ni 9.0Ð12.0 9.0Ð13.0 9.0Ð12.0 9.0Ð13.0 9.0Ð13.0 9.0Ð13.0 11.0Ð15.0 11.0Ð13.0 11.0Ð15.0 11.0Ð15.0 10.0Ð14.0 12.0Ð16.0 19.0Ð22.0 19.0Ð22.0 19.0Ð22.0 19.0Ð22.0 10.0Ð14.0 34.0Ð37.0 30.0Ð35.0 31.0Ð35.0 12.0Ð15.0 12.0Ð16.0 19.0Ð22.0 19.0Ð22.0 19.0Ð22.0 20.5Ð23.5 19.0Ð22.0 10.0Ð14.0 10.0Ð14.0 10.0Ð14.0 10.0Ð14.0 10.0Ð14.0 13.5Ð17.5 13.5Ð17.5 19.0Ð21.0 14.5Ð16.5 17.0Ð19.0 34.0Ð36.0 Cr 22.0Ð24.0 22.0Ð24.0 24.0Ð26.0 24.0Ð26.0 18.0Ð20.0 17.0Ð19.0 22.0Ð24.0 19.0Ð21.0 24.0Ð26.0 24.0Ð26.0 24.0Ð26.0 24.0Ð26.0 24.0Ð26.6 24.0Ð26.0 23.0Ð26.0 16.0Ð18.0 16.0Ð18.0 16.0Ð18.0 16.0Ð18.0 16.0Ð18.0 16.0Ð18.0 16.0Ð18.0 18.0Ð20.0 18.0Ð21.0 18.0Ð20.0 17.0Ð20.0 17.0Ð19.0 17.0Ð19.0 17.0Ð20.0 19.0Ð23.0 20.0Ð23.0 18.0Ð20.0 17.0Ð19.0 17.0Ð19.0 17.0Ð19.0 12.5Ð14.5 15.0Ð17.0 24.0Ð26.9 N ...... 0.10 0.10Ð0.16 0.10Ð0.20 0.11 0.15Ð0.35 0.09Ð0.15 0.11 0.10Ð0.16 0.10 0.10 0.10 0.10 0.10 0.10 0.005 0.12Ð0.18 0.10 0.10Ð0.22 0.10 C 0.04Ð0.10 0.20 0.03 0.08 0.08 0.08 0.04Ð0.10 0.04Ð0.10 0.25 0.08 0.08 0.25 0.04Ð0.10 0.08 0.08 0.03 0.03 0.08 0.03 0.08 0.04Ð0.10 0.03Ð0.05 0.08 0.08 0.04Ð0.10 0.04Ð0.10 0.04Ð0.10 0.02 0.08 0.08 0.03 0.03 0.03 0.40Ð0.10 0.10 0.08 0.08 0.08 DIN 1.4828 DIN 1.4841 S30941 S31042 S31609 S31703 S37000 S30909 S30940 S31000 S31008 S31009 S31040 S31041 S31050 S31400 S31600 S31620 S31603 S31653 S31651 S31635 S31700 S31753 S31725 S31726 S32100 S32109 S33000 S35125 S33400 S34700 S34709 S34800 S34809 S38400 S35315 Designation(a) N08800 Name 309HCb 309Si 310 316H 316Ti 317 317L 347 347H 348 370 All compositions include Fe as balance. Single values are maximum unless otherwise noted. (a) Unified Number System, UNS numbers are S or N followed by 5 digits. Notes: Table A1.1 (continued) Table 309H 309Cb 310S 310H 310Cb 310HCb 310HCbN 310MoLN 310Si 314 316 316F 316L 316LN 316N 317LN 317LM 317LMN 321 321H 330 332 332Mo* 348H 334 384 353MA 272 / Stainless Steels for Design Engineers ...... Other Nb+Ta 0.6Ð1.4 Cu W 2.5-3.5 0.4-0.7 Ti W 1.0-3.0 Al 0.025 0.15Ð0.60 Al 0.25Ð0.60 Nb Cu 3.0Ð4.0 Cu 0.5Ð1.5 2.0Ð4.0 Cu 0.5 Cu 8xC to 1.00 ...... Other 0.040 P 2.5Ð3.5 Cu 0.05Ð0.10 Ce 0.15Ð0.60 Ti 0.4Ð1.0 Ti 0.1Ð0.35 Ti 8xC to 1.00 Nb Nb 0.15-0.35 8xC Nb 0.030 P 8xC to 0.5 Nb 1.0Ð2.0 Cu 0.5Ð1.0 Cu 0.10 Nb 0.3Ð0.6 Cu 3.0Ð4.0 Cu V 2.5 Co,0.35 0.5Ð1.5 Cu 1.0Ð1.4 Cu 1.0Ð2.0 Cu 0.5Ð3.0 Cu 3.0Ð4.0 Cu S ...... 0.030 0.030 0.035 0.030 0.030 0.030 0.015 0.030 0.035 0.030 0.030 0.035 0.010 0.005 0.035 0.015 0.015 P ...... 0.050 0.045 0.045 0.030 0.040 0.045 0.045 0.0035 0.030 0.030 0.030 0.045 0.030 0.030 0.045 Si 1.00 0.40 1.5 0.03 1.0 0.6Ð1.0 0.25 0.70Ð1.25 0.69Ð0.90 0.03Ð0.80 1.00 0.50 0.50 1.00 1.00 1.00 0.80 1.0 0.5 1.0 0.08 0.50 0.75 0.30 1.0 0.5 0.25 1.00 1.0 Mo ...... 2.0Ð3.0 6.0Ð7.0 5.0Ð6.7 3.0Ð4.0 4.0Ð5.0 4.0Ð5.0 7.0Ð8.0 2.5Ð3.0 4.5Ð6.5 5.0Ð7.0 2.5Ð3.5 2.0Ð3.0 4.0Ð6.0 1.0Ð1.75 3.5Ð5.0 6.0Ð7.0 4.3Ð5.0 6.0Ð6.5 2.0Ð3.0 6.0Ð7.0 6.0Ð7.0 4.0Ð4.8 12.5Ð14.5 Composition, % Mn . 0.75 2.0 1.0 1.0 8.0Ð10.0 1.0Ð1.5 0.75Ð1.50 1.0 2.0 2.0 5.0Ð7.0 2.0Ð3.0 1.0 1.5 1.5 7.0Ð9.50 1.5Ð3.5 0.75Ð1.50 1.0 2.5 2.0 2.0 1.0 2.0 2.0 2.0 2.0 2.5 Ni 3.25Ð4.50 1.50Ð2.75 7.9Ð9.0 8.0Ð11.0 balance 30.0Ð38.0 16.0Ð18.0 24.0Ð26.0 31.0Ð33.0 32.0Ð37.0 32.0Ð38.0 33.0Ð37.0 29.9Ð32.5 21.0Ð23.0 23.0Ð27.0 27.5Ð30.5 25.0Ð27.0 10.5Ð12.5 35.0Ð40.0 23.5Ð25.5 23.5Ð25.5 24.0Ð26.0 23.0Ð28.0 17.50Ð18.50 32.0Ð38.0 24.0Ð26.0 26.0Ð28.0 24.0Ð26.0 21.0Ð24.0 Cr 25.0Ð29.0 18.0Ð21.0 22.0Ð26.0 26.0Ð28.0 19.0Ð21.0 20.0Ð22.5 26.0Ð28.0 20.0Ð23.0 19.0Ð22.0 26.0Ð28.0 20.0Ð25.0 20.0Ð22.0 19.25Ð21.50 20.0Ð22.0 22.0Ð24.0 19.0Ð21.0 22.5Ð25.0 20.0Ð22.0 20.0Ð22.0 19.0Ð23.0 19.0Ð23.0 19.50Ð20.50 23.0Ð25.0 24.0Ð25.0 19.0Ð21.0 22.0Ð24.0 24.0Ð26.0 23.0Ð26.0 21.0Ð23.0 N ...... 0.15Ð0.25 0.28Ð0.50 0.15Ð0.25 0.28Ð0.38 0.18Ð0.25 0.18Ð0.22 0.40Ð0.60 0.45Ð0.55 0.15Ð0.25 0.15Ð0.25 0.35Ð0.60 0.20Ð0.40 C 0.15Ð0.25 0.030 0.07 0.04Ð0.08 0.06Ð0.10 0.08 0.48Ð0.58 0.07 0.03 0.03 0.02 0.04 0.02 0.02 0.03 0.02 0.015 0.04 0.015 0.02 0.03 0.03 0.50Ð0.60 0.28Ð0.38 0.28Ð0.36 0.035 0.07 0.02 0.05 . S63017 S63012 S63018 S31254 S34565 S63008 S63198 S32654 S31266 S32200 N08020 N08024 N08367 N08700 N80904 N06022 N08932 N08007 N08026 N08028 N08366 N08020 N08926 N08031 20Mod S33228 S25045 S35135 Designation(a) 1925 hMo 3127HMo Name Incoloy 864 Table A1.1 (continued) Table AC66 20Cb-3 AL-6X 904L B66 N08320 Incoloy 803 21-4N 21-2N 21-12N 23-8N 19-9DL 20Mo-4 20Mo-6 Sanicro 28 AL-6XN JS-700 254SMO 4565 654SMO AL 20 AL 22 Cronifer Cronifer 2328 Nicrofer URSB-8 NIC 25 CN-7M All compositions include Fe as balance. Single values are maximum unless otherwise noted. (a) Unified Number System, UNS numbers are S or N followed by 5 digits. Notes: Appendix 1: Compositions / 273 ...... Other Al 1.0 Al 0.10Ð0.30 Al Al 0.25 Al 0.15 SE 0.15 AL 1.2 AL 3.0 AL 4.0 AL 0.15 ...... Nb 0.35 0.17 0.60 0.55 Ti to 0.75 to 0.15Ð0.50 Ti+Nb to 0.75 Nb to 1.10 to 1.10 to 0.75 Ti+Nb 4x(C+N) to 1.10 . . 6x(C+N) 6x(C+N)to 0.5 8x(C+N) 8x(C+N) 0.8+ 8x(C+N) 10xC . . 0.40 0.40 . . . . 0.20+4x(C+N) 0.20+4x(C+N) 0.20+4x(C+N) 0.35 Ti+Nb:0.20+ 0.25 0.30 . min S ...... 0.030 0.030 0.030 0.045 0.030 0.030 0.030 0.030 0.060 0.030 0.030 0.150 0.030 0.030 P ...... 0.040 0.045 0.040 0.040 0.040 0.040 0.040 0.040 0.060 0.040 0.040 0.040 0.040 0.060 . Si 1.0 1.30 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.0 1.00 1.00 1.0 0.03 1.0 1.00 1.00 1.0 0.45 0.45 1.00 0.03 1.00 Composition, % Mo ...... (continued) . Mn 1.00 1.00 1.00 1.00 1.25 1.00 1.00 1.00 1.00 1.00 0.75 1.00 1.00 0.25 0.035 0.035 0.70 1.00 1.25 1.00 0.35 1.00 0.30 Ni ...... 0.50 0.75 0.60 0.50 0.50 0.05 0.50 0.5Ð1.0 0.20 0.75 0.50 0.50 0.2 0.50 0.50 1.00 Cr 10.5Ð11.75 18.0 11.5Ð14.5 10.5Ð11.75 10.5Ð11.75 10.5Ð11.7 11.35 12.0 13.0 13.5 14.0Ð16.0 16.0Ð18.0 16.0Ð19.5 17.5 12.0Ð13.0 10.5Ð11.7 10.5Ð11.7 10.5Ð11.7 13.0 16.0Ð18.0 16.0Ð18.0 17.0Ð19.0 17.0Ð19.0 18.0Ð20.0 N ...... 0.030 0.030 0.020 0.03 0.015 0.015 0.04 0.04 0.03 0.01 0.020 C 0.03 0.020 0.10 0.07 0.03 0.01 0.08 0.05 0.08 0.03 0.02 0.02 0.06 0.03 0.010 0.025 0.025 0.08 0.12 0.12 0.12 0.12 0.030 0.020 UNS typical typical alloys typical typical typical typical designation AK alloy AK alloy alloy ATI S43020 S43036 AK ATI, AK alloy S40500 S40900 S40910 S40920 AK alloy S40930 S40940 S40975 AK alloy alloy ATI Outukumpu S42900 S43000 S43023 S43035 S43932 S46800 Table A1.2 Table Composition of ferrite stainless steels 409 409 466 409Cb CrÐCb 11 Alfa I 4724 430 439 439LT 439 HP 439 ultraform 18 CrÐCb All compositions include Fe as balance. Single values are maximum unless otherwise noted Notes: Name 405 400 409 409 ultraform 409Ni 12 SR Alfa II 429 430F 430Se 430Ti 468 274 / Stainless Steels for Design Engineers ...... Other 0.10 REM 0.5 Cu+Ni Al 1.7 Al 1.0 Al 1.5 0.60 Al 0.2 Cu Nb ...... 9xC 5xC:0.70 0.3Ð1.0 0.5Ð0.20 Nb+Ta 10x(C+N) Ti ...... 4x(C+N) to 0.80 4x(C+N) to 0.80 4x(C+N) to 0.80 4x(C+N) to 0.80 0.25 0.1Ð0.6 Ti+Nb:0.20+ 0.02 8x(C+N) min Ti+Nb:0.20+ Ti+Nb:0.20+ Ti+Nb:0.20+ S ...... 0.030 0.030 0.030 0.030 0.030 0.030 0.020 0.030 0.030 0.030 P ...... 0.040 0.040 0.040 0.040 0.040 0.040 0.040 0.020 0.040 0.040 Si 1.4 0.3 1.00 1.3 1.0 1.0 1.0 1.0 1.0 0.4 0.40 0.75 0.75 0.4 1.00 . .. Composition, % Mo ...... 0.75Ð1.25 1.2 2.5Ð3.5 0.75Ð1.25 0.75Ð1.25 0.75Ð1.25 3.5Ð4.5 3.5Ð4.5 Mn 0.7 1.0 0.30 0.7 0.30 1.0 1.00 1.0 0.20 1.0 0.3 0.40 1.00 1.00 1.00 1.50 Ni . . . . . 1.5Ð3.5 0.3 1.0 0.25 0.50 3.5Ð4.5 0.5 0.6 0.3 0.6 0.25 Cr 18.0 17.30 17.5Ð18.5 17.3 24.5Ð26.0 23.0Ð27.0 16.0Ð18.0 16.0Ð18.8 18.0Ð23.0 17.5Ð19.5 20.0 24.0 22.0 25.0Ð27.5 25.0Ð27.0 28.0Ð30.0 N ...... 0.035 0.035 0.25 0.015 0.015 0.035 C 0.12 0.01 0.20 0.08 0.12 0.030 0.20 0.01 0.025 0.08 0.03 0.025 0.025 0.025 0.015 0.01 UNS typical typical typical typical typical designation AK alloy typical Outukumpu S43600 Outukumpu alloy ATI S43400 S44200 alloy ATI alloy ATI S44627 S44635 S44660 S44600 S44100 S44400 S44735 4509, 430J1L 190-EM Table A1.2Table (continued) 18SR 453 446 All compositions include Fe as balance. Single values are maximum unless otherwise noted. Notes: Table A1.2 Table Composition of ferrite stainless steels Name 4742 434 436 441, 442 436S 444, YUS 433 4762 E-Brite, 26-1 Monit SeaÐcure 29-4C Appendix 1: Compositions / 275 ...... 5.5 V 1.2 V 4.0 V 9.0 V Other 1.5 Cu 1.5 Cu 0.5 Cu Cu 0.30 CU 2.0Ð3.0 Se 0.15 min Se 0.15 min Se 0.15 min W 2.50Ð3.50 0.75Ð1.25 W Nb 0.05Ð0.30 S ...... 0.030 0.030 0.06 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.005 0.030 0.15 0.30 0.060 0.030 0.030 0.15 0.030 0.030 0.010 0.030 0.030 0.030 0.10Ð0.35 0.060 P ...... 0.040 0.060 0.040 0.040 0.040 0.060 0.040 0.040 0.040 0.040 0.040 0.040 0.040 0.040 0.060 0.040 0.040 0.040 0.040 0.040 0.040 0.040 0.040 0.040 0.040 0.040 0.040 Si ...... 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.35 0.3 0.50 1.00 1.00 1.00 0.60 0.60 1.00 1.00 0.50 1.00 1.00 0.75 0.30Ð0.6 1.00 1.00 1.00 0.3 0.3 Mo ...... 1.00 1.00 1.00 0.75 1.00 1.00 1.00 1.00 0.35 0.3 0.50 1.00 1.00 1.00 0.60 0.60 1.00 1.00 0.50 1.00 1.00 0.30Ð0.60 1.00 1.00 1.00 1.00 1.00 0.3 0.3 Composition, % Mn ...... 0.45 1.00 1.00 0.50Ð1.0 1.25 0.50 1.25 1.00 1.00 1.00 2.0 1.00 1.00 1.00 1.25 0.4 0.45 1.00 1.00 1.00 1.50 1.50 0.5Ð1.0 1.25 1.00 1.25 0.50Ð1.0 1.00 0.45 1.45 1.00 1.25 0.5 0.40 Ni ...... 1.5 1.25Ð2.50 3.50Ð5.50 1.80Ð2.20 0.50Ð1.00 1.00Ð2.00 5.0 5.8 1.25Ð2.50 1.5 4.0Ð7.0 3.50Ð4.50 0.50 4.0 4.5 2.4 5.5 0.75 0.75 Cr 13.0 12.7 12.3 11.0 14.0 13.0 14.5 13.0 12.0 14.5 14.0 14.0 14.0 17.0 11.5Ð13.5 11.5Ð13.5 11.5Ð14.0 11.0Ð13.5 11.5Ð13.5 11.5Ð13.5 11.5Ð13.5 10.5Ð12.5 12.0Ð14.0 12.0Ð14.0 14.0Ð16.0 10.5Ð12.5 12.0Ð15.0 12.0Ð14.0 12.0Ð14.0 12.0Ð14.0 12.0Ð14.0 12.0Ð14.0 13.0Ð14.0 12.0Ð14.0 15.0Ð17.0 16.0Ð18.0 16.0Ð18.0 16.0Ð18.0 16.0Ð18.0 16.0Ð18.0 N ...... 0.06 0.12 0.040 0.015 0.010 0.010 C 0.03 0.05 max 0.15 max 0.15 min 0.20Ð0.25 0.025 2.15 2.20 0.15 max 0.15 max 0.08 0.18 max 0.030 max 0.15 max 0.05 0.15 max 0.15Ð0.20 0.15 min 0.50 0.15 min 0.06 max 0.08Ð0.20 0.50Ð0.55 0.30 max 0.025 0.03 0.02 0.01 0.01 0.20 max 0.60Ð0.75 0.75Ð0.95 0.95Ð1.20 0.95Ð1.20 0.95Ð1.20 1.15 1.05 1.05 1.05 1.45 UNS designation S41500 S42023 JFE nominal S43100 Nominal PM S41000 S41003 S41040 S41400 S41425 S41600 DIN 1.4116 nominal S42200 . S44003 Nominal PM S40300 S41008 S41003 S41623 S41800 S42000 S42020 S42400 S42500 . JFE nominal Nippon nominal Nippon nominal JFE nominal JFE nominal S44002 S44004 S44020 S44023 Nominal PM Nominal PM Nominal PM Nominal PM Nominal PM 410 410S 420FSe 440C 440FSe ATS-34 Table A1.3 Table Composition of martensitic stainless steels Name 403 410 410Cb 412 414 414 mod 415 416 416Se 418 420 4116 420F 422 424 425 425 mod Trinamet HP13Cr-1 HP13Cr-2 NT-CRS NT-CRSS KL-12Cr KL-HP 12Cr 431 440A 440B 440F BG-42 14-4 CrMo CPM S30V CPM S90V All compositions include Fe as balance. Single values are maximum unless otherwise noted Notes: 154 CM CPM S60V 276 / Stainless Steels for Design Engineers W ...... 0.8Ð1.2 0.1Ð0.5 0.5Ð1.0 1.5Ð2.5 . . . Other Al 0.4 Nb .015Ð0.45 Nb 0.15Ð0.45 Al 0.90Ð1.35 Nb 8XC Al 1.0Ð1.5 Al 0.75Ð1.5 Al 0.75Ð1.5 Al 0.35 B 0.001Ð0.010 W 1, V 0.3 Nb +Ta 0.1Ð0.5 Nb +Ta Cu 1.0 1.4Ð2.0 1.0 0.8 1.0 2.0 0.6 0.75 0.75 1.0 1.0 1.0 1.0 1.0 0.8 0.5 0.8 0.8 Other S Ti 0.4Ð1.2 Ti Cu 2.5Ð4.5 Cu 1.25Ð1.75 Co 8.0Ð9.0 . . Cu 3.0Ð5.0 . Cu 1.5Ð2.5, 0.08Ð1.4 Ti 1.50Ð1.80 Ti . . 1.9Ð2.35 Ti V 0.10Ð.050 Co 14 0.010 0.030 0.020 0.030 0.020 0.030 0.030 0.040 0.030 0.030 0.020 0.020 0.030 0.030 0.020 0.020 0.020 0.010 S . 0.030 0.008 0.030 0.010 0.030 0.030 0.030 0.030 0.030 0.030 0.010 0.030 0.030 P 0.030 0.040 0.030 0.040 0.045 0.030 0.030 0.040 0.040 0.030 0.040 0.035 0.040 0.035 0.030 0.035 0.030 0.030 P . 0.040 0.010 0.015 0.040 0.040 0.040 0.040 0.030 0.040 0.015 0.040 0.040 0.040 Si 0.75 1.0 1.0 1.0 0.5 0.8 1.0 0.75 1.4Ð2.0 1.0 1.0 2.0 0.8 1.0 1.0 0.8 0.6 0.8 Si 1.00 1.00 0.10 0.25 0.50 1.00 0.50 1.00 0.10 1.00 1.00 0.50 1.00 0.50 Mo . . 1.2Ð2.0 1.0Ð2.5 2.5Ð3.0 0.1Ð0.8 3.0Ð3.5 3.0Ð5.0 3.0Ð4.0 1.0Ð2.0 2.5Ð3.5 2.5Ð3.5 1.5Ð2.0 2.9Ð3.9 3.0Ð5.0 3.0Ð5.0 1.5Ð2.6 2.5Ð3.5 Mo 2.0 . . . 0.5 . 2.0Ð2.5 0.5Ð1.0 4.5Ð5.5 2.0Ð3.0 1.0Ð1.5 2.5Ð3.25 2.5Ð3.25 0.75Ð1.25 Composition, % Composition, % Mn 2.0 2.0 1.0 1.0 1.0 1.2Ð2.0 2.0 4.0Ð6.0 1.0 2.5 1.5 1.5 2.0 1.0 0.8 4.0Ð6.0 1.2 0.8Ð1.5 Mn 1.00 1.00 1.0 1.0 1.0 0.25 0.50Ð1.25 0.50Ð1.25 1.0 0.2 0.50 0.50 2.00 0.10 Ni Ni 5.5Ð6.0 5.5Ð7.5 4.25Ð5.25 2.5Ð3.5 1.35Ð1.70 4.5Ð6.5 5.5Ð8.0 6.0Ð8.0 6.0Ð8.0 3.5Ð5.2 6.5Ð8.0 2.5Ð5.0 1.0Ð3.0 3.0Ð4.0 3.0Ð5.0 6.0Ð8.0 5.8Ð7.5 6.0Ð8.0 7.5Ð8.5 4.0Ð5.0 5.5 6.0Ð7.5 3.5Ð5.5 7.5Ð9.5 10.75Ð11.25 7.5Ð8.5 6.5Ð7.75 14.0Ð27.0 3.0Ð5..0 5.0Ð7.0 6.5Ð7.75 4.0Ð5.0 Cr Cr 24.0Ð26.0 24.0Ð26.0 18.0Ð19.0 21.0Ð23.0 19.5Ð21.Ð 21.0Ð22.0 22.0Ð23.0 21.5Ð23.5 24.0Ð26.0 24.0Ð27.0 24.0Ð26.0 24.0Ð26.0 26.0Ð29.0 24.0Ð26.0 24.0Ð26.0 23.0Ð28.0 19.5Ð21.5 28.0Ð30.0 14.0Ð16.0 15.0Ð16.0 15.0Ð17.0 14.0Ð15.5 12.25Ð13.25 11.0Ð12.50 10.5Ð11.50 14.0Ð16.0 13.5Ð16.0 115.5Ð17.5 11.0Ð12.50 16.0Ð18.0 16.0Ð17.0 10.0 N . N ...... 0.14Ð0.20 0.10Ð0.30 0.05Ð0.10 0.20Ð0.25 0.20Ð0.35 0.08Ð0.20 0.05Ð0.17 0.14Ð0.20 0.14Ð0.20 0.05Ð0.20 0.10Ð0.25 0.20Ð0.30 0.20Ð0.30 0.30Ð0.40 0.15Ð0.35 0.24Ð0.32 0.23Ð0.33 0.010 0.07Ð0.13 0.07Ð0.13 C C 0.03 0.03 0.03 0.03 0.03 0.03 0.04 0.03 0.03 0.03 0.03 0.08 0.30 0.03 0.04 0.03 0.03 0.025 0.08 0.08 0.07 0.07 0.05 0.05 0.05 0.02 0.01 0.09 0.09 0.07Ð0.11 0.10Ð0.15 0.21 . S31200 S32003 S32101 S32205 S32906 S32950 S32900 S31260 S31500 S31830 S32001 S32304 S32520 S32550 S32750 S32760 S39274 S39277 S15500 S45500 S46500 S15700 S17600 S17400 S13800 S45000 S17600 S17700 S35000 S35500 S66286 Designation Designation 15-5 PH 255 DP3W Table A1.4 Table stainless steels Composition of selected precipitation–hardenable Name Stainless W 17-4 PH 13-8 PH Custom 450 Custom 455 Custom 465 Custom 475 17-7 PH 15-7 PH AM-350 AM-355 A-286 Ferrium S53(a) All compositions include Fe as balance. Single values are maximum unless otherwise noted. (a) Nominal value Notes: A1.5 Table stainless steels Composition of selected duplex Name 329 44LN DP3 3RE60 2205 (old) 19 D 2003 2205 2304 Uranus 52N+ 2507 2906 7-Mo Plus AF 918 All compositions include Fe as balance. Single values are maximum unless otherwise noted Notes: 2101 Zeron 100 Appendix 1: Compositions / 277 Other ...... 0.20Ð0.40 S V 0.1Ð0.3 Nb, (10xC)Ð1.2 Nb Se 0.2Ð0.35 Cu 1.0 Cu 1.4Ð1.9 Cu 2.75Ð3.25 Cu 1.75Ð2.5 Cu 0.5Ð1.0 Cu 0.5Ð1.0, WCu 0.5Ð1.0, 0.5Ð1.0 Cu 2.75Ð3.25 2.5Ð3.2 Cu, 0.2Ð0.35Nb 0.9Ð1.25 W, 0.2Ð0.3V 0.9Ð1.25 W, 2.5Ð3.2 Cu, 0.2Ð0.35Nb Nb 8XC min S 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 P 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.17 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 Si 0.65 1.50 1.50 1.00 1.00 1.00 1.10 1.00 1.00 1.00 1.50 2.00 2.00 1.50 2.00 2.00 3.5Ð4.5 2.00 1.00 1.50 2.00 2.00 1.00 1.50 1.50 1.50 1.00 1.00 1.00 1.00 1.00 2.00 2.00 2.00 1.50 1.50 2.00 2.00 1.50 2.00 2.00 1.00 1.00 Mo ...... 1.75Ð2.25 0.4Ð1.0 2.9Ð3.8 1.75Ð2.25 2.0Ð3.0 2.0Ð3.0 1.5 6.0Ð7.0 6.0Ð7.0 0.50(b) 0.15Ð1.00 0.50(b) 0.9Ð1.25 2.5Ð3.5 3.0Ð4.0 1.75Ð2.25 1.75Ð2.25 4.0Ð5.0 3.0Ð4.5 2.0Ð3.0 2.0Ð3.0 2.0Ð3.0 1.5Ð3.0 4.5Ð5.5 Composition, % Mn 0.50 0.70 1.00 1.00 1.00 1.50 1.50 1.50 1.50 1.50 1.50 7.0Ð9.0 1.50 4.0Ð6.0 1.50 1.50 1.20 2.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.5Ð1.0 0.70 1.50 1.20 1.00 1.50 1.00 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 2.00 2.00 Ni 1.0 3.5Ð4.5 3.6Ð4.6 4.5Ð5.5 9.0Ð12.0 9.0Ð12.0 8.0Ð11.0 1.0 1.0 2.0 4.0 4.5Ð6.5 5.6Ð6.7 4.75Ð6.0 4.0Ð6.0 6.0Ð8.0 8.0Ð11.0 8.0Ð11.0 8.0Ð12.0 8.0Ð11.0 8.0Ð11.0 9.0Ð12.0 9.0Ð12.0 9.0Ð13.0 1.0 6.0Ð8.0 0.5Ð1.0 6.5Ð8.5 4.7Ð6.0 8.0Ð12.0 9.0Ð13.0 9.0Ð12.0 8.0Ð9.0 11.5Ð13.5 12.0Ð15.0 17.5Ð19.5 19.0Ð22.0 23.0Ð27.0 23.0Ð27.0 13.0Ð16.0 10.0Ð13.0 12.0Ð15.0 12.0Ð15.0 (continued) Cr 11.0Ð12.5 11.5Ð14.0 11.5Ð14.0 11.5Ð14.0 11.5Ð14.0 11.5Ð14.0 15.5Ð17.7 21.0Ð23.5 24.5Ð26.5 24.0Ð27.0 24.0Ð26.0 22.5Ð25.5 18.0Ð21.0 22.0Ð26.0 19.5Ð20.5 20.0Ð22.0 26.0Ð30.0 10.5Ð12.5 14.0Ð15.5 24.0Ð26.7 24.5Ð26.5 26.0Ð30.0 17.0Ð21.0 17.0Ð21.0 18.0Ð21.0 18.0Ð21.0 18.0Ð21.0 18.0Ð21.0 15.0Ð18.0 16.0Ð18.0 18.0Ð21.0 18.0Ð21.0 20.5Ð23.5 20.0Ð23.0 22.0Ð26.0 23.0Ð27.0 20.0Ð22.0 18.0Ð22.0 24.0Ð26.0 17.0Ð21.0 18.0Ð21.0 18.0Ð21.0 22.0Ð26.0 N ...... 0.10Ð0.30 0.15Ð0.25 0.10Ð0.30 0.18Ð0.24 0.05 0.05 0.20Ð0.30 0.10Ð0.25 0.10Ð0.20 0.08Ð0.18 0.22Ð0.33 0.10Ð0.30 0.20Ð0.40 0.18Ð0.24 C 0.30 0.06 0.10 0.12 0.15 0.15 0.40 0.2Ð0.4 0.06 0.20Ð0.28 0.07 0.07 0.03 0.03 0.04 0.04 0.06 0.03 0.08 0.30 0.03 0.03 0.08 0.08 0.08 0.04Ð0.10 0.10 0.12 0.16 0.20 0.06 0.08 0.04Ð0.10 0.20 0.025 0.20 0.03 0.03 0.30 0.03 0.03 0.04Ð0.10 0.08 . UNS J91153 J91154 J91150 J91151 J92110 J92900 J92602 J93401 J94653 J94651 J91803 J92613 J91540 J91422 J92180 J92205 J93373 J93370 J93371 J93345 J93423 J92500 J92800 J92700 J92600 J92590 J92901 J92971 J93790 J93000 J93001 J93402 J94652 J91650 J93380 J93372 J93404 J92710 J92972 J92701 J93400 J94202 designation ¨ Wrought equivalent(a) 309S 410 . 420F . 422 (S32760) . 316L 316LN 304 347 304H Nitronicª60 303 309 ALÐ6XN 420 431,442 446 S41500 17Ð4PH 15Ð5PH 2205 (S32205) 255 (S32550) . . 2507 (S32750) . 312 304L 316 316H 316H 316 302 Nitronicª50 317 308 309H 254SMOª 310 904L Table A1.6 Table casting alloys heat– and corrosion–resisting Casting Institute (ACI) Alloy Composition of CD-3MCuN CE-8MN CF-8M CF-10SMnN CF-20 CH-20 CK-20 CN-3MN Table A1.6Table (continued) Name alloys CorrosionÐresisting CA-15 CA-15M CA-40 CA-40F CB-30 CC-50 CA-6N CA-6NM CA-28MWV CB-7Cu-1 CB-7Cu-2 CD-3MN CD-3MWCuN CD-4MCu CD-4MCuN CD-6MN CE-3MN CE-30 CF-3 CF-3M CF-3MN CF-8 CF-8C CF-10 CF-10M CF-10MC CF-16F CG-6MMN CG-8M CG-12 CH-10 CK-3MCuN CN-3M The wrought equivalent composition is not the same as cast. (b) Mo an intentional addition. All compositions include Fe as balance. Single values are maximum unless otherwise noted. (a) Notes: CF-12M CH-8 278 / Stainless Steels for Design Engineers Other ...... W 4.0Ð6.0, Zr 0.1Ð1.0 Cu 3.0Ð4.0 Nb 0.5Ð1.5 Cu 1.5Ð2.0 S 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.035 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 P 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.035 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 Si 1.50 0.50Ð1.5 1.5 2.0 2.0 2.0 2.0 3.50 0.35Ð0.65 1.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 Mo . 0.9Ð1.2 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) 2.0Ð3.0 2.5Ð3.0 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) 0.5(b) Composition, % Mn 0.15Ð1.5 2.0 2.0 2.5 2.5 2.5 2.5 1.50 1.50 1.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.5 2.5 Ni . 4 max 4Ð7 8Ð11 9Ð12 11Ð14 19.0Ð22.0 33.0Ð37.0 58Ð62 22.0Ð25.0 31.0Ð34.0 18Ð22 19.0Ð22.0 18Ð22 23Ð27 33Ð37 33Ð37 27.5Ð30.0 14Ð18 33Ð37 37Ð41 64Ð68 Cr 8Ð10 19.0Ð22.0 18.0Ð20.0 19.0Ð21.0 23.0Ð27.0 13.0Ð17.0 26Ð30 26Ð30 19Ð23 24Ð28 26Ð30 23.0Ð27.0 28Ð32 19Ð23 24Ð28 24Ð28 13Ð17 10Ð14 15Ð19 26Ð30 24Ð38 17Ð21 N ...... 0.2 C 0.20Ð0.40 0.20Ð0.50 0.35Ð0.75 0.35Ð0.75 0.07 0.07 0.05Ð0.15 0.50 0.20Ð0.50 0.20Ð0.50 0.25Ð0.35 0.20Ð0.60 0.20Ð0.50 0.45Ð0.55 0.35Ð0.75 0.25Ð0.35 0.35Ð0.75 0.20 0.50 0.20Ð0.60 0.35Ð0.45 0.35Ð0.75 . UNS J94203 J94650 J92605 J93005 J93403 J93505 J94224 J94204 J82090 J92603 J94003 J94213 N08605 N08007 N08151 N08604 N08705 N08005 N08006 N06050 N08603 designation Wrought equivalent(a) 504 312 302B 309 . . 320 . . 446 327 310 . . . . . 330 . . . . Name CT-15C HA Table A1.6Table (continued) CN-7M CN-7MS alloys Heat resisting HC HD HE HF HI HK HK-30 HK-40 HL HN HP HP-50WZ HT HT-30 HU HW HX The wrought equivalent composition is not the same as cast. (b) Mo an intentional addition. All compositions include Fe as balance. Single values are maximum unless otherwise noted. (a) Notes: HH Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 279-280 All rights reserved. DOI: 10.1361/ssde2008p279 www.asminternational.org

APPENDIX 2

Physical and Mechanical Properties of Select Alloys

Table A2.1 Physical properties of major stainless steel engineering alloys

Coefficient of Thermal Electrical Modulus of thermal exp., conductivity, Speciﬁc heat, resistivity, UNS Density, kg/dm3 elasticity, GPa 10Ð6 × KÐ1 W/Má¡K J/kgá¡K Ωámm2/m 201 S20100 7.86 207 16.6 16.3 502 0.67 301 S30100 8.03 193 16.6 16.3 500 0.73 304 S30400 7.90 200 16.6 16.3 500 0.72 304L S30403 7.90 200 16.6 16.3 500 0.72 305 S30400 7.90 200 16.6 16.3 500 0.72 316L S31603 8.00 200 16.5 14.6 480 0.74 321 S32100 7.92 193 16.6 16.3 500 0.72 904L N08904 7.95 190 15.3 13.2 460 0.95 AL6-XN© N08367 8.06 200 15.3 11.8 474 0.89 409 S40920 7.76 200 10.5 25.0 477 0.60 430 S43000 7.70 200 10.3 23.9 460 0.60 439 S43035 7.70 200 10.2 24.2 460 0.63 468 S46800 7.76 200 10.5 25.0 477 0.60 410 S41000 7.65 200 10.5 24.9 460 0.56 2101 S32101 7.8 200 13.5 17.0 500 0.80 2003 S32003 7.72 210 13.5 17.0 510 0.80 2205 S32205 7.8 200 14.6 16.5 500 0.80 2507 S32750 7.8 200 12.5 13.5 500 0.80 280 / Stainless Steels for Design Engineers

Table A2.2 Typical minimum mechanical properties of representative stainless steel engineering alloys

Tensile strength, Name Condition UNS Yield strength, MPa MPa Elongation, % Hardness

201 Annealed S20100 260 min 550 min 40 min 100 Rb max

201F 2B S20100 330 700 51 89 Rb

301 Annealed S30100 205 min 515 min 40 min 95 Rb max

301 tensile 2D S30100 320 850 49 88 Rb 1/ hard 301 4 S30100 580 900 32 25 Rb 1/ hard 301 2 S30100 815 1150 23 35 Rc 3 301 /4 hard S30100 1000 1270 17 40 Rc

301 Full hard S30100 1160 1380 12 42 Rc

301 sink 2D S30100 270 690 57 82 Rb

304 Annealed S30400 205 min 515 min 40 min 92 Rb max

304 Hot rolled S30400 335 640 51 86 Rb

304 2D S30400 265 625 55 81 Rb

304 2B S30400 305 635 52 85 Rb

304 #4 polish S30400 325 650 51 85 Rb

304 2BA S30400 315 640 53 85 Rb 1 304 /4 hard S30400 705 890 23 29 Rc

304L Annealed S30403 170 min 485 min 40 min 92 Rb max

304L 2D S30403 255 590 53 80 Rb

304LT 2D S30403 255 600 51 81 Rb

304DD 2D S30400 270 610 55 82 Rb

304EDD 2D S30400 260 600 56 78 Rb

305 Annealed S30500 170 min 485 min 40 min 88 Rb max

305 2D S30500 245 560 52 73 Rb

316L Annealed S31603 170 min 485 min 40 min 95 Rb max

316L 2B S31603 310 595 51 82 Rb

321 Annealed S32100 205 min 515 min 40 min 95 Rb max

321 2B S32100 285 570 49 78 Rb

904L Annealed N08904 220 min 490 min 35 min 90 Rb max

904L 2B N08904 270 605 50 79 Rb © AL6-XN Annealed N08367 310 min 690 min 30 min 100 Rb max © AL6-XN 2B N08367 365 745 47 88 Rb

409 Annealed S40920 170 min 380 min 20 min 88 Rb max

409 2D S40920 260 440 31 60 Rb

430 Annealed S43000 205 min 450 min 20 min 89 Rb max

430 2B S43000 345 515 27 67 Rb

439 Annealed S43035 205 min 415 min 22 min 89 Rb max

439 2D S43035 315 455 32 76 Rb

468 Annealed S46800 205 min 415 min 22 min 90 Rb max

468 2D S46800 205 415 32 76 Rb

29-4C Annealed S44735 415 min 550 min 18 min 25 Rc max

29-4C 2D S44735 550 650 20 20 Rc

410 Annealed S41000 205 min 450 min 20 min 96 Rb max

410 2B S41000 320 515 28 81 Rb 2101 Annealed S32101 530 min 700 min 30 min ... (a) Finish conditions: 2D is cold rolled, annealed, and pickled; 2B is 2D with an added temper mill pass (approximately 0.5% reduction); 2BA is cold rolled, bright annealed, and temper passed. Stainless Steels for Design Engineers Copyright © 2008 ASM International® Michael F. McGuire, p 281-283 All rights reserved. DOI: 10.1361/ssde2008p281 www.asminternational.org

APPENDIX 3

Introduction to Thermo-Calc and Instructions for Accessing Free Demonstration Version

WITHIN THE MAIN BODY of this text- Thermo-Calc is used in conjunction with book, a number of diagrams have been plotted thermodynamic databases containing polyno- and attributed to a software package called mial functions that describe the Gibbs energies Thermo-Calc. The purpose of this appendix is of the different phases according to certain to give a brief introduction to Thermo-Calc, ex- models that take into consideration nonideal plain what it is, and what are its uses. Also pro- chemical interactions in solution phases. These vided are instructions for accessing a demon- databases are based on the critical evaluation of stration version of the software. thermodynamic and phase equilibria data for binary, ternary, and some higher-order systems, which are then assembled into self-consistent What Is Thermo-Calc? databases. Different databases are available for different broad classifications of materials, sys- Thermo-Calc (Ref 1) is a powerful, flexible tems, or applications. For example, there are software package available from Thermo-Calc databases for steels and iron-based alloys; iron- Software AB for performing various kinds of based slags; nickel superalloys; aluminum, thermodynamic and phase diagram calculations magnesium, titanium, and zirconium alloys; ce- for multicomponent systems. mented carbides; nuclear materials; and more. The software is based on the so-called CAL- Further information on the different databases PHAD (CALculation of PHAse Diagrams) available can be found at the Thermo-Calc Web method (Ref 2), which describes mathemati- site: www.thermocalc.com cally the thermodynamics of a system through a The thermodynamic database for steels (Ref representation of the Gibbs energies of the dif- 3), as developed by Thermo-Calc Software AB, ferent crystalline phases relevant to that system was used in conjunction with Thermo-Calc for and defined by the chemical composition of the all the calculations made during the preparation system. Thermo-Calc minimizes the total Gibbs of this book. The version of the database used energy of the system with respect to various for these calculations contains data for 20 ele- constraints such as temperature, pressure, and ments and 85 phases. chemical composition and thus predicts the Although the databases are based primarily most stable energy state (or equilibrium state) on the critical assessments of binary, ternary, that can form. By suspending certain phases and some quaternary systems, the CALPHAD (i.e., manually removing certain selected phases methodology provides a theoretical framework from the system and thus restricting the forma- on which extrapolations can be made to predict tion of such phases), Thermo-Calc can also be the phase equilibria for higher-order, multi- used to investigate meta-stable equilibria-type component systems (the higher the order of the problems. system, the weaker the nonideal interaction 282 / Stainless Steels for Design Engineers

parameters become). Thermo-Calc can there- be used to perform calculations for most appli- fore be used in conjunction with such databases cations involving phase equilibria, meta-stable to make predictions for multicomponent sys- equilibria, phase transformations, phase dia- tems and alloys of industrial relevance as illus- grams, and various thermodynamic properties, trated by some of the examples given in the as well as critical assessments and data evalua- main body of this book. These calculations can tions for multicomponent systems. be validated against real alloy data if this infor- While many types of calculations can be mation is available but is not based on (or ad- made using Thermo-Calc, the software typically justed to fit) such higher-order alloy data. is used to predict: Higher order in this sense means more than four ¥ elements (i.e., larger than a quaternary system). Stable and meta-stable phase equilibria for Four specific types of calculation can be per- binary, ternary, and higher-order systems formed using Thermo-Calc, although the range (calculations for alloy compositions with 6, of problems to which these can be applied is 10, 15 elements are not uncommon, as illus- broader: trated by some of the examples in the main body of this book). 1. Single-Point Equilibria: The temperature, ¥ Amounts of phases (mass, volume and mole pressure, composition/activity of a compo- fractions) formed (phase balance) as a func- nent (or the amount of a phase) are fixed and tion of temperature, pressure, and composi- the stable or meta-stable equilibrium for the tion and also the chemical compositions of specified conditions is calculated. the phases formed 2. Step: The amount of one state variable pa- ¥ Phase transformation temperatures such as rameter (or condition) can be changed, while liquidus, solidus, and solvus temperatures. the other conditions remain fixed. For exam- Phase transformation temperatures can be ple, to see how the different phases and their predicted based on the actual chemistry (not amounts and compositions would vary with nominal chemistry). temperature for a given alloy, one would ¥ Thermochemical data such as enthalpies, “step in” temperature. Alternatively, one can heat capacity, and activities vary the composition of one of the compo- ¥ Driving forces for precipitation nents/elements and calculate how the phase ¥ Phase diagrams (isothermal and isoplethal amounts change for a fixed temperature or sections for multicomponent, multiphase predict how the solidus or liquidus would systems as illustrated in this book) change with varying alloy composition. ¥ Molar volume, density, and thermal expan- 3. Map: Two axis variables (such as tempera- sion ture, pressure, composition, or activity of ¥ Scheil-Gulliver (nonequilibrium) solidifica- the components) are changed at the same tion simulations time. Isoplethal sections are generated by varying temperature and composition of one Thermo-Calc is not restricted just to model- of the components. Isothermal sections are ing the alloy. Complex systems representing the result of varying the amounts of two of processing, for example, can also be consid- the components for a fixed temperature. Ex- ered. For example, another application is to cal- amples of each of these kinds of diagrams culate the carbon potential of multicomponent are given in the main body of the text. gas phase systems as a function of composition, 4. Scheil: Thermo-Calc includes a Scheil- temperature, and pressure and then predict what Gulliver model for nonequilibrium solidifi- phases an alloy might form at a given tempera- cation and a modified Scheil model that con- ture when exposed to such a carbon potential. siders partial equilibrium for components Thermo-Calc can thus be applied to a number that are selected by the user. of practical problems related to metallurgy, processing, in-service performance, etc. as summarized by: Applications of Thermo-Calc ¥ Alloy Design: Modification of alloy chemistries to improve properties or reduce costs Thermo-Calc is used around the world within using calculations to guide which composi- academia, in government research laboratories, tions may be most suitable before preparing and by commercial industry. The software can them for testing Appendix 3: Introduction to Thermo-Calc / 283

¥ Heat Treatment: Prediction of formation and are supplied with only certain small data- of problematic phases prior to thermal pro- bases that are for demonstration purposes and cessing the evaluation of the software only. ¥ Casting: Calculation of liquidus and solidus A link to register and download the demon- temperatures; calculation of thermodynamic stration version of the software can be ac- properties of the alloy for input into casting cessed via a link on the Thermo-Calc web site modeling codes at www.thermocalc.com. All fields in the regis- ¥ Welding and Joining: Prediction of the tration form should be completed before contin- phases formed at the joining of two dissimilar uing to the download page, where further in- materials or the interaction with filler material structions regarding installation of the software ¥ Quality Control: Investigation of properties will be provided. and phase balance within designated compo- On installation of the software, additional sitional tolerances documentation, including a Users Guide/Exam- ples manual in the form of PDF files will also be More examples are available in the literature installed. Technical support for the demo ver- (search on key terms Thermo-Calc or CAL- sions of the software is limited, but problems PHAD). A list of published articles citing related to installation or general inquiries can be Thermo-Calc is available at www.thermocalc. addressed by visiting www.thermocalc.com and com. linking to their support. The demo version will run for approximately 1 month on a single computer, and installation How to Obtain a Free Demonstration on a network system is not supported. If you Version of Thermo-Calc wish to run the software after the demo license has expired, it can be downloaded again (i.e., Thermo-Calc is available in two formats: obtaining a new demo license). Thermo-Calc Classic, which has a command line interface and can be run under a number of REFERENCES different operating systems (including Mi- crosoft Windows and Linux/Unix), and 1. J.O. Andersson, T. Helander, L. Höglund, Thermo-Calc for Windows, which has an easy- P.F. Shi, and B. Sundman, Thermo-Calc and to-learn graphical user interface but only oper- DICTRA, Computational Tools for Materials ates in the Microsoft Windows environment. Science, Calphad, Vol 26, 2002, p 273Ð312, Demo versions are available for both of these 2002 versions of the software. 2. N. Saunders and A.P. Miodownik, Pergamon The demo versions are free to use, subject to Materials Series, CALPHAD (Calculation of the terms outlined in the Thermo-Calc Software Phase Diagrams): A Comprehensive Guide, End User License Agreement. It should be 1, Elsevier, 1998 noted that the demo versions are limited to 3. TCFE5—TCS Steel/Fe-Alloys Database, using just three elements (whereas in the full Version 5.0, 2007, Thermo-Calc Software product the current upper limit is 40 elements) AB, www.thermocalc.com Copyright © 2008 ASM International®. All rights reserved. Stainless Steels for Design Engineers (#05231G) www.asminternational.org

Index

A isocorrosion curves for, in sulfuric acid, 33(F) isocorrosion curves for, sulfuric acid with chlorides, 34(F) acetic acid alloy systems corrosion rates for various alloys of, plus formic acid, alloying elements, 1 103(F) austenitic alloys, 69, 72 duplex alloys, 102 austenitic stainless family, 71(F) isocorrosion curves in, 36(F) body-centered cubic (bcc) phase, 1 isocorrosion performances of various alloys, 102(F) families in perspective, 69Ð72 acids, corrosion in, and bases, 31Ð36 most widely used, 1 adsorption-induced brittleness, 51, 53 SchaefflerÐDelong stainless steels constitution diagram, adsorption-induced plasticity, 51, 52Ð53 5(F), 70(F) aeration cells, differential, 38Ð39 alpha prime aerobic bacteria, influencing corrosion, 55 formation kinetics for duplex alloys, 94, 96(F) aesthetic finishes influence of, formation on hardness, 116(F) Chrysler Building in New York City, 196, 197(F), 213(F) iron-chromium phase, 8 considerations, 217Ð219 martensite, 7Ð8, 73 flatness, 219 transition temperature change with, formation with surface, 217Ð219 aging, 98(F) aging treatments, precipitation hardening stainless alteration, surface, 199 steels, 168, 170 aluminum aggressive chemical agents, refinery applications, 255 inclusions and pitting, 40, 41 Alloy Casting Institute (ACI) influence on thermodynamic activity of C, N, S and O, composition of, heat- and corrosion-resisting casting 157(T) alloys, 277(T), 278(T) metal migration into acetic solution from, 234(T) naming system, 147 migration into acetic solution from stainless, aluminum alloy design or carbon steel, 234(T) avoiding unwanted phases, 94 oxidation resistance, 79, 226, 228 duplex alloys, 92Ð94 protective layer formation, 64Ð65 Thermo-Calc software, 93, 282 aluminum alloys, 1 alloying, sensitization, 47Ð48 aluminum oxide, 191 alloying elements aluminum-titanium-nitride (AlTiN), 191 alloys, 1 aluminum/titanium precipitates, possible, 138(F) influence on alloy families, 147, 149 American Iron and Steel Institute (AISI), 240 influence on corrosion rate in contaminated sulfuric acid, American Petroleum Institute (API), 135 33 (F) American Society for Testing and Materials (ASTM), influence on thermodynamic activity of C, N, S and O, 238Ð239 157(T) American Society of Mechanical Engineers (ASME), 266 influence on uniform corrosion, 29–30 anisotropy martensitic precipitation-hardening stainless steels, deep drawing, 174 141(T) ferritic stainless steels, 120Ð121 martensitic stainless steels, 130, 131(T) Lankford ratio, 120, 175 alloy oxidation, behaviors, 66 stainless long products, 179 alloys annealing influence of, content on corrosion rate in hydrochloric austenitic stainless steels, 162Ð164 acid, 35(F) bright, 198 286 / Index

annealing (continued) atmospheres, oxidation, 66Ð67 deep drawing, 177 atmospheric corrosion duplex stainless steels, 170Ð171 uniform corrosion, 36Ð37 ferritic stainless steels, 165Ð166 atomic rearrangements, 2 long-term, of welds, 43 attraction, interatomic, 2 martensitic stainless steels, 166 austenite precipitation hardening steels, 139 alloying elements, 5 anode carbide precipitation, 9 electrochemical reactions, 12 carbon and nitrogen, 6, 9 polarization, 20Ð21, 23 diffusion rates, 6 appliances face-centered cubic (fcc), 5 facades, 239Ð240 γ-austenite in precipitation hardening alloys, 138Ð139 kitchen, 237Ð240 interstitial elements, 6 laundry, 241 lattice expansions, 6(F) stainless steels commonly used for, 239(T) lean alloy of martensite and, 73Ð74 architecture and construction mechanical properties, 7 aesthetic considerations, 217Ð219 metastable state, 6 average chloride concentration in rainwater in United phase in duplex alloy at room temperature, 91 States, 217(F) SchaefflerÐDelong constitution diagram, 5(F), 70(F) balancing corrosion resistance, processing and economy, Schaeffler diagram, 202(F) 214Ð215 secondary type, 7 balancing service environment, design and maintenance, semiaustenitic precipitation-hardenable stainless steel, 215, 217 143 cleaning methods for uncoated stainless steel, 220(T) sulfur and oxygen, 6 concrete reinforcing bar, 222 austenite conditioning, 169 corrosion resistance, 213Ð214 austenitic-ferritic “C” alloys, 151–152 design considerations, 216(F) austenitic “H” alloys ecological considerations, 222 high temperature HEÐHP, 152Ð154 environment, 216(F) precipitation hardening stainless steel, 170 fabrication and joining, 221 austenitic precipitation-hardenable stainless steels fabrication considerations, 221 cold work and aging, 146(F) fire resistance, 221–222 composition, 145(T) flatness, 219 corrosion resistance, 145Ð146 grades recommended by expert system, 217(F) mechanical properties, 145 graphic depicting low release of metal ions from 304 austenitic stainless steels and 316 stainless steels, 222(F) alloy families, 69Ð72 local weather pattern, 216(F) annealing, 162Ð164 maintenance, 220Ð221 automotive structural components, 229 maintenance schedule, 216(F) carburization, 82 roof, 219(F) composition of, 270(T), 271(T) ranking common stainless steels by pitting resistance composition of high-temperature, 82(T) equivalent number (PREN), 214(T) compositions of commonly used lean, 72(T) repair, 221 compositions of corrosion-resistant, 86(T) rolled-on stainless steel finishes, 218(F) corrosion resistance ratings, 87(T) salt exposure, 216(F) corrosion-resistant alloys, 84Ð89 special finishes, 218(F) corrosive environments, 88Ð89 stainless steel selection expert system, 216(F) critical pitting temperature (CPT), 43, 44(F) surface finish aesthetics, 217–219 drawability, 176 surface finish and corrosion resistance, 215 ductility, 180 argon oxygen decarburization (AOD) face-centered cubic (fcc), 174 adoption, 70 family, 71(F) alloy adjustment, 157 forming limit diagram of carbon steel and, 176(F) automotive industry, 225 halogens, 82 cleanliness, 184 heat exchangers, 246 control of nitrogen in refining by, 92 high-temperature alloys, 79Ð83 ferrite, 4 high-temperature mechanical properties, 82Ð83 first commercial use, 109 impact strength variation with temperature, 75(F) foundry practice, 154 intermetallic phases, 82, 203 inclusions in steel, 40 isocorrosion curves for, in hydrochloric acid, 34(F) production process, 155 kitchen appliances, 239 Index / 287

lean alloys, 72Ð78 tailpipe, 227(T), 228 machinability, 185 tensile properties of carbon and stainless steels for auto- machining setup recommendations, 183(T) mobiles, 229(T) martensite and austenite, 73Ð74 trucks, 231 mechanical properties, 74Ð76, 82Ð83 mechanical properties after cold work and annealing, 163 B metal migration into acetic solution from, 234(T) nitriding, 82 bacteria influencing corrosion, 55 oxidation resistance, 79Ð81 bacterial retention, food contact materials, 236(F) petroleum industry applications, 249(T) bases, corrosion in acids and, 31Ð36 pitting resistance equivalent number (PREN), 43, 78, 85 basic oxygen furnace (BOF), 156 precipitation of carbides and nitrides, 76Ð78 Bauschinger effect, 164 recommended thermal processing temperatures, 162(T) biocorrosion, 55Ð56 resilience and toughness of carbon steel vs. for automo- biological neutrality, food contact, 235 tive components, 229(T) bleaching, pulp, 266 SCC (stress corrosion cracking), 49Ð50 body-centered cubic (bcc) phase secondary phases in, 82(T) carbon and alloy steels, 1 sensitization, 46Ð47 change to face-centered cubic (fcc), 127(F) soaking, 161Ð162 ferrite, 4, 110 stabilization, 78 ferritic material, 174 stainless steel in shipping, 245 metals, 2 SCC performance, 262 boron stress-strain curve for single crystals of stable, 53(F) additions to ferritic stainless steels, 121 surface finish, 89 ferrite, 4 tensile properties of carbon steel vs. for automotive brazing, 211 components, 229(T) bright annealing, 198 thermal processing, 161Ð164 brightening stainless steels, 196 water vapor, 81Ð82 buffing, 197 weaknesses, 69 built-up edge (BUE) welding characteristics, 201Ð204 austenitic stainless grades, 185 welding parameters, 207(T) carbon and, 182 austenitizing, martensitic stainless steels, 131, 132(F), coolants minimizing, 191 166Ð167 copper and, 183 automotive and transportation ferritic stainless steels, 185 alloy selection for exhaust systems, 226 grain sizes, 184 alloys for major elements of automotive exhaust sys- iron and tool, 182 tems, 227(T) nickel and, 183 automotive emission standards, 225 nitrogen and, 184 bus bodies, 231(F) precipitation hardening stainless steels, 185 car manufacturers, 230 bus catalytic converter, 227(T), 228 life-cycle cost calculation for stainless vs. carbon steel, center pipe, 227(T), 228 231(T) decorative to highly engineered applications in automo- stainless steel body, 230, 231(F) biles, 225 ButlerÐVolmer equation, 20Ð21 exhaust manifold and high-temperature, 227 exhaust systems, 225Ð228 ferritic stainless, 226 C flexible pipe, 227(T), 228 calcium front pipe, 227(T), 228 effect on machinability of 303, 188Ð189, 189(F) fuel tanks, 231 inclusions and pitting, 41 life-cycle cost calculation for stainless vs. carbon steel calcium-fluoride based slag for bus, 231(T) electroslag remelting (ESR), 158 microcar frame, 231, 232(F) “C” alloys muffler, 227(T), 228 austeniticÐferritic alloys, 151Ð152 rail transport, 232 corrosion resisting, 147 resilience and toughness of carbon and stainless steels duplex alloys, 151 for automobiles, 229(T) mechanical properties of corrosion resisting cast, 150(T) stress-strain curves for 301 variants vs. duplex steels and metallurgy of, 149, 151Ð152 transformation steel, 230(F) precipitation hardening, 151 structural components, 229Ð231 carbide, precipitation kinetics, 7(F) 288 / Index

carbides naming system, 147 duplex alloys, 94 precipitation hardening, 151 flatware, 240 room temperature mechanical properties of corrosion precipitation, 76Ð78 resisting stainless, 150(T) stainless steel, 9 welding, 154 tooling, 191 catalytic converter, 227(T), 228 carbon catastrophic oxidation, 65 alloying element, 1 cathode austenite, 5 effect of, polarization, 24(F) ferrite, 3Ð4 electrochemical reactions, 12 influence of alloying elements on thermodynamic activ- mass transfer limitations, 24 ity of, 157(T) polarization, 20Ð21, 23 influence on thermodynamic activity of C, N, S and O, caustic solutions, 50 157(T) center pipe of exhaust systems, 227(T), 228 influence on uniform corrosion, 29 cerium interstitial atoms of, in austenite, 6 inclusions and pitting, 41 machinability of stainless steels, 182Ð183 oxidation resistance, 79Ð80 precipitation of carbides, 76Ð78 protective layer formation, 65 precipitation rates by, content, 76(F) Charpy V toughness steel content, 156 high-temperature austenitic alloys, 83(F) variation of martensite hardness with, 128(F) niobium-stabilized alloy, 119(F) welding of austenitic stainless steels, 201 titanium-stabilized alloy, 119(F) carbon diffusion, 115 chemical agents, 255 carbon dioxide, wet, 248 chemical and process industry. See also corrosion types carbon solubility, austenitic stainless, 76(F) corrosion table for fuming sulfuric acid, 261(T) carbon steel corrosion table for sulfuric acid, 259(T), 260(T), 261(T) activities and activity coefficients of elements in, 40(T) corrosion types, 258Ð262 body-centered cubic (bcc) phase, 1 erosion, 262 corrosion rates of stainless vs., 135(F) forms of corrosion, 258 metal migration into acetic solution from, 234(T) intergranular corrosion, 261Ð262 resilience and toughness of, vs. stainless steel for auto- isocorrosion chart for sulfuric acid, 258(F) motive components, 229(T) pitting and crevice corrosion, 258Ð259, 261 tensile properties of, vs. stainless steel for automotive single- and dual-environment systems, 257Ð258 components, 229(T) specific corrosives, 262 carburization, 82 stress corrosion cracking (SCC), 257, 262 carburizing, 199 chemical neutrality, food contact, 233Ð235 casting chemistry, machinability of stainless steels, 182Ð184 stainless steel processing, 158Ð159 chi Thermo-Calc software, 283 intermetallic phase, 9 casting alloys precipitation kinetics, 7(F) Alloy Casting Institute (ACI), 147 chip breaking, 187(F) austenitic-ferritic alloys, 151Ð152 chloride concentration in rainwater, 217(F) austenitic HEÐHP alloys, 152Ð154 chloride-containing solutions, 50Ð51 chromium alloys, 148(T) chloride ion, aggressive against stainless steel, 85 chromium-nickel alloys, 148(T) chlorinated oils or waxes, 179(T) composition of cast heat-resistant stainless and nickel chromium base alloys, 149(T) alloying element, 1 compositions of cast stainless corrosion resisting alloys, austenitic alloys, 72 148(T) austenitic stainless steels, 85Ð86 duplex alloys, 151 chromium-oxygen system volatility vs. temperature and ferritic HA, HC, HD alloys, 152 oxygen pressure, 64(F) foundry practice, 154 corrosion resistance, 228 high-temperature mechanical properties of “H” alloys, depletion from austenite near grain boundaries, 77(F) 153(T) ferrite, 3 influence of alloying elements, 147, 149 ferritic stainless alloys with low, medium and high, 110 mechanical properties of heat-resistant stainless, at room inclusions and pitting, 42, 43 temperature, 152(T) influence on thermodynamic activity of C, N, S and O, metallurgy of “C” alloys, 149, 151–152 157(T) metallurgy of “H” alloys, 152–154 influence on uniform corrosion, 29, 30(F) molten metal transfer, 153 ion release from stainless steel grades, 222(F) Index / 289

machinability of stainless steels, 182 construction. See architecture and construction migration into acetic solution from stainless, aluminum contamination, 235 or carbon steel, 234(T) continuous slab casting, 158 oxidation resistance, 71, 79, 80(F), 228 cookware, 237 paralinear oxidation from evaporation of chromium coolants, 191 superoxide, 64(F) copper Pourbaix diagram, 17(F) acid resistance, 71 sulﬁde formation, 186 machinability of stainless steels, 183 thermodynamics of oxidation, 57Ð59 copper sulfate. See also sulfuric acid plus copper sulfate

volatile nature of Cr2O3, 63, 64(F) corrosion of stainless steel and titanium in, plus sulfuric chromium alloys, cast stainless, 148(T) acid, 31(F), 32(F) chromium-nickel alloys, 148(T) corrosion Chrysler Building combating, in alloys for petroleum industry, 248, 250 architecture using stainless steel, 213 definition, 11 highly polished surface, 196, 197(F), 213(F) erosion, 262 cleaning intergranular, 261Ð262 passivation, 195Ð196 isocorrosion chart for sulfuric acid, 258(F) recommended methods, 195(T) pitting and crevice, 258Ð259, 261 stainless steel, 194Ð196 pitting resistance equivalent number (PREN), 258 cleaning methods single- and dual-environment systems, 257Ð258 stainless steels, 235Ð236 stress corrosion cracking (SCC), 257, 262 uncoated stainless steel, 220(T) table for sulfuric acid, 259(T), 260(T), 261(T) cleanliness tendency, 15Ð16 food contact materials, 235Ð236 types, 258Ð262 machinability of stainless steels, 184 corrosion cost, 252Ð253 coastal climates, 237Ð238 corrosion kinetics coatings ButlerÐVolmer equation, 20Ð21 cookware, 237 introduction, 19Ð20 tooling, 191 mass transfer control, 21 cold heading, 179Ð180 migration and ionic diffusion, 21Ð22 cold work, 75 mixed potential theory and polarization diagrams, coloring of stainless steels, 196 22Ð23 commercial use passivation, 23Ð25 applications, 237Ð242 Tafel regime: electrode-kinetics control, 21 cookware, 237 corrosion rate flatware and cutlery, 240Ð241 alloys in simulated evaporator liquid, 37(F) food contact, 233Ð237 influence of alloying element on, in contaminated heating and water heating, 241Ð242 sulfuric acid, 33(F) kitchen appliances, 237Ð240 stainless oil country tubular goods, 135(F) laundry appliances, 241 stainless vs. carbon steel, 135(F) stainless steel, 233 Tafel slope, 23(F) composition vs. surface roughness, 215(F) Alloy Casting Institute (ACI) heat- and corrosion- corrosion resistance resisting casting alloys, 277(T), 278(T) architecture, 213Ð214 austenitic precipitation-hardenable (PH) stainless steel, austenitic precipitation-hardenable (PH) stainless steel, 145(T) 145Ð146 austenitic stainless steels, 270(T), 271(T) balancing, processing and economy, 214Ð215 duplex alloys commercially available, 97 (T) duplex alloys, 99Ð106, 204 duplex stainless steels, 276(T) ferritic stainless steels, 109, 110, 121Ð122 ferrite stainless steels, 111(T), 112(T), 273(T), 274(T) function of salinity and temperature, 244(F) martensitic PH stainless steels, 140(T) martensitic PH stainless steels, 141Ð142 martensitic stainless steels, 124(T), 125(T), 275(T) material selection for desalination, 244Ð245 PH stainless steels, 276(T) pulp-and-paper industry, 265Ð267 semiaustenitic PH stainless steel, 143(T) rail transport applications, 232 tool and cutlery martensitic stainless steels, 134(T) ratings of austenitic stainless steels, 87(T) concrete reinforcing bar, 222 semiaustenitic PH stainless steel, 144 constitution diagram stainless steel for refinery equipment, 254–255 SchaefflerÐDelong stainless steels, 5(F), 70(F) stainless steel in shipping, 245 Schaeffler diagram, 202(F) sulfur hurting, 188 Welding Research Council’s 1992, 203(F) surface finish and, 215 290 / Index

corrosion resisting alloys creep rupture strength, 83, 84(F) austenitic stainless steels, 84Ð89 creep strength, 83(F) “C” alloys, 149, 151–152 crevice corrosion composition of, austenitic stainless steels, 86(T) austenitic stainless steels, 85, 87 composition of Alloy Casting Institute (ACI), 277(T), corrosion type, 214, 258Ð259, 261 278(T) critical, temperature with alloy content, 45(F) compositions of cast stainless, 148(T) critical crevice temperature (CCT) and critical pitting duplex alloys, 91 temperature (CPT), 105(F) mechanical properties of stainless, 150(T) dissimilar metals and differential aeration cells, corrosion theory 38Ð39 corrosion tendency, 15Ð16 duplex alloys, 103Ð104 electrochemical reactions, 11Ð12 geometry, 45 Faraday’s law, 12 preventing, 45Ð46 galvanic vs. electrochemical cells, 14 critical crevice temperature (CCT), 105(F) Nernst equation, 12Ð14 critical current density, 29, 30(F) Pourbaix diagrams, 16Ð17 critical pitting temperature (CPT) standard half-cell reduction potentials vs. normal hydro- austenitic steels, 43, 44(F) gen electrode, 14(T) critical crevice temperature (CCT) and CPT, 105(F) corrosion types duplex alloys, 103, 104(F), 105(F) atmospheric, 36Ð37 stainless steels for unwelded and welded material, biocorrosion and microbiologically induced, 55Ð56 44(F) chromium influence, 29 vs. pitting resistance equivalent number (PREN), 85(F), corrosion fatigue, 55 104(F) corrosion in acids and bases, 31Ð36 cryogenic containers, 245 corrosion with fatigue or fraction, 48Ð55 current density, 23(F) crack initiation, 48Ð49 cutlery crack propagation, 49, 52(F) flatware and, 240–241 crevice, 38Ð39, 45Ð46 martensitic stainless steels, 133Ð134, 240 critical current density, 29, 30(F) stainless steels commonly used for, 241(T) dissimilar metals and differential aeration cells, 38Ð39 cutting tools, 133Ð134 environmental variables, 50Ð51 environmental variables influencing uniform corrosion, 28Ð29 D grain boundary, 46Ð48 deep drawing hydrochloric acid, 33Ð34, 35(F) anisotropy, 174 hydrogen embrittlement, 54Ð55 forming stainless steel, 173, 174Ð179 influence of alloying elements, 29, 30(F) geometry, 174Ð175 localized, 37Ð38 hydroforming, 177Ð178 material variables, 29Ð31, 49Ð50 intermediate annealing, 177 molybdenum role, 29Ð30 materials composition, 175(T) nickel, 30 schematic, 174(F) nitric acid, 34, 35(F) strain rate, 177 nitrogen, 30 texture, 174 organic acids, 35, 36(F) tooling, 176, 178Ð179 phosphoric acid, 34Ð35, 36(F) defects, 160 pitting, 39Ð45 delignification of pulp, 266 pitting resistance, 43Ð45 demand for steel, 247 preventing crevice, 45Ð46 desalination SCC (stress corrosion cracking), 48Ð54 materials selection for, 244Ð245 SCC mechanisms, 51Ð54 multi-stage flash (MSF), 243 sensitization, 46Ð48 reverse osmosis (RO), 243Ð244 sodium chloride/carbon dioxide environment, 30 technology, 243Ð244 strong bases, 35 design, balance, 215, 217 sulfuric acid, 31Ð33 designers sulfuric acid plus copper sulfate, 31(F), 32(F) car manufacturers, 230 uniform, 27Ð37 pitting corrosion, 39 corrosive environments development austenitic stainless steels, 88Ð89 precipitationÐhardening stainless steels, platforms, 254 137Ð138 refinery equipment, 254–255 welding, 211Ð212 Index / 291

differential aeration cells soaking, 170 active alloys, 38Ð39 sodium hydroxide, 101 microfouling, 56 stainless steel for line pipe, 253 schematic, 20(F) stainless steel in shipping, 245 differential aeration corrosion cell, 12 strength, 96 diffusion stress corrosion cracking (SCC) performance, 262 atomic rearrangements, 2 stress-strain curves for 301 variants vs., 230(F) ionic transport, 21Ð22 structure and alloy design, 92Ð94 diffusion rates, austenite vs. ferrite, 6 sulfuric acid, 100 digesters thermal processing, 170Ð171 first kraft from alloy 2205, 266(F) umbilical tubing and risers, 253Ð254 pulp-and-paper industry, 265 variations of ferrite, austenite, and duplex with tempera- disinfection, 236Ð237 ture, 98(F) dissimilar metals, 38Ð39 welding characteristics, 204Ð205 dissociated ammonia, 198Ð199 welding parameters, 207(T) dissolution wrought 2205 duplex microstructure, 91(F) equation, 37 term, 27 dryers, laundry appliances, 241 E dry film, forming stainless, 179(T) earing dual-environment system, 257 deep drawing, 178 duplex alloys measuring tendency, 178 acetic acid, 102 ecological considerations, 222 annealing, 170Ð171 electrochemical cell composition of selected, stainless steels, 276(T) closed circuit, 11Ð12 compositions, 97(T) potential, 38 concept, 91Ð92 electrochemical corrosion, 19 corrosion resistance, 99Ð106 electrochemical reactions, 11Ð12 corrosion-resistant “C” alloys, 151 electrode-kinetics control Tafel regime, 21 crevice corrosion, 103Ð104 electrodes, polarization, 20Ð21 deep drawing, 178 electrolysis cell, 14(F) fastest-growing stainless steel family, 91Ð92 electrolyte resistance, 22 fatigue, 98 electrolytic cells, galvanic vs., 14 Fe-Cr-Ni phase diagrams, 92(F) electrolytic pickling, cold-rolled stainless, 194 formation kinetics, 96(F) electromotive force, 13 formic acid, 103(F) electropolishing, 196 forming and machining, 99 electroslag remelting (ESR), 157Ð158 heat exchangers, 246 embrittlement. See also hydrogen embrittlement (HE) hot forming, 180 alpha prime, 8 hydrochloric acid, 100Ð101 high-temperature, 114 impact strength, 97Ð98 σ phase at higher temperatures, 151 impact strength variation with temperature, 75(F) engineering alloys iron-nickel diagrams, 93(F) minimum mechanical properties of stainless steel, machinability, 186 280(T) machining setup recommendations, 183(T) physical properties of major stainless steel, 279(T) mechanical properties, 94Ð98 environment, stainless steel selection expert system, nitric acid, 101 216(F) organic acids, 102, 103(F) environmental variables partitioning of elements, 93Ð94 stress corrosion cracking, 50Ð51 petroleum industry applications, 248(T) uniform corrosion, 28Ð29 phosphoric acid, 101Ð102 epsilon martensite, 7Ð8, 73Ð74 photomicrographs, 95(F) equilibrium, argon oxygen decarburization (AOD), pitting corrosion, 102Ð103 155Ð156 pitting resistance equivalent number (PREN), 43 equivalent weight (EW), 19 PREN influencing fatigue, 98, 99(F) erosion, corrosion, 258, 262 pulp-and-paper industry, 265, 266Ð267 expert system recommended annealing and stressÐrelieving tempera- recommended stainless steel grades, 217(F) tures, 170(T) stainless steel selection, 216(F) SCC (stress corrosion cracking), 49, 104Ð106 exposure to salt, stainless steel selection expert system, sensitization, 47 216(F) 292 / Index

F sensitization, 47 soaking, 165 fabrication, 221 stabilization, 109, 115, 118Ð120 facades of appliances, 239Ð240 stress corrosion cracking (SCC), 49 face-centered cubic (fcc) phase stress relieving, 166 aluminum alloys, 1 superferritics, 113 austenite, 5 texture and anisotropy, 120Ð121 austenitic materials, 174 time-temperature-transformation (TTT) curve for 430, change to body-centered cubic (bcc), 127(F) 115(F) metals, 2 titanium and niobium, 118 Faraday’s law, 12, 21 titanium for carbide and nitride formation, 115 fatigue toughness, 116(F), 117(F), 118Ð119 corrosion, 55 welding characteristics, 205Ð206 duplex alloys, 98, 99(F) welding parameters, 207(T) fatty oils and blends, suitability in forming stainless ferromagnetism, 5 steel, 179(T) fingerprints, cleaning methods for uncoated stainless, ferrite 220(T) carbon and nitrogen, 3Ð4 fire resistance, stainless steel, 221Ð222 carbon diffusion rate, 115 flatness, surface aesthetic, 219 chromium, 3 flatware, 240Ð241 diffusion rates in austenite vs., 6 flexible pipe, 227(T), 228 δ-ferrite in precipitation hardening alloys, 138Ð139 flow lines, 252Ð253 hydrogen and boron, 4 flux cored wire (FCW) welding, 210 mechanical properties, 4 food contact molybdenum, 4 bacterial retention by material and cleaning time, 236(F) phase diagram of iron chromium, 3(F) biological neutrality, 235 phase in duplex alloy at room temperature, 91 chemical neutrality, 233Ð235 SchaefflerÐDelong constitution diagram, 5(F), 70(F) cookware, 237 Schaeffler diagram, 202(F) flatware and cutlery, 240Ð241 stabilization with titanium, 4 heating and water heating, 241Ð242 thermal conductivity and thermal expansion, 4Ð5 kitchen appliances, 237Ð240 ferritic “H” alloys, 152 material cleanliness, 235Ð236 ferritic stainless steels metal migration into acetic solution, 234(T) alpha prime formation, 116, 117(F) qualifications, 233–237 annealing, 165Ð166 stainless steels commonly used for appliances, 239(T) automotive exhaust systems, 226 stainless steels commonly used for cutlery, 241(T) body-centered cubic (bcc), 174 surface disinfection, 236Ð237 carbon diffusion rate in, 115 formability, ferritic stainless steel, 120 composition, 111(T), 112(T), 273 (T), 274 (T) formic acid corrosion and oxidation resistance, 109, 121Ð122 austenitic stainless steels, 89 deep drawing, 178 corrosion in, 35 embrittling phenomenon, 116 corrosion rates for various alloys of acetic plus formic forming limit diagrams, 176 (F) acid, 103(F) groups of low, medium and high chromium, 110, 113 duplex alloys, 102 heat exchangers, 245Ð246 isocorrosion curves in, 36(F) high-temperature properties, 121 forming limit diagram (FLD), 175Ð176 hot rolling, 159 forming technology impact strength variation with temperature, 75(F) deep drawing, 173, 174Ð179 intermetallic phases, 116 deep drawing materials composition, 175(T) iron-chromium phase diagrams, 113(F), 114(F) deep drawing schematic, 174(F) kitchen appliances, 239 duplex alloys, 99 lowest cost and simplest stainless, 109Ð110 duplex stainless steel, 178 machinability, 185 ferritics, 120, 178 machining setup recommendations, 183 (T) flat, rolled stainless steel, 173–179 mechanical behavior, 116Ð117 forces for hot working, 180(F) metallurgy, 113Ð116 forming limit diagram of carbon steel vs. austenitic metal migration into acetic solution from, 234(T) stainless steel, 176(F) petroleum industry applications, 247(T) forming limit diagrams for stainless steel categories, pitting resistance equivalent number (PREN), 43 176(F) recommended annealing temperatures, 165(T) hot, of stainless steel, 180 Index / 293

hydroforming, 177Ð178 grit sizes, 197(T) hydrogen embrittlement, 177 GuinierÐPreston (GP) zones, 138 limiting drawing ratio (LDR) vs. Lankford ratio, 175(F) optimized 409 for forming vs. normal 409, 177(F) orange peel, 178 H stainless long products, 179Ð180 half-cell reactions stainless steel, 173 reduction potentials, 13Ð14 stretch forming, 174 vs. normal hydrogen electrode, 14(T) suitability of lubricants for use in, 179(T) “H” alloys surface ﬁnish, 178 austenitic HEÐHP alloys, 152Ð154 tooling, 176, 178Ð179 corrosion resisting, 147 foundary practice, casting alloys, 154 ferritic HA, HC, HD, 152 free energy, phases, 2 high-temperature mechanical properties of, 153(T) friction stir welding, 212 mechanical properties of heat-resistant, 152(T) front pipe, alloys in automotive exhaust systems, metallurgy, 152Ð154 227(T), 228 halogens, 82 fuel tanks, 231 hardening fuming sulfuric acid, 261(T) austenitic stainless steels, 75 furnace, stainless steels in, 241Ð242 ferritic stainless steels, 116Ð117 heat-affected zone (HAZ) austenitic stainless steel, 207 G chromium carbide formation in, 201 Gallionella, 55 duplex stainless steels, 204 galvanic cell ferritic stainless steels, 109, 205 electrochemical reaction, 38 laser welding, 210 schematic, 14(F) martensitic stainless steels, 206 vs. electrolytic cells, 14 secondary austenite, 7 gas metal arc welding (GMAW) heat exchangers, 245Ð246 joint design, 209(F) heat-resistant alloys process, 210 composition of Alloy Casting Institute (ACI), 278(T) gas tungsten arc welding (GTAW) compositions, 149(T) joint design, 209(F) “H” alloys, 152–154 process, 208Ð210 mechanical properties of cast stainless, 152(T) geometry heat tint, coloring stainless steels, 196 crevice corrosion, 45 heat tinting, cleaning method, 220(T) deep drawing, 174Ð175 heat treatment, Thermo-Calc software, 283 pitting corrosion, 39 heat treatment and conditioning, 168Ð170 Gibbs free energy heavy-duty emulsions, 179(T) electrochemical reactions, 12Ð13 heavy metals, elimination, 157 oxidation, 57, 58(F) high-frequency induction welding, 211 grade selection, corrosion resistance, processing and high-speed tool steels, 190Ð191 economy, 214Ð215 high-temperature alloys graffiti, cleaning methods for uncoated stainless, austenitic stainless steel, 82(T) 220(T) intermetallic phases of austenitic stainless steel, 82 grain boundaries martensitic stainless steels, 133, 134(F) austenite, 6 mechanical properties of austenitic, 82Ð83, 84(F) austenite, of martensite, 8 oxidation resistance of austenitic, 79Ð81 boron additions to ferritics, 121 water vapor, 81Ð82 carbide precipitation, 9, 76Ð77, 77(F) high-temperature embrittlement, 114 corrosion, 46Ð48 high-temperature properties, ferritic stainless steels, 121 defects in stainless steel, 160 hopper cars, 232 depletion of chromium from austenite near, 77(F) hot ductility defects, 160 ferrite-austenite, 8 hot forming, 180 grain size hot mill defects, 160 austenitic stainless steel annealing, 163 hot rolling, 159Ð160 martensitic stainless steels and toughness, 131, hot Steckel mills, 159 132(F) hot strip tandem mills, 159 material structure, 184 hydrochloric acid graphite, suitability in forming stainless steel, 179(T) austenitic stainless steels, 88, 89(F) grinding, coarse polishing, 197 corrosion in, 33Ð34 294 / Index

hydrochloric acid (continued) phase diagrams with varying carbon, 114(F), duplex alloys, 100Ð101 130(F) influence of alloy content on corrosion rate in, 35(F) phase diagrams with varying chromium, 130(F) isocorrosion curves for austenitic stainless steels in, 34(F) iron dissolution, 20(F) isocorrosion curves for stainless steels in, 34(F) iron reduction, 15(F), 28(F) isocorrosion performance of duplex, 101(F) hydrofluoric acid, 193Ð194 J hydroforming, 177Ð178 hydrogen/argon atmosphere, bright annealing, 198Ð199 JFI Steel, oil country tubular goods and line pipe hydrogen embrittlement (HE) alloys, 135(T) corrosion fatigue, 55 joining crack growth, 49 stainless steel, 221 ferritic stainless steels, 121Ð122 Thermo-Calc software, 283 mechanisms, 54Ð55 joint design, 208, 209(F) stress corrosion cracking, 51, 52Ð54 hydrogen ion reduction, 15(F), 28(F) K hypochlorite bleaches, 195 kinetics, alpha prime formation, 94, 96(F) kitchen appliances I austenitic stainless steel, 239 impact strength coastal conditions, 238 duplex alloys, 97Ð98, 204 exposure of stainless samples to North Carolina beach, variation with temperature for stainless steels, 75(F) 238(F) inclusion-related defects, 160 facades, 239Ð240 inclusions ferritic stainless grades, 239 chip breaking at sulfides, 187(F) food contact, 237Ð240 lead, selenium, tellurium, 186 interior or working parts, 239 oxides, 188Ð190 knife-line attack, 48, 202 pitting corrosion, 40Ð43 kraft process role in machining stainless steels, 186Ð190 paper-making, 265Ð267 stainless steel, 10 pulp-and-paper industry, 36, 265 sulfur, 186Ð188 induction welding, high-frequency, 211 ingot method, 158, 159 L initiation Lankford r, earing tendency, 178 pitting, 39Ð40, 43 Lankford ratio stress corrosion cracking, 48Ð49 anisotropy measure, 120, 175 interatomic attraction, thermodynamics, 2 limiting drawing ratio as function of, 175 intermetallic phases lanthanum, protective layer, 65 austenitic stainless steel, 82, 203 laser welding, 210 ferritic stainless steels, 116 lattice expansions, 6(F) stainless steel, 8Ð9 laundry appliances, 241 International Nickel Company (INCO) process, 196 laves, precipitation kinetics, 7(F) interstitial elements, 6 laves phase, 9 ionic current, 11Ð12 leaching, elements from stainless to foods, 234 ionic diffusion, 21Ð22 lead, stainless steel machinability, 186 iron lean alloys body- and face-centered cubic transformations, 2 austenitic, 72Ð78 electrochemical corrosion, 19 compositions of austenitic, 72(T) ferrite, 3 martensite and austenite, 73Ð74 ion release from stainless steel grades, 222(F) lime content, electroslag remelting, 158 machinability of stainless steels, 182 limiting current migration into acetic solution from stainless, aluminum electrode reaction kinetics, 21 or carbon steel, 234(T) increasing mass transfer, 23, 24(F) penetration rates, 19(T) limiting drawing ratio, 175 Pourbaix diagram, 16(F) line pipe pseudo-binary-phase diagram for, and sulfur, 41(F) martensitic stainless, 134Ð135 iron-chromium stainless steel application, 252Ð253 phase diagram, 3(F) liquefied natural gas (LNG) vessels, 254 phase diagram from Thermocalc, 113(F) localized corrosion, 37Ð38 Index / 295

lubricants machinability of stainless steels, 183 oxides, 189Ð190 sulfide formation, 186 suitability in forming stainless steel, 179(T) manganese sulfides lubrication, 197Ð198 inclusions, 41 stress risers, 188 M x-ray examination, 189(F) Marangoni effect, 207 machinability, material’s, 181(F) marine systems machining stainless steels corrosion resistance vs. salinity and temperature, 244(F) austenitic, 185 desalination, 243Ð245 carbides, 191 heat exchangers, 245Ð246 carbon, 182Ð183 materials for desalination, 244Ð245 chromium, 182 shipping, 245 cleanliness, 184 typical analyses and properties of marine alloys, 245(T) coatings, 191 martensite coolants, 191 carbon and nitrogen, 9 copper, 183 composition range, 7 cross-section size, 185 formation, 126Ð127 duplex, 186 forms, 7Ð8, 73Ð74 duplex alloys, 99 lattice expansions, 6(F) ferritic, 185 lean alloy of, and austenite, 73Ð74 high-speed tool steels, 190Ð191 platelets from surface, 126(F) introduction, 181Ð182 reversion of, formed by cold work, 75(F) iron, 182 SchaefflerÐDelong constitution diagram, 5(F), 70(F) lead inclusions, 186 Schaeffler diagram, 202(F) machinability of stainless steel families, tempering, 7 185Ð186 varying hardness with carbon content, 128(F) manganese, 183 martensitic alloys martensitic, 185 “C” alloys, 149, 151 material’s machinability, 181(F) composition of, precipitation hardening (PH) alloys, molybdenum, 183 140(T) nickel, 183 corrosion resistance of, PH alloys, 141Ð142 niobium, 184 mechanical properties of, PH alloys, 139(T) nitrogen, 184 microstructures of, PH alloys, 140(F) oxide inclusions, 188Ð190 precipitation hardening stainless steels, 139Ð142 physical and mechanical properties, 182Ð185 martensitic stainless steels precipitation hardening, 185Ð186 annealing, 166 process, 184 applications, 133Ð135 role of chemistry, 182Ð184 austenitizing, 166Ð167 role of inclusions, 186Ð190 composition, 124(T), 125(T) selenium inclusions, 186 composition of, 275(T) setup recommendations for turning wrought stainless compositions of tool and cutlery, 134(T) steels, 183(T) corrosion rates of stainless oil country tubular goods structure, 184 (OCTG) alloys, 135(F) sulfur, 183 corrosion resistance, 123 sulfur inclusions, 186Ð188 distinction from other alloys, 123, 126 super stainless steels, 186 expanding austenite stability range with nickel, 131(F) tellurium inclusions, 186 flatware and cutlery, 240Ð241 titanium, 184 hardness variation with carbon content, 127, 128(F) tooling and coolants, 190Ð191 high-temperature use, 133, 134(F) maintenance hot rolling, 159 balancing service, design and, 215, 217 influence of alloying elements, 130, 131(T) stainless steel, 220Ð221 iron-chromium phase diagrams, 130(F) stainless steel selection expert system, 216(F) machinability, 185 manganese machining setup recommendations, 183(T) alloying element, 1 martensite formation, 126Ð127 austenite, 5 OCTG and line pipe, 134Ð135 inclusions and pitting, 41 passivation, 195 influence on thermodynamic activity of C, N, S and O, petroleum industry applications, 247(T) 157(T) phase structure, 127Ð128, 130Ð131 296 / Index

martensitic stainless steels (continued) metal oxides photomicrographs, 129(F) parabolic rate constants for growth, 59(F) recommended annealing, austenitizing, and tempering standard Gibbs free energy of, formation vs. tempera- temperatures, 166(T) ture, 57, 58(F) sensitization, 47 metals smallest stainless steel category, 123 oxidation, 57 soaking, 166 with oxide scale, 61(F) strain energy, 126(F), 127(F) metastable condition, 2 stress relieving, 167Ð168 metastable pitting, 40 tempering, 167 microbiologically induced corrosion, 55Ð56 tempering and toughness, 132, 133(F) microcar frame, 231, 232(F) thermal processing, 131Ð133, 166Ð168 microorganisms, food contact, 235 tool and cutlery alloys, 133Ð134 microstructures, martensitic precipitation-hardening toughness by austenite grain size and phosphorus, stainless steels, 140(F) 132(F) migration, ionic transport, 21Ð22 welding characteristics, 206 mineral resin, bacterial retention, 236(F) welding parameters, 207(T) mischmetal, 41 mass transfer, 23, 24(F) mixed potential theory, 22Ð23 mass transfer control, 21 mold powder, continuous casting, 158 mass transport, 24Ð25 molybdenum material selection, welding, 206Ð208 alloying element, 1 material structure, 184 austenitic alloys, 72 material variables carbide precipitation, 9 stress corrosion cracking (SCC), 49Ð50 corrosion resistance, 228 uniform corrosion, 29Ð31 disulfide, 179(T) mechanical behavior, ferritic stainless steels, 116Ð117 ferrite, 4 mechanical properties influence on resistance to stress corrosion cracking, austenite, 7 50(F) austenitic precipitation-hardenable (PH) stainless steel, influence on thermodynamic activity of C, N, S and O, 145 157(T) corrosion resisting cast stainless alloys, 150(T) influence on uniform corrosion, 29–30 deep-drawing stainless steels, 175(T) influencing critical pitting temperature in welded vs. duplex alloys, 94Ð98 unwelded austenitic grade, 208(F) ferrite, 4 machinability of stainless steels, 183 heat-resistant cast stainless alloys, 152(T) oxidation resistance, 226, 228 high-temperature, of austenitic stainless steels, 82Ð83, protective layer formation, 64Ð65 84(F) muffler, automotive exhaust systems, 227(T), 228 high-temperature, of “H” alloys, 153(T) multistage flash (MSF), 243 lean austenitic alloys, 74Ð76 machinability of stainless steels, 182Ð185 marine alloys, 245(T) N martensite, 8 naming system, Alloy Casting Institute (ACI), 147 martensitic PH stainless steels, 139(T) National Association of Corrosion Engineers (NACE) minimum, of stainless steel engineering alloys, 280(T) alloy listing, 247 semiaustenitic PH stainless steel, 144(T) diagram showing alloy suitability, 250, 251(F) stainless steels, 10 regulating high-strength alloys, 142 mechanisms restrictions in use recommendations for stainless steels, pitting corrosion, 39Ð40 252(T) precipitation-hardening, 138 natural gas, vessels for liquefied, 254 stress corrosion cracking (SCC), 51Ð54 Nernst equation melting production process, 155Ð157 open circuit potential, 20 metal dusting, oxidation, 67 thermodynamics of electrochemical reactions, 12Ð14 metal flow directions, 208(F) neutrality metal ions release, 222(F) biological, in food contact, 235 metallurgy chemical, in food contact, 233Ð235 “C” alloys, 149, 151–152 New York City’s Chrysler Building, 196, 197(F), ferrite stainless steels, 113Ð116 213(F) ferritic stainless, 226 nickel “H” alloys, 152–154 alloying element, 1 introduction, 1Ð2 austenite, 5 Index / 297

austenitic alloys, 72 O carbide precipitation, 9 corrosion rates for stainless steels and, base alloys, Occupational Safety and Health Administration 80(F) (OSHA), 211 influence on thermodynamic activity of C, N, S and O, Ohm’s law, 22 157(T) oil and grease marks, cleaning methods, 220(T) influence on uniform corrosion, 30 “oil-canning,” 219 influencing oxidation of iron-chromium alloys, 79, oil country tubular goods (OCTG) 80(F) influence of chromium on corrosion rate of steel, ion release from stainless steel grades, 222(F) 249(F) machinability of stainless steels, 183 influence of copper and nickel on corrosion rate of migration into acetic solution from stainless, aluminum martensitic alloys, 249(F) or carbon steel, 234(T) martensitic stainless, 134Ð135 resistance to stress corrosion cracking, 50(F) stainless steels in petroleum industry, 250Ð252 nickel base alloys, 149(T) oleum, 33 niobium open circuit potential, 13 carbide former, 78 orange peel, 178 creep resistance, 71Ð72 organic acids high-temperature martensitic stainless, 133, 134(F) austenitic stainless steels, 89 machinability of stainless steels, 184 corrosion in, 35 replacing titanium, 205 duplex alloys, 102 role in sensitization, 47 isocorrosion curves in, 36(F) stabilization, 226 overpotentials, 20 stabilization of ferritic stainless steels, 118Ð119 oxidation Nippon Steel, oil country tubular goods and line pipe effect of chromium, 57Ð59, 60(T) alloys, 135(T) effect of rare earth additions, 65 nitric acid effect of silicon, aluminum, and molybdenum, austenitic stainless steels, 88 64Ð65 corrosion behavior of high-silicon alloys in concen- electrochemical nature of, 60Ð61 trated, 35(F) influence of nickel on, of iron-chromium alloys, 79, corrosion in, 34 80(F) duplex alloys, 101 iron-chromium-oxygen phase diagram, 59(F) isocorrosion curve for, 35(F) kinetics and, rates, 61Ð63 pickling oxide scale, 193Ð194 metal dusting, 67 nitrides metal with oxide scale, 61(F) duplex alloys, 94 oxidation-resisting grades of stainless steel, 60(T) stainless steel, 9Ð10 parabolic rate constants for growth of oxides, 59(F) nitriding paralinear, from evaporation of chromium superoxide, austenitic stainless steel and, 82 64(F) surface alteration, 199 quasi-steady-state approximation of moving boundary nitrogen problem of internal, 66(F) austenite, 5 reaction at anode, 12 austenitic alloys, 72 schematic predicting thermal stresses, 65(F) austenitic stainless steels, 71, 86 spalling and cracking of scale, 63Ð65 delay in carbide precipitation by, 78(F) standard Gibbs free energy of formation of metal oxides ferrite, 3Ð4 vs. temperature, 58(F) influence of alloying elements on thermodynamic activ- temperature dependence of metal dusting of iron, ity of, 157(T) 67(F) influence on thermodynamic activity of C, N, S and O, thermodynamics of, 57Ð60 157(T) transient, 60 influence on uniform corrosion, 29, 30 under less-oxidizing atmospheres, 66Ð67 interstitial atoms of, in austenite, 6 volatile nature of Cr2O3, 63, 64(F) machinability of stainless steels, 184 Wagner’s theory, 61Ð63 solubility in austenite, 77Ð78 oxidation resistance stainless steel for line pipe, 253 austenitic stainless steels, 79Ð81 nondestructive evaluation (NDE), 211 ferritic stainless steels, 109, 121Ð122 normal hydrogen electrode (NHE) isooxidation curves, 80(F) half-cell reduction potential vs., 14(T) oxide film standard hydrogen electrode, 14 coloring stainless steels, 196 North Carolina, exposure of stainless steel, 238(F) parameters for, coloring stainless steel, 196(T) 298 / Index

oxides ferritic stainless steels for, 247(T) effect calcium on machinability of 303, 188Ð189, 189(F) line pipe and flow lines, 252–253 elongated, 190(F) liquefied natural gas (LNG) vessels, 254 inclusions, 10 martensitic stainless steels for, 247(T) inclusions and pitting, 41 molybdenum influence on SCC susceptibility, 248, machinability of stainless steel, 188Ð190 250(F) metal with, scale, 61(F) National Association of Corrosion Engineers (NACE), pickling to remove, layer, 25, 193Ð194 247, 248, 250, 252(T) removal of oxide scale, 193Ð194 OCTG, 250Ð252 stabilization of austenitic alloy, 78 platforms, 254 Ugima, in 303 matrix, 189(F) precipitation-hardening stainless steels for, 248(T) un-deformed, 190(F) presence of wet carbon dioxide, 248 x-ray examination showing Ugima, and manganese sul- refinery equipment, 254–255 fides, 189(F) restrictions in stainless steel use recommended by oxyfuel gas welding (OFW), 210 NACE, 252(T) oxygen stainless steels for refinery processes, 254(T) austenite impurity, 6 stress corrosion cracking (SCC), 248 ferrite impurity, 3 umbilical tubing and risers, 253Ð254 impurity, 1 pH, corrosion tendency, 15Ð16 inclusions and pitting, 40, 41Ð42 phase diagrams influence of alloying elements on thermodynamic activ- computer models, 2 ity of, 157(T) Fe-Cr-Ni, 92(F) influence on thermodynamic activity of C, N, S and O, iron-chromium, 3(F), 73(F), 113(F), 114(F), 130(F) 157(T) iron-chromium-oxygen, 58, 59 (F) influence on uniform corrosion, 29 iron-nickel, 93(F) steel content, 156 pseudo-binary-, for iron and sulfur, 41 (F) oxygen gas reduction, 15(F), 28(F) phases oxygen pressure, 64(F) alloy systems, 1 ferrite, 3Ð5 free energy, 2 P intermetallic, of stainless steel, 8Ð9 paint, cleaning method for uncoated stainless, 220(T) structure of martensitic stainless steels, 127Ð128, paper-making processes 130Ð131 bleaching pulp, 266 phosphoric acid digestion, 265Ð266 austenitic stainless steels, 88 kraft process, 265 corrosion in, 34Ð35 process equipment, 266Ð267 duplex alloys, 101Ð102 washing and screening, 266 electropolishing solution, 196 partitioning elements, 93Ð94 isocorrosion curves in, 36(F) passenger trains, 232 minimum temperatures for wet, with duplex alloys, passivation 102(F) effect on polarization diagrams, 23Ð25 phosphorus removing surface contamination, 195Ð196 ferrite impurity, 3 stainless steel, 25 impurity, 1, 156Ð157 theory, 23 martensitic stainless steels and toughness, 131, 132(F) transpassive regime, 24 photomicrographs passive behavior, 27 duplex alloys, 94, 95(F) passivity, 27 martensitic stainless steels, 127, 129(F) penetration equation, 19 physical properties penetration rates, 19(T) major stainless steel engineering alloys, 279(T) petroleum industry stainless steels, 10

alloy suitability vs. H2S and CO2 partial pressure, 251(F) pickling austenitic stainless steels for, 249(T) oxide layer removal, 25, 193Ð194 chromium inﬂuence on corrosion rate in environments uniform corrosion, 28 by oil country tubular goods (OCTG), 249(F) pigmented pastes, 179(T) combating corrosion in applications, 248, 250 PillingÐBedworth ratio (PBR), 63 copper and nickel inﬂuencing corrosion rate of marten- pitting sitic stainless alloys for OCTG, 249(F) activities and activity coefficients in liquid steels, 40(T) demand for steel, 247 austenitic stainless steels, 85, 88 duplex stainless steels for, 248(T) corrosion type, 39Ð40, 258Ð259, 261 Index / 299

critical, temperatures, 44(F) powder metallurgy, 159 geometry, 39 precipitated phases inclusions, 40Ð43 GuinierÐPreston zones, 138 influence of sulfur level on, resistance, 42(F) stainless steels, 8(T) metastable, 40 precipitation mischmetal, 41 carbides, 9 passive anode polarization curve, 40(F) carbides and nitrides, 76Ð78 pit initiation, 39Ð40 possible aluminum/titanium, 138(F) pseudo-binary-phase diagram for iron and sulfur, 41(F) precipitation-hardening stainless steels resistance, 43Ð45 advantage over martensitic, 137 “weakest link” phenomenon, 214 annealed condition, 139 pitting corrosion austenitic, 144Ð146 CPT (critical pitting temperature) vs. NaCl concentra- austenitic alloys, 170 tion, 103, 105(F) cast PH alloys, 151 CPT vs. pH, 103, 105(F) cold work influence on aging of AÐ286, 145, 146(F) CPT vs. pitting resistance equivalent number (PREN), composition, 276(T) 103, 104(F) composition of austenitic, 145(T) duplex alloys, 102Ð103 composition of martensitic, 140(T) varying pitting potential with temperature, 103, 104(F) compositions of semiaustenitic, 143(T) pitting resistance equivalent number (PREN) corrosion resistance of martensitic, 141Ð142 austenitic alloys, 43, 78, 85 corrosion resistance of semiaustenitic, 144 corrosion, 37, 258 development, 137Ð138 critical pitting temperature vs., 85(F) influence of alloying elements, 141(T) duplex alloys, 43, 92, 102 machinability, 185Ð186 ferritic alloys, 43 machining setup recommendations, 183(T) influence on duplex alloy fatigue strength, 98, 99(F) martensitic, 139Ð142 pitting corrosion, 43Ð45, 214 martensitic grades, 168Ð170 ranking stainless steels by PREN, 214(T) mechanical properties of martensitic PH alloys, 139(T) umbilical tubing and risers, 253 mechanical properties of semiaustenitic, 144(T) Pittsburgh Convention Center, 219(F) mechanism of PH, 138 platforms, stainless steel, 254 microstructures, 140(F) polarization passivation, 195 anode, 23 petroleum industry applications, 248(T) cathode, 23 phases in stainless steel, 10 influence on uniform corrosion, 29, 30(F) possible aluminum/titanium precipitates, 138(F) overpotentials, 20 presence of δ-ferrite and γ-austenite, 138Ð139 passivating alloys, 39(F) processing routes for S15700, 142, 143(F) passive anode, curve, 40(F) properties of A-286 vs. test temperature, 145(F) stainless steel in chloride-containing solution, 40 recommended annealing and stress-relieving tempera- polarization diagrams tures for martensitic grades of, 169(T) effect of cathode polarization, 24(F) semiaustenitic, 142Ð144 effect of mass transport, 24Ð25, 25(F) specialized family, 137 mixed potential theory and, 22Ð23 stress corrosion cracking (SCC), 141, 142 passivation, 23Ð25 thermal processing, 168Ð170 schematic, 22(F) welding characteristics, 206 schematic of passive anode polarization curve, 23, 24(F) welding parameters, 207(T) polishing precipitation kinetics, 7(F) grit sizes for target surface roughness, 197(T) prevention, crevice corrosion, 45Ð46 surface finishing, 197–198 production processes polycarbonate, bacterial retention, 236(F) basic oxygen furnace (BOF), 156 polythionate, 50 casting, 158Ð159 polythionic acid, 255 defects, 160 porosity, 52 electroslag remelting (ESR), 157Ð158 Porsche, auto components, 230 hot rolling, 159Ð160 postweld stress relief, 205 hot Steckel mills, 159 potassium hydroxide, 88Ð89 hot strip tandem mills, 159 Pourbaix diagrams impurities, 156Ð157 chromium, 17(F) influence of alloying elements on thermodynamics, construction of, 16Ð17 157(T) iron, 16(F) melting and refining, 155–157 300 / Index production processes (continued) S remelting, 157Ð158 semiaustenitic precipitation-hardenable stainless steel, Saab, auto components, 230 142, 143(F) safety, welding, 211 stainless steel, 155 salinity, corrosion resistance vs., 244(F) thermodynamics, 156 salt exposure, stainless steel selection expert system, vacuum arc remelting (VAR), 157Ð158 216(F) vacuum induction melting (VIM), 157 sanitation, cleaning stainless steels, 195 vacuum oxygen decarburization (VOC), 156 scale propagation metal with oxide, 61(F) crack, rates of metals vs. current density, 52(F) spalling and cracking of, 63Ð65 stress corrosion cracking (SCC), 49 SchaefflerÐDelong constitution diagram, 5(F), 70(F) pulp-and-paper industry Schaeffler diagram, 202Ð203 duplex stainless steels, 265, 266Ð267 seawater. See also marine systems kraft process, 36, 265 desalination, 243Ð245 paper-making processes, 265Ð267 secondary austenite, 7 pulsed arc transfer, 210 secondary phases, 82(T) selenium, 186 semiaustenitic precipitation-hardenable stainless steels Q austenite-stabilizing elements, 142 quality control, Thermo-Calc software, 283 compositions, 143(T) corrosion resistance, 144 mechanical properties, 144(T) R processing by T route, 142, 143(F) rain, cleansing action of, 220 sensitization rainwater, average chloride concentration, 217(F) austenitic, 46Ð47 rare earth metals duplex steels, 47 inclusions and pitting, 41 effect of alloying, 47Ð48 protective layer formation, 65 ferritic, 47 reduction, 12 ferritic stainless steels, 115 reduction potential heat treatment vs. time, 46(F) iron and hydrogen ion reductions vs. pH, 15(F), intergranular corrosion, 9 28(F) knife-line attack, 48, 202 iron and oxygen gas reductions vs. pH, 15(F), martensitic steels, 47 28(F) schematic of, due to chromium-rich precipitates, 46(F) reference electrode, 13Ð14 welding, 48 refinery equipment, 254Ð255 service, design, and maintenance, 215, 217 refining production process, 155Ð157 shielded metal arc welding (SMAW) repair, 221 joint design, 209(F) residential applications process, 210 cookware, 237 shielding gas domestic goods, 233 welding austenitic stainless steel, 203Ð204 flatware and cutlery, 240Ð241 welding parameters for various stainless steels, 207(T) heating, 241Ð242 shipping, 245 kitchen appliances, 237Ð240 short-circuiting transfer, 210 laundry appliances, 241 sigma water heaters, 241, 242 intermetallic phase, 8Ð9 resistance precipitation kinetics, 7(F) Ohm’s law, 22 silicon pitting, 43Ð45 alloying element, 1 resistance welding, 210Ð211 content in cast alloys, 147 resistivity, 22(T) corrosion of high- austenitic steels in nitric acid, 34, 35(F) reverse osmosis, 243Ð244 inclusions and pitting, 41 ridging, ferritics, 178 influence on thermodynamic activity of C, N, S and O, rolled finishes 157(T) applications, 198Ð199 oxidation resistance, 71, 79, 226, 228 benefits, 198 protective layer formation, 64Ð65 roping, ferritics, 178 single-environment system. See also chemical and rust staining, cleaning method for uncoated stainless, process industry 220(T) aggressive chemical species, 258 Index / 301

slabs, 158 corrosion table for, in sulfuric acid plus copper sulfate, slip dissolution, 51 31(F) soaking deep drawing, 173, 174Ð179 austenitic stainless steels, 161Ð162 defects in, hotÐrolled bands, 160 duplex stainless steels, 170 flat, rolled, 173–179 ferritic stainless steels, 165 hot rolling, 159Ð160 martensitic stainless steels, 166 inclusions, 10, 186Ð190 soap-fat pastes, 179(T) isocorrosion curves for, in sulfuric acid plus copper sul- sodium chloride/carbon dioxide environment, 30 fate, 32(F) sodium hydroxide lubricants for forming, 179 (T) austenitic stainless steels, 88Ð89 machinability, 185Ð186 corrosion in, 35Ð36 machining setup recommendations, 183 (T) corrosion rates of duplex alloys, 101(F) melting and refining, 155–157 corrosion rates of duplex alloys with contaminated envi- minimum mechanical properties of, engineering alloys, ronment, 101(F) 280 (T) duplex alloys, 101 nitrides, 9Ð10 sodium hypochlorite oxidation-resisting grades, 60(T) cleaning stainless steels, 195 passivation, 25 disinfecting stainless, 236Ð237 penetration rates, 19(T) software package, Thermo-Calc, 2, 281Ð283 physical properties of major, engineering alloys, 279 (T) soldering, 211 precipitated phases, 8 (T) solution treatment, precipitation-hardening stainless precipitation-hardening process, 10 steels, 168Ð170 precipitation kinetics in 316, 7 (F) specialization, precipitation-hardening stainless steels, properties, 10 137 ranking by pitting resistance equivalent number (PREN), Sphaerotilus, 55 214(T) spots, cleaning method for uncoated stainless, 220(T) ranking common, by PREN, 214(T) spray transfer, 210 refinery processes, 254(T) stability remelting, 157Ð158 expanding austenite, with nickel, 131(F) resilience and toughness of carbon steel vs. for lean alloy of martensite and austenite, 73Ð74 automotive components, 229(T) stabilization SchaefflerÐDelong constitution diagram, 5 (F) ferrite, 4 selection expert system, 216(F) ferritic stainless steel, 109, 115, 118Ð120 tensile properties of carbon steel vs. for automotive ferritic steels for exhaust systems, 226 components, 229(T) lean austenitic alloys, 78 thermodynamics, 2 stacking fault, 73Ð74 welding parameters, 207(T) stainless long products stains, cleaning method for uncoated stainless, 220(T) cold heading, 179Ð180 standard Gibbs free energy, 57, 58(F) hot forming, 180 standard hydrogen electrode, 14 stainless steel alloys Steckel mills, hot, 159 austenite, 5Ð7 Steel Founder’s Society of America, 147 ferrite, 3Ð5 strain energy, martensitic stainless steels, 126(F), 127(F) stainless steels. See also casting alloys strain rate, deep drawing, 177 activities and activity coefficients of elements in, 40(T) stress corrosion cracking (SCC) bacterial retention by material and cleaning time, advantages of duplex alloys, 105Ð106 236(F) austenitic stainless steels, 87Ð88 carbides, 9 corrosion form, 258 casting, 158Ð159 crack initiation, 48Ð49 casting alloys, 147, 149 crack propagation, 49 categories for oxidation resistance, 59Ð60 crack propagation rates of metals vs. current density, classifications by sulfur content, 187–188 52(F) cleaning methods for uncoated, 220(T) debating mechanisms, 105 composition of austenitic, 270 (T), 271 (T) dilation of austenite due to hydrogen in solution, 53, 54(F) composition of duplex, 276(T) duplex alloys, 91, 104Ð106 composition of ferrite, 273(T), 274(T) environmental variables, 50Ð51 composition of martensitic, 275(T) ferritic stainless steels, 121 composition of precipitation-hardenable (PH), 276(T) influence of molybdenum on resistance, 50(F) concrete reinforcing bar, 222 martensitic precipitation-hardening (PH) stainless steels, corrosion rates of, vs. carbon steel, 135(F) 141, 142 302 / Index

stress corrosion cracking (continued) sulfuric acid material variables, 49Ð50 austenitic stainless steels, 88 mechanisms, 51Ð54 corrosion in, 31Ð33 petroleum industry, 248 corrosion table, 259(T), 260(T), 261(T) resistance, 257, 262(F) corrosion table for fuming, 261(T) stress-strain curve for single crystals of austenitic steel duplex alloys, 100 with and without hydrogen, 53(F) electropolishing solution, 196 susceptibility of martensitic stainless steels, 123 influence of alloying element on corrosion rate in con- susceptibility to, with oxygen and chloride content for taminated, 33(F) 304 stainless, 51(F) isocorrosion, 88(F) theory, 262 isocorrosion chart for, 258(F) threshold stress for, for various alloys, 88(F) isocorrosion curves for various alloys in, 33(F) varying resistance to, with nickel content, 50(F) isocorrosion curves for various alloys in, with chlorides, water heaters, 242 34(F) zones of susceptibility, 48(F) isocorrosion curves of duplex grades, 100(F) stress relief annealing (SRA), 48 isocorrosion rates for various stainless steels, 32(F) stress relieving oleum, 33 austenitic stainless steels, 164 pickling oxide scale, 193Ð194 ferritic stainless steel, 166 sulfuric acid plus copper sulfate martensitic stainless steels, 167Ð168 corrosion table for stainless steels and titanium in, 31(F) stress risers, 188 isocorrosion curves for stainless steel and titanium in, stress sorption, stress corrosion cracking, 51, 53 32(F) stretch forming, 174 sulfurized or sulfochlorinated oils, 179(T) strip casters, 158Ð159 superaustenitic stainless steels, 164 strip tandem mills, 159 superferritics, 113 strong bases super stainless steels, 186 austenitic stainless steels, 88Ð89 surface finishing corrosion in, 35Ð36 aesthetics, 217Ð219 structure aesthetic surface finishes, 196–199 duplex alloys, 92Ð94 austenitic stainless steels, 89 machinability of stainless steels, 184 bright annealing, 198 phase, of martensitic stainless steels, 127Ð128, 130Ð131 brightening, 196 submerged arc welding (SAW) cleaning, 194Ð196 joint design, 209(F) coloring, 196 process, 210 and corrosion resistance, 215 submerged entry nozzle, 158 deep drawing, 178Ð179 sulfate process, pulp-and-paper, 36 function of surface treatments, 193Ð196 sulfides introduction, 193 inclusions, 10 parameters for oxide film coloring of stainless, 196(T) inclusions in stainless steel, 186Ð187 passivation, 195Ð196 size and shape and machinability, 187 pickling, 193Ð194 stabilization of austenitic alloy, 78 polished finishes, 197–198 sulfite process, 265 recommended cleaning methods, 195(T) sulfur removal of oxide scale, 193Ð194 austenite impurity, 6 rolled finishes, 198–199 comparing machinability, 190(F) rolled-on finishes, 218 (F) effect on stainless machinability, 187, 188(F) surface alteration, 199 ferrite impurity, 3 surface roughness, 215(F) impurity, 1, 156 surface treatments inclusions and pitting, 40, 41 brightening, 196 influence of alloying elements on thermodynamic activ- cleaning, 194Ð196 ity of, 157(T) coloring, 196 influence on thermodynamic activity of C, N, S and O, passivation, 195Ð196 157(T) removal of oxide scale, 193Ð194 machinability of stainless steels, 183 metal flow directions in weld pool with and without, T 208(F) pitting resistance of unannealed welds, 42(F) Tafel slope, 23(F) pseudo-binary-phase diagram for iron and, 41(F) tailpipe, automotive exhaust systems, 227(T), 228 stainless steel machinability, 186Ð188 tellurium, 186 Index / 303

temperature thermodynamics chromium-oxygen system volatility, 64(F) argon oxygen decarburization (AOD), 156 corrosion resistance vs., 244(F) influence of alloying elements on, activity of C, N, S, critical crevice corrosion, with alloy content, 45(F) and O, 157(T) critical pitting, (CPT), 43, 44(F) oxidation, 57Ð60 impact strength variation with, for stainless steels, stainless steel, 2 75(F) thiosulfate, 50 partitioning ratio varying with, 93(F) time-temperature-transformation (TTT) diagram standard Gibbs free energy of metal oxide formation vs., high-alloy stainless steel, 94, 96(F) 58(F) unstabilized 430-type alloy, 115(F) variation of pitting potential with, for duplex alloys, titanium 104(F) carbide and nitride formation, 115 tempering carbide former, 78 influencing martensitic stainless hardness, 132, corrosion table for, in sulfuric acid plus copper sulfate, 133(F) 31(F) martensite, 7 deoxidizer in chromiumÐiron alloys, 156 martensitic stainless steels, 167 ferritic alloy stabilization, 4, 205 tensile properties inclusions and pitting, 41 austenitic precipitation-hardenable stainless steel, influence on thermodynamic activity of C, N, S and O, 145(F) 157(T) austenitic stainless steels, 75 isocorrosion curves for, in sulfuric acid plus copper sul- tensile strength equation, 74 fate, 32(F) texture isocorrosion curves in phosphoric acid, 36(F) deep drawing, 174 machinability of stainless steels, 184 ferritic stainless steels, 120Ð121 possible aluminum/titanium precipitates, 138(F) thermal conductivity role in sensitization, 47 duplex alloys, 205 stabilization of ferritic stainless steels, ferrite, 4Ð5 118Ð119 thermal cutting, 211 stabilization of ferritic steels, 226 thermal expansion titanium-aluminum-nitride (TiAlN), 191 austenitic stainless steels, 202 titanium carbonitride (TiCN), 191 duplex alloys, 205 titanium nitride (TiN), 191 ferrite, 4Ð5 tooling thermal processing carbides, 191 annealing, 162Ð164, 165Ð166, 170Ð171 coatings, 191 austenitic stainless steels, 161Ð164 coolants, 191 austenitizing, 166Ð167 costs in deep drawing, 176 duplex stainless steels, 170Ð171 high-speed tool steels, 190Ð191 ferritic stainless steels, 165Ð166 lubricants, 189, 190 martensitic stainless steels, 131Ð133, 166Ð168 materials in deep drawing, 178Ð179 precipitation-hardening stainless steels, 168Ð170 tools, martensitic stainless steels, 133Ð134 soaking, 161Ð162, 165, 166, 170 toughness stress relieving, 164, 166, 167Ð168 austenitic stainless steels, 75Ð76 tempering, 167 duplex alloys, 97Ð98, 204 thermal stresses, predicting, 64, 65(F) ferritic stainless steels, 117, 118(F), 118Ð119 Thermo-Calc high-temperature austenitic alloys, 83(F) alloy design, 93, 282 martensitic stainless steels, 131, 132(F) applications, 282Ð283 trains, 232 casting, 283 transient oxidation, 60 free demonstration version, 283 transpassive dissolution, 27 heat treatment, 283 transpassive regime, 24 iron-chromium phase diagram, 113(F) transportation. See automotive and transportation map, 282 trucks, 231 phase determination program, 2 tubular goods, 134Ð135 quality control, 283 tungsten ScheilÐGulliver model, 282 carbides for flatware, 240 single-point equilibria, 282 influence on thermodynamic activity of C, N, S and O, software package, 281Ð282 157(T) step, 282 tungsten inert gas (TIG), 208Ð210 welding and joining, 283 tuyeres, oxygen injection, 156 304 / Index

U ferritic stainless steels, 205Ð206 flux cored wire (FCW), 210 Ugima oxide gas metal arc welding (GMAW), 210 coating cutting tool and lubricant, 189Ð190 gas tungsten arc welding (GTAW), 208Ð210 comparing 304L chips with and without, 190(F) high-frequency induction, 211 machinability by sulfur levels with and without, 190(F) joint design, 208, 209(F) x-ray showing, 189(F) laser, 210 umbilical tubing and risers, 253Ð254 martenistic stainless steels, 206 uniform corrosion. See also corrosion types material selection and performance, 206Ð208 environmental variables influencing, 28–29 metal flow directions in weld pool, 208(F) material variables, 29Ð31 new developments, 212 stainless steel, 27Ð28 nondestructive evaluation (NDE), 211 United States, chloride concentration in rainwater, oxyfuel gas welding (OFW), 210 217(F) parameters for various stainless steels, 207(T) unmixed zone, 208 practices, 211Ð212 precipitation-hardening (PH) stainless steels, 206 V processes, 208Ð211 recent developments, 211Ð212 vacuum arc remelting (VAR), 157Ð158 resistance, 210Ð211 vacuum induction melting (VIM), 157 safety, 211 vacuum oxygen decarburization (VOD) Schaeffler diagram, 202(F) cleanliness, 184 sensitization, 48 refining process, 156 shielded metal arc welding (SMAW), 210 vanadium soldering and brazing, 211 carbides for flatware, 240 submerged arc welding (SAW), 210 high-temperature martensitic stainless, 133, 134(F) thermal cutting, 211 Volvo, auto components, 230 Thermo-Calc software, 283 tungsten inert gas (TIG), 208Ð210 W Welding Research Council’s 1992 constitution diagram, 203(F) Wagner’s theory, 61Ð63 weld shielding gas composition and crevice corrosion washer tubs and drums, 241 resistance, 204(F) water heaters, 241, 242 Welding Research Council, constitution diagram, waterline corrosion, 38 203(F) water marking, cleaning method for uncoated stainless, welds 220(T) influence of sulfur on pitting resistance of unannealed, water vapor, 81Ð82 42(F) wax-base pastes, 179(T) long-term annealing, 43 wax or soap plus borax, 179(T) wet carbon dioxide, 248 weather pattern, stainless steel selection expert system, 216(F) weldability, 253 Y welding yield strength austenitic stainless steels, 201Ð204 austenitic precipitation-hardenable stainless steel, 145(F) cast stainless alloys, 154 equation, 74 characteristics of stainless steels, 201Ð206 high-temperature austenitic alloys, 83, 84(F) duplex stainless steels, 204Ð205 yttrium, 65 ASM International is the society for materials engineers and scientists, a worldwide network dedicated to advancing industry, technology, and applications of metals and materials.

ASM International, Materials Park, Ohio, USA www.asminternational.org

Publication title Product code Stainless Steels for Design Engineers #05231G