stainless steels for design engineers

STAINLESS STEELS FOR DESIGN ENGINEERS

MICHAEL MCGUIRE

ASM International®

Materials Park, Ohio 44073-0002www.asminternational.org

Copyright © 2008 ASM International®. All rights reserved. Stainless Steels for Design Engineers (#05231G) www.asminternational.org

Copyright © 2008by

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 anymeans, electronic, mechanical, photocopying, recording, or otherwise, without the written permission of thecopyright 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 WITHTHIS PUBLICATION. Although this information is believed to be accurate by ASM, ASM cannot guaranteethat favorable results will be obtained from the use of this publication alone. This publication is intended for useby persons having technical skill, at their sole discretion and risk. Since the conditions of product or material useare outside of ASM’s control, ASM assumes no liability or obligation in connection with any use of this infor-mation. No claim of any kind, whether as to products or information in this publication, and whether or notbased on negligence, shall be greater in amount than the purchase price of this product or publication in respectof which damages are claimed. THE REMEDY HEREBY PROVIDED SHALL BE THE EXCLUSIVE ANDSOLE REMEDY OF BUYER, AND IN NO EVENT SHALL EITHER PARTY BE LIABLE FOR SPECIAL,INDIRECT OR CONSEQUENTIAL DAMAGES WHETHER OR NOT CAUSED BY OR RESULTINGFROM THE NEGLIGENCE OF SUCH PARTY. As with any material, evaluation of the material under end-useconditions 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 reproduc-tion, in connection with any method, process, apparatus, product, composition, or system, whether or not cov-ered by letters patent, copyright, or trademark, and nothing contained in this book shall be construed as a de-fense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liabilityfor 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 ServiceDevelopment; Steven R. Lampman, Technical Editor; Eileen De Guire, Associate Editor; Ann Britton, EditorialAssistant; Bonnie Sanders, Manager of Production; Madrid Tramble, Senior Production Coordinator; DianeGrubbs, Production Coordinator; Patty Conti, Production Coordinator; and Kathryn Muldoon, Production Assistant

Library of Congress Control Number: 2008934669ISBN-13: 978-0-87170-717-8

ISBN-10: 0-87170-717-9SAN: 204-7586

ASM International®

Materials Park, OH 44073-0002www.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

iviv

Preface

The rate of growth of stainless steel has outpaced that of other metals and alloys, and by 2010 maysurpass aluminum as the second most widely used metal after carbon steel. The 2007 world produc-tion 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 becomemore consolidated. One result of this is a more widespread need to understand stainless steel withfewer resources to provide that information. The concurrent technical evolution in stainless steel andincreasing volatility of raw material prices has made it more important for the engineers and design-ers who use stainless steel to make sound technical judgments about which stainless steels to use andhow to use them.

This book provides design engineers with an up-to-date source of information at a level useful forboth 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 pro-vide theory that is difficult to put into practice. The content of the book is selected for utility for theuser of stainless steel. The first section gives elementary metallurgy and identification of constituentsof stainless, the effects of alloying elements and a significant section on corrosion. A second sectionis oriented toward processes important to users of stainless steel. The third section is about each fam-ily of stainless alloys and includes the most recent additions that have come to the market. The fourthsection deals in some depth with the major applications for stainless steel. This last part is presentedwithout the promotional bias which is found in many steel producers’, alloy producers’, and trade as-sociations’ literature. While a number of steel producers have provided assistance to the author, therehas been no attempt to unfairly bias information in their favor. To the contrary, those producers re-sponsible for generating factual, useful data for the user community are those who should benefit themost by books such as this. The author is particularly indebted to Allegheny Ludlum and JohnGrubb, and his many colleagues who assisted him, for technical assistance throughout the writingand 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 oth-ers 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, EileenDeGuire, Charlie Moosbrugger and Steve Lampman. I would also like to thank the many membersof my forum at Eng-tips.com who have contributed much collective knowledge and perspective tothis 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

This publication is copyright © ASM International®. All rights reserved.

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

Mail Customer Service, ASM International 9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA

Email [email protected]

In Europe

American Technical Publishers Ltd. 27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX, United Kingdom Telephone: 01462 437933 (account holders), 01462 431525 (credit card) www.ameritech.co.uk

In Japan Neutrino Inc. 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.

CHAPTER 1

Metallurgy

Summary

COMPARED TO ALLOY STEELS, stainlesssteels are chemically complex. The large numberof alloying elements makes possible a largerrange of possible phases or basic crystal struc-tures. The large amount of the alloying elementsmakes the deviation from the behavior of pureiron greater; consequently, the calculations thatpredict which phases will exist are more difficult.The three basic phases of stainless steels are fer-rite, austenite, and martensite. The wide varietyof alloys that exist is based on:

• Combinations of these phases• Altering the composition of these phases• Adding secondary phases for particular

purposes

Metallurgy, as discussed in this chapter, fo-cuses on phases normally encountered in stain-less steels and their characteristics. In subsequentchapters on types of stainless steel, there aremore detailed treatments of the alloys made ofthese phases and their properties.

Introduction

Most widely used alloy systems, such as car-bon steels, alloy steels, and aluminum alloys, arerelatively dilute solutions of several elements inthe parent matrix. Carbon and alloy steels, withvery few exceptions, are principally of the mag-netic body-centered cubic (bcc) phase or aslightly distorted version of it. Aluminum alloysshare the face-centered cubic (fcc) structure ofpure aluminum. A given structure, which canhave a certain range of compositions, is what ismeant by a phase, just as a gas or liquid is aphase. In solid metals, there can be a number of

phases coexisting simultaneously. Stainlesssteel is an exceptional alloy system in that it isnot a dilute solution. Alloy steels may containseveral percent of alloying elements, such ascarbon, manganese, nickel, molybdenum,chromium, and silicon, in addition to the impu-rities sulfur, oxygen, and phosphorus. Alloysteels typically contain very small amounts oftitanium, niobium, and aluminum. The totalamount* of these alloying elements seldom ex-ceeds 5%. The same is true for most aluminumalloys. In contrast, stainless steels contain noless than about 11% chromium alone. Moststainless alloys have manganese, silicon, car-bon, and nickel in thermodynamically mean-ingful amounts as well as large concentrationsof nickel and/or molybdenum.

The result of the large number of alloying ele-ments in relatively high concentrations is thatstainless steel can have many stable phases con-currently. In almost every case, having phasesother than the principal one or two phases forwhich the alloy was designed is undesirable be-cause of the possibility of undesirable variationsin mechanical or corrosion performance. Theproducer of stainless steel controls the chemicalcomposition and thermomechanical processing,so that when the processor or end user receivesthe product it is usually in the correct condition.However, subsequent processing or service con-ditions may alter the carefully established phasestructure. Therefore, it is necessary to discussthe phases that can exist in stainless steel andthe conditions under which they form so that theenlightened user will know which phases toavoid and how to avoid them.

It is possible to use thermodynamics to calcu-late which phases may exist at a given tempera-

* All compositions are given in weight percent unlessstated otherwise.

Stainless Steels for Design Engineers Michael F. McGuire, p 1-10 DOI: 10.1361/ssde2008p001

Copyright © 2008 ASM International® All rights reserved. www.asminternational.org

2 / Stainless Steels for Design Engineers

ture for a given composition. It is not remotelyfeasible, however, to give an adequate treatmentof the thermodynamics required to do this. Thetopic alone requires a book. The necessaryknowledge has been embedded in proprietarycomputer programs that will be used instead.

Thermodynamics of Stainless Steel

Pure metals, from a practical viewpoint, areeither liquid or solid depending on temperature,with the possibility of some trivial small gasvapor pressure. A law of thermodynamics is thatthe number of possible condensed (i.e., solid)phases equals the number of elemental con-stituents plus one. The solid has a crystallo-graphic structure that may vary with tempera-ture. Many metals have a less-dense bccstructure at high temperature and transform to adenser fcc structure at lower temperatures. Irondoes this. Iron has the curious characteristic oftransforming from fcc back to the low-densitybcc at still lower temperatures. This is a result ofthe unpaired 3d orbital electrons (those that giverise to ferromagnetism) that are not given up asvalence electrons, causing repulsive forces be-tween atoms and requiring a more widely spacedstructure.

All thermodynamic properties are based oninteratomic attractions. In metals, the metalatoms give up valence electrons to the entiremass. These electrons are of varying energystates and highly mobile. They are responsiblefor the ability of metals to conduct heat andelectricity well. The attraction, the strength ofthe bond, is proportional to the charge differ-ence and distance. The attraction determinessuch macroscopic properties as melting temper-ature, density, and elastic modulus.

In this book, the main concern with thermo-dynamics is predicting which phases are presentboth at equilibrium and in the quite frequentmetastable state. The prediction involves calcu-lating the free energy of the various possiblephases. The phase with the lowest energy ismost favored, but others may have free energiesthat permit them to exist. The difference be-tween these two is that the equilibrium state,that of the lowest free energy, may requireatomic rearrangements to occur for equilibriumcompositions to be reached on an atomic scale.If diffusion is too sluggish for these rearrange-ments to take place, the structure may retain theprior metastable structure indefinitely. This is

not a small, pedantic point. Most stainless steelsare used in the metastable condition. For exam-ple, the common alloy 304 (also called 18-8) isnormally used in the fully austenitic condition.It would “rather” be partly ferritic, but the sub-stitutional diffusion of chromium in austenitethat is required to form a ferrite phase of a sepa-rate composition is so slow that it cannot occurin terrestrial time frames. However, if energy isapplied by mechanical shear, the austenite cantransform without diffusion to the lower free-energy martensite phase, a quasi-bcc structureof lower free energy.

The calculation of which phases exist underequilibrium conditions proves to be extraordi-narily difficult in complicated alloy systems.This is because thermodynamic values can bemeasured accurately only in the liquid state, sothe values for the solid state are extrapolations.Also, the interaction between elements is veryimportant in nondilute alloys such as stainlesssteel. Consequently, most published phase dia-grams are experimentally derived. To determinewhich phases exist at a given composition andtemperature, a sample is made, equilibrated atthe appropriate temperature, and quenched toroom temperature. It is assumed that the charac-teristic equilibrium phases have been frozen andare then identified by various techniques forstructure, composition, and the like. This impor-tant work is obviously tedious and susceptibleto experimental error and applies only to spe-cific compositions. Any “what if” extrapolationto a different alloy composition carries the riskof error.

A practical tool has been developed that per-mits phase diagrams to be calculated for arbitrarycompositions. These are computer simulated,mathematical models that can perform the com-plex thermodynamic calculations. To do this withaccuracy requires databases of thermodynamicvalues. These values must be derived from com-puter analysis of experimental phase equilibriumdiagrams. They are expensive to derive and vali-date, and only a few exist. Hence, they are pro-prietary. In Appendix 3, a license to one suchprogram, Thermo-Calc, can be found. The ver-sion has a reduced three-element capability butuses the same proprietary thermodynamic data-base of the full version. The program allows de-termination of which phases can exist for anycomposition and temperature. Whether thephases will form depends also on kinetic factors.First, however, it is good to become familiar withthe principal phases found in stainless steel.

Chapter 1: Metallurgy / 3

Phases

Ferrite

The basis of stainless alloys is, of course,iron. Iron, as stated, solidifies as a bcc alloy be-fore transforming to the denser fcc austenite atlower temperatures. At still lower temperatures,it reverts to the bcc structure. It is accurate tosurmise from this that the free energy of bothstructures is close. Alloying elements that pro-mote one structure over the other can thereforechange which one predominates. The elementthat produces the ability to form the passive filmthat makes stainless corrosion resistant,chromium, has the characteristic of stabilizingthe bcc structure. As chromium is added to iron,the temperature range over which austenite isstable grows smaller until, at about 12%chromium, ferrite is stable at all temperatures.This is, coincidentally, the approximate level ofchromium needed to keep alloys from rustingunder ambient conditions, but this effect is notrelated to whether the structure is bcc or fcc.The iron-chromium phase diagram (Fig. 1)shows the composition and temperature regionswhere ferrite (a), martensite (α' ), austenite (γ),and sigma phase (σ) are stable.

While chromium is the principal ferrite-pro-moting alloying element, other elements havesimilar effects, but none produces the quality ofstainlessness. Silicon, aluminum, molybdenum,tungsten, niobium, and titanium all favor ferrite.Carbon, nitrogen, manganese, nickel, and cop-per do not and expand the temperature rangeover which austenite exists. Elements that areinsoluble in iron at austenite-forming tempera-tures, such as the impurities phosphorus, sulfur,

and oxygen, have no influence on which phaseis favored. Again, it must be emphasized thatthe influence of an alloying element on structurehas zero bearing on its influence on corrosionresistance.

The elements that promote ferrite over austen-ite also have the effect, at still lower tempera-tures, of promoting intermetallic compoundsgenerally composed of iron, chromium, andsome of those alloying elements. These are dis-cussed separately.

Metals are effective solvents in both the liq-uid and solid states. An important part of steel-making is refining the molten metal to removethe undesired impurities dissolved in it. Thenormal technique is to add elements that reactselectively with the targeted impurities to forman immiscible reactant that can become part ofthe slag and physically separated from the re-fined alloy. This is done for the primary impuri-ties oxygen and sulfur. A third common impu-rity, phosphorus, is not so easily removed andmust be excluded from raw materials to be keptunder control.

In stainless steel, carbon and nitrogen can bedetrimental impurities. Both are quite solublein molten iron-chromium alloys and are fairlysoluble in ferrite at high temperatures. This sol-ubility decreases exponentially with tempera-ture so that it is essentially zero at room tem-perature. These elements have small atomicsizes compared to iron and chromium and,when dissolved, squeeze into interstitial siteswithin the bcc matrix. Such interstitial soluteatoms profoundly distort the structure. They aremuch more soluble in the fcc structure, which,while denser, has roomier interstitial spaces, sothey stabilize that structure. To preserve the fer-rite structure, carbon and nitrogen must beeliminated.

There are additional reasons to eliminate car-bon and nitrogen. During cooling as these ele-ments become less and less soluble, they mustprecipitate. The most thermodynamically favor-able form in which they can precipitate is as acompound of chromium, with which they arevery reactive. This occurs at the grain bound-aries, where nucleation is favored, and depletesthose regions of chromium, rendering them lesscorrosion resistant. A second effect is a loss oftoughness due to these precipitates. The diffusionrates of carbon and nitrogen in ferrite are toohigh to prevent this precipitation by quenching.

Modern refining methods can reduce carbonplus nitrogen to under 0.020%, but even this is

Fig. 1 The iron chromium phase diagram. Courtesy ofThermo-Calc Software


too high. So, to avoid the detrimental effects ofchromium carbide and nitride formation in fer-rite, other benign carbides and nitrides such asthose of titanium or niobium are allowed to formpreferentially. This approach is called stabiliza-tion and is used for most ferritic alloys today. Theolder approach, as characterized by alloy 430, isto permit chromium carbides and nitrides to formbut then to perform a subcritical anneal to reho-mogenize the chromium and coarsen the pre-cipitates so that they have only a small negativeeffect on mechanical properties.

Hydrogen and boron are other elements thatcan be interstitially dissolved in ferrite. Boron isnormally found at levels of around 5 to 10 ppm.At higher levels, boron substitutes for carbon incarbides. Hydrogen is soluble to several partsper million by weight. It does not cause hydro-gen embrittlement in annealed ferrite. If the fer-rite is cold worked, the solubility of hydrogenincreases as the defect structure accommodateshydrogen atoms. In this condition, ferrite maybe embrittled by hydrogen, especially if it en-ters the metal through corrosion processes likepitting. This is one explanation of, and the mostlikely explanation for, stress corrosion cracking.While hydrogen is easily removed by argonoxygen decarburization (AOD), assuming ab-solutely dry blowing gases and additions areused, it can be picked up during pickling, weld-ing, or annealing as well as by corrosion.

All stainless alloys rely on having a uniformlevel of chromium and the other element,molybdenum, which assists in corrosion resist-ance, distributed throughout the matrix. If thereare locally low levels of these elements, local-ized resistance to corrosion is reduced, and lo-calized corrosion can occur. This can occur bythe precipitation of any phase that is richer inchromium or other corrosion-resisting ele-ments. Because chromium is a reactive ele-ment, its success depends to a great degree onmaintaining the homogeneity required forproper corrosion-resistant performance. Incor-rect thermal processing is the main way homo-geneity can be lost. Stabilizing makes it mucheasier to keep chromium from segregating inferritic alloys.

A by-product of stabilization with titanium isthat oxygen and sulfur are also eliminated ascompounds of titanium along with carbon and nitrogen. These impurity elements would other-wise also precipitate as compounds containingsome chromium, potentially depleting chromiumin the vicinity of their precipitation.

The bcc structure of ferrite allows morerapid diffusion than does the fcc structure ofaustenite. This is true for both the interstitialdiffusion of the elements helium, boron, car-bon, nitrogen, and oxygen and the substitu-tional diffusion of all other elements. The rateof diffusion of all elements, both interstitialand substitutional, in ferrite is about two orthree orders of magnitude higher than inaustenite. The practical implication of this isthat precipitation reactions generally cannot besuppressed by quenching in ferrite if they in-volve interstitial elements, whereas they canbe in austenite. Intermetallic phases can formmore rapidly in ferrite. This becomes an issueonly when total chromium plus molybdenumexceeds about 20%, above which the sigmaphase appears. This is thus only an issue forsuperferritic (high chromium content) alloys orfor the ferrite phase of duplex (ferrite-austen-ite) alloys.

The mechanical properties of the ferrite phaseare discussed extensively in Chapter 8, “FerriticStainless Steels.” Here, it is only necessary tonote that ferrite in stainless steel closely resem-bles low-carbon steel in mechanical behavior. Itshares the following characteristics:

• A toughness transition that occurs aroundroom temperature

• Notch sensitivity• A yield point phenomenon• Pronounced crystallographic anisotropy of

mechanical properties• High stacking fault energies and low work-

hardening rates

These issues are dealt with in the same way asin carbon steel when these characteristics be-come an issue. The first two are controlled byreduction of interstitial levels and refining ofgrain size. The yield point is eliminated byslight elongation by temper rolling or elimina-tion of interstitial carbon and nitrogen, whoseinteraction with dislocations causes the yieldpoint. The anisotropy is either utilized to advan-tage by maximizing it, as in the case of deep-drawing alloys, or minimized by refining grainsize and randomizing grain orientation by spe-cial thermomechanical processing.

Ferrite has a greater thermal conductivityand lower thermal expansion than austenite. Itsstrength decreases with temperature more thanthat of austenite, but the good match in thermalexpansion between the ferrite and its oxidestill makes it an excellent high-temperature


material. Ferrite has very nearly the same cor-rosion resistance as austenite, but since ferritecan hold no nitrogen in solution, it cannot ben-efit from this element. In duplex alloys, theferrite is generally the more corrosion resistantphase because it is richer in chromium andmolybdenum.

Austenite

The second major constituent phase of thestainless steel alloy system is austenite. Austen-ite has an fcc atomic structure. The fcc structureis common in many transition metals to the rightof iron in the periodic table. As stated, the fccstructure should be considered normal for metalswell below their melting temperature as it is adenser structure. The presence of the bcc struc-ture relates to the unpaired 3d electrons, whichprovide ferromagnetism. Adding elements toiron that causes pairing of the 3d electrons di-minishes ferromagnetism and promotes the fccstructure. Nickel and manganese are the mostprominent alloying elements that do this, but theinterstitials carbon and nitrogen are the mostpowerful austenite stabilizers on a percentage

basis. Their use is limited by their solubility andtheir tendency to form precipitating compoundswith chromium. Manganese acts largely throughits ability to promote nitrogen solubility. Super-austenitic stainless steels, such as S34565, use 4to 6 % manganese to permit nitrogen levels of0.4 to 0.6% to be achieved, resulting in higherpitting corrosion resistance.

Since all stainless steels contain principallyiron and chromium, the addition of a substantialamount of austenitizing elements is necessary totransform the structure to austenite. As a rule ofthumb, iron alloys require about 17% chromiumand 11% nickel (or its equivalents) to remainaustenitic at room temperature. One percentnickel can be replaced by about 2% manganeseas long as nitrogen is present to maintain thesame phase stability. The omnipresent carbonand nitrogen have an effect 30 times that ofnickel, so even in the small amounts in whichthey are normally present, they have a signifi-cant effect. These stabilizing factors are mappedin the Schaeffler diagram of Fig. 2 (Ref 1),whose purpose is to predict the phase makeupof weld metal. Since welds solidify relativelyrapidly, no carbides or intermetallic phases

Fig. 2 Schaeffler-Delong constitution diagram showing phases present in as-solidified stainless steels at room emperature as a func-tion 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 PhasesDiagram for Stainless Steel Weld Metals, Met. Prog., Vol 77, Feb 1960, p 98


form, and only ferrite, austenite, and martensitewill be present. Thus, they provide useful infor-mation about the compositional effects on phasedevelopment in nonequilibrium situations. Thenickel equivalent (vertical axis) summarizeshow nitrogen, carbon, and other elements com-bine to create a nickel-like effect. The horizon-tal axis does the same for chromium and thoseelements that have a similar effect.

In most common stainless steels, austenite isnormally present in the metastable state, for ex-ample, the retained austenite in alloy steels. Thosewith carbon above 0.02% would eventually breakdown into austenite plus carbides, and those withless than about 30% chromium plus nickel willform martensite if deformed sufficiently. But inthe annealed state, the austenite in standardaustenitic stainless steels will remain indefinitelyas fully austenitic without precipitates unlessheated above 400 °C (750 °C) for protracted peri-ods of time or deformed extensively.

Interstitial elements are much more soluble inaustenite than in ferrite. Of these, only nitrogenis considered a beneficial alloying element. Itboth strengthens and improves the pitting corro-sion resistance of austenite. Carbon has a paral-lel effect, but its tendency to form chromiumcarbides limits its use and in fact leads to itsminimization in most alloys. Before the AODwas developed and carbon levels in stainlesssteels were higher, austenitic stainless steelswere sometimes stabilized by titanium or nio-bium to counter the effects of carbon. Both car-bon and nitrogen stabilize the austenite phase,permitting lower levels of nickel to be used inaustenitic alloys.

Interstitial atoms of carbon and nitrogen dis-tort the fcc lattice, causing it to expand about1% linearly per 1 wt% of solute (Fig. 3) (Ref 2).This produces solid solution hardening of theaustenite. The work hardening of austenite is in-creased by nitrogen. A third interstitial solute,hydrogen, produces the same effect but to alesser degree. Austenite is not embrittled by hy-drogen to the extent ferrite or martensite is, buthydrogen does raise its flow stress and hardnesswhile lowering its work-hardening rate.

Sulfur and oxygen are considered impuritiesbecause they form inclusions, usually chrome/manganese silicates and sulfides. If present insufficient amounts, sulfur and oxygen precipitateas primary inclusions before or during solidifica-tion. In most austenitic stainless alloys, the re-mainder of these elements are near saturation inthe as-solidified ferrite at very high temperatures

and then frozen in a state of supersaturation inthe austenite when it forms on cooling. The sul-fur and oxygen then precipitate during cooling orsubsequent hot working as isolated inclusions.The interface between these inclusions and thematrix is the locus of corrosion pit initiation,quite probably because of chromium depletionoccurring during and as a result of inclusiongrowth. When an alloy solidifies as austenite,sulfur immediately segregates to the grainboundaries because of its low solubility inaustenite, and it forms a low-strength film with alow melting temperature. This causes poor hotworkability and hot cracking of welds.

The diffusion rates in austenite are quite lowcompared to ferrite, so even interstitial elementscannot move quickly enough to precipitate belowabout 400 °C (750 °F). This permits carbon andnitrogen to exist in very high degrees of supersat-uration if introduced below this temperature, asis done by various proprietary processes. Thelow diffusion rates restrict such colossally super-saturated zones to thin surface layers, but theycan reach phenomenal hardness of over Rc 70.The austenite structure does not discourage theformation of intermetallic compounds such assigma, but it does, fortunately, make their forma-tion very sluggish, as seen in Fig. 4. The differ-ence of three orders of magnitude for carbideformation reflects the difference between the dif-fusion of carbon and that of substitutional ele-ments. The formation of sigma in ferrite is about100 times faster than in austenite. Sigma is al-most never seen in commercial 316 alloys.

Fig. 3 Lattice expansions due to carbon. Source: Ref 2


The mechanical properties of austenite arequite different from those of ferrite. Austenite ischaracterized by:

• Low stacking fault energies leading to highwork-hardening rates

• Good toughness even at very low tempera-tures

• Low notch sensitivity• Lack of a sharp elastic limit• Good high-temperature strength• Fairly isotropic mechanical properties

While there is not a great deal of differencein the yield strengths of austenitic and ferriticalloys of similar alloy levels, austenitic alloysare more ductile, have high work-hardeningrates, and therefore have higher tensilestrengths. Austenite can be cold worked to ex-tremely high strengths, around a maximum of2000 MPa (290 ksi). Chapter 3, “AusteniticStainless Steels,” gives a more thorough andquantitative treatment of the mechanical prop-erties of austenite.

In duplex stainless steels, a secondary austen-ite, γ2, can form from ferrite below 650 °C(1200 °F). At this temperature, it has the samecomposition as the ferrite from which it formsand is called type 1. In the 650 to 800 °C (1200to 1470 °F) range, a range that can be encoun-tered in the heat-affected zone (HAZ) at γ/δboundaries during welding, another type forms.This so-called secondary austenite, γ2, type 2, issomewhat enriched in nickel over the ferritefrom which it forms but poorer in nitrogen thanthe primary austenite, giving it poorer corrosionresistance. Secondary austenite can also coformwith sigma as γ/δ grain boundaries are depletedof chromium. This secondary austenite is calledtype 3 and is also poor in chromium.

The physical properties of austenite com-pared to ferrite include lower thermal and elec-trical conductivity and greater thermal expan-sion. It is also, of course, nonmagnetic.

Martensite

Martensite is a phase that forms from the dif-fusionless shear of austenite to a distorted cubicor hexagonal structure. This transformation canoccur spontaneously on cooling or isothermallywith externally applied deformation. It is essen-tially ferrite that has been formed with a super-saturation of carbon. The resulting structure isvery fine and highly faulted, making it quitehard. As in carbon steel, the hardness of themartensite increases dramatically with intersti-tial content because of the huge strain intersti-tial elements impose on the bcc lattice, distort-ing it into tetragonality.

Martensite in stainless steels is restricted toalloy levels at which austenite can form athigher temperatures but at which the austenite isunstable at ambient temperatures. This givesmartensite a fairly narrow composition range.The lowest alloy level is that of the basic 12%chromium steels with 0.1 to 0.2% carbon. Themost highly alloyed martensites are found in theprecipitation-hardening grades. Thus, marten-sitic stainless steels are inherently limited incorrosion resistance to a level no better than a17 or 18% chromium alloy and often barelyqualify as stainless after the chromium tied upas chromium carbide is recognized as not con-tributing to the corrosion resistance.

The as-formed martensite to the degree it hassignificant carbon content is hard and requirestempering to give it adequate toughness. Thetempering reaction is the precipitation of car-bon in the form of carbides with the concurrentloss of internal strain in the martensite lattice.The complexities of tempering require its dis-cussion in detail to be found in Chapter 3,“Martensitic Stainless Steels.” It is worth noting,however, that all tempering involves carbide for-mation, thus losing some corrosion-fightingchromium.

There are two forms of martensite, the ε, ep-silon, and the α' , alpha prime. Epsilon is formedin steels with low stacking fault energy, whichare primarily the leaner austenitic alloys. Thus, itforms at cryogenic temperatures or by coldworking. It appears in martensitic alloys of theprecipitation-hardening type. It is nonmagnetic,has a hexagonal close-packed (hcp) structure,Fig. 4 Precipitation kinetics in 316 stainless steel. Source: Ref 3


and is very difficult to identify microscopically.The a' martensite is the familiar magnetic vari-ety known in alloy steels that forms both byquenching and by deformation.

The mechanical properties of stainlessmartensite are parallel to those of alloy steels.The high quantity of alloying elements instainless give an extreme depth of hardening,so there is no concern with ancillary phasessuch as bainite. The physical properties arevery close to those of ferrite of the same com-position.

Intermetallic Phases

The number of phases that can coexist in analloy is proportional to the number of alloyingelements in the alloy. Table 1 lists data on themore common precipitates found in stainlesssteel. It is not surprising that stainless steel withiron, chromium, nickel, manganese, silicon, andoften molybdenum, titanium, and niobiumshould have numerous ancillary phases. Inter-metallic phases are normally hard and brittle.They can render the bulk alloy brittle when theyform along grain boundaries. The other concernarising from intermetallic phase formation is thedepletion from the surrounding matrix of

chromium or molybdenum, causing localizedlower corrosion resistance. Intermetallic phasesform by diffusion of substitutional alloying ele-ments, which makes their precipitation slowerthan that of carbides, but they can form in amatter of minutes in alloy-rich grades. Defor-mation, which enhances substitutional diffu-sion, accelerates their formation. The principalintermetallic phases are described next.

Alpha Prime. Not to be confused withmartensite, alpha prime is an ordered iron-chromium phase (i.e., iron and chromium atomsoccupy specific, rather than random, sites ontwo intersecting superlattices). This structure isquite brittle. It forms at relatively low tempera-tures, between 300 and 525 °C (570 and 980°F). Before its true nature was understood, itspresence was known through its causing thephenomenon called 475 embrittlement, origi-nally called 885 °F embrittlement. This is some-times confused with temper embrittlement,which occurs in the same temperature range butis caused by phosphide precipitation on prioraustenite grain boundaries of martensite. Alphaprime precipitation can cause 475 embrittle-ment in ferritic or duplex stainless steels andlimits their use in this temperature range but notat higher temperatures, at which the phase dis-solves. This phase forms at chromium contentsas low as 15%, but fortunately it takes a rela-tively long time to form, on the order of hours,so it will not occur inadvertently during thermalprocessing such as welding or annealing.

Sigma. Sigma is a brittle tetragonal phasericher in chromium and molybdenum than ei-ther the ferrite or austenite matrix around it. Itforms preferentially at ferrite-austenite bound-aries in the temperature range 600 to 1000 °C(1110 to 1470 °F) in alloys with more thanabout 18% chromium plus molybdenum. Itscomposition is sometimes given as (CrMo)35(FeNi)65, but examination of the iron-chromiumphase diagram shows that it is archetypically anequiatomic iron chromium compound. It isstrongly promoted by silicon and suppressed bynitrogen. Stabilized alloy grades show morerapid sigma formation than unstabilized alloygrades (e.g., 347 versus 304). In unstabilized al-loys the prior precipitation of carbides destabi-lizes austenite, leading to subsequent sigma for-mation. This makes alloys like 310H,essentially 25Cr-20Ni, especially prone tosigma formation.

Sigma forms much more rapidly from ferritethan from austenite because of the 100-fold

Table 1 Precipitated phases found in stainlesssteelsPrecipitate Structure Parameter, A 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 = 10.57–10.68 Cr16Fe5Mo2C (e.g.)

M23(C,B)6 fcc a = 10.57–10.68 Cr23(C,B)6

M6C Diamond a = 10.62–11.28 (FeCr)21Mo3C; cubic Fe3Nb3C; M5SiC

M2N Hexagonal a = 2.8 c = 4.4 Cr2N

MN Cubic a = 4.13–4.18 CrN

Gamma fcc a = 3.59 Ni3(Al,Ti)prime

Sigma Tetragonal a = 8.80 c = 4.54 Fe, Ni, Cr, Mo

Laves Hexagonal a = 4.73 c = 7.72 Fe2Mo, Fe2Nb phase

Chi phase bcc(b) a = 8.807–8.878 Fe36Cr12Mo10

G-phase fcc a = 11.2 Ni16Nb6Si7,Ni16Ti6Si7

(a) fcc, face-centered cubic. (b) bcc, body-centered cubic.


higher diffusion rate of alloy elements in ferrite.This makes it a much larger issue in superfer-ritic and duplex alloys, which have highchromium and/or molybdenum levels. Chapter7, “Duplex Stainless Steels,” contains an in-depth discussion of sigma.

Chi. Chi, χ, is similar to sigma except it con-tains more molybdenum and less chromium andhas a cubic structure. It can coexist with sigmaand forms in the same temperature range. It alsoprecipitates at ferrite-austenite boundaries andhas the same deleterious effects.

Laves Phase. The laves phase has the struc-ture A2B where A is iron or chromium and B ismolybdenum, niobium, titanium, or silicon. Itforms at 550 to 650 °C (1020 to 1200 °F) overthe course of hours. Thus, although its effectwould be deleterious, it seldom becomes a prac-tical problem. It is possible for it to form at tem-peratures below sigma and above alpha prime,but the long times for formation make it rare.

Carbides, Nitrides, Precipitation Hardening,and Inclusions

Carbon and nitrogen are very important in allsteels, but they take on a special significance instainless steel because chromium, the essentialalloying element of stainless steel, reacts morevigorously with carbon and nitrogen than irondoes. Except for its role in hardening martensiteand strengthening austenite at high tempera-tures, carbon is almost universally a detrimentalimpurity from a corrosion point of view and isminimized. Its beneficial effect on corrosion re-sistance when it is in solution is negligible because so little of it can be held in solution. Nitrogen has a lesser tendency to form com-pounds with chromium, so it is considered abeneficial alloying element in austenite but notin ferrite, in which it has essentially zero solu-bility. Common carbide and nitride precipitatingphases are also listed in Table 1.

Carbides. M23C6 is the main carbide found instainless steel. Its structure is orthorhombic, andit contains both iron and chromium. It can format any temperature at which the host austenite orferrite becomes saturated with carbon. It ismainly chromium carbide, but iron can substitutefor chromium up to about 50%. Other elements,such as tungsten, vanadium, and molybdenum,can also dissolve in this carbide. The ratio ofchromium to iron in the carbide increases withtime and temperature, as chromium diffusionpermits, up to a maximum of 4 or 5 to 1.

The precipitation of the carbide from ferriteoccurs at grain boundaries, is extremely rapid,and cannot be suppressed by quenching. Lessthan 20 ppm carbon content is required to pre-vent its precipitation from ferrite, although upto 50 ppm can be effectively kept in solution byvery vigorous quenching. From austenite, car-bide precipitation occurs below about 900 °C(1650 °F) for carbon levels under 0.10% and at650 °C (1200 °F) for carbon levels below0.03%. For practical purposes, precipitationceases below 500 °C (930 °F) due to the slow-ing diffusion of carbon. While carbon is essen-tially insoluble in austenite at room tempera-ture, quenching can easily preserve up to 0.10%in supersaturation, as is commonly seen in type301 stainless.

The carbide precipitation occurs first at grainboundaries. The chromium that combines withthe carbon comes from the matrix in the imme-diate vicinity and therefore decreases thechromium content of that region, giving rise tothe phenomenon of sensitization, which comesfrom the original phrase “sensitization to inter-granular corrosion.” Nickel and molybdenumdecrease the solubility of carbon and thus accel-erate the precipitation. Nitrogen retards precipi-tation. Cold work accelerates precipitation. Thecarbide has a hardness of about Rc 72. Thismakes the phase a useful constituent in wear resistance in martensitic alloys.

In higher carbon grades such as the marten-sitic stainless alloys, additional, more carbon-rich, carbides may form. These include M7C3and M3C. The latter carbide forms during thelow-temperature tempering of martensite, whilethe former precipitates at higher temperatures.

Stabilizing carbides are those that are formedby the intentional addition of elements such astitanium and niobium. These elements form car-bides of the type MC (metal carbide). The car-bon in these compounds may be replaced by ni-trogen or, in the case of titanium, sulfur. Thesecarbides form preferentially over chromium car-bides and thus prevent sensitization. They pre-cipitate in both the liquid and solid states. In thesolid state, the precipitate normally forms withingrains. The Ti(CN) appears as a cube of goldTiN surrounded by gray TiC. The Nb(C,N) isless regularly shaped. They affect mechanicalproperties in ferrite both by their influence onrecrystallization and by their ability to act as nu-cleation sites for brittle fracture

Nitrides. At low levels, nitrogen can substi-tute for carbon in M23C6. At higher nitrogen


levels, Cr2N can form. This can occur in duplexalloys if they are heated to a solution annealingtemperature at which the alloy has high solubil-ity for nitrogen. Cooling from these tempera-tures can cause the excess nitrogen to precipi-tate as needles of Cr2N. Another nitride CrNcan form in HAZs of welds.

Precipitation-Hardening Phases. Phasesthat have a very similar lattice match to the par-ent phase can precipitate coherently, that is,without changing the continuity of the crystallattice. In these cases, the slight mismatchcauses a strain that can significantly restrict dis-location movement and thereby strengthen thematrix. One such precipitate is gamma prime,an intermetallic, ordered, fcc phase with thecomposition Ni3(AlTi). Copper forms the ep-silon phase, essentially pure copper, whichcauses precipitation hardening. The secondaryhardening of martensite due to the precipitationof molybdenum nitride or carbide is also a pre-cipitation-hardening reaction.

Inclusions. Inclusions are principally oxidesand sulfides that form in the melt (type I), at theend of solidification (type II), or in the solid(type III). Type I inclusions are the largest andare globular. Except when they are deliberatelykept to improve machinability, they are physi-cally removed by various steelmaking practices.Type II inclusions form in interdendritic spacesas the solubility of oxygen and sulfur drop onsolidification. Type III inclusions precipitate theremaining oxygen and sulfur, up to 100 ppm fornormal manganese-silicon killed stainlesssteels, in the solid state either on preexisting in-clusions or as micron-size particles. Inclusionsare mainly oxides and sulfides of silicon andmanganese. If more reactive elements, such asaluminum or titanium, are present, their oxidesand sulfides can also be present.

Sulfides and oxysulfides can be beneficial formachining as solid-state lubricants and chipbreakers. Otherwise, their presence is detrimen-

tal as inclusions have been shown to be the initi-ation sites for corrosion pits, which have beenlinked to both their sulfur ions disrupting thepassive layer and their chromium content caus-ing slight local chromium depletion.

Properties of Stainless Steels

Physical and mechanical properties of repre-sentative stainless steel alloys are summarizedin Appendix 2. Properties are also discussed inchapters specific to each alloy family. Thereader is referred to primary sources, such ascompany web sites, such as Ref 4 and 5.

REFERENCES

1. D.J. Kotecki, Welding of Stainless Steels,Welding, Brazing, and Soldering, Vol 6,ASM Handbook, ASM International, 1993,p 677–707

2. G.E. Totten, M. Narazaki, R.R. Black-wood, and L.M. Jarvis, Failures Related toHeat Treating Operations, Vol 11 ASMHandbook, ASM International, 2002, p192–223

3. High Performance Stainless Steels, Refer-ence Book Series 11 021, Nickel Develop-ment Institute, p 16

4. ASM Handbook, Vol 1, Properties and Selection, ASM International, 1990

5. ASM Speciality Handbook, Stainless Steels,ASM International, 1996

SELECTED REFERENCES

• D.J. Kotecki and T.A. Siewert, WRC 1992Constitution Diagram, Welding Journal, Vol5, 1992, p 171s–178s

Summary

THIS CHAPTER INTRODUCES THE funda-mentals of electrochemical theory as it pertainsto corrosion. Topics covered include an overviewof electrochemical reactions, Faraday’s law, theNernst equation, galvanic versus electrochemicalcells, and Pourbaix diagrams. The examples pro-vided relate these fundamentals to the corrosionresistance of stainless steels.

Introduction

Corrosion—the environmental degradationof materials through electrochemical reac-tions—is a key subject for more or less allclasses of alloys that fall within the broad defi-nition of stainless steels because these alloyswere developed with the intention of prevent-ing corrosion. This chapter aims first to providean introduction to the fundamentals of electro-chemical theory as it pertains to corrosion.Thermodynamics are presented in light of elec-trochemical potentials as opposed to purelychemical ones. Chapter 3 introduces the formalterms needed to describe electrode reaction ki-netics. Chapter 4 describes the various forms ofcorrosion and how they are related to alloymetallurgy, chemistry, and structure. Chapter 5focuses on oxidation. For an in-depth study ofelectrochemical kinetics and electroanalyticalmethods, Ref 1 is recommended. For a broaderstudy of corrosion, the reader is referred totexts by Jones (Ref 2), Uhlig and Revie (Ref 3),and Fontana (Ref 4) and to ASM Handbook,Volume 13A (Ref 5).

Electrochemical Reactions

In electrochemical reactions, charge is trans-ferred across interfaces of species of differentchemistry. Consider, for example, the reaction:

(Eq 1)

An inspection of this reaction suggests thatthree phases must be present for the reaction toproceed: an ion-conducting phase (water-basedsolution), a metallic phase (iron), and a gasphase O2(g). Second, electrons have been trans-ferred from the metallic phase, iron to O2 +H2O. Figure 1(a) shows the arrangement of anexperimental setup in which Reaction 1 couldproceed.

On the left, iron is allowed to dissolveaccording to:

(Eq 2)

resulting in Fe2+ ions that dissolve in the water-based solution and electrons that are car-ried to the right side, where they participate inthe reaction:

(Eq 3)

Inside the water-based solution, ions (Fe2+,OH�, H+, or any others) migrate, thereby con-stituting a so-called ionic current. This currenttogether with Reactions 2 and 3 and the trans-port of electrons from left to right form a closedcircuit called an electrochemical cell. The cell ismade up of four parts: the two electrodes wherethe charge transfer Reactions 2 and 3 take place

O H O e OH2 22 4 4+ + →− −

2 2 42Fe s Fe e( ) → ++ −

2 2 2 42 22Fe s O g H O Fe OH( ) ( )+ + → ++ −

CHAPTER 2

Corrosion Theory




(the anode and cathode, respectively), an elec-trolyte, and an electron pathway. It should benoted that electrodes are interfaces that requireseveral phases to be in contact. Oxidation, Re-action 2, occurs at the anode and reduction, Re-action 3, occurs at the cathode. The electrolyteis the medium through which the ions migrate;in the case of corrosion reactions, this is mostcommonly a water-based solution, but at hightemperatures it could be a solid oxide. The finalconstituent of the electrochemical cell is a path-way through which electrons can migrate fromthe anode to the cathode.

As a shorthand notation, electrochemical cellsare written by separating components within aphase by a comma and separating phases by aslash; gaseous species are written next to theirconducting electrode. For example, the cell de-scribed in Fig. 1(a) would be recorded as

. This cell is an exampleof a differential aeration corrosion cell, which isdiscussed later. Processes at a single electrode

often are described as half cells, for example,and .

Faraday’s Law

If the cell in Fig. 1(a) was allowed to proceedand thermodynamics favored to proceed ac-cording to the direction in Reaction 1, then acurrent i will flow from the anode to the cath-ode, and the amount of charge passed per unittime as a result of this current will be linked tothe amount of iron dissolved per unit time or theamount of oxygen reacted per unit time byvirtue of Eq 2 and 3. This is given by Faraday’slaw:

(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 isthe ratio of electrons to consumed/producedspecies, which in the case of Reaction 2 will be2; and F is Faraday’s constant, which is essen-tially the charge in coulombs corresponding to 1 mole of electrons.

The Nernst Equation

Electrochemical reactions require a transferof charge; hence, there is a coupling betweenchemical and electrical energy. Consider the hy-pothetical setup in Fig. 1(a) with the addition ofa variable resistor and a voltmeter, resulting inthe arrangement shown in Fig. 1(b). Thermody-namically, the Gibbs free energy of the cell isthat of Reaction 1:

(Eq 5)

where �G is the Gibbs free energy, H is the en-thalpy, S is the entropy, R is the gas constant,and T is the absolute temperature. If this is neg-ative, the reaction would be expected to proceedspontaneously as written in Reaction 1. Let usassume that this is the case. The thermal heatproduced by the system can be divided into twoparts: the thermal heat produced by the cell Qtand the heat produced at the resistor QRes. QRes

Δ Δ Δ

Δ

G H T S

G RTa a

a P

= −

= +( ) ( )−0

4 2

ln OH Fe

H O O

2+

2 2

nNF it=

Fe Fe/ 2+Fe O OH/ /2 −

Fe O OH Fe Fe/ / , /2 2− +

Fig. 1 Schematic illustration of (a) a differential aeration cellinvolving iron dissolution and (b) the same cell with a

variable resistor and voltmeter

Chapter 2: Corrosion Theory / 13

in this case is heat, but in essence it representsthe available energy or work, which in the caseof a resistance is given by the product of chargepassed times potential difference. If the resist-ance approached infinity ( ), Reaction 1would proceed through infinitesimal steps andcan be considered thermodynamically reversible.In this case, the thermal heat produced by thecell is minimized and according to thermody-namics is given as Qt = Qrev = T�S1. On theother hand, the net work gained QRes is maxi-mized and constitutes the rest of the free energy:

(Eq 6)

As mentioned, the energy dissipated throughthe resistance is charge passed times potentialdifference, and in this case the potential differ-ence is the reversible potential difference E;thus, in an absolute sense:

(Eq 7)

Here, n is the number of electrons passed peratom of iron reacted, and F = 96,485 C per moleelectrons, is Faraday’s constant. The reversiblepotential difference E represents the potentialdifference between the two electrode reactions(cathode and anode), and as such they are asso-ciated with Reaction 1 rather than a physicalcell. The potential difference is referred to asthe electromotive force (emf) of the cell. It isalso referred to as the open circuit potential be-cause it is the potential measured by the volt-meter in Fig. 1(b) when a negligible currentflows. It is defined here as Erxn. By convention,this potential is positive for a spontaneous reac-tion (as opposed to the chemical free energy,which is negative); hence, Eq 7 becomes:

(Eq 8)

and if all elements have unit activities:

(Eq 9)

Equation 8 is the Nernst equation. By virtueof Eq 8 and 9 and the expression for Gibbs freeenergy of a reaction (e.g., Eq 5), an expressionfor Erxn is obtained:

(Eq 10)

Here, pi and ri are the concentrations of reac-tant and products, respectively, and αi and βi arethe numbers that are needed to balance the reac-tion stoichiometrically. In the case of Reaction1, Eq 10 would be:

(Eq 11)

If the emf according to Eq 11 is positive, thismeans that the free energy is negative (accordingto the Nernst equation); hence, the net reaction isthermodynamically favored as it is written inReaction 1. By inspection of Eq 11, it can beseen that it is the difference between two hypo-thetical half reactions, defined as:

(Eq 12)

which corresponds to the reduction Reaction 3and:

(Eq 13)

which corresponds to the reverse of Reaction2, that is, if it was a reduction reaction. Thepotentials as written in Eq 12 and 13 are called reduction potentials, and because

has to be positive for thereaction to be thermodynamically favored aswritten in Eq 1, the reduction potential EO2 /OH−

has to be larger than . If it was not, thenReaction 1 would proceed in the reverse direc-tion, which means that the electrode Reactions2 and 3 would be reversed and thus so wouldthe anode and cathode of the cell.

It is useful to list reduction potentials for half-cell reactions, just as it is useful to list free en-ergy data. However, half-cell potentials (likeany electrical potentials) cannot be measured inan absolute sense; only potential differences canbe measured. ( can be

measured because it is a difference.) Therefore,half-cell potentials are measured with respect toa reference electrode. Reference electrodes areconstructed such that they have a stable potential;this is discussed further in Chapter 3. A com-mon reference electrode in aqueous solutions is

E E Erxn

= −O /OH Fe /Fe2

– 2+

EFe /Fe2+

E E Erxn

= −−O /OH Fe /Fe22+

E ERT

FFe /Fe Fe /Fe

Fe

Fe

2+ 2+

2+

= −( )⎛

⎝⎜⎜

⎞

⎠⎟0

4

4ln

a

a ⎟⎟

E ERT

F PO /OH O /OH

0 OH

H O O2 2

2 2

− −

−

= −( )⎛

⎝

⎜⎜

⎞

⎠4

4

lna

a⎟⎟⎟

( )E E Erxn

= −−O /OH Fe /Fe22+

E ERT

nF

p p p

r r rrxn rxn= −0 1 2 3

1 2 3

1 2 3

1 2ln

...α α α

β β β33 ...

⎛

⎝⎜

⎞

⎠⎟

ΔG nFE0 0= −rxn

ΔG nFE= −rxn

ΔG nFE=

Q G H T SRes

= = −Δ Δ Δ

R→

E ERT

F Prxn rxn

a a

a= −

( ) ( )⎛

⎝

⎜⎜

−0

4 2

4ln OH Fe

H O O

2+

2 2

⎞⎞

⎠

⎟⎟


the normal hydrogen electrode (NHE), alsoknown as the standard hydrogen electrode(SHE), with a potential set (arbitrarily) as zeroat all temperatures. The NHE is schematicallyshown in Fig. 2. In shorthand notation, it is:

, and the half-cell reac-tion is:

(Eq 14)

Table 1 (Ref 6) lists half-cell reduction stan-dard potentials ( ) versus NHE that are aresult of the emf of the following types of cells(for Reaction 2, as an example):

Galvanic versus Electrochemical Cells

When reactions in a cell occur spontaneouslyin the direction dictated by the open-circuit po-tential of a cell that is positive , a cur-rent flows as shown in Fig. 3(a). This is the casein environmentally caused electrochemical cor-rosion reactions. It also is the case in fuel cellsand batteries (under discharge), in which thecurrent is used as electricity. These types ofcells are called galvanic cells, in which chemi-cal energy is converted to electrical energy.Most of the discussion in the following chaptersconcerns these types of cells. In electrolyticcells (Fig. 3b), an imposed electrical potentialcounters the “natural” cell potential to drive areaction in a desired direction. These types ofcells are used for many metallurgical processes,such as electroplating, electrorefining and elec-troextraction (e.g., the Hall-Heroult aluminumsmelting cell), and for other applications, suchas charging batteries. In the case of corrosion,the principle is used for protection against cor-rosion. In electrolytic cells, electrical energy isconverted to chemical energy.

( )Erxn

> 0

Pt a a a/ ( ) / ( ), ( ) /H H Fe Fe221 1 1= = =+ +

EOx/Re0

2 2 2H H+ −+ =e

Pt a a/ ( )/ ( )H H2 1 1= =+

Fig. 2 The normal hydrogen electrode (NHE)

Table 1 Standard half-cell reduction potentialsversus the normal hydrogen electrode

ReactionStandard half-cell reduction

potential vs. NHE(a) (V)

Fe Fe3 2+ − ++ =e 0.771

O H O OH pH2

–2

2 4 4 14+ + = =−e ( ) 0.401

2 2 2H H+ + =−e 0.000

Ni Ni2 2+ −+ =e –0.250

Fe Fe2+ + =−2e –0.447

Cr Cr3+ + =−3e –0.744

2 2 2 142H O H OH pH2 + = + =− −e ( ) –0.828(a) NHE, normal hydrogen electrode. Source: Ref 6

Fig. 3 Schematic of (a) galvanic cell and (b) electrolysis cell


Corrosion Tendency

The tendency to corrode, that is, whether asystem consisting of anode, cathode, and elec-trolyte can react thermodynamically, is deter-mined by evaluating . If this is positive, thenthere is thermodynamically a possibility for cor-rosion. The rate of corrosion, which is in mostcases determined by corrosion kinetics, is dis-cussed in Chapter 3. Consider, for example, acase of iron in aerated water. Figure 1 (withelectrode Reactions 2 and 3) can be viewed asan idealized equivalent cell for this situation. Itshould be noted, however, that the locations ofanode(s) and cathode(s) on the iron surface can-not be identified with ease. At room temperature(298 K), 1 atm oxygen partial pressure, andusing Table 1, Eq 12 can be written by assumingunit activity for water and unit activity coeffi-cient for OH−:

(Eq 15)

Here, the following definition of pH has beenused: pH = –log CH+, pOH = –log COH− and pH+ pOH = 14. Similarly, the iron dissolution Re-action 2 will have a reduction potential accord-

ing to Eq 13, which, assuming a Fe2+ activity of10−6 (this is an arbitrary value but is usuallytaken to represent a low ion concentration), be-comes at room temperature (using Table 1 forthe standard potential):

(Eq 16)

Figure 4(a) shows a schematic plot of the tworeduction potentials (Eq 15 and 16) versus pH.Because a spontaneous reaction requires Erxnto be positive, if the only pertinent reactionswere Eq 2 and 3, this means that corrosion (dueto iron dissolution to Fe2+ and oxygen reduc-tion) is possible when the line representingEO2/OH− (Eq 15) lies above the line representing

(Eq 16). This is indicated by the regionshaded in gray in Fig. 4(a). Hydrogen reductionis another possible cathode reaction in water:

(Eq 17)

and its reduction potential is (using the defini-tion of pH):

(Eq 18)

E ERT

nF

P

a

pH

H H H H

H

H2

+2

2

+

V vs

+ = −

= −

/ /ln

.

02

0 0 059 .. NHE

2 2 2H H+ −+ →e

EFe Fe2+ /

E aFe /Fe Fe2+ 2+

–0.624 V (

= − −

=

0 447 0 0295. . log( )

vvs. NHE)

E pOH

pH

O OH2–/

. .

. . ( )

= +

= + −=

0 401 0 059

0 401 0 059 14

1.. .227 0 059− pH V (vs. NHE)

Erxn

Fig. 4 Reduction potential versus pH for iron and (a) oxygen gas reduction and (b) hydrogen ion reduction


Figure 4(b) shows the condition in which corro-sion under deaerated conditions (due to iron dis-solution to Fe2+ and hydrogen ion reduction) ispossible as a gray shaded region. In Fig. 4(a)and (b), the regions where iron is stable aredenoted as immunity (corresponding to immu-nity from corrosion). When comparing thesetwo figures, it is noteworthy that hydrogen ionsare able to cause corrosion only under relativelylow pH conditions, whereas oxygen gas is ableto corrode iron in the entire pH range.

The Construction of Pourbaix Diagrams

Figures 4(a) and (b) are types of phase dia-grams that show the stable phases in an areabounded by pH and potential. In reality, severalelectrochemical and chemical reactions need tobe considered when constructing these types ofdiagrams. Each reaction is represented by a line.In the case of iron, the following chemical reac-tions will have to be considered (the pH de-pendency of these reactions is listed next tothem [Ref 7] and since they are not electro-chemical, they are evaluated from the equilib-rium constants):

(Eq 19a)

(Eq 19b)

(Eq 19c)

Since these are independent of potential, theywill appear as vertical lines (see lines 19a to 19cin Fig. 5a).

The following pH-independent electrochemi-cal reactions need to be considered, and theywill result in horizontal lines (Fig. 5a):

(Eq 20a)

(Eq 20b)

The following electrochemical reactions willdepend on pH and thus will be sloped depend-ing on this dependence (Fig. 5a).

HFeO H O Fe OH2

Fe(OH) HFeO32–

2

32

0 8

− −+ = ( ) +

= −

e

E/

,

. 110 0 0591− ( ). log aHFeO2

–

Fe Fe2+

Fe /Fe Fe2+ 2+

+ →

= − + (−2

0 447 0 0295

e

E a

,

. . log ))

F

a

e H O Fe(OH) +3H

pH 1.613– (1/3) log(3

+

F

323+ + =

=,

ee3+ )

Fe(OH) HF O +H ,

pH

2 2+

HFeO2

=

= + ( )−

−

e

14 30. log a

Fe H O OH H

pH

2

Fe

222 2

6 65 0 5 2

+ ++ = +

= − ( )+

Fe

a

( ) ,

. . log

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


(Eq 21a)

(Eq 21b)

(Eq 21c)

(Eq 21d)

For the pH-dependent reactions (chemicaland electrochemical), one can readily label theregions depending on what iron species increas-ing pH favors. If iron would be an anode andthe tendency to corrode were to be evaluated,then the reduction potential for a possible cath-ode reaction would be placed on this diagram. Ifthis point were to be, for example, in A in Fig.5(a), this means that the reduction potential forthis assumed cathode lies below any reductionpotential of iron, and hence under these condi-tions iron is immune (since Erxn is negative). Infact any Fe2+ ions present could plate as iron.On the other hand, if the reduction potential ofthe assumed cathode reaction were to lie inpoint B, then there is a tendency to dissolve ironto Fe2+ since Erxn is positive. Finally, if the re-duction potential of the assumed cathode was atpoint C, corrosion would occur, resulting inFe(OH)3, but when oxides or hydroxides areformed there is a possibility that this productcould form a solid protective layer that kineti-cally hinders further corrosion. These types ofdiagrams are called Pourbaix diagrams. Figure5(b) shows the Pourbaix diagram for iron over-laid with the common cathode reactions inwater, Eq 15 and 18 (Ref 8). The ionic activitywas previously arbitrarily set at 10–6, but fromthe Pourbaix diagram it can be seen that changesin ion activity do not have dramatic effects onthe boundaries. It can be seen that both theoxygen gas reduction reaction and hydrogen ion

reduction are able to cause corrosion throughthe entire pH region. Unfortunately, Fe-OH cor-rosion products are generally not passivating.Iron or carbon steel alloys are therefore not par-ticularly corrosion resistant in water solutions.Figure 6 shows the Pourbaix diagram forchromium (Ref 8). While chromium oxidizeseven more readily than iron, it forms Cr2O3 overa significantly large region that is of relevanceto pH values in water solutions. Since Cr2O3 isprotective, it prevents further corrosion. Whenchromium is added to iron as an alloying ele-ment, it corrodes selectively due to its low re-duction potential, but this means that it also pro-tects the iron alloy due to the properties ofCr2O3. This is the basic design principle behindiron-chromium-based stainless steels.

REFERENCES

1. A.J. Bard and L.R. Faulkner, Electrochemi-cal Methods: Fundamentals and Applica-tions, 2nd ed., Wiley, 2001

2. D.A. Jones, Principles and Prevention ofCorrosion, 2nd ed., Prentice Hall, 1996

3. H.H. Uhlig and R.W. Revie, Corrosion andCorrosion Control: An Introduction to Cor-rosion Science and Engineering, 3rd ed.,Wiley, 1985

4. M.G. Fontana, Corrosion Engineering, 3rded., McGraw-Hill, 1986

Fe(OH) H O Fe(OH) H2

Fe(OH) /Fe(OH)3 2

2 3+ = + +=

+ −e

E

,

00 271 0 0591. .− pH

Fe H O Fe OH H2+2 3

Fe(OH) Fe32+

+ = ( ) + +

=

+ −3 3

1 0

e

E

,

./

557 0 1773

0 0591

−

− ( ).

. log

pH

aFe2+

Fe H O HFeO H2+

HFeO /Fe2-

+ = + += −

− −2 3 2

0 495 0 02 e

E

,

. . 8886

0 0295

pH

a+ ( ). logHFeO2

–

Fe H O Fe OH H2 2

+

Fe(OH) Fe2

+ = ( ) + +

= −

−2 2 2

0 0470

e

E/

,

. −− 0 0591. pH

Fig. 6 Pourbaix diagram for chromium in water. Source: Ref 8


5. ASM Handbook, Vol 13A, Corrosion: Fun-damentals, Testing, and Protection, S.D.Cramer and B.S. Covino Jr., Ed., ASMInternational, 2003

6. Handbook of Chemistry and Physics, 71sted., CRC Press, 1991

7. D.A. Jones, Principles and Prevention ofCorrosion, 2nd ed., Prentice Hall, 1996, p. 59

8. S.A. Bradford, Corrosion Control, 2nd ed.,CASTI Publishing, Inc., 2001, p 41

SELECTED REFERENCE

• M. Pourbaix, Atlas of Electrochemical Equi-libria in Aqueous Solutions, NACE, 1974

Summary

CORROSION INVOLVES chemical reac-tions with equilibrium that is known throughthermodynamics. In practice, the rate at whichcorrosion reactions occur is the most importantconsideration. This chapter deals with corro-sion kinetics, which allows us to understand therates of corrosion.

Introduction

Consider the differential aeration cell dis-cussed in the Chapter 2 on corrosion theory, Fe/O2/OH–, Fe2+/Fe. If the thermodynamic condi-tions favor electrochemical corrosion of iron,that is, is positive, then a net corrosion current i will flow, resulting iniron dissolution and consumption of oxygen gasaccording to the net reaction, 2Fe (S) + O2 +2H2O → 2Fe2+ + 4OH–. The magnitude of thiscurrent will determine the rate or iron dissolu-tion according to Faraday’s law, which wasintroduced in Chapter 2: nNF = it. Because n =2 and F = 95,485 C per mole electrons, themoles of dissolved iron are given as a functionof time as N = i* t/(2* 95,485). Practically, thiscan be readily converted to lost mass m, whichin the case of iron loss becomes m = N* MFe =MFe* i * t/(2* 95,485), or thickness reduction r,which in the case of iron becomes r = MFe* i*t/(2* 95,485* A* �Fe). Here, MFe and �Fe aremolar mass and density of iron, respectively. Itis often the thickness loss (referred to aspenetration per unit time) that is useful;therefore, i/A is often replaced by j, which isdefined as current density and has the unitsamperes/square meters. A general equation of

the penetration due to dissolution of element ibecomes:

(Eq 1)

The penetration rates for iron and variousstainless steels are listed in Table 1 (Ref 1) inunits of mils (0.001 in.) per year, or mpy. In thecase of alloys, the ratio of Mi /ni is computed asan equivalent weight (EW) according to:

(Eq 2)

where fi, ni, and Mi are the weight fraction, va-lence, and molar mass of element i,respectively.

As stated, the amount of corroded (dis-solved) iron is determined by the current i, andthe magnitude of this current is determined by

EWf nM

i i

i

=∑

1

r jt M

n Fi

i

= ⋅⋅ ⋅ ρ

E E Erxn

=O OH Fe Fe– 2+

2 / /–

CHAPTER 3

Corrosion Kinetics

Table 1 Penetration rates for a current of 1µA/cm2 (mpy)

Alloy Element/oxidation stateDensity,

g/cm3 EW(a)Penetrationrate, mpy

Fe Fe/2 7.87 27.92 0.46304 Fe/2,Cr/3,Ni/2 7.9 25.12 0.41321 Fe/2,Cr/3,Ni/2 7.9 25.13 0.41309 Fe/2,Cr/3,Ni/2 7.9 24.62 0.41316 Fe/2,Cr/3,Ni/2,Mo/3 8.0 25.50 0.41430 Fe/2,Cr/3 7.7 25.30 0.42446 Fe/2,Cr/3 7.6 24.22 0.4120Cb3 Fe/2,Cr/3,Mo/3,Cu/1 7.97 23.98 0.39

(a) Equivalent weight. Source: Ref 1




the corrosion potential. The corrosion potentialis determined by the reaction potential (whichwas discussed in Chapter 2) and the kinetics ofthe various steps involved in completing theelectrochemical circuit depicted in Fig. 1.These involve: (a) electrode reactions at thecathode and anode, (b) conduction of ions inthe electrolyte, and (c) conduction of electronsfrom the anode to the cathode. The conductionof electrons is generally not a problem in stain-less steels because the corroding metal (iron)and scale (Cr2O3) provide an easy path for elec-trons. The other two kinetic processes are dis-cussed briefly in this chapter.

The Butler-Volmer Equation

For the case study 2Fe (S) + O2 + 2H2O→ 2Fe2

+ + 4OH–, the cathode and anode reac-tions are:

(Eq 3)

and

(Eq 4)

The Nernst equation predicts an open circuitpotential of

where

and

When a cell is not under open circuit (i.e., anet current passes through it), the cathode andanode potentials deviate from the half-cell po-tentials, and the electrode states are then definedas being polarized. The polarization is quantifiedas overpotentials �, which are defined by the de-viation from the equilibrium half-cell potentials,that is, for the cathode, and η

cE E=

cathode– O / OH2

E ERT

F

a

aFe Fe Fe Fe

Fe

Fe

2+ 2+

2+/ /

ln= −( )⎛

⎝⎜

⎞

⎠⎟0

4

4

E ERT

F

a

a PO /OH O /OH

0 OH

H O O2

–2

–

–

2 2

= −( )⎛

⎝

⎜⎜

⎞

⎠4

4

ln ⎟⎟⎟

E E Erxn

=O OH Fe Fe– 2+

2 / /–

2 2 42Fe Fe( )s e→ ++ −

O H O OH2 22 4 4+ + →− −e

Fig. 1 Schematic illustration of a differential aeration cellinvolving iron dissolution. Kinetic steps: (1) electrode

reactions, (2) ion conduction, (3) electron conduction

for the anode, . Effectively, theoverpotential reduces the activation energy forthe electrode Reactions 3 and 4. In the case of areduction reaction at a cathode, such as Reac-tion 3, the overpotential is negative, and drivingan electrode toward a lower potential driveselectrons from the electrode into the solution,resulting in a net cathodic current ic at this elec-trode. Similarly, at the anode the overpotentialis positive, which results in electrons that are fa-vored to be removed from the solution andtransferred into the electrode, thus producing anet anodic current ia.

If the magnitudes of cathode and anode polar-ization are large, as would be expected in a gal-vanic cell, the relation between each electrodecurrent/current density and overpotential isgiven by the following equations (for a thor-ough derivation of the current overpotentialequation, Ref 2 is recommended):

(Eq 5)

and

(Eq 6)

where the jo,i terms are the exchange currentdensities and represent the equally large cath-ode and anode currents at equilibrium (zerooverpotential) at the electrodes. The exchangecurrent densities are a measure of the electrocat-alytic ability of the surface to promote/demotethe electrode charge transfer reactions; as such,they can vary over many orders of magnitudedepending on the surface chemistry and struc-ture and on electrode reaction. The α-terms arefractions that define the amount to which the

ji

Aj

nF

RTaa

aa

a= =−⎛

⎝⎜⎞⎠⎟0

1, exp

( )α η

ji

Aj

C t

C

nF

RTcc

cc

c= = −⎛⎝⎜

⎞⎠⎟0

2

2

0,

( , )exp

*

O

O

α η

ηa

E E=anode

– Fe + Fe2

Chapter 3: Corrosion Kinetics / 21

activation energies are lowered. They do nothave to be the same for the anode and the cath-ode, but due to the uncertainty in evaluatingthem, they are often taken as 0.5. The concen-tration terms represent the ratios between the re-actant concentration at the electrode/electrolyteinterface and bulk, which could deviate fromunity as a result of consumption/production ofspecies at the interface. In an iron-based alloy,this ratio for the anode would be close to unitybecause the reactant is iron itself, and noconcentration gradient would be expected as aresult of the corrosion reactions. When a netcorrosion current flows, icorr = ia = ic. If the cath-ode and anode areas are assumed to be equal,then jcorr = ja = jc and Eq 5 and 6 can be rewrit-ten (using βc = 2.3RT/(αnF) and βa = 2.3RT/((1 � α)nF) as:

(Eq 7)

and

(Eq 8)

Tafel Regime: Electrode-Kinetics Control.If the electrode charge transfer reactions arerate limiting, the supply of oxygen to the reac-tion site would be rapid enough to maintain aconcentration at the electrode close to that ofthe bulk. In this case, Eq 7 and 8 would both re-sult in a linear dependence of the overpotentialsversus log jcorr:

(Eq 9)

and

(Eq 10)

Mass Transfer Control. In Eq 7, the term:

stands for the ratio of oxygen gas concentrationat the electrode/electrolyte interface and theconcentration in the bulk, sufficiently far awayfrom the interface. If the electrode reaction ki-netics are very fast, the depletion of oxygen willlead toward a zero oxygen concentration at the

electrode, and the rate of cathode reaction willdepend on how rapidly oxygen molecules dif-fuse to the electrode/electrolyte interface. As alimiting case, when the oxygen concentration isactually zero at the interface, the corrosion cur-rent can, through Faraday’s law, be coupled tothe steady-state flux of diffusive oxygen supplythrough a boundary layer δ. This limiting casecurrent is called the limiting current (iL or jL)and can be expressed as:

(Eq 11)

In a nonlimiting case, the correspondingequation would be:

(Eq 12)

Combining Eq 11 and 12, one obtains:

(Eq 13)

Thus, Eq 7 can be written:

(Eq 14)

The slower the diffusion (small d), the lowerthe limiting current and thus a larger contribu-tion from the mass-transfer-dependent secondterm on the overpotential.

Migration and Ionic Diffusion

The ionic transport in the electrolyte phase,the flux of an ion i under an electric field φacross a distance L, can be shown to be:

(Eq 15)

In an electrolyte with many different ions, anion current through an area A can be computedby multiplying Eq 15 with zi * A and summingthe contribution from all ions:

i FA z DC x

x

F A

RTi i

i=∂

∂++

+∑( ) 2

J Dc

x

z F

RTD c

Li ii i

i i= −

∂∂

− Δφ

η β βc c

c

corrc

corr

i

j

j

j

j= + −

⎛

⎝⎜⎞

⎠⎟log log,0 1

C t

C

j

jcorr

l

O

O

2

2*

( , )01= −

jD nF C C t

corr=

−O O O2 220( ( , ))*

δ

j jD nFC

corr l= = O O2 2

*

δ

C t

CO

O

2

2*

( , )0

η β βa a corr a a

j j= −log log ,0

η β βc c c c corr

j j= −log log,0

η βa a

corr

a

j

j= log

,0

η β βc c

c

corrc

j

j

C t

C= +log log

( , ),0 2

0O

O2*


(Eq 16)

Because the first term is important only at theregions near the electrodes (where consump-tion/creation of species occur), the current inthe majority region of the electrolyte can be es-timated as:

(Eq 17)

Using Ohm’s law (R = U/i), the electrolyte re-sistance can be computed as:

(Eq 18)

The resistivities of some test solutions areshown in Table 2.

Mixed Potential Theory and Polarization Diagrams

Viewing the electrochemical cell as an elec-trical circuit, Kirchoff’s law can be used todesign a so-called polarization diagram. Con-sider, as a case study, a steel corroding underdeaerated conditions, in a water solution, asshown in Fig. 2. Assume that the pH is such thata passive layer does not form (see the discus-sion of Pourbaix diagrams in Chapter 2). Thecathode and anode reactions, respectively, are:

(Eq 19)

(Eq 20)

and the respective equations describing theoverpotentials will be:

(Eq 21)

and

(Eq 22)

Now, the potentials of anode and cathodewhen current is flowing are in each case theequilibrium potential plus overpotential, that is:

(Eq 23)

(Eq 24)

A polarization diagram is now constructed byplotting the anode and cathode potentials versuslog jcorr. Strictly speaking, to close the circuit,

a= − +0 624. β llog

,

j

jcorr

a0

E aanode a

= − − ( ) +0 447 0 0295. . logFe2+ η

+ −⎛

⎝⎜⎞

⎠⎟β

ccorr

i

j

jlog 1

= − pH.0 0 059 ++ βc

c

corr

j

jlog ,0

E ERT

nF

P

acathode c= − +

H H

H

H

+2

2

+/

ln02

η

η βa a

corr

a

j

j= log

,0

η β βc c

c

corrc

corr

i

j

j

j

j= + −

⎛

⎝⎜⎞

⎠⎟log log,0 1

Fe Fe( )s e→ ++ −2 2

2 2 2H H+ −+ →e

R LF A

RTz C Delectrolyte i i i=

⎛⎝⎜

⎞⎠⎟∑/

22

iF A

RTz C D L

i i i≅ ∑

22 Δφ /

z C D Li i i

× ∑ /2 Δφ

Table 2 Test solution resistivity

Test solutionRatio by volume

Resistivity, ohm-cm

Natural Seawater . . . 25Fresh (tap) water adjusted with seawater

28:1 500

Fresh (tap) water adjusted with seawater

68:l 1,000

Fresh (tap) water adjusted with seawater

950:1 3,000

Deionized water adjusted with fresh (tap) water

21:10 10,000

Fig. 2 Schematic polarization diagram


the potential drop across the electrolyte needs tobe included, which simply equals icorr * Relectrolyte(the electrolyte resistance is evaluated from Eq18); however, in many cases, this term can beneglected. A schematic polarization diagram isshown in Fig. 3. The anode polarization is linearwith decade current as predicted by Eq 24 be-cause the overpotential has only a Tafel regimeand no mass transfer dependence. On the otherhand, the cathode polarization deviates from theTafel behavior as a result of the effect of themass transfer (hydrogen ion supply), dependenton the limiting current in Eq 23. It is notewor-thy that the cell shown in Fig. 1 does not have amacroscopic anode and cathode. Different mi-croscopic regions on the surface are assumed toact as cathodes and anodes, and in the lack ofmore detailed knowledge, the cathode andanode areas are assumed to be equal. The over-all mixed potential of the surface would be at acorrosion potential Ecorr, defined in Fig. 2.

In effect, the corrosion current resulting fromthe cell depends on the equilibrium half-cell po-tentials (Ecathode and Eanode), the Tafel slopes (βcand βa), the exchange current densities (jo,c andjo,a), and any limiting current density (jl). Figure3 shows schematically how decreasing any of theTafel slopes and increasing an exchange currentdensity increases the corrosion rate. The effect ofthe electrolyte resistance has been ignored; thatis, corrosion current is where the two polariza-tion curves intersect. Figure 4 shows the effect ofincreased mass transfer, which would result in anincrease in the limiting current. In an active (non-passive) alloy, this results in an increased corro-sion current up to a point.

Passivation

Theory. In Chapter 2, it was identifiedthrough the Pourbaix diagrams that there wereconditions under which an alloy could bepassive. In the case of stainless steels, the rangeof pH and other conditions under which thiswould occur has been increased thanks to thechromium content, which readily forms a Cr2O3scale. In general, a passive layer constituted ofadsorbed molecules or thin oxide/hydroxidelayers decreases the corrosion current. Re-searchers (Ref 3) have reported that the con-stituents of the passive film are alpha Cr2O3 andCr(OH)3nH2O. The structure is reported to be ananocrystalline spinel, epitaxial to the surface.The grain size may decrease with increasingchromium content. This protection bychromium requires a threshold level of 11 to12% chromium.

Effect on Polarization Diagrams. The polar-ization diagram for a passive alloy is quite dif-ferent from those discussed for active alloys. Aschematic of a typical polarization curve isshown in Fig. 5. When a passive alloy is anodi-cally polarized, it initially behaves like an activealloy (i.e., with a Tafel slope, etc., as the pas-sive layer is building up). The building up isactually a selective dissolution of iron, whichcauses a greater remaining surface concentra-tion of chromium and other alloying elements.Once the passive layer is formed and offersprotection against further dissolution, the po-tential-decade current relation drops to lowercurrents. This happens at potentials beyond thepassivation potential Epp. At some high enough

Fig. 3 Corrosion rate and the effect of (a) Tafel slope and (b) exchange current density


polarization level, the passive layer breaksdown, and the metal becomes active again; thisregion is called the transpassive regime. Thedesign of a structure involving a passive metalshould aim at forming a corrosion cell in whichthe cathode polarization curve intersects the an-odic one in the passive regime.

Consider, for example, an alloy that exhibitsthe behavior shown in Fig. 6 in deaeratedacidic solutions with different pH. If masstransfer limitations due to hydrogen ion supplyare neglected, then the cathode polarization isgiven by:

(Eq 25)

This will result in a straight line as shown inFig. 6, which will be shifted vertically depend-ing on the pH. The dashed circles indicate theintersection between anode and cathode polar-ization curves that would yield the corrosioncurrent. At a sufficiently high pH (= pH1), thealloy is clearly not optimal because intersectionoccurs in the active regime, and the passiveproperties are not utilized. This is what occurswhen a reducing acid is too strong for a givenstainless steel, such as with concentrated hy-drochloric acid. At pH = pH2, on the otherhand, a low-corrosion current is obtained as aresult of intersection at the passive regime.This is the benevolent case when stainless steelis correctly matched to the environment, andlow rates of uniform corrosion occur. Finally, atpH = pH3, the resulting corrosion current isagain high as a result of intersection occurringat the transpassive region. This could occurwith some stainless steels exposed to a verystrong alkali solution. Using a similar argu-ment, the readers can themselves deduce theeffects of cathode exchange current density andTafel slopes. In the discussion of active anodepolarization, it was found that increasing thetransport rate of cathode reactants throughagitation, for example, would increase the cor-rosion rate up to a point but beyond that haveno further effect (see Fig. 4). In the case of apassive/active behavior, the effect of mass

= − pH.0 0 059 ++ βc

c

corr

j

jlog ,0

E ERT

nF

P

acathode c= − +

H H

H

H

+2

2

+/

ln02

η

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 fur-ther effect

Fig. 5 Schematic of a passive anode polarization curve Fig. 6 Effect of cathode polarization


transport is somewhat different, as shownschematically in Fig. 7.

Increasing mass transport, such that the limit-ing current increases, results initially in an in-creased corrosion current (e.g., increasing jlfrom 1 to 2). It should be noted that there areseveral intercepts possible (both in the activeand passive regime), but assuming there aredefects present, it is likely that there will be cor-rosion corresponding to the higher current. In-creasing the limiting current beyond the kneecorresponding to Epp, however, results in a dropin the current because now the only corrosionpotential possible is at the intersection in thepassive regime. This is the case for jL3.

In the normal use of stainless steel, achievingpassivity takes on several forms. What is oftencalled passivation is actually a cleaning processin which contaminants, such as tramp iron, areremoved from the surface. Dilute nitric acid isan excellent vehicle to achieve this. Thismedium has the additional benefit of forming apassive film on an active stainless surface. This

is the actual passivation; the iron removal isreally a chemical cleaning operation, which hap-pens to be called passivation. During the produc-tion of stainless steel, after a final anneal anotherversion of passivation is carried out. The oxidefrom annealing in air is dissolved by a strongmixture of nitric and hydrofluoric acids, whichdoes not allow passivation. This treatment,called pickling, removes by dissolution both theoxide layer and the chromium-depleted layerbelow the oxide formed during annealing. Thedepleted layer can extend a number of micronsin depth and would seriously degrade corrosionresistance if not removed (Ref 4). This is thenfollowed by a straight nitric acid immersion,which ensures complete passivity. This is theprocedure that should be performed on the ox-ides formed during welding if full corrosion re-sistance is to be restored. Simply removing theoxide through mechanical means leaves achromium-depleted layer that corrodes morereadily than is expected of the alloy.

REFERENCES

1. D.A. Jones, Principles and Prevention ofCorrosion, 2nd ed., Prentice Hall, 1996

2. A.J. Bard and L.R. Faulkner, Electrochemi-cal Methods: Fundamentals and Applica-tions, 2nd ed., Wiley, 2001

3. M.P. Ryan et al., Critical Factors in Focal-ized Corrosion, Proc. Electrochem. Soc.,Vol 150, 2003, p 583–594

4. J. Grubb and J. Maurer, “Corrosion of theMicrostructure of a 6% Molybdenum Stain-less Steel with Performance in a Highly Ag-gressive Test Medium,” paper 300 pre-sented at Corrosion 95, NACEInternational, 1995

Fig. 7 Effect of mass transport

Summary

STAINLESS STEEL is unusual among alloysystems in that its corrosion resistance derivesfrom the passivating ability of a minor con-stituent, chromium. Thus, while stainless steelscan be made to be essentially immune to corro-sion in many environments, it can also experi-ence various debilitating forms of localized cor-rosion, which stem from the failure of thispassive film. This chapter explores the behaviorof stainless steel in media that promote uniformcorrosion and the various mechanisms of local-ized corrosion, such as pitting and crevice cor-rosion.

Introduction

To most designers, the most recognized char-acteristic of stainless steel is corrosion resist-ance. Stainless, unlike noble metals such asgold, does not obtain its excellent corrosion re-sistance from inertness. Instead, it is the reactiv-ity of chromium that allows the surface layer ofcorrosion product to become sufficiently adher-ent and impenetrable, which effectively stopsfurther corrosion by isolating the base materialfrom the environment. This resistance to corro-sion is called passive behavior or passivity.Other metals, such as aluminum and titanium,form similar layers and also exhibit passivity.The important difference in the case of stainlesssteel is that chromium is still a minor con-stituent, never more than 30% by weight, some-times little more than 10%. How muchchromium there is and how uniformly it is dis-tributed have a profound effect on corrosion re-sistance by virtue of its ability to concentrateinto the surface film. Stainless steels are com-

plex in their behavior because the influence ofprocessing and alloying variables changes theability of this layer to form and remain stable inthe face of aggressive environments. The be-havior of stainless steel is further affected by itsmicrostructural complexity. Stainless steel al-loys may have many constituent elements andmany thermodynamically possible phases, andnone of these are necessarily uniform in theircomposition. Yet, it is the composition of thealloy in contact with the specific environment atany microscopic point that determines the cor-rosion resistance of that particular point.

Uniform Corrosion

When all parts of a corroding surface haveequal access to the corroding atmosphere andthe structure of the corroding metal is relativelyuniform, a uniform thinning of the material isexpected. Stainless steels are materials of choicebecause, by virtue of their passive behavior, theyshow very low rates of uniform corrosion inmany environments. The metallurgy and pro-cessing of a particular grade are designed to pro-vide passivity in a given environment. The envi-ronment can be too aggressive to allow passivityto be maintained either by being too reducing, aswith some acid media, so that passivatingspecies cannot form or by being too oxidizing sothat the oxidized species that normally affectpassivity are no longer stable. The former iscalled dissolution in the active state, while thelatter is termed transpassive dissolution.

Intelligent design and knowledge of the envi-ronmental variables for a stainless steel compo-nent ensure that the alloy is used in the passivestate, at which uniform corrosion occurs at avery low rate.

CHAPTER 4

Corrosion Types




Among the important media with which weencounter uniform, but acceptably controlled,corrosion in stainless steel are atmospheric andmarine environments and chemical environ-ments such as sulfuric acid, phosphoric acid, ni-tric acid, strong bases, and organic acids, suchas acetic and formic.

Pickling is an example of controlled, acceler-ated uniform corrosion. This is typically donewith 10 to 20% hot sulfuric acid or a mixture ofhydrofluoric and nitric acids.

Environmental Variables Influencing Uniform Corrosion

The corrosion of stainless steels is usually theresult of contact with an electrolyte, allowing acomplex set of partial electrochemical reac-tions, which may occur sequentially or concur-rently. The corrosion rate depends on thecurrent exchanged between the negative andpositive electrode (anode and cathode). Thesemay be on a macroscopic or microscopic level.The main consideration is normally ionic trans-por tthrough the passive film, which after all is what makes stainless so effective againstcorrosion.

The chemical parameters that influence themedia with respect to uniform corrosion rate arethe acidity and the oxidation-reduction (redox)potential of the electrolytic medium, both ofwhich act through their influence on the stabil-ity of the passive film, rendering it active, pas-

sive, or transpassive; Fig. 1 illustrates the effectof redox potential on a solution.

Certain anions have strong effects in mediathrough their well-known, if not well under-stood, disruption of the passive film. Halides arewell known for this effect, but sulfides areactive. These anions seem to intervene in theadsorption of the hydroxyl ions. In acid media,these anions accelerate uniform corrosion,while in neutral media they may result in local-ized corrosion. Anions that form soluble com-plexes with elements in stainless, such asamines, formates, or acetates, can also disruptthe stability of the passive film and thus pro-mote active corrosion.

Of the physical variables of the environment,it should be obvious that temperature is para-mount since all the reactions are thermally acti-vated. Increasing temperature may speed theformation of the passive film when thermody-namic conditions are favorable, but in generalone expects increasing temperature to increasecorrosive attack. Access to passivating species,such as oxygen, is important in establishing andmaintaining passivity.

Increased diffusion of reacting species in theliquid will normally accelerate the partial reac-tions, but if passivity is stable, the rate-limitingtransport through the passive film will not be af-fected. Therefore, increasing the flow rate of acorrosive fluid does not automatically acceler-ate corrosion. The reduction of concentrationgradients 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 anincreased supply of passivating species.

Increased flow rate in a fluid medium is dele-terious if it induces mechanical damage to thepassive film by erosion, abrasion, or cavitation.These are complex mechanisms, but it shouldbe apparent that the success of a stainless steelto a given flow condition will depend mainly onits ability to form and re-form its passive film.Somewhat counterintuitively, thinner passivefilms are more protective than thicker filmsamong stainless alloys. The tenacity of the thinpassive films on stainless (and titanium) makethese alloys quite resistant to flow-acceleratedcorrosion, as contrasted to copper and alu-minum alloys, which have soft, thick corrosionproduct films.

Material Variables

Stainless steels have a great variety of alloy-ing elements and microstructures. As a general-ization, we can say that corrosion resistance is afunction of composition rather than structure.Then, we must quickly add the qualifiers to thisstatement. On an undisrupted, stress-free sur-face, local composition does quite precisely de-termine corrosion resistance. But, stainlesssteels are seldom homogeneous or at thermody-namic equilibrium. Impurities such as oxygenand sulfur are usually present, mainly as inclu-sions since they have diminishingly small solu-bility at room temperature. At high tempera-tures after solidification, as in welds, they canbe present in supersaturation, ready to precipi-tate as inclusions that alter local composition.The tendency of carbon and nitrogen to formprecipitates is controlled by diffusion rates,which if elevated by increasing temperature cancause debilitating, composition-altering precipi-tation. The even more slowly diffusing substitu-tional alloying elements, such as chromium,molybdenum, and nickel, have strong tenden-cies to form phases that disturb their uniformityin the austenite or ferrite matrix in which theyare intended to work. So, any discussion of theinfluence of alloying element on corrosion re-sistance of a phase like austenite or ferrite mustrecognize that alloying elements exert their ef-fect when they are in solution in that phase. Thesame element may under some conditions notbe in solution and have a contrary effect. An ex-ample is molybdenum, which is obviously agreat enhancer of corrosion resistance when insolid solution. When it precipitates as a con-

stituent of sigma phase, which it promotes,however, it combines with chromium. If thishappens at relatively low temperatures, the sur-rounding matrix is depleted of both chromiumand molybdenum, and the corrosion resistancein that region is diminished. Nitrogen also is ef-fective when in solid solution in austenite butcan precipitate as a chromium nitride under cer-tain conditions and cause depletion of the re-maining matrix. Local structure and composi-tion are paramount. This becomes moreimportant to localized corrosion, as discussedlater, but it should be remembered in examininguniform corrosion because corrosion will ceaseto be uniform when composition becomesnonuniform.

The compositional material variables that in-fluence uniform corrosion are not exactly thesame as those that will be seen to influence lo-calized corrosion. The foremost element is, ofcourse, chromium. Researchers (Ref 1) have re-ported that the constituents of the passive filmare alpha Cr2O3 and Cr(OH)3nH2O. The struc-ture is reported to be a nanocrystalline and epi-taxial to the surface. The grain size may de-crease with increasing chromium content. Thisprotection by chromium requires a thresholdlevel of 11 to 12% chromium. This thresholdhas been attributed most convincingly to theminimum chromium content that permitschromium atoms on surface sites to be linked byadsorbed oxygen atoms (Ref 2). In any event,the mechanism by which this thin, several-nanometer-thick, film forms is the subject of on-going debate, but we do know that it is enrichedin chromium, and that it is thinner for higherchromium alloys. The critical current density j,as measured during polarization, is also smalleras chromium content increases (Fig. 2). This isconsistent with the lower dissolution of noncon-tributing elements required to achieve a criticalsurface chromium concentration. Increases inchromium can also be seen to lower the currentdensity in the passive region. This is manifest inalloy performance as a reduction in the uniformcorrosion rate in a given medium. From an elec-trochemical point of view, this is explained as amanifestation of the stability of the Cr(OH)3nH2O.

The role of molybdenum is less clear. The ob-served action of molybdenum is to greatly re-duce the critical current density required forpassivation. This is also seen as accelerating theformation of the passive films and as increasingthe resistance of the alloy to depassivation at


lower pH. The role of molybdenum is not to en-rich in the passive film itself, although it can befound in the film. Its potency is far more than itspresence can take into account. Pure molybde-num is itself not passive. Its action does not ap-pear to be via a product of reaction. Instead, itseems to reduce the dissolution rate of elementsother than iron, which would promote a surfacericher in chromium (Ref 3). The action ofmolybdenum as an alloying element is compli-cated by the fact that molybdate ions are knownto impede pit growth as a separate effect fromtheir action within the alloy matrix (Ref 4).Copper has a similarly complicated effect, withcopper ions gettering sulfide ions and redeposit-ing as metallic copper (Ref 5).

Nickel also lowers the critical current densityfor passivation without contributing directly tothe passive film’s stability. This also may be theresult of the stronger bond between nickel andchromium reducing the anodic dissolution rateof the alloy by permitting the anodic enrichingof the surface by selection iron dissolution.Nickel does not actively help passive film for-mation and can actually hinder film stability inhighly acidic/oxidizing environments.

Nitrogen, however, appears to be more likemolybdenum in its effect. While nickel and cop-per provide no benefit to the stability of the pas-sive film once it is formed, both nitrogen andmolybdenum do, and to a degree that cannot beexplained by their presence in the film. Thismay then relate to their thermodynamic actionwithin the alloy itself. Molybdenum and nitro-gen act both to enhance the enrichment of

chromium in the passive layer and to decreaseactive dissolution of noniron alloying elements,thereby promoting both the formation and sta-bility of the passive film. A summary of theknown major alloying effects in acidic chloridemedia is shown in Fig. 3 in acidic chlorides. Al-loying elements provide benefits in the part ofthe chart where they appear (Ref 6). From this,it can be seen that chromium, molybdenum,nickel, copper, and nitrogen all assist in the ac-tive region, while chromium, molybdenum, andnitrogen expand the region of passivity and di-minish the corrosion current.

An example of the influence of these alloyingelements on the uniform corrosion rate of stain-less steels in a sodium chloride/carbon dioxideenvironment is shown in Fig. 4 (Ref 7). Notethe alloying composition is measured by acrevice corrosion index (CCI), which is dis-cussed in the section Localized Corrosion.

Fig. 2 Schematic illustration of polarization behavior for apassive alloy with and without pitting occurring

Fig. 3 Influence of alloying element on corrosion rate as ex-plained by the effect on polarization.Source: Ref 6

Fig. 4 Influence of alloying elements on uniform corrosionrate in 20% sodium chloride solution with carbon

dioxide pressure of 20 MPa. Source: Ref 7


Unfortunately, it cannot be assumed that thisrelationship is true for other environments, al-though other empirical relationships exist or canbe generated. Because the influence of alloyingelement varies with environment, we need todiscuss some of the more commonly encoun-tered severe environments.

Corrosion in Acids and Bases

The examples discussed in Chapter 3, “Cor-rosion Kinetics,” refer mostly to corrosion inthese aqueous solutions, in which the slow thin-ning rate of the chosen alloy can be determinedthrough the mixed potential theory and polar-ization diagrams. In the case of stainless steels,the alloy chemistry is chosen such that thepassive-active behavior favors corrosion in thepassive regime. The corrosion rate of the vari-ous stainless steels in the myriad possible envi-ronments has been measured in probably allpractical cases. These data can be obtained froma number of sources, such as the National Asso-ciation of Corrosion Engineers (NACE) andASM Handbook volumes. None is more accessi-ble than the Web site of Outukumpu, whichcontains a “Steel Professional Tool,” a lookuptable in which the corrosion rate of many stain-

less steels in a great number of environmentscan be obtained. Figure 5 shows an example ofone such table. Many of the isocorrosion chartsin this book are reprinted from this source,http://www.outokumpu.com/applications/docu-ments/start.asp (Ref 8).

These tables are supplemented by isocorro-sion diagrams such as that shown in Fig. 6.These diagrams show constant corrosion behav-ior under varying environmental conditions suchas temperature and solution composition. Thisinformation is available to guide the designer inselecting appropriate steels for various environ-ments, and it is highly recommended that it beused. Free sites tend to promote proprietary al-loys, as these charts suggest. The serious engi-neer will consult multiple sources and unbiasedsources before making alloy decisions.

The influence of alloying element is by nomeans the same in all environments. So, while itis useful and necessary to have these experi-mental data, it is also helpful to understand thepeculiarities of some of the major alloy-envi-ronment pairings.

Sulfuric Acid. Stainless steels require morethan a minimum amount of alloying to resist sul-furic acid. Straight 16% chromium grades suchas 430 fare poorly, while the nickel-containing

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


304 has more than an order of magnitude bettercorrosion rate in either dilute or concentratedsulfuric acid at ambient temperatures. Figure 7(Ref 9) shows the isocorrosion rate curves forseveral common alloys. Alloying with molyb-denum is also very effective, as is alloying withcopper. If passivity cannot be established, in-creasing chromium content actually increasescorrosion rate.

The corrosion behavior of sulfuric acid variesgreatly with concentration. At low concentra-tions, sulfuric is a classic reducing acid. It dis-sociates in water to create hydrated hydrogenions (H3O

+) that release hydrogen gas bubblesduring the corrosion reaction. As the acid con-centration increases, the solutions become morecorrosive, and progressively more highly al-loyed stainless steels are required to provide ad-equate corrosion resistance. At about 50% acid,only the most highly alloyed stainless alloys(alloy 20, AL-6XN, C-276, etc.) can provide ac-ceptable corrosion rates, and even these alloysare restricted to use at near ambient tempera-tures. As acid concentration increases beyond50%, the solution begins to show oxidizing be-havior. At acid concentrations above 80%,nickel-molybdenum-copper-bearing stainlesssteels begin to exhibit useful corrosion resist-ance. In the 93 to 98% sulfuric acid concentra-tion range, carbon steel can be used to hold sul-furic acid at ambient temperatures, althoughstainless steels provide better performance at el-evated temperatures or if flow-erosion can occur.In the 96 to 100% sulfuric acid concentrationrange, at elevated temperatures, the oxidizingcharacter is quite pronounced, and oxidation-

resistant high-chromium (type 310S) and high-silicon (MECS ZeCor UNS S38815 and Sand-vik SX S32615) stainless steels are frequentlyused, especially in sulfuric acid-manufacturing

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

Fig. 7 Isocorrosion rates of various stainless steels in sulfuricacid. Source: Ref 9


equipment. Sulfuric acid-containing dissolvedsulfur trioxide is called oleum, and such solu-tions are often identified as sulfuric acid ofgreater than 100% concentration. High-chromium stainless steels (i.e., type 310S) areamong the very few materials that exhibit corro-sion resistance in oleum. (See MTI MaterialsSelector Volume 3—Sulfuric Acid at www.mti-global.org for more information.)

Aeration has a major influence on corrosionrates because oxygen stabilizes the passive film.Molybdenum-alloyed stainless has dramaticallylower corrosion rates in aerated solutions than

non-molybdenum-bearing alloys. Their superi-ority in deaerated solutions is much lessmarked. Studies (Ref 10) have shown that insulfuric acid molybdenum is highly enriched inthe passive film, and when molybdenum is analloy, chromium also is enriched. This is a man-ifestation of selective dissolution of other ele-ments in the matrix.

Oxidizing impurities, such as ferrous ions, actlike aeration to diminish the corrosive attack,but reducing impurities such as halides have anextremely negative effect, as the corrosion ta-bles will show. These effects are not linear andunderscore the value of these tables.

The uniform corrosion rate in contaminatedsulfuric acid may be more important than inpure acid since this represents a potentiallylikely failure mode because contamination is aconstant hazard.

Figure 8 shows the corrosion rate of variousalloys in sulfuric acid contaminated with chlo-rides and iron. These researchers (Ref 11) foundthat the resistance to attack correlated to thealloy content by the formula shown.

Figures 9 and 10 show how isocorrosion ratesvary with alloy and contamination level.

Hydrochloric acid is very destructive of thepassive film on stainless. An alloy like 304 isnot suitable even in a deaerated 1% HCl solu-tion at room temperature. Chromium additions

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

Fig. 9 Isocorrosion curves for various alloys in sulfuric acid


Fig. 10 Isocorrosion curves for various alloys in sulfuric acidwith chlorides

Fig. 11 Isocorrosion curves for various stainless steels in hydrochloric acid. Source: Ref 8. Courtesy of Out-

okumpu Stainless

Fig. 12 Isocorrosion curves for austenitic AL-6XN (UNSN08367) and 904L (UNS N08394) stainless steels

in hydrochloric acid. Source: Ref 12

Nitric acid is strongly oxidizing. This actu-ally promotes the passive film formation; con-sequently, even low-chromium alloys remainpassive at all concentrations at ambient temper-ature (see Fig. 14) (Ref 8).

The addition of molybdenum, which is sogenerally helpful, is deleterious in this case be-cause it forms soluble compounds. It is usefulto keep carbon, silicon, and phosphorus as lowas possible.

Silicon is unusual in that normal levels (0.4 to1.0 %) are worst, with very low (0.05%) or veryhigh levels (4.0%) beneficial. The low levels ofsilicon contents of these alloys are useful fortheir action in minimizing grain boundary seg-regation, which is the usual locus of attack.High silicon levels contribute to a general pro-tective silica surface layer in concentrated acid,which augments the true passive layer. This su-periority appears above the azeotropic composi-tion of about 67%, which is a common com-mercial concentration, as shown in Fig. 15 forhigh-silicon austenitic stainless steels (Ref 13).

Phosphoric Acid. This oxidizing acid be-haves more like sulfuric acid in that simpleiron-chromium alloys have only moderate re-sistance to uniform corrosion in them, while al-loying with molybdenum and copper producesmajor improvements. This can be seen in Fig. 16,in which alloys with increasing nickel (18-10)show clear benefits over a chromium-molybde-num alloy (18-2), and the added molybdenumin 317 (17-14-4) is better, while 904L withnickel, molybdenum, and copper is even better(Ref 8).

In the commercial production of phosphoricacid, halide impurities may be present, in which

are only modestly helpful, while nickel, copper,and molybdenum are more beneficial. Stainlesssteels are not good materials for contact withhydrochloric acid. Figures 11 (Ref 8) and 12(Ref 12) show how even the most highly al-loyed grades can withstand only dilute concen-trations and low temperatures.

While a stainless steel vessel may not be in-tended to be used for hydrochloric acid, resist-ance to lesser amounts of chlorides is importantbecause of the possibility that an acidic environ-ment may be contaminated with chlorides. Whenthis is a possibility, then proper alloy selectionmust guard against it. Figure 13 shows a correla-tion between alloy content and resistance to thegeneral corrosion (GI) by sulfuric acid contami-nated with hydrochloric acid (Ref 11).


case alloys with higher molybdenum, chromium,copper, and nitrogen may be required.

Organic Acids. The weakly dissociating or-ganic acids are normally not aggressive againststainless steels. The exceptional dangerous en-vironments are those that include high tempera-ture and the presence of chloride contamination.It should be noted that in formic acid, whichdoes dissociate more strongly, nickel is detri-mental. This phenomenon is also seen in theproduction of urea via the intermediary ammo-nium carbamate. The difficulty lies in the high-temperature solubility of nickel complexes andis best addressed by the use of ferritic or duplexalloys.

Alloying with molybdenum seems to providethe greatest resistance to uniform corrosion instrong organic acids, as illustrated in Fig. 17(Ref 8). If halides are present in organic acidsand liberated by contact with water, then pHand chloride concentration will govern the cor-rosive attack, which could then become nonuni-form.

Strong Bases. In strong bases, the stainlesssteels are generally quite resistant to uniformcorrosion. Straight chromium (17%) alloys areusable at any concentration up to 50 °C. Addingmolybdenum and nickel does little to furtherimprove performance as the underlying resist-ance is due to chromium. Increasing chromiumlevels provide increased resistance. Attackwhen it does occur can be manifested as grainboundary attack. Figure 18 shows isocorrosioncurves for sodium hydroxide (Ref 8).

Fig. 14 Isocorrosion curve for nitric acid. Courtesy of Outokumpu Stainless

Fig. 15 Corrosion behavior of high-silicon alloys in concen-trated nitric acid. Courtesy of Outokumpu Stainless

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


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

Fig. 18 Isocorrosion curves for various materials in sodiumhydroxide. SCC, stress corrosion cracking. Courtesy

of Outokumpu Stainless

In the pulp-and-paper industry, chemicalpulping is called the kraft or sulfate process. Inthe presence of sulfur, nickel can be quite detri-mental, and ferritic or duplex alloys arepreferred. This again is caused by the solubilityof nickel complexes formed in the presence ofsulfur-containing compounds. This can be seenin Fig. 19, which shows a 26-1 (chromium-molybdenum) alloy significantly outperforminghigher alloys that contain nickel and molybde-num (Ref 12).

Atmospheric Corrosion

Atmospheric corrosion is an example of uni-form corrosion that occurs when a thin layer ofwater condenses on a metal surface and as such


depends on humidity, temperature, and other at-mospheric conditions. The rate of corrosionmeasured as defined in Chapter 3, “CorrosionKinetics,” as dissolution r of element i as

where j is current density, t is time in seconds,M is molar mass, n is valence, F is the faradayconstant, and ρ is the density), and therefore hastwo contradicting effects of temperature. Ingeneral, temperature increases the exchangecurrent density and transport properties and thusthe kinetic rates involved in corrosion. On theother hand, increasing temperatures may re-duce the concentration of dissolved oxygen inthe electrolyte and eventually will dry the sur-face and thus limit the electrochemical corro-sion due to the access to an electrolyte. In steels,the corrosion products are (a) an outermostlayer of porous rust (FeOOH) characterized bylow water content but easy access to oxygenand (b) an inner layer of magnetite (Fe2O3) inwhich pores are filled with water. The access ofoxygen to the bare metal limits the cathode re-duction reaction rate, and in relatively pure at-mospheres, the corrosion rate is negligible duethe protective nature of the oxide. However, sul-fur dioxide impurities in the atmosphere reactwith water to form sulfuric acid, which tends todissolve the protective oxide. In the case ofstainless steels, the passive region is extendeddue to chromium, to a wide enough region interms of pH (see Pourbaix diagrams in Chapter

2) that atmospheric corrosion can in effect beprevented.

The most deleterious impurity in the atmos-phere for stainless is the chloride ion. Chloridesare pervasive. Borne from oceans by normal cli-matological processes, they are found far inland.In many colder climates, they are also seen inhigh concentrations from road salts. Withoutwashing or the natural rinsing by rain, surfacechloride concentrations can become very high.Thus, the rules of corrosion of aerated aqueoussolutions are followed by stainless with respectto atmospheric corrosion. The difficulty is accu-rately estimating the solution that constitutes theaqueous solution. Much experience has shownthat if coastal and road salt effects are minimal,then 18% chromium alloys such as 304 experi-ence such negligible visible corrosion that theycan be used for exposed, unrinsed architecturalpurposes. If the same alloy is used in an unrinsedcoastal environment, red rust stain will occur.This is the corrosion product from metastableand possibly stable pitting. In Japan, wherecoastal conditions prevail throughout, much re-search has been done that has shown that a pit-ting resistance equivalent number (PREN) of 25is necessary for freedom from corrosion (i.e.,zero pitting) (Ref 14). This contrasts to a re-quirement of about PREN 35 to resist pitting inseawater. Pitting is a form of localized corro-sion, and PREN is an index to pitting resistance.These concepts are examined in the next section.

Localized Corrosion

Localized corrosion is in general moredamaging from a structural integrity point ofview than uniform corrosion since the corro-sion current is limited to a small area and thepenetration distance is large. Often, localizedcorrosion involves a large-area cathode and asmall-area anode, which means that for a givencorrosion current, the corrosion current densityat the anode is very large. In localized corro-sion, unlike uniform corrosion, the anode andcathode are clearly identifiable locations, andthe reason that certain structural features as-sume the roles of cathode and anode can beused to categorize and exemplify differentcases. Interestingly, the cathode and anodes,while identifiable, can vary across scales, thatis, from distinct macroscopic components orparts to microstructural features.

In Chapter 2, the tendency for corrosionwas introduced as a positive value for an

r jt M

n Fi

i

=⋅⋅ ⋅ρ

Fig. 19 Corrosion rates of various alloys in simulated evapo-rator liquid. Source: Ref 12


electrochemical cell potential (Erxn) correspon-ding to a spontaneous electrochemical reactionforming a galvanic cell. Erxn is obtained as:

(Eq 1)

Here, pi and ri are the concentrations of reac-tant and products, respectively, and αi and βi arethe numbers that are needed to balance the reac-tion stoichiometrically. Any time that Erxn ispositive, there is thermodynamically a tendencyfor an electrochemical reaction, in our case acorrosion reaction. The rate of corrosion, as dis-cussed in Chapter 3, is dependent on the polar-ization behavior.

Dissimilar Metals and Differential Aeration Cells

The case of dissimilar metals and differentialaeration cells is perhaps more important in ac-tive alloys than for stainless steels, which aregenerally passive, and occurs when twometals/alloys are in contact that have elementsin them that are dissimilar in the electromotiveforce (emf) series (see Chapter 2) and there is anelectrolyte present. For example, if nickel andiron pipes are connected and water flows thoughthem containing some traces of Ni2+ ions, then:

(Eq 2)

In this case, the corrosion tendency is prima-rily caused by the first two terms on the rightside of Eq 2, the dissimilarity in the standardhalf-cell reduction potentials:

This tendency is caused by the galvanic dis-similarity between the metals. This is normallyimportant for alloys joined to stainless that arethemselves less noble. Less-noble alloys, suchas carbon steel, can fail rapidly if coupled tostainless. A classic example is the use of carbonsteel fasteners for joining stainless sheets. Dif-ferent stainless steel alloys have minor differ-ences when passive, but if the environment issuch that one alloy is active while another ispassive, then the galvanic differential could beharmfully large. In many cases, a situation

arises in which access of oxygen is not the sameto different areas of a sample. In effect, this re-sults in that the cathode reaction:

(Eq 3)

is limited from proceeding in some areas but notothers. This gives rise to a differential aerationcell. For example, consider Fig. 20(a), in whicha metal is partially immersed in water. Trans-port distance of oxygen increases with depth;the limiting current would then vary with depth,such as at locations 1 and 2, and cathode polar-ization curves as a result of this are schemati-cally plotted in Fig. 20(b). Near the surface ofthe water, where oxygen is readily replenished,passivation is likely to be fast, and thus anodiciron dissolution is slow. This region assumes therole of the cathode, and reaction 1 occurs. Suffi-ciently far away from the surface, if there areregions where passivation is incomplete (e.g.,surface defects or scratches) or has brokendown as a result of, for example, Cl− (see sec-tion on pitting), repassivation does not readilyoccur since oxygen transport is too slow. Theseregions become anodes where the following re-action occurs:

(Eq 4)

The distance at which this occurs is balancedby being large enough to limit the rate ofoxygen transport but not too long to bestrongly influenced by ion transport that isneeded to complete the electrochemical cell.Resulting corrosion currents are shown in Fig.20b. This type of degradation is called water-line corrosion.

Crevice Corrosion. In stainless, the moresignificant occurrence of this type of cell occurswhen a crevice, from whatever cause, exists,and reactions within the crevice or pit cause theaccumulation of iron ions by:

(Eq 5)

The regions adjacent to the drop that main-tained their passive layer and have access tooxygen act as cathodes where the oxygen reduc-tion reaction takes place:

(Eq 6)

This reaction maintains an alkali solution. Asa result of the geometry, Fe2+ ions remain and

O H O OH2 22 4 4+ + →− −e

2 2 42Fe Fe( )s e→ ++ −

2 2 42Fe Fe( )s e→ ++ −

O H O OH2 22 4 4+ + →− −e

E E VNi Ni Fe Fe2 2

0 0 0 250 0 447 0 197+ +− = − + =/ /

. . .

E E ERT

F

a

arxn= − −

⎛

⎝+ +

+

+Ni Ni Fe Fe

Fe

Ni

2 2

2

2

0 0

2/ /ln⎜⎜

⎞

⎠⎟

E ERT

nF

p p p

r r rrxn rxn= −0 1 2 3

1 2 3

1 2 3

1 2ln

...α α α

β β β33 ...

⎛

⎝⎜

⎞

⎠⎟


enrich in the water-filled pit; to maintain chargeneutrality, Cl– migrates into the pit. This causesthe following reaction:

(Eq 7)

which has several consequences: (a) Hy-drochloric acid further acidifies the pit andincreases the rate of iron dissolution since de-creasing pH increases cathode half-cell poten-tial, which increases corrosion rate (see polar-ization diagram construction in Chapter 3,“Corrosion Kinetics.”) (b) The formation ofporous Fe(OH)2 further helps to isolate the pit,thereby separating anode and cathode regions inthe differential aeration cell. (c) The presence ofCl– prevents repassivation.

As a result of increased acidification, the dis-solution rate becomes autocatalytic, and as a re-sult the pit grows in depth.

At the outside, the reaction:

(Eq 8)

further consolidates the isolation of the pit andimpedes the ingress of oxygen.

Pitting Corrosion

Pitting corrosion is the most intensely studiedand debated form of corrosion of stainless steel.

Pitting corrosion is important to designers be-cause it is corrosion under conditions at whichcorrosion may not have been anticipated. Thus,it is both a materials selection and an environ-mental control problem. Its consequences maybe only cosmetic, such as on a building or appli-ance facade, or potentially catastrophic, such asif leaks of toxic materials were to result fromperforation. Stainless steels are designed to bepassive, and localized corrosion is the local lossof passivity. Whether the consequences aremajor or not, it is always undesirable, and gooddesign allows it to be avoided.

What do we know for certain about pitting?We know quite a lot, really. Experts now con-clude that since the early 1970s the local chem-istry of pitting has been understood (Ref15).The greatest contributions to this field havebeen electrochemical studies. The tools of elec-trochemistry have been especially successful inelucidating the mechanism involved in pitgrowth and pit stability (Ref 16). The local en-vironment within pits has been sufficientlymeasured and correlated with cavity geometrythat some experts can say, “In a sense, all pittingis crevice corrosion” (Ref 15). This is to say thatthe electrochemistry of cavities such as pits andcrevices is quite similar and has been well mod-eled. These same tools, however, have beenmuch less successful in clarifying the mecha-nism of pit initiation, which is still the subject

2 22 2 2 3

Fe OH O H O Fe OH( ) + + → ( )

Fe H O Cl Fe OH HCl22 2

2 2 2+ −+ + → ( ) +

Fig. 20 Schematic illustration of (a) sample partially immersed in water; (b) resulting polarization behavior for two different passi-vating alloys (A and B polarization curves)


of debate, possibly indicating that the rootcauses are more metallurgical than electro-chemical.

Figure 21 depicts a polarization curve forstainless steel in a chloride-containing solution.Pitting occurs in the zone in which passivity isexpected. As potential increases, small spikes incorrosion current occur. These spikes measurelocal dissolution, called metastable pitting.Some such sites complete their dissolution andrepassivate, while others continue to grow asstable pits. The potential at which stable pittingoccurs is the pitting potential, while metastablepitting can occur at much lower potentials. Pit-ting events, stable or not, cause the generationof iron ions and local pH reduction. To the ex-tent these remain concentrated in a small vol-ume, they will affect subsequent events. Thedissolution during metastable pitting is locatedat the matrix-inclusion interface. Different re-searchers assume dissolution of the inclusion,while others assume dissolution of the matrix.The dissolution parameters, as measured bycurrent transients, depend on variables not ofthe inclusion chemistry but of the matrix com-position, notably molybdenum and nitrogenlevels (Ref 17), which is in keeping with the re-duction in dissolution of the matrix that thesealloying elements confer.

Inclusions. The question of what causes theinitial dissolution that causes both stable andmetastable pits focuses on inclusions, whichmost authorities (Ref 18) have concluded areassociated in some way with pit initiation. Inthe absence of inclusions, metastable pittingevents are not noted, and the potential at whichpitting occurs is the beginning of the transpas-sive regime.

What are the typical inclusions in stainlesssteel? Inclusions in steel are normally theresidue of normal deoxidation and desulfuriza-tion taken during steel refining usually done inan argon oxygen decarburization (AOD). Afterremoval of the carbon, the subsequent objectiveis to remove or render less harmful the dis-solved oxygen and sulfur, which if left in solu-tion would later precipitate as low-melting-point iron compounds that would make the steelfragile and unworkable at high temperatures.Inclusions in stainless steel are typically oxidesand sulfides. A key point to understand whenconsidering inclusions as initiation sites for pit-ting is that inclusions are not simply inert debrisbut precipitates that are seeking thermodynamicequilibrium with the steel in which they havepreviously been dissolved. The reactions instainless steel differ thermodynamically fromthose in carbon steel because of the presence ofhigh chromium concentrations. This lowers theactivity of oxygen and sulfur, making themmore soluble, as Table 1 indicates (Ref 19). Italso alters the efficiency of deoxidizing ele-ments. Aluminum is a powerful deoxidant incarbon steel but is less effective in stainless,while titanium becomes a stronger deoxidizer instainless. Their effect on sulfur is similar to thaton oxygen.

The bottom line is that oxygen and sulfur aregenerally removed by silicon/manganese deoxi-dation, but that this process occurs in both theliquid and solid states. That it carries over sig-nificantly into the solid state means that diffu-sion has a major role in determining if equilib-Fig. 21 Schematic of a passive anode polarization curve

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, 1600 °C

%i . . . 0.05 0.45 0.02 0.01 0.3 0.05 . . . . . . . . . . . . . . .

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,1600 °C

%i . . . 0.05 0.45 0.02 0.01 0.3 0.05 . . . . . . . . . 18 8

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


rium reactions occur and whether they go tocompletion. We will see that they do not.

Oxide inclusions also are common. They areformed as the products of the reactions of sili-con and manganese with dissolved oxygen. Thethermodynamics of the reactions determine atany time how much oxygen can be dissolved inthe steel at equilibrium. That equilibrium is eas-ily achieved in the molten state, in which diffu-sion is very rapid, but achieved more slowlyonce the material has solidified. The inclusionsin the solid state grow by the diffusion of oxy-gen to inclusion sites, where it precipitates as anoxide of silicon or manganese to the extent thatthese are locally present or of chromium whenits local concentration (or more properly, its ac-tivity) makes it more favorable. These oxidesare often the nucleation sites for manganese sul-fide inclusions.

Sulfur is a very surface active impurity that as-sists in weld penetration in stainless by virtue ofits effect on weld pool circulation. Otherwise, itis a detrimental impurity, forming low-meltingoxysulfides that diminish hot workability. Man-ganese is a strong sulfide former, and it is themain agent used to tie up sulfur. Manganese sul-fide precipitates as an inclusion as a function ofmanganese and sulfur concentrations and tem-perature. Inclusions form not only in the moltenmetal but also in the solidified metal. The solubil-ity, which is high in the liquid state, decreases onsolidification, as seen in Fig. 22. Only resulfur-ized free-machining stainless steels have suffi-cient sulfur to precipitate manganese sulfide in

the liquid. At high sulfur and manganese concen-trations, some manganese sulfides can precipitateduring solidification interdendritically, while nor-mal alloys with less than 100 ppm of sulfur formtheir inclusions after solidification. The distinc-tion is important because precipitation in the liq-uid state permits rapid diffusion, which results inthe most thermodynamically favorable species,manganese sulfide, to form. It may, and oftendoes, nucleate on a preexisting inclusion, such assilicate present from the deoxidation process. Inaustenitic steels, manganese is generally presentat a level of around 1.5% as a deoxidant and as asubstitute for some nickel. An inclusion formedin the molten metal does not cause alloy deple-tion around it. One that forms or grows in thesolid state does cause depletion of the elementsthat are precipitating, causing its growth.

If manganese is lowered to very low levels,the supersaturation of sulfides is pushed to alower temperature, at which lower diffusionrates hinder or prevent the precipitation. Thus,low-manganese alloys can be free of manganesesulfide inclusions even at somewhat high sulfurlevels. Such alloys have elevated resistance topit initiation. Lower manganese levels also ther-modynamically reduce the chromium sulfide co-precipitation in inclusions, lowering chromiumdepletion around manganese sulfide/chromiumsulfide inclusions.

Elements more effective than silicon andmanganese are now in use for deoxidation anddesulfurization. These include aluminum, cal-cium, cerium, and other rare earth metals(REMs), and titanium. The action of calcium isnotable. In a well-deoxidized and well-stirredmelt and with a basic slag, calcium dissolved inthe metal will react with dissolved sulfur toform calcium sulfide, which will be incorpo-rated into the slag phase. Aluminum, while apotent deoxidizer, is less effective directly indesulfurization, but it can act indirectly by re-ducing a small amount of Ca2+ in the slag, al-lowing the formation of calcium sulfide. Tita-nium can sequester some sulfur as titaniumcarbosulfide precipitates. The greatest amountof sulfur removal is obtained by the addition ofcerium or other REMs, usually in the form ofthe alloy mischmetal. These reactive elementstypically form oxysulfide particles in the meltthat may be trapped in the slag before metalsolidification.

Oxygen is normally dissolved in solidifyingstainless steel, also at amounts in the neighbor-hood of 100 ppm depending on deoxidation

Fig. 22 Pseudo-binary-phase diagram for iron and sulfur at1.8% manganese and 18% chromium


methods. Inclusions based on oxygen and sulfurformed in the liquid or during solidification arerelatively large, greater than 1μ. As the alloycools after solidification, precipitation continuessince sulfur and oxygen are decreasingly solu-ble with temperature, to virtually nil at roomtemperature. This causes existing inclusions togrow and new ones to nucleate. This precipita-tion is similar to that which carbon undergoes instainless, except carbon is generally not super-saturated until below 1200 °C at the highest inmost alloys, whereas sulfur and oxygen are nor-mally near saturation even at freezing or almostalways when the solidifying ferrite transformsto austenite. Thus, inclusions grow via diffusionof oxygen and sulfur, which, as interstitials, dif-fuse much more rapidly than the silicon or man-ganese with which they have the greatest ther-modynamic affinity. But precipitate they must,even if the silicon and manganese in the vicinityof their inclusion are exhausted. Thus, inclu-sions can grow with chromium substituting foreither silicon or manganese as the precipitatingpartner for oxygen and sulfur. The inclusiongrowth necessarily depletes the surrounding re-gion of reactants, silicon, manganese, andchromium (Ref 20). Inclusions thus formed arenonequilibrium in nature, and thermal cycles ofsteel production are rarely sufficient for theequilibrium to be attained. The chromium en-richment of such inclusions and correspondingchromium depletion of surrounding regions hasbeen measured (Ref 21) and corresponds to thedepletion seen next to chromium carbide pre-cipitates at grain boundaries in sensitized al-loys. These zones are altered in size and shapeby thermomechanical processing in wrought al-loys but exist fairly undistorted in welds. Hotrolling and cold rolling followed by annealingelongate manganese sulfide inclusions and flat-ten them, allowing depleted zones around theinclusion in the reduced dimension to be morerapidly homogenized during annealing. Thus,wrought material has better pitting resistancethan cast or welded material. Inclusions thatprecipitate from the liquid, as is more the casefor alloys solidifying in an austenitic mode, areat equilibrium with the surrounding matrix byvirtue of the faster diffusion in liquids, do littleto diminish the chromium content around them,and have a small effect on lowering pitting re-sistance. Pitting resistance is still affected to adegree by alloy depletion due to solidificationsegregation. However, if the alloy solidifies in aferritic mode (FA, i.e., ferrite forming first on

solidification as opposed to austenite first, AF),as is almost always the case with commercialalloys, more sulfide precipitation happens in thesolid state, pitting resistance is lowered propor-tionately to the sulfur level (Ref 22), and thereis little negative effect from solidification segre-gation, as is shown in Fig. 23 and 24 (Ref 23).Solidification can also occur in a mixed ferritic-austenitic mode, in which case each microstruc-tural component behaves according to the chart.The ratio of chromium and chromium-like ele-ments molybdenum and silicon to nickel andnickel-like elements carbon, nitrogen, man-ganese determines the mode of solidification. It

Fig. 23 Influence of sulfur level on pitting resistance ofunannealed welds for different solidification modes.

Source: Ref 23

Fig. 24 Influence of sulfur level on pitting resistance of weldswithout homogenizing anneal. FA, ferrite forming

first on solidification as opposed to austenite first, AF. Source:Ref 23


can also be altered by freezing rate. Faster cool-ing favors austenitic solidification.

Long-term annealing of welds has shownthat sufficient time and temperature to achievesome rehomogenization the alloy result in bet-ter pitting resistance (Ref 24), approaching thatof the wrought alloy. Examination of the de-creasing solubility of sulfur in stainless in Fig.22 indicates that the precipitation of sulfidesthat cause chromium depletion occurs in deltaferrite on freezing when sulfur exceeds 0.007%and in austenite when sulfur exceeds 0.003%.Oxygen behaves in a parallel manner and isusually present in sufficient quantities, about0.01% in manganese/silicon deoxidized steels,to cause the same phenomenon. This funda-mentally is due to the high ratios of the diffu-sivities of oxygen and sulfur to chromium,which are about 10,000 and 680, respectively.Whenever fast-diffusing elements such as oxy-gen, sulfur, carbon, and nitrogen, which have astrong affinity for chromium and a solubilitythat decreases strongly with temperature, arepresent in steel, their precipitation will result insome degree of chromium depletion around theprecipitation site because chromium diffusestoo slowly to be replenished.

The low chromium around inclusions is a suf-ficient condition for the local dissolution meas-ured as metastable pitting, and if the depletionzone shape and size are favorable, then stablepitting would ensue.

Certain other types of inclusions/precipitatesare less harmful in this regard. Titanium, for in-stance, which is often added to form carbidesand nitrides, also forms sulfides and oxidesmore strongly than manganese and thereforedoes so at higher temperatures. Such precipi-tates have a much lower tendency to allowchromium to join in the precipitation since thehigher the temperature of precipitation the morethat diffusion allows the more favorable reac-tion to occur. Rare earths also behave the sameway. Metastable pitting is diminished by thepresence of these elements.

The initiation of pitting is also affected bystress and inclusion orientation (Ref 25), whichthe researchers correlated to the dimensions ofthe inclusion-derived cavity being able to sus-tain a sufficiently low pH due to iron dissolu-tion to maintain stable pitting. The influence ofstress was to cause cracking at otherwise unfa-vorably shaped inclusions, which then provideda crevice capable of sustaining stable pitting.This will be relevant to later discussions of

stress corrosion cracking (SCC). There havebeen numerous proposed mechanisms for thebreakdown of a passive film in chloride-con-taining media; these have been summarized inother publications (Ref 23). These hypothesesdeal with how a passive film on a homogeneoussurface could break down. They include:

• Adsorption of chloride ions• Penetration of the passive film by chloride

ions• Film breakdown by electrostriction• Formation of stable metallic chlorides• Coalescence of cationic vacancies• Random localized thinning of the passive

film• Local variations in the composition of the

corrosive medium

By and large, these mechanisms presupposea stainless steel surface that is homogeneouslypassive and try to explain the observed inho-mogeneous behavior of the passive film. How-ever, since it is clear that the surface is not ho-mogeneous, especially with regard to thepassive film, these hypotheses are not neces-sary to explain the behavior of everyday stain-less steels, which unfortunately have abundantinclusions and chemical inhomogeneities capa-ble of locally diminishing the integrity of thepassive film. More research in understandingthe exact nature of the inhomogeneity of stain-less steel surfaces is necessary for a completeunderstanding of pit nucleation and thereforeprevention.

Pitting Resistance. Pitting has been exten-sively correlated with environment and compo-sitional variables. The most well-known anduseful correlations are between the PREN andthe critical pitting temperature (CPT) and by ex-tension to the pitting potential.

For austenitic alloys:

PREN = % Cr + 3.3 % Mo + 30 % N (Eq 9)

For ferritic alloys, which hold no nitrogen insolution:

PREN = % Cr + 3.3 % Mo (Eq 10)

For duplex alloys, which have two phases,neither of which matches the bulk composition:

PREN = % Cr + 3.3 % Mo + 16 % N (Eq 11)

These equations are useful, if approximate,and their correlation is shown in Fig. 25 (Ref26) They do not include tungsten, which, if


present, has half the effectiveness of molybde-num. They neglect carbon, which seldom variesenough to have a visible effect but has beenshown when in colossal supersaturation to havea factor of about 10, not unlike nitrogen, an-other interstitial that it resembles in solutionthermodynamically (Ref 27). It also does not in-clude the negative influence of elements such assulfur. Likewise, the equations cannot deal withinhomogeneity issues, so welded alloys havedifferent CPTs for the same PREN (Fig. 26)(Ref 13). These equations are all-other-things-being-equal equations and are useful for gross

alloy behavior predictions. It is noteworthy thatthe elements copper and nickel, which are bene-ficial against uniform corrosion and which slowthe growth of pits by this same action, do notcontribute to increasing the resistance to theonset of pitting. This is another manifestation ofpitting initiated by the local stability of the pas-sive film, which is primarily a function of localchromium content. Molybdenum and nickelthus seem to bolster local chromium content inthe passive film. Nitrogen seems to act by con-centrating at the passive film-alloy interfacerather than by buffering the solution by ammo-nia formation, which has been proposed (Ref36). Research on very pure sputtered films ofiron-chromium alloys have demonstrated thatboth titanium and niobium in solution diminishactive dissolution, assist repassivation, and im-prove pitting resistance (Ref 29). In most practi-cal cases, these elements are not found in solu-tion because of their affinity for oxygen, sulfur,carbon, and nitrogen, with which they formcompounds.

It should also be noted that the critical PRENvalues vary with crystallographic structure. Fer-ritic alloys require somewhat lower PREN val-ues to exhibit similar pitting resistance asaustenitic alloys of somewhat higher PREN.

While pitting is of great theoretical and prac-tical interest, there are significant problems inactually conducting good tests. Monitoring ofthe electrochemical potential during the test isconsidered mandatory by most researchers.How is a metallic sample suspended in a solu-tion without creating any crevices and withoutexposure at the liquid-gas interface? The devel-opment of the flooded gasket technique (used inASTM G150) was a milestone, but it also hassome problems—most notably the potential fordilution of the test solution, especially duringprolonged testing. FeCl3 testing benefits fromthe fact that the solution creates a reproduciblepositive potential.

While the PREN approximates the pitting re-sistance of an alloy, there is a standard test bywhich the CPT is measured. Pitting in a givenmedium capable of causing pitting does notoccur below a temperature that is characteristicof the medium and the material, with the myriadexceptions of stress state, surface finish, mi-crostructure, etc. The most commonly used testmedia are the unacidified 10% FeCl3, which isused in the ASTM G 48 practice B, and the 3.5%NaCl solution of the ASTM G 150. The latter, ifmodified to 0.1N NaCl, allows the ECPT, the

Fig. 25 Variation of critical pitting temperature with pit-ting resistance equivalent number (PREN) of

austenitic steels in water plus 6% FeCle. Source: Ref 26

Fig. 26 Differential variation of critical pitting temperatureof several stainless steel alloys for unwelded

wrought and welded material. Source: Ref 13


electrochemical pitting potential, of lower alloyssuch as 304 to be measured (Ref 30).

Crevice Corrosion

In the case of pitting, the geometry thatmakes up the pit is essential in creating the dif-ferential aeration cell and to cause the autocat-alytic dissolution process. In many cases, ageometry that retains and acidifies water is al-ready present in crevices in different types ofstructures such as gaskets, under faulted coat-ings, under bolt or screw heads, etc. Crevicecorrosion occurs because zones have restrictedaccess of reactants and restricted exit of corro-sion products. It is especially the inhibition ofthe cathodic reaction inside the crevice by thedearth of oxygen, which sets up a more aggres-sive environment within the crevice than with-out. The interior reactions become increasinglyanodic, and the aggressiveness of the environ-ment can reach a threshold at which active cor-rosion occurs, while the situation exterior to thecrevice is safely passive. Crevice corrosion oc-curs at lower temperature than pitting in thesame environments, so it is a greater danger inthat sense.

The relationship between the alloy content,given as the crevice corrosion resistance equiv-alent number (CCREN), and critical crevicecorrosion temperature (CCT), shown in Fig. 27(Ref 11), is similar to that of PREN (PI) to CPTexcept for the molybdenum factor being moreimportant:

Cl = %Cr + 4.1%Mo + 27%N (Eq 12)

Since a crevice has a preexisting favorablegeometry for pit growth, any pitting event,metastable or stable, can initiate ongoingcrevice corrosion. Crevices are thus incubatorsfor corrosion triggered by metastable pittingevents. The dissolution of iron during passiva-tion itself as well as the differential oxygen cellcreated by the crevice contribute to the process.It is logical to think that alloying the elementsthat contribute to lowering the critical currentdensity for passivation and the uniform corro-sion rate, such as nickel, would reduce the cre-ation of the reactants that start the crevice corro-sion process, but this presumed effect is notstrong enough to be reflected in this actual be-havior Eq 12 represents, although it is generallyacknowledged that austenitic steels perform bet-ter than ferritic steels in the absence of molyb-denum. Materials are characterized as having acritical depassivation pH. If crevice conditionsare such that the reactions over time allow thepH to be reduced to this level, then active corro-sion will begin within the crevice. Thus, passivefilm stability seems to be the critical factorrather than corrosion rate after initiation.

Preventing Crevice Corrosion. The coun-termeasures against crevice corrosion are ca-thodic protection, design, maintenance, and, ofcourse, alloy selection. Designing to avoidcrevices should include maximizing the volumeof unavoidable crevices, engineering flow to en-hance transport in and out of crevices, and

Fig. 27 Variation of critical crevice corrosion temperature with alloy content


avoiding stagnation. Any maintenance or designprocedure that prevents formation of deposits isbeneficial. Welds are particularly vulnerablesurface sites, so any combination of welds andcrevices or crevices caused by poor weld geom-etry must be avoided. S32205 is a benchmarkalloy of sorts. It has just sufficient alloying toresist pitting in seawater, but it is susceptible tocrevice corrosion.

As a practical matter, crevices are almost im-possible to eliminate. Threaded fasteners andjoints represent severe crevices and should beavoided in aggressive environments if possible.Gasketed joints are another severe crevice loca-tion, and their usage should be curtailed to theminimum practical extent. In these situations,judicious use of very expensive, highly corrosionresistant materials is justified. The use of smoothwelded joints is thus generally preferred. In amore general consideration, deposition andfouling create crevice sites, and design and op-erational controls to preclude the formation ofdeposits and the prompt removal of sludge andthe like are necessary. But in some situations,such as marine exposures, biofouling will createcrevice sites. This fouling may be macroscopic,such as from shellfish and barnacles, or it maybe microscopic. Microscopic biofouling causesthe special form of crevice corrosion called mi-crobiologically influenced corrosion (MIC) dis-cussed in a separate section).

Sensitization/Grain Boundary Corrosion

The maintenance of a passive layer in a widerange of pH conditions in stainless steels is de-pendent on the alloying elements, primarily

chromium. In the various grades of stainlesssteels, there are many intermetallic phases that arethermodynamically stable but kinetically slow toprecipitate that are enriched in chromium. An ex-ample of such a phase is chromium carbide (Fe,Cr)23C6. These phases tend to form at grainboundaries where nucleation is favored, resultingin a depletion of chromium in the adjacent re-gions, as shown in Fig. 28. Thus, the chromium-depleted regions near the grain boundaries aresensitized in that they behave as active anodescompared to the larger interior of the grains thatare still passive. In an aerated corrosive environ-ment, the smaller chromium-depleted nonpassiveanodes dissolve, whereas the larger cathodes re-duce oxygen, resulting in a localized corrosionalong grain boundaries. Any heat-treating orwelding procedure of stainless steels should thusbe tailored to avoid sensitization.

When a stainless steel is heat treated, there isa risk that the unwanted phases may form, de-pending on the time-temperature history andprecipitation kinetics of the unwanted phase.Figure 29 shows schematically the temperatureversus time due to welding and the resultingsensitization. Figure 29 shows a TTT (time-temperature-transformation) curve for precipi-tation of the unwanted phase. Near the weld(A), the time spent in the temperature regionwhere precipitation occurs is too short, whereasfar away from the weld (C) the temperature ex-perienced is too low. At location B, there is,however, a risk for sensitization.

Austenitic. Sensitization can occur at anytemperature at which carbon is supersaturatedin an alloy. Current austenitic stainless steelshave carbon levels of under 0.10% normally

Fig. 28 Schematic illustration of sensitization due tochromium-rich precipitates that deplete adjacent

regions of chromium. GB, grain boundary

Fig. 29 Schematic illustration of how a heat treatment re-lates to sensitization due to precipitation kinetics.

TTT, time-temperature-transformation


and under 0.03% for low-carbon L grades.Thus, normal grades sensitize below around800 °C. The supersaturation increases withdecreasing temperature, but below about 500 °Cdiffusion of carbon is too slow for carbon tomove to grain boundaries and cause the damag-ing combination with chromium that causessensitization. Low-carbon grades avoid sensiti-zation because they are not sufficiently super-saturated at temperatures at which carbon ismobile enough to diffuse to grain boundaries.

Ferritic. Another situation exists in ferriticstainless steels, in which carbon is much lesssoluble but is much more mobile. Annealingover 900 °C can put enough carbon in solution tocause sensitization even at the lowest carbon lev-els attainable in an AOD and even at the fastestpossible quench rates. The damaging chromiumdepletion caused by this very rapid precipitationcan be undone by a simple rehomogenizationanneal of the remaining chromium. This is theo-retically possible with austenitic alloys also, butthe diffusion rates of chromium in austenite asso slow that it is impractical in most real cases.

Duplex steels have a subtle near immunity tocarbide sensitization. While they are typicallylow carbon anyway, the carbides that do formdo so at ferrite-austenite grain boundaries. Here,chromium is consumed from the chromium-richferrite phase, leaving the austenite intact. Theirlarge grain boundary area keeps carbide con-centration per unit area low, and the fast diffu-sion in the ferrite keeps austenite from becom-ing depleted. However, the rapid formation ofintermetallic phases at the ferrite-austenite in-terfaces can lead to a rapid loss of corrosion re-sistance and a severe loss of toughness if expo-sure to temperatures within the intermetallicprecipitation range is not controlled.

Martensitic steels are quenched as austeniteto and through the Ms temperature without timefor carbon to precipitate in austenite. The car-bon in the martensite can precipitate and causesensitization if reheated to the 300 to 700 °C re-gion. Fortunately, heating to above 700 °C re-homogenizes the chromium and eliminates sen-sitization.

Effect of Alloying. Besides determiningbasic phase structure, alloying plays a role insusceptibility to sensitization. Those elementsthat reduce the tendency of chromium carbidesto form also reduce the susceptibility to sensiti-zation. This is a purely thermodynamic effect.Molybdenum, silicon, and nickel promote car-bide formation by increasing the thermodynamic

activity of carbon and make alloys more suscep-tible. Nitrogen lowers the tendency for carbideformation and slows sensitization. Nonthermo-dynamic effects are those of austenite grain sizeand prior cold work. Decreasing grain size andtherefore increasing grain boundary surfacearea decreases the amount of precipitate per unitarea of grain boundary and therefore the amountof chromium depletion per unit area. Cold workaccelerates diffusion and makes precipitationmore rapid, thus aggravating sensitization.

The thermodynamic affinity tool can be usedto prevent chromium carbide formation in an-other way. Introducing alloying elements thatcombine with carbon more strongly and rapidlythan chromium can exhaust the supply of car-bon available to precipitate as chromium car-bide. There are a number of candidate elements,zirconium, vanadium, tantalum, niobium, andtitanium, most prominently. Of these, the diffu-sivity and affinity for carbon of niobium and ti-tanium make them the best for this purpose.Each forms stable carbides at much higher tem-peratures than chromium, starving chromium ofsufficient carbon to form damaging precipitates.The caveat with titanium is that it forms oxides,sulfides, and nitrides preferentially to carbides.Therefore, sufficient quantities must be used toaccommodate the prior formation of thesephases. Niobium tends more toward carbidethan nitride formation but is a weaker carbideformer than titanium. The solubility products ofthese precipitation reactions are:

(Eq 13)

(Eq 14)

These equations follow the form of the gen-eral equation for precipitation reactions:

(Eq 15)

in which A is a constant, H is the heat of disso-lution, R is the gas constant, and T is the ab-solute temperature. If the amount of titanium orniobium is stoichiometrically sufficient, no car-bon will form chromium carbides under equilib-rium conditions. It is possible to defeat the sta-bilization reactions by quenching the alloysfrom temperatures at which titanium carbide orniobium carbide is dissociated. If free carbon is

log[ ][ ] /M X A H RT= −

log[ ][ ] .Nb CT

= −4 559350

log [ ][ ] .Ti CT

= −2 976780


left free in the matrix by quenching, then on re-heating it may form carbides with the most lo-cally accessible favorable element, such aschromium, rather than the most thermodynami-cally favorable element, which would be tita-nium or niobium. This can occur when a stabi-lized alloy such as 321 is welded. A zone awayfrom the weld may experience a high enoughtemperature to put carbon into solution and thencool just rapidly enough to not form only theequilibrium titanium carbide but also Cr23C6 atgrain boundaries, causing the type of sensitiza-tion called knife-line attack. This problem hasnearly ceased to exist as modern 321 has lowlevels of carbon and nitrogen for economic rea-sons; this effectively precludes this chromiumcarbide precipitation in most cases.

Welding. Many of the most severe problemsof sensitization arise when stainless steels arewelded to carbon or low-alloy steels. In thesesituations, construction code rules usually re-quire that the carbon steel component be given astress relief annealing (SRA) treatment. SuchSRA treatments are typically in the sensitizationtemperature range for austenitic stainless steels.Use of low-carbon or stabilized grades is neces-sary in such cases. Even then, use of the lowestallowable temperature SRA treatment for theshortest allowable time is preferred.

Corrosion Combined with Fatigue or Fracture

Environmentally induced failure occurs whenbrittle failure under tensile mechanical loadingoccurs at a lower stress when a material is sub-jected to a corrosive environment than whatwould happen in a noncorrosive environment.This introduces us to what is perhaps the mostcontroversial technical subject in all of stainlesssteel research, SCC.

Stress Corrosion Cracking

The key cause for SCC is the cooperating ef-fects of tensile stress and a corrosive environ-ment. Such cases can be identified in most alloysystems, and even pure metals, which werethought to be more or less immune, also havehad cases of SCC reported.

In passive metals, two sensitive potential re-gions for the occurrence of SCC have beenidentified and are shown in Fig. 30: region I,where pitting and breakdown of the passive

layer occur, and region II, where the protectivelayer is not fully developed, suggesting an ap-preciable electrochemical effect. The latter is azone that exists in alloys that have zones ofchromium depletion.

Stress corrosion cracking has always beenamong the most controversial subjects amongmetallurgists and electrochemists. The debatecenters on whether the critical mechanism isdissolution or fracture, and if a fracture, by whatmechanism. Is the cracking zone locally soft-ened, locally hardened, transformed, to a morebrittle phase or embrittled by hydrogen? As ofthis writing, there is no general agreement onwhich type mechanism is the fundamentalcause, but there is room for convergence. Obvi-ously, elements of many may come into play. Itis likely, as in most prolonged arguments, thatno hypothesis is completely correct. We will tryto fairly set out what is known and agreed on asfact and then present researchers’ views in anunbiased manner, but since we concern our-selves only with stainless steel, no attempt ismade to address an all-encompassing theory.

Crack Initiation. In stainless steels, crackscan be seen to initiate at surface defects and ir-regularities. In stainless steel, it must be agreedby all that the preponderant initiation site is acorrosion pit or, in some cases, a crevice. Inter-granular corrosion sites, as are seen in sensi-tized material, can also provide the conditionsfor SCC initiation. The interrelationship be-tween pits and SCC cracks has been studied(Ref 25). Stress lowers the anodic potential atwhich pitting occurs and permits metastable pits

Fig. 30 Zones of susceptibility to stress corrosion cracking


to become stable via the generation of cracks.Cracks, once formed, presumably have favor-able geometry to duplicate pit internal chemicalreactions and must be considered to be de-scribed by the models that apply to pits andcrevices.

The stress at which SCC initiates has athreshold, which has been reported as between25 and 50% of the yield strength in austeniticstainless steel.

The temperature at which SCC is initiatedranges from ambient to under 100 °C formartensitic materials, while austenitic alloysbegin their sensitivity above room temperatureand increase in susceptibility with increasingtemperature. The ferritic steels, while consid-ered nearly immune to SCC, have their maxi-mum susceptibility in the same range as marten-sitic steels. In environments of mixed chloridesand sulfides, however, SCC can occur in alltypes of stainless at room temperature. This hasbeen seen in the SCC of austenitic stainlesssteel in swimming pool environments, in whichchloride ions can condense on the stressed steeland cause pitting and SCC.

Cracks propagate very slowly below spe-cific certain stress intensity levels, but once thatintensity is reached, they have a plateau ratethat is fairly constant until the stress level atwhich catastrophic failure occurs at very highpropagation rates. Rates of crack propagationare exponentially increased by increasing tem-perature. The crack propagation rate has beenseen across a range of alloys to be linearly pro-portional to the average current density thatalloy experiences when its surface is strained,indicating that reactions at the crack tip arestrain sensitive, and overall rate limiting, butnot necessarily the mechanism of cracking.Crack growth is discontinuous with individualsteps of growth many times the average rate,which is similar to that seen with gaseous hy-drogen embrittlement (HE). The crack growthgives off acoustic emissions as cracking stepsoccur. These steps of growth are brittle and areseen as facets on fractographs with cleavagescorresponding to crystallographic planes. Thecrack facets match with high perfection, show-ing almost no evidence of plastic deformationor dissolution.

The propagation path may be intergranular ortransgranular. Grain boundary propagation instainless steels usually corresponds to condi-tions under which grain boundaries are less cor-rosion resistant because of either material or

environmental variables. The most common ex-ample is that of sensitized 304 in high-tempera-ture water or caustic media. The relevance ofthis to the normal case of stainless steels mustbe questioned since, by definition, the sensitizedgrain boundaries themselves can be depleted ofchromium to a degree they are not stainless andhave a much less stable austenitic structure,having their martensite start temperature Ms,raised by the loss of chromium.

Material Variables. Martensitic stainlesssteels and martensitic precipitation hardenedstainless steels are quite susceptible to SCC.This susceptibility increases with hardness,yield strength, and embrittling heat treatments.They will crack at threshold stresses equal to50% of yield strength. These alloys can be tem-pered at sufficiently high temperatures that theybecome soft and tough enough to have verygood resistance.

Ferritic stainless steels of low and mediumchromium are generally not susceptible to SCC.Ferritic alloys, which can have a martensiticstructure, should be considered martensitic forSCC purposes. If purely ferritic alloys are al-loyed with copper, molybdenum, and nickel,they can become susceptible. The presence ofα'( or high-temperature embrittlement also in-creases susceptibility, as does cold work.

Despite the controversy surrounding themechanism of SCC in austenitic stainless steels,there is almost complete agreement that SCC ofbody-centered cubic (bcc) stainless steels,martensitic, ferritic, and pH is simply a mani-festation of HE, with hydrogen provided by ei-ther anodic (e.g., active corrosion within a pit)or cathodic reactions.

Duplex stainless steels have low susceptibil-ity to SCC. Their dual-phase microstructure en-sures that under conditions that crack austenite,ferrite remains as a crack-arresting phase, whileunder conditions that cause SCC in highly al-loyed ferrite, the austenite is a crack arrester.

A second explanation of the resistance of du-plex alloys to SCC is that their two phases havedifferent corrosion potentials, and that themixed potential that arises because they are inintimate contact is outside the potential rangefor SCC on either phase. This fits with the re-sistance to SCC of wrought alloys with a lamel-lar structure and the lesser resistance of cast al-loys that lack that structure.

Austenitic stainless steels are the type of stain-less steel generally associated with SCC, andthey vary in their degree of susceptibility to


SCC. All other things being equal, alloyingelements that delay or prevent localized corro-sion do the same to delay SCC. This is simplythe delay of initiation. However, if pitting can bedelayed indefinitely, then SCC can also, assum-ing, of course, more harmful localized corrosion,such as that due to intergranular chromium de-pletion, is not occurring. Molybdenum, whichwe already know helps prevent pitting andcrevice corrosion, also increases the thresholdstress for SCC, as shown in Fig. 31 (Ref 31).

But, if metastable or stable pitting is occurring,the threshold stress has been reached, and thetemperature is sufficient, then SCC will proceed.It is mitigated by material variables such as coldwork and by alloying elements that increaseaustenite stability. Many publications cite nickelas beneficial in enhancing resistance to SCC,often referring to the data from Fig. 32. However,its role seems mainly to be as an austenite stabi-lizer and as a retarder of active corrosion. Theminimum in the curve corresponds to the nickellevel at which the structure is entirely austenitic,but least stably so. Lower nickel levels producebetter immunity through the duplex structure,while higher levels promote austenite stabilityand correspond to alloys having more alloyingelements, such as chromium and molybdenum,

which are probably the greater cause of resist-ance to SCC.

Environmental Variables. There are threekey types of environments in which SCC occursin stainless:

• Chloride-containing solutions• Caustic solutions• Polythionate and thiosulfate solutions

The cases of polythionate and thiosulfate solu-tions are industrially important but can be ade-quately explained as simply the stress-assistedintergranular corrosion of sensitized material.Hot caustic solutions are aggressive againststainless steels. Certain combinations of concen-trations, temperature, impurity, and dissolvedoxygen can cause SCC as well as other undesir-able corrosive attack. Resistance to general cor-rosion is proportional to nickel content, but fer-ritics and duplex alloys are less prone to SCC.304 has been reported to have no meaningfulthreshold stress for SCC in hot caustic solutions,leading one to question whether such a failureshould even be classified with SCC of the typi-cal chloride-induced type or belong with the pre-vious polythionate and thiosulfate solutions.

Chloride containing environments are themain ones that induce SCC. Water can cause

Fig. 31 Influence of molybdenum on resistance to stresscorrosion cracking (SCC) in austenitic steels

Fig. 32 Variation of resistance to stress corrosion crackingwith nickel (and other) content and structure


SCC at sufficiently high temperatures (i.e.,above 100 °C) if there are even very low com-bined concentrations of chloride (greater than0.1 ppm) and oxygen (greater than 0.1 ppm)dissolved (see Fig. 33) (Ref 32).

Failure times decrease exponentially with de-creasing chloride content. Crack growth rate in-creases by a factor of ten with each 30 °C rise intemperature. Decreasing pH lowers the temper-ature at which SCC occurs in a given time.

Mechanisms. There have been many mecha-nisms proposed for SCC in stainless steel. Wefocus only on those that address the failure inchloride-containing media, the main concern forusers of stainless steel.

The models that have found some supportare:

• Slip dissolution• Adsorption-enhanced plasticity• Adsorption-induced brittleness• Hydrogen embrittlement

Slip dissolution (anodic dissolution) was theearliest proposed model for SCC. It simply pro-poses that at a crack tip a passive film forms, andafter time it fractures by an unspecified mecha-nism. The fresh active surface may or may notrepassivate, after which the process repeats itself.

The strength of this model is that it actuallydoes describe what is happening. The crack

does advance discontinuously, and after eachadvance there is fresh surface, which comesinto equilibrium with the solution within thecrack. So, any experiment, such as that shownin Fig. 34 (Ref 33), that tests crack propagationagainst electrochemical events will absolutelysupport this model.

It is axiomatic that films must rupture and re-form as cracks advance discontinuously. Theweakness of this model is that it does not pro-vide a mechanism for brittle fracture, and thevery brittle features of transgranular SCC frac-ture surfaces do not show any supporting evi-dence of dissolution. Research (Ref 34) show-ing that metal dissolution at the crack tip isisotropic rather than crystallographically ori-ented make dissolution models incapable ofbeing reconciled with the crystallographic frac-ture surface facets.

Adsorption-induced brittleness, also knownas stress-sorption, looks to the parallels be-tween liquid metal embrittlement and SCC toexplain the mechanism of SCC as the action ofadsorbed species weakening atomic bonds onthe crack tip surface. If the action is on the sur-face, however, the mechanism cannot producethe observed discontinuous, brittle cracks whichcharacterize SCC. Only in alloys such as Fe-3Siare steps small enough to make this mechanismplausible.

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


Adsorption-enhanced plasticity/hydrogen em-brittlement encompasses a number of modelsthat observe that adsorbed species enter the lat-tice in the vicinity of the crack tip and thencause failure by one of several mechanisms:

• Dealloying and porosity• Adsorption-induced brittleness• Coalescence of voids formed by cross slip

enhanced by the adsorbed species

Since hydrogen is the only species that is pro-duced in quantity and is capable of diffusing intothe lattice, HE is implicit in all these models.

All of the models have support in that theyhave some experimental observations that showthat the phenomena they propose as causal

actually take place, but none is specific enoughto have been tested by critical experiments toprove or disprove it.

It has been demonstrated that hydrogen is ab-sorbed into the material at the crack tip. Themain question is whether it causes damage bycreating porosity, altering dislocation mobility,or causing lattice decohesion. There is supportfor each.

It has been observed that where SCC occursthere is a large concentration of vacancies. Thishas led to speculation that porosity is a weaken-ing mechanism responsible for SCC (Ref 35). Ithas been proposed and supported by calcula-tions that hydrogen lowers the energy requiredor vacancy formation. The lowest energy

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


configuration is calculated as two hydrogenatoms per vacancy. This pairing of hydrogensolute atoms to dislocations is very reasonablegiven the major distortion the interstitial hydro-gen causes to the lattice, so there is no basis tochallenge the enhanced vacancy formation.Whether the effect is large enough to cause fail-ures has not been demonstrated. The largestmeasurable effect of hydrogen has been a slightacceleration of stress relaxation in martensite.The relevance of hydrogen-induced vacancy ag-glomeration as the principal cause of failure mustbe considered questionable until some furthercritical experiments link the vacancies to the ob-served instances of failure quantitatively, andmore important, to show how this mechanismcould account for the temperature and stress de-pendence observed. The major influence of va-cancy formation due to hydrogen may be to en-hance the volume expansion due to hydrogen.

Adsorption-induced brittleness, also knownas stress sorption, looks to the parallels be-tween liquid metal embrittlement and SCC toexplain the mechanism of SCC as the action ofadsorbed species weakening atomic bonds onthe crack tip surface. If the action is on the sur-face, however, the mechanism cannot producethe observed discontinuous, brittle cracks thatcharacterize SCC. Only in alloys such as Fe-3Si are steps small enough to make this mech-anism plausible. In stainless steels, thereseems to be nothing to support this proposedmechanism.

Adsorption-enhanced plasticity has becomeknown recently as HELP or hydrogen-enhancedlocalized plasticity. The underlying mechanismat work in this model is the hydrogen-inducedshielding between microstructural defects. Thishas been observed distinctly in single crystals ofaustenitic stainless alloys. The Cottrell atmos-phere of hydrogen around dislocations causesmutual repulsion, causing strain to be localizedon certain slip systems. This has been observedto occur and has caused deformation to becomeconcentrated in Luders bands in austenitic al-loys, which of course do not show such behav-ior without hydrogen (Ref 36). This also pro-duces ε-martensite in austenitic alloys, whichwould be considered stable without hydrogenand deformation.

This theory encounters a problem, however,with the fact that the same studies showed thathydrogen actually strengthens the matrix bysolid solution hardening. It acts in much thesame way as carbon and nitrogen do, as shown

in Fig. 35 (Ref 36). All these interstitials strainthe lattice and therefore harden in proportion totheir atomic size. Hydrogen, as the smallest ofthem, has about half the distorting effect andhalf the hardening effect. But, its small sizemakes it mobile at ambient temperatures, so itcan diffuse to sites where it can alter mechani-cal properties.

But, while hydrogen causes dislocation mo-tion and lower work hardening, it does notweaken austenite, so this theory by itself cannotaccount for the role of hydrogen in SCC and, byinference, in HE in the more general case.

The quandary of hydrogen finally havingbeen shown to have a clear effect on mechanicalproperties but having that not account for eitherSCC or HE may be put to rest by the additionalobservations of hydrogen’s role as a lattice dis-torter (Ref 37). While not formalized as a pro-posed hypothesis for SCC, the role of hydrogenas a generator of very high stresses has beenpointed out as a factor that cannot be neglectedwhen evaluating other proposed mechanisms.Hydrogen has been shown to distort the latticein proportion to its concentration. The effect isnot small, accounting for about 1% strain per0.1% concentration by weight, as shown inFig. 36 (Ref 38). At hydrogen levels of over1000 ppm, which are thought to exist aroundgrowing SCC crack tips, there could thereforebe hydrogen concentration gradients capable ofproducing additional tri- or biaxial stresses onthe matrix ahead of the crack tip that may ap-proach the yield stress and account for some orall of the difference between the normal fracturetoughness KI and the KISCC, that for SCC. Thisalso precludes the necessity of hypothesizinghydrogen-induced phase changes, although

Fig. 35 Stress-strain curve for single crystals of stableaustenitic stainless steel with and without hydrogen.

Source: Ref 36


were they to exist, they would result in the samelattice expansion. In both cases, the failurewould occur at a region ahead of the crack tipand beyond the highest hydrogen concentration,which is what is observed to occur. The growthof this stress over time with increasing hydro-gen-producing corrosion would account for theobserved kinetics, locus, and stress dependenceof SCC.

If nothing else, the main models for SCC andthe experimental results on which they arebased should be reexamined in view of the factthat the stresses induced by hydrogen are notnegligible and, in fact, may account for much ofthe observed SCC behavior of stainless steels.

The next few years may finally see the resolu-tion of the lengthy debate over the causes ofSCC. If it comes, it will be from critical experi-ments, which can quantitatively differentiateamong the above effects and measure the con-tribution of each.

Hydrogen Embrittlement

Like SCC, there has been debate about HEthat has produced more heat than light. This in-

volved distinguishing among the same mecha-nisms, namely:

• Decohesion• Enhanced local plasticity• Adsorption embrittlement• Void coalescence

The identification of the operative mechanismfor HE involves again distinguishing what roleeach of the above contributes to HE in a givensituation since all are known to be real metallur-gical phenomena.

The main difference between HE and SCC instainless steel is that HE is limited to ferrite,which is hardened by cold work or alloying, andmartensite. Austenite is somewhat diminishedin ductility by hydrogen, but not subject to thecompletely brittle, discontinuous cracking ofbcc stainless. The observations that make agiven model plausible as a mechanism for SCClack traction for the same materials in HE. Ithard to envision enhanced plasticity involved inthe completely brittle fracture of high-strengthmartensitic stainless steels, whereas void coa-lescence by vacancy creation seems more likelyto account for the observed behavior.

Fig. 36 Dilation of austenite due to hydrogen in solution. Source: Ref 38


The resolution of mechanism here also mustaccount for the contribution of hydrogen-in-duced stress as well as hydrogen effects on me-chanical processes, especially since the ob-served susceptibility to HE is proportionalhardness, therefore to the amount of hydrogen agiven material can hold both in normal intersti-tial solution and the amount it can trap at latticedefects (Ref 39), especially the dislocationswithin the plastic zone at the crack tip, whichprovide enhanced hydrogen solubility where itcan aggravate the applied crack opening with awedge effect from hydrogen dilation.

Corrosion Fatigue

Just like SCC, corrosion fatigue causes brittlefailure under a combined environment of corro-sion and a tensile stress component. The stress,however, is cyclic and in a test of stress versusnumber of cycles (S vs. N), failure will occur at alower N under the corrosive environment. Thecracks are transgranular, and the collaborative ef-fect of corrosion and fatigue is that corrosion ac-celerates the plastic deformation that accompa-nies the evolution of extrusions and intrusions.

In corrosion fatigue, an obvious pit corrosionsite may not be necessary because of the com-bined action of cyclic stresses and the environ-ment. However, an initiation site that is theweakest link in a combined mechanical andmetallurgical sense will be the initiation pointafter which conditions that may not cause SCCcan help propagate fatigue cracking at lowerstresses than would be expected in more benignenvironments and in environments that may notcause SCC or pitting under static loads.

The importance of the environmental interac-tion is reflected in the sensitivity to frequency ofstress application. High-frequency loadinggives less time for corrosive attack and bringscrack propagation rates down closer to those inair. In some materials, crack propagation ratesare elevated above those in air at all stress lev-els, while in others a threshold stress intensitymust be reached before an acceleration is noted.Some materials show a combination of both.The first case seems to be merely fatigue as-sisted by corrosion, while the last two seem toindicate an SCC–type behavior.

The same uncertainties that cloud our under-standing of SCC necessarily disguise the pre-cise mechanism of corrosion fatigue, whichmust be viewed as a combination of SCC andfatigue.

Biocorrosion and MicrobiologicallyInduced Corrosion

There are many cases for which biological or-ganisms contribute to initiating or enhancingrates of corrosion. This can occur in natural en-vironments such as ground or seawater as wellas domestic and industrial environments such asthe nuclear and chemical processing industries,for example. This is called biocorrosion orMIC, microbiologically induced corrosion.

The bacteria that are known to influence cor-rosion can be sorted as aerobic bacteria that liein aerated water and anaerobic bacteria. Amongthe anaerobic bacteria that are known to [16](Ref 40) affect stainless steels can be counted:Desulfibrio and Desulfotomaculum. Both ofthese are so-called sulfate-reducing bacteria(SRB), which means that they promote the reac-tion:

(Eq 16)

which in turn accelerates the cathode reaction:

(Eq 17)

Aerobic bacteria flourish under oxygen (Ref40). Examples are the iron-oxidizing Gal-lionella and Sphaerotilus, which increase theanode dissolution reaction:

(Eq 18)

by converting the ferrous iron-ion product(Fe2+) to less soluble ferric (Fe3+). Due to this,macroscopic so-called tubercules form that cancause crevice-type shelters where differentialaeration and pit initiation can occur.

Countering MIC with biocides can causeproblems in manganese-containing waters. Oxi-dizing biocides, such as ozone, chlorine, or per-oxide, can cause manganese to be oxidized tomanganese dioxide. The precipitated deposits ofmanganese dioxide can accelerate pitting corro-sion even in low-chloride waters in which al-loys such as 316 would otherwise be safe frompitting attack.

Biocorrosion is most commonly encounteredin ambient aqueous environments, which are theenvironments in which most microorganismshave evolved to thrive. So, it tends to be a prob-lem for the medium-alloyed steels such as 304and 316, which are used in these environments.

Fe Fe→ =+ −2 2e

2 2 2H H= −+ →e

SO S O42 2 4− −→ +


The more chemically or thermally hostile envi-ronments in which higher alloyed grades areused are also hostile to bioorganisms and thusminimize the problem.

The development of microbiological consortiaallow anaerobes to flourish under biofilms thatform in an aerated environment. These representa differential aeration cell that acts just like a se-vere crevice. Also, the action of microbes inraising the corrosion potential is key to under-standing why natural seawater is so much morecorrosive than sterile sodium chloride or syn-thetic seawater solutions. And, macrofouling or-ganisms are important. They create crevices andsites where microfouling can start early. At thesame time, they are sources of turbulence inflowing systems, and this turbulence can causeflow erosion in copper materials, making use ofstainless steels more attractive.

REFERENCES

1. M.P. Ryan et al., Critical Factors in Local-ized Corrosion, Proc. Electrochem Soc., Vol150, 2003, p 284–294

2. K. Sieradski and R.C. Newman, J. Elec-trochem. Soc., Vol 133, 1986, p 1980

3. L. Brewer, Science, Vol 161, 1968, p 1154. W.J. Tobler and S.Virtanen, Critical Factors

in Localized Corrosion, Proc. ElectrochemSoc., 2003, p 583–594

5. B. Baroux et al., Corros. Sci., Vol 47 (No.5), 2005, p 1097–1117

6. http://www.alleghenyludlum.com/pages/products/xq/asp/T.1/qx/productCategory.html

7. K. Kimura et al., High Cr Stainless OCTGwith High Strength and Superior CorrosionResistance, JFE Technical Report 7, Jan 2006

8. http://www.outokumpu.com/applications/documents/start.asp

9. J.E. Truman, Corrosion: Metal/Environ-ment Interaction, Vol 1, Newness-Butter-worths, 1976, p 352

10. J.P. Audouard, Stainless Steels, Les Editionsde physique, 1993, p 268

11. H. Mimura et al., Nippon Steel Tech. Report90, July 2004, p 94–99

12. http://www.alleghenyludlum.com/ludlum/Documents/AL-6XN_sourcebook.pdf

13. http://www.alleghenyludlum.com/ludlum/Documents/al610_611.pdf

14. F. Tagashi et al., Kawasaki Technical Re-port 31, 1994

15. R.C. Newman, Corrosion, Dec 2001, p 1030–1041

16. N.J. Laycock and R.C. Newman, Corros.Sci., Vol 39, 1997, p 1771

17. Y. Kobyashi, S.Virtanen, and H. Bohni,Proc. Electrochem. Soc., 1999, p 533–540

18. Z. Szlarska-Smialowska, Pitting Corrosionof Metals, NACE, Houston, TX, 1986

19. E.T. Turkdogan, Fundamentals of Steelmak-ing, Institute of Materials, 1996

20. H.S. Kim and H. Lee, Met. Trans. A, Vol 32A, June 2001, p 1519

21. M.P. Ryan, D.E. Williams, et al., Nature,Vol 415, Feb 2002, p 770–777

22. A.J. Grekula et al., Corrosion, 40, 1984, p 569

23. Stainless Steels Les Editions de physique, 199324. N. Suutala and M.Kurkela, Stainless Steel

‘84, Metals Institute, 1985, p 240–24725. T.Suter, E.G. Webb, H. Bohni, and

R.C. Alkire, J. Electrochem. Soc., Vol 148(No. 5), 2001, B174–B185

26. M.O. Spiedel, Stainless Steels ‘87, Instituteof Metals, London, 1988, p 247–252

27. Y. Cao, F. Ernst, and G.M. Michal, ActaMater., Vol 51, 2003, p 4171.

28. G. Lothongkum et al., Corros. Sci., Vol 48,2006, p 137–153

29. S. Fujimoto , Sci. Technol. Adv. Mater., Vol5, 2004, p 195–200

30. J.D. Fritz, J.F. Grubb, B.W. Parks, and C.P. Stinner, Stainless Steel World, KCI,P01488, 2001

31. M.O. Spiedel, Met. Trans. A, Vol 12A,1981, p 779

32. A.J. Sedricks, Corrosion of Stainless Steels,Wiley, 1979, p 158

33. R.N. Parkins, Br. Corros. J., Vol 14, 1979, p 534. S. Tahtinen, H. Hahhinen, and T. Hakkarainen,

Stainless ‘84, Metals Institute, 1985, p 143–14835. M. Nagumo et al., Met. Trans. A, Vol 32A,

Feb 2001, p 33236. H. Hanninen et al., Hydrogen Effects on

Materials Behavior, TMS, 2003, p 201–21037. V.J. Gadgil, Scr. Metal., Vol 28, 1993,

p 1489–149438. M. Hoelzel et al., Mater. Sci. Eng. A,

Vol 384, 2004, p 255–26139. B.G. Pound, Hydrogen Effects on Materials

Behavior, TMS, 2003, p 93–10340. S.C. Dexter, Microbiologically Influenced

Corrosion, Corrosion: Fundamentals, Test-ing, and Protection, Vol 13A, ASM Hand-books, 2003, p 398–413

Summary

STAINLESS STEEL, often considered mainlyas a corrosion-resisting material, plays an impor-tant role as a heat-resisting material. This ispartly due to its ability to retain strength at highertemperatures at which many otherwise useful al-loying systems, such as aluminum, copper, andeven titanium, soften. Stainless steel retainsstrength and has excellent oxidation resistancefrom room temperature to nearly 1000 °C, atwhich other economical alternatives are lacking.

Introduction

High-temperature oxidation is a form of envi-ronmental degradation of metals and alloys thatresults from the following chemical reaction inwhich metal atoms M react with gaseous oxidants:

(Eq 1)

Due to the high temperatures involved, thesereactions are generally rapid and thus are a con-cern for high-temperature applications such ascomponents for power generation. The elec-tronegative gaseous oxidant X could be sulfur,chlorine, etc., but the discussion here mainly islimited to oxidation by oxygen or water vapor(in the latter case, hydrogen would be added asa product in Eq 1. For a thorough study of oxi-dation, referred to Ref 1 to 3).

Thermodynamics of Oxidation

As discussed in Chapter 2, Corrosion Theory, areaction will be possible when the net free energy

is negative. In Eq 1, the free energy G is decreasedby a lower nobility of the metal (or a higher activ-ity a of a metallic alloying element), a lower tem-perature T, and a higher partial pressure P of theoxidizing gas according to:

(Eq 2)

In the case of alloy oxidation, for which tem-peratures are high enough to form mixed oxidesor spinels, the activities of the oxide speciesalso need to be considered.

The standard Gibbs free energy ΔG0 is oftenpresented in Richardson-Jeffes (Gibbs free energy-temperature) diagrams such as the oneshown in Fig. 1 (Ref 4).

It is evident from Fig. 1 that the major alloyingelement in stainless steels, chromium, forms athermodynamically significantly more stableoxide, Cr2O3, than those of the base alloy iron(FeO, Fe3O4, and Fe2O3) or the major ternary el-ement nickel (NiO), and to a great extent, thechromium content determines the oxidation be-havior of stainless steels.

The Effect of Chromium. The oxidation ofmulticomponent alloys is a complex processfrom both thermodynamic and kinetic points ofview. A range of oxides may form with variousdegrees of thermodynamic stabilities and stoi-chiometries (including complex ones with dif-ferent cations), and there might be degrees ofsolubilities of oxides in one another. Kinetics oftheir growth is complex because metal solutediffusion in the metal phases varies, as do metaland oxygen ion mobilities in the different oxidephases.

Birks, Meier, and Pettit distinguished betweentwo basic types of behavior: (a) a noble matrix

Δ ΔG G RTa

a P

MeX

M Xy

y= +⎛

⎝⎜⎜

⎞

⎠⎟⎟0

2

0 5ln

.

M s yX g MXy

( ) . ( )+ →0 5 2

CHAPTER 5

Oxidation




metal with less-noble alloying elements and (b)both matrix element and alloying elements arenonnoble. The concept of nobility is decided bythe thermodynamic conditions of Eq 2; that is,an element for which the free energy defined byEq 2 is negative is nonnoble. We first discuss themore common case (b), in which oxidation takesplace under significantly oxidizing conditions,such as air. In such a situation, it can be seenfrom Fig. 1, that the matrix iron is nonnoble andso are many of the solutes (chromium, molybde-num, aluminum, silicon, manganese, etc.).

A high-temperature Fe-Cr-O phase diagram isshown in Fig. 2. It can be seen here that Fe2O3and Cr2O3 are soluble in each other, and that the

spinels Fe1.5Cr1.5O4 (with a solid solubility withFe3O4) and FeCr2O4 form. The progressivechange in oxidation behavior as chromium isadded to iron has been described in the litera-ture (Ref 1).

At lower chromium contents and above a min-imum temperature, an iron-chromium alloywould behave as pure iron, where FeO wouldform next to the metal, then gradually Fe3O4 andFe2O3 would form toward the gas as oxygen po-tential increases. Isolated pockets of spinel mayform within the FeO layer. The oxidation of ironproceeds predominantly due to the rapid ionicdiffusion of Fe2+ cations on the FeO layer,which leads to growth of this layer. If chromium

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

Chapter 5: Oxidation / 59

in the base alloy is increased, the spinel pocketsincrease, and the mobility of Fe2+ decreases. Aschromium content is increased further, a mixed-spinel scale is formed. Iron diffusion through themixed spinel is significant, and thus the scale isnot yet protective. As chromium content is in-creased further, an outer layer of Cr2O3 isformed, and the oxidation behavior becomessimilar to that of chromium. A chromium limitof roughly 20% is needed to achieve a perma-nent Cr2O3 scale. This amount decreases ifnickel is added. While the thermodynamic driv-ing force is important, the chromium content ofthe alloy can override because the supply ofchromium to the interface becomes dominant.At chromium contents less than about 16 wt%,the oxidation rate is influenced by the rate ofsupply of chromium from the alloys beneath theoxide. Above 16%, the supply of chromium isfast enough that chromium gradients are lowenough that instead transport in the oxide layercontrols the rate. The rate of oxidation then fol-lows a so-called parabolic law (this is explainedin the next section), by which the mass changeper unit area due to oxidation (incorporation ofoxygen) is given by:

(Eq 3)

Here, m is the added mass, A is the areaexposed to the oxidizing atmosphere, t is thetime exposed, and km is the parabolic rate con-stant. The subscript “m” is added here to denotethat the reaction is measured as added mass (itcan also be defined for oxide thickness X). Asshown in Fig. 3, the parabolic rate constant forthe oxides of chromium, silicon, and aluminumare low compared to others, and this is the rea-son that these elements are used as alloying ele-ments to reduce oxidation rates for alloys inhigh-temperature applications.

The solubility of Fe3O4 in the spinel willeventually result in continued iron oxide for-mation, as the iron-chromium system is not anoptimal basis for high-temperature oxidationresistance, although it might be an option froman economical standpoint compared to otheralloy systems such as superalloys. The highmobility of both iron and manganese in thespinel structure is also an important factor.

The major stainless steels used for oxidationresistance fall into two categories: the ferriticstainless steels and the austenitic. Table 1 listssome of the more significant alloys commonlyencountered in applications for which oxida-tion resistance is paramount. The value of fer-ritic alloys (such as 409, 439, and 446) is thatthey are relatively inexpensive, and that theyhave a thermal expansion coefficient that iscloser to that of the oxide than do austenitic al-loys (such as 302B, 309, and 310). This givesthem an advantage in cyclic oxidation applica-tions even though their strength at high temper-atures does not rival that of austenitic alloys.The ferritic stainless steels are the most widelyused alloys based on their low cost, which hasm A k t

m/( ) =

2

Fig. 2 The iron-chromium-oxygen phase diagram at 1300 °C.Source: Ref 5

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


made them the standard alloys for automotiveexhaust systems.

Transient Oxidation

The oxidation of a clean metal surface on ex-posure to an oxidizing environment will ini-tially lead to all the nonnoble components of thealloy being oxidized together, forming mixedoxides having composition similar to the basealloy. In stainless steels, these initial oxides aretypically Fe-Cr-Ni-Mn mixed oxides. This iscalled transient oxidation. As these oxidesthicken, the partial pressure of oxygen at thescale-metal interface falls until only the mostreactive element present in high concentrationcan be oxidized. For stainless steels, this meansthat a layer of Cr2O3 is eventually established incontact with the alloy.

The Electrochemical Nature of Oxidation

Once an inner scale of Cr2O3 is formed, asshown in Fig. 4(a), the oxidizing gas is reducedat the gas-scale interface, and the chromium isoxidized at the metal-scale interface. The Cr2O3scale serves as both electrolyte, through whichions are transported, and electron lead, throughwhich electronic defects are transported. Inprinciple, either or both metal or oxygen ionscan migrate. If oxygen ion mobility dominates,then the scale would continue to grow at theoxide-metal interface, whereas if chromium ionmobility dominates, the oxide will grow at the

scale-gas surface. Similarly, the electric chargecan be carried by either n-type (electrons) or p-type (electron holes) electronic defects. Thecase will be determined by the equilibrium de-fect structure of the oxide, which depends ontemperature and oxygen partial pressure. In thecase of Cr2O3, chromium cations are the pre-dominantly mobile defects (Fig. 4b) as a resultof a very small degree of deviation from stoi-chiometry in the cation lattice, that is, Cr2-5O3,leading to metal deficiency. The contribution ofchromia grain boundary diffusion is large andprobably dominates the process at temperaturesof interest. The defect can be described as an in-teraction with oxygen, at high oxygen poten-tials, through Kroger-Vink notations as:

(Eq 4)

The electron holes that form as charge-com-pensating defects serve as the “electron lead” inthe electrochemical cell in Fig. 4b. The free en-ergy of Reaction 4 determines the concentrationof mobile defects and thus the diffusion coeffi-cient and electrochemical mobility (Be) of thecation according to:

(Eq 5)

D VG

RT

V

Cr Cr

m

Cr

3 3

3

2+ +

+

= ⎡⎣ ⎤⎦ ⋅ −⎛⎝⎜

⎞⎠⎟

= ⎡⎣

γ υΛΔ

exp

⎤⎤⎦ ⋅ ⋅ −⎛⎝⎜

⎞⎠⎟

cons tH

RTmtan exp

Δ

V p K p V K K PCr O Cr o''' / '''⎡⎣ ⎤⎦ = → ⎡⎣ ⎤⎦ =

−3

23 4

1

32

22 2

334

3

2

3

232O O( ) '''g V h

Cr Ox= + + •

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

UNS Name

Composition, %

C N Cr Ni Mn Si Ti Nb Other

S40900 409 0.08 10.5–11.75 0.5 1 1 6X(C + N) to1.10

. . . . . .

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 Al4742(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 CeS38150 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 CeS30900 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


(Eq 6)

where zi is the ion charge, F is Faraday’s con-stant (96,457 C.eq–1). Compared to wustite(FeO), the equilibrium constant of Eq 4 is quitelow, resulting in a low degree of nonstoichiome-try in Cr2–δO3 compared to Fe1–δO (where δ canbe as large as 0.05), and thus the transport ofCr3+ through its scale is much slower than thetransport of Fe2+ through FeO and thus the dif-ference in parabolic rate constants in Fig. 3.

Kinetics and Oxidation Rates: Wagner’sTheory

The parabolic oxidation rate was introducedwithout explanation in Eq 3. It was first de-scribed in terms of oxide defect structure and re-sulting transport properties by Wagner (Ref 6),and the theory is explained in most of the mono-graphs on oxidation, such as Chapter 3 in Ref 3and Chapter 4 in Ref 1. This treatment followsthe derivation in Ref 1. Consider a general case,as shown in Fig. 4 under the assumptions that(a) the scale is compact and adherent, (b) elec-trode reactions are rapid enough to be in equi-librium at the interface and surface, (c) nonstoi-chiometry is small and uniform throughout thescale (i.e., defects are in thermal equilibriumthroughout the scale), and (d) double-layer

effects are ignored (i.e., scales are relativelythick compared to range of space charge ef-fects). As a case study, let us assume that themobile ion defect is cations due to metal vacan-cies in the scale.

The molar flux J (moles/m.s) of a particle i inan electrolyte subjected to an electrochemicalpotential gradient was shown to be:

(Eq 7)

where zi is the ion charge, F is Faraday’s con-stant (96,457 C/gram equivalent), and φ is theelectric field (V). The electronic or ionic con-ductivity κ in an electrolyte can be computedthrough:

(Eq 8)

The contribution of a given ion specie type orelectron defect type to this conductivity is de-noted as the partial conductivity and computedas:

(Eq 9)

Inserting Eq 9 into Eq 7 yields:

(Eq 10)JF z

z F

xii

i

i i= −∂ +( )

∂κ μ φ2 2

κi i i i

F z B c= 2 2

κ = ∑F z B ci i i

2 2

J c Bz F

xi i ii i= −

∂ +( )∂

μ φ

BFD

RTCr

e Cr3

33+

+=

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


Now, if the mobile particles are a single typeof metal cations (e.g., Cr3+) and electrons, thentwo fluxes are present:

(Eq 11)

and

(E q 12)

Electrical neutrality requires that:

(Eq 13)

And at the oxide-scale/metal interface, theanode reaction is in equilibrium, that is:

and therefore,

(Eq 14)

Combining Eq 11 to 14, the potential gradientis eliminated, and the chemical potential gradi-ents can be replaced by the metal (M) potentialgradient, and the following equation results:

(Eq 15)

To obtain an explicit function for the cationflux, Eq 15 needs to be integrated after variableseparation, keeping in mind that conductivitiesand metal chemical potential may vary withinthe scale. Integrating from the gas-scale surface(x = 0, μM = μ′′

M) to the scale-metal interface (x = X, μM = μ′′

M) one obtains.

(Eq 16)

Now, if the growth of the oxide scale is con-trolled by the flux of cations:

(Eq 17)

The concentration drop across the scale isconstant since interface and surface reactionsare at equilibrium. If quasi steady state is

assumed, that is, a linear drop across X, at alltimes:

(Eq 18)

The constant k is the parabolic rate constant.A mass balance can be written where the flux ofcations for a period of time dt is equated to theamount of metal being accumulated as cationsinside the scale of thickness dx:

(Eq 19)

By combining Eq 18 and 19 and inserting Eq16 for the flux, an expression for the parabolicrate constant is obtained:

(Eq 20)

If the mobility and thus partial conductivityof electrons is significantly higher than that ofthe ions (a reasonable assumption), then Eq 20can be simplified as:

(Eq 21)

Since from diffusion theory we know that Dc = BcRT, and inserting this in Eq 9, one obtains

and inserting this into Eq 21 results in:

(Eq 22)

It is often more convenient to express Eq 22in terms of oxygen potentials rather than themetal potentials. It was assumed at the onset ofthis analysis that the deviation from stoichiom-etry is small and constant throughout the scale.Therefore, the oxide potential is constant and:

(Eq 23)z

cO M MOzC4 2 2

μ μ μ+ = =/

constant

kRT

D dc M

M

M

= ∫1 μ

μ

μ

''

'

κc c

cc

F zD

RTc= 2 2

kz F C

dc C

c MM

M

= ∫1

2 2κ μ

μ

μ '

kz F C

dc C

c e

c eM

M

M

=+∫

12 2

κ κκ κ

μμ

μ

''

'

J dt C dxc C⋅ =

dx

dt

k

X t=

( )

dx

dtJ D

C

xC CC∝ = −

∂∂

Jz F x

dc

c

c e

c eM

M

M

= −+∫

12 2

κ κκ κ

μμ

μ

'

"

Jz F xc

c e

c c e

M= −+( )

∂∂

κ κκ κ

μ2 2

μ μ μM c c e

z= +

M M z ez

cc= ++ −

J Z J Zc c e e

+ = 0

JF z

z F

xee

e

e e= −∂ +( )

∂κ μ φ2 2

JF z

z F

xcc

c

c c= −∂ +( )

∂κ μ φ2 2


And thus, . Therefore,

Eq 22 can be written:

(Eq 24)

In the case of Cr2O3, if bulk diffusion is dom-inating, Eq 4 and 5 inserted into the diffusioncoefficient in Eq 24 for C = Cr3+, results in:

(Eq 25)

Thus, the parabolic rate constant would be pre-dicted to vary with the power of three-quarters ofthe external oxygen partial pressure.

Grain boundary diffusion has however beenidentified to be important in the case of Cr3+

transport (Ref 8). The observed growth rate ofCr2O3 polycrystalline films is far too fast to beaccounted for by bulk diffusion of chromiumions; instead, grain boundary diffusion wouldbe expected to dominate (Ref 9).

The Volatile Nature of Cr2O3

At high enough temperatures and highenough oxygen partial pressures, the formationof a gaseous hexavalent chromium oxide CrO3*could lead to thinning of the Cr2O3 scale ac-cording to the following reaction:

(Eq 26)

Figure 5 shows the vapor pressure of the su-peroxide as a function of temperature and par-tial pressure of O2.

The effect of this reaction on the oxidation ki-netics can be described as follows: The thick-ness change described through the parabolicrate constant in Eq 18 is corrected for by the

thickness loss due to evaporation, which is de-scribed by a first-order reaction kinetics expres-sion with rate constant ke. Thus, the thicknesschange becomes:

(Eq 27)

This results in a so-called paralinear (as op-posed to parabolic) rate for the oxide thickening(Fig. 6), and at a critical oxide scale thicknessX( the rate of thinning due to evaporation equalsthe rate of thickening due to oxidation. In Eq27, this means that dx/dt = 0; consequently, X =k/ke. While at first this seems to suggest that itdoes not affect the oxidation process in that therate of oxidation does not increase, the forma-tion and evaporation of chromium oxides re-sults in greater chromium consumption in thealloy compared to what would be the case ifevaporation did not occur. As a result of evapo-ration losses, stainless steels that depend on aprotective chromium oxide layer are limited inuse to temperatures up to 900 to 1000 °C.

The presence of water vapor promotes theformation of even more volatile oxyhydroxides(e.g., CrO2(OH)2 ) (Ref 11, 12).

Spalling and Cracking of the Scale

At elevated temperatures or during tempera-ture cycling, there are multiple ways in whichstresses can develop that may crack and blisterthe scale, rendering it nonprotective. The differ-ent causes of stress generation are described inChapter 5 in Ref 1. So-called growth stressesarise due to changes caused by the oxidationprocess itself. These include differences in lat-tice mismatch, alloy depletion in the metal,point-defect gradients in scales containing ox-ides such as FeO, with large deviation from sto-ichiometry, recrystallization, and volume differ-ences between the oxide and metal. The last isperhaps the most commonly mentioned and ischaracterized by the Pilling-Bedworth ratio, ab-breviated as PBR (Ref 13).

(Eq 28)

Here, the subscript m stands for molar volume,and υ is the number of metal atoms needed toform a stoichiometric unit of the oxide (in the

PBRV

V

V

VOxide

Metal

mOxide

mMetal

= =* ν

dx

dt

k

X tk

e= −

( )

Cr O O CrO2 3 2 3

3

22( ) ( ) ( )s g g+ =

kP

PdP P dP

OO

P

P

O

P

P

O

O

O

O

∝ =∫ ∫ −3 4

1 4

2

2

2

2

2

2

2//

'

''

'

''

''/

'/

= ( ) − ( )⎡⎣⎢

⎤⎦⎥

3

4 2 2

3 4 3 4

P PO O

k D d PD

PdP

c O

P

P

c

OO

P

P

O

O

O

O

∝ =∫ lnln

ln

'

''

'

''

2

2

2

2

2

2

2

∫∫

d dRT P dO O Me

μ μ2 2

= ∝ −ln

* Hexavalent chromium is now considered a human car-cinogen and is rigorously regulated by both the Occupa-tional Safety and Health Administration (OSHA) and theU.S. Environmental Protection Agency (EPA).


case of Cr2O3, υ is 2, and in the case of FeO, itis 1). When PBR is greater than 1, then theoxide is expected to be in compression and islikely to be protective, whereas if it is less than1, the oxide is in tension and thus nonprotective.There are, however, many exceptions to this,partly because the stress state often dependsmore on the mechanisms and conditions of theoxidation process rather than the properties ofmetal and oxides.

Thermal, stresses are caused by differences inthermal expansion between the oxide and metal,and the stresses generated in oxide scales can beestimated:

(Eq 29)

The equation is written for a case shownschematically in Fig. 7, where both sides on ametal undergo oxidation. Here, σ is the stress,υp is Poisson’s ratio (it has been assumed thatthere is no mismatch), α is the coefficient ofthermal expansion, E is the modulus of elastic-

ity, t is thickness, and ΔT is the temperaturechange. In general, α is larger for the metal thanthe oxide; thus, during cooling the stresses areexpected to be compressive and during heatingtensile. Thermal stresses can cause spalling ofthe protective oxide layer, and it is most severeunder cyclic conditions.

Effect of Silicon, Aluminum, and Molyb-denum. Due to concerns about the cost ofchromium and its (former) classification as a

σα α

νOxide

Oxide Oxide Metal

pOx

E T

t=

− −( )−( ) +

Δ

1 1 2 iide Oxide

Metal Metal

E

t E

⎛

⎝⎜⎞⎠⎟

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

Fig. 6 Schematic of paralinear oxidation as a result of evapo-ration of chromium superoxide


strategic material, there were efforts to try tosubstitute less-expensive elements such as alu-minum and silicon that also are known to formprotective layers (Ref 14), even though theyhave significant metallurgical and mechanicaldrawbacks.

Silicon additions of 1.5 wt% or more have theeffect of forming a continuous amorphous sub-surface layer in iron-silicon and Fe-Cr-Si alloysthat is relatively impervious to transport of ions.The mechanism for the evolution of such a layeris as follows: (1) The more readily availableiron or chromium first forms a surface layer,and this causes an enrichment of silicon at theoxide-metal interface. (2) As sufficient silicon isenriched, the SiO2 layer is formed. It has beenreported that alloys with chromium content aslow as 6 wt% and silicon content of 1.5 wt%perform in terms of oxidation as well as com-mercial stainless steels. Also, an addition of 4wt% Si to a Fe14wt%Cr14wt%Ni alloy resultedin a 200-fold reduction in weight gain at 900 °CHowever, this SiO2 layer seems to promoteoxide spalling, especially in cyclic service.

Aluminum forms a very stable thin outerlayer of Al2O3 that initially reduces the oxida-tion rate. Alumina is among the most stable anddefect-free oxides, giving it an extremely lowdiffusion rate. In consequence, if sufficient alu-minum is present to maintain the protective alu-mina scale, the aluminum-bearing alloys pro-vide the greatest oxidation resistance attainablein engineering alloys. If, however, the alu-minum content is not sufficient to force aluminascale formation at the scale-metal interface, an

internal Cr2O3 layer forms and thickens. Even-tually, the surface aluminum-oxide layer flakesoff. Also, due to the low oxygen potentialneeded to form Al2O3, internal oxidation mayresult below the metal-scale interface in alloysin which formation of a continuous aluminascale film does not occur.

Molybdenum is suggested to strain the lat-tice due to its larger size and consequently in-crease the rate of bulk diffusion of elements(Ref 11), which can enhance the rate of initialCr2O3 formation. Molybdenum is usually con-sidered detrimental for oxidation resistance.Molybdenum normally forms MoO2 oxide, butthis can oxidize further to form the low-melt-ing and volatile MoO3. If the MoO3 evapo-rates, there is little problem, but if itsvolatilization is inhibited by low atmospherecirculation, liquid MoO3 can accumulate anddissolve the protective Cr2O3 scale, leading tocatastrophic oxidation.

Effect of Rare Earth Additions. Cerium, lan-thanum, and yttrium additions are known to im-prove oxidation resistance of high temperaturenickel- and iron-based alloys (Ref 6). Rare earthadditions have been suggested to have a multi-tude of beneficial effects, such as reducing thegrowth kinetics of Cr2O3 scales, stabilizingCr2O3 scales at lower chromium levels, increas-ing adhesion, and preventing spalling of theoxide scale during thermal cycling. The expla-nation for any of this does not seem clear, butsome hypotheses have been suggested. The ef-fect on the growth kinetics could be because thereactive element ions collect at grain boundariesand block fast path diffusion.

The improved adherence could be becausethese elements getter tramp elements such assulfur and suppress void formation at the inter-face. Furthermore, they might form so-calledoxide pegs at the interface (Ref 16). The pre-cise role of the rare earth additions to Cr2O3oxide protection and the mechanism by whichthey are incorporated into the scale during thesurface treatment processes remain unknown.An understanding of these fundamental issueswould help to develop optimum alloychemistries for selected high-temperature and -pressure applications and to further develop thesurface infusion process. The lack of funda-mental understanding of how rare earths im-prove oxidation resistance has not stopped thedevelopment of several alloys that benefit fromthe effect.

tox toxtm

Oxide OxideMetal

Fig. 7 Schematic of a cross section of oxidized sample indi-cating dimensions in Eq 29 for predicting thermal

stresses


Oxidation Under Less-OxidizingAtmospheres

The two types of alloy oxidation behaviors,(a) a noble matrix metal with less-noble alloy-ing elements and (b) both matrix element andalloying elements are nonnoble, were men-tioned. When designing against oxidizing envi-ronments, case b is perhaps the most relevant,and most of the discussion has been devoted tothis. However, during annealing for microstruc-tural control, steels are exposed to furnace gasesat high temperatures that have relatively lowoxygen or steam contents, for which case a willapply, that is the atmosphere does not cause iron(or nickel) to oxidize but chromium (and alu-minum, silicon, molybdenum, etc.) does.

For simplicity, assume a binary system A-Bof “noble” iron and “reactive” chromium. Inthis case, depending on the concentration ofthe reactive element and atmosphere, theoxide of the reactive element can, in principle,form either on the surface (as has been dis-cussed so far) or internally as discrete oxideparticles in a metal matrix through oxygendiffusion into the metal. Both cases are shownschematically in Fig. 8. Let us discuss theconditions that promote one or the other ofthese by starting with a situation in which (a)no surface oxide exists and (b) the oxygen at-mosphere is such that the solubility of oxygenwithin a distance X is enough to thermodynam-ically render Cr2O3 stable according to Eq 2but none of the iron oxides. The derivation isdone in terms of both a generic system A-Bcausing an oxide BOν and for iron-chromiumcausing Cr2O3. Assume for a start that the DO>> DCr, and thus while oxygen diffuses intothe alloy, chromium does not counterdiffuse.The flux of oxygen inward into the metal isthen the cause of increased mass. Assuming a

quasi-steady-state situation as shown in Fig. 9,the flux can be written:

(Eq 30)

Within the depth X, the oxygen solubility issuch that Eq 2 is negative enough that Cr2O3forms. Beyond X, it is not. At the distance X, theoxygen concentration is negligibly low com-pared to the surface composition, which is inequilibrium with the gas phase. The molar vol-ume Vm is used to obtain the flux in units ofmoles per square meter. Within the layer 0 < x <X, all the chromium is assumed to be oxidized;therefore, the accumulated mass due to oxygenaddition:

(Eq 31)

Differentiating Eq 31 with time, equating toEq 3, and separating variables results in:

(Eq 32)

Integrating Eq 32 from x = 0 to x = X results inan expression for the internal oxidation depth X:

(Eq 33)XN D

Nt

N D

NtO

SO

B

OS

O

Cr

=⎛

⎝⎜⎞

⎠⎟=

⎛

⎝

⎜⎜⎜

⎞

⎠

2 2

32

0

1 2

0υ

/

⎟⎟⎟⎟

1 2/

XdXN D

Ndt

N D

N

OS

O

BO

OS

O

Cr

= =υ 3

20

mN X

V

N X

VB

m

Cro

m

= =0

32υ

Jdm

dtD

N N

V XD

N

V XO OOX

OS

mO

OS

m

= = −−

=

Fig. 8 Schematic of two cases in a less-oxidizing atmosphere. (a) Adsorption of oxygen leading to internal oxidation

and (b) external oxidation as the B element migrates.

Fig. 9 Quasi-steady-state approximation of the movingboundary problem of internal oxidation. Counterdif-

fusion of B is assumed to be negligible


This expression predicts a parabolic depend-ence of X with time, just as Eq 3 did for the ex-ternal oxidation. For a more rigorous derivation(without assuming quasi steady state), refer toAppendix B in Ref 1. Now, let us see whatcauses this to transition into an external scale. Itwas assumed in the derivation of Eq 33 thatcounterdiffusion of chromium does not occur.When considering Eq 33, is clear that the rate ofpenetration of the internal oxidation front willdecrease with (a) increasing NB

0, (b) decreasingNO

S, and (c) decreasing DO. If DCr was not negli-gible, there would be a gradual change in oxida-tion morphology if the ratio (NB

0DB)/(NOSDO).

Gradually, if the ratio were increased therewould be a slowing of the inward penetration ofthe internal oxide front and an enrichment ofBOυ in the internally oxidized zone. Wagner de-veloped a model (Ref 16), based on that at somepoint, when the volume fraction of BOυ versusvolume metal in the internally oxidized zonereaches a critical value g*, there is a transitionfrom internal to external oxidation; specifically,this happens when:

(Eq 34)

If more than one reactive elements were pres-ent (such as is the case in stainless steels inwhich aluminum, silicon, molybdenum, nio-bium, etc., may be present), this will decreasethe inward flux of oxygen, and thus the transi-tion to external may occur at a lower solute(CCr

0) concentration than what is predicted byEq 34.

Metal Dusting. Under reducing conditions,in products of combustion atmospheres, oxida-tion and carburization may occur simultane-ously and at a higher rate than exhibited in pureoxidation. Under even more reducing condi-tions, the condition called metal dusting mayoccur. Metal dusting is often characterized bythe generation of large, smooth pits that look asif metal had been scooped from the surface. Theunderlying phenomenon is the formation ofmetal carbides, which manifests itself as thebreakup of bulk metal to metal powder. This oc-curs at temperatures at which the carbide ismost stable (Fig. 10).

During oxidation in air, if large surface area ispresent under conditions of restricted air supply,oxygen can be depleted to the point that oxida-tion essentially ceases. Oxide films inhibit ni-

trogen absorption, but if the oxygen is depletedbefore all surfaces are oxidized, the remainingmaterial can be rapidly nitrided by the residual,essentially pure, nitrogen atmosphere.

REFERENCES

1. N. Birks, G.H. Meier, and F.S. Pettit, Intro-duction to the High-Temperature Oxidationof Metals, 2nd ed., 2006, Cambridge Uni-versity Press, New York

2. P. Kofstad, High Temperature Corrosion,1988, Elsevier Applied Science, London

3. K. Hauffe, Oxidation of Metals, PlenumPress, New York, 1965

4. F.D. Richardson and J.H.E. Jeffes, J. IronSteel Inst., Vol 160, 1948, p 261

5. C. Wagner and K. Grünewald, Z. Phys.Chem., Vol 40B, 1938, p 455

6. J.H. Park, W.E. King, N.L. Peterson, andS.J. Rothman, The Effect of Reactive Ele-ment on Self-Diffusion in Cr2O3, NormanL. Peterson Memorial Symposium, Oxida-tion of Metals and Associated Mass Trans-port, edited by M.A. Dayananda, S.I. Roth-man, and W.E. King, AIME, Warrendale,PA, 1998, p 103–107

7. C. Wagner and K. Grünewald, Z. Phys.Chem., Vol 40B, 1938, p 455

8. D. Caplan and G.I. Sproule, Effect of OxideGrain Structure on the High TemperatureOxidation of Cr, Oxid. Met., Vol 9, 1975, p 459–472

9. B.B. Ebbinghaus, Combust. Flame, Vol 93,1993, p 119–137

Ng

ND V

D VB OS O M

B Ox

0

1 2

2>

⎡

⎣⎢

⎤

⎦⎥

πν

*/

Fig. 10 Temperature dependence of metal dusting of iron.Source: Ref 18


10. K. Hilpert et al., JECS, Vol 143/11, 1996, p3642–3647

11. N.B. Pilling and R.E. Bedworth, J. Inst.Met., Vol 29, 1923, p 529

12. J.K. Tien and J.M. Davidson, Oxide Spalla-tion Mechanisms, Stress Effects and the Ox-idation of Metals, ed. J.V. Cathcart, AIME,New York, 1975, p 200

13. J. Rawers, Understanding the OxidationProtection of Fe-Cr-Si Alloys, Norman L.Peterson Memorial Symposium, Oxidation

of Metals and Associated Mass Transport,edited by M.A. Dayananda, S.I. Rothman,and W.E. King, AIME, Warrendale, PA,1998, p 323–340

14. E.J. Felten, J. Electrochem. Soc., Vol 108,1961, p 490

15. C. Wagner, J. Electrochem. Soc., Vol 103,1956, p 571

16. C.M. Chun, J.D. Mumford, and T.A. Rama-narayanan, J. Electrochem. Soc., Vol 147,2000, p 3680

CHAPTER 6

Austenitic Stainless Steels

Summary

AUSTENITIC STAINLESS STEELS are themost common and familiar types of stainlesssteel. They are most easily recognized as non-magnetic. They are extremely formable andweldable, and they can be successfully usedfrom cryogenic temperatures to the red-hot tem-peratures of furnaces and jet engines. They con-tain between about 16 and 25% chromium, andthey can also contain nitrogen in solution, bothof which contribute to their high corrosion re-sistance. Were it not for the cost of the nickelthat helps stabilize their austenitic structure,these alloys would be used even more widely.

Introduction

Austenitic stainless steels have many advan-tages from a metallurgical point of view. Theycan be made soft enough (i.e., with a yieldstrength about 200 MPa) to be easily formed bythe same tools that work with carbon steel, butthey can also be made incredibly strong by coldwork, up to yield strengths of over 2000 MPa(290 ksi). Their austenitic (fcc, face-centeredcubic) structure is very tough and ductile downto absolute zero. They also do not lose theirstrength at elevated temperatures as rapidly asferritic (bcc, body-centered cubic) iron base al-loys. The least corrosion-resistant versions canwithstand the normal corrosive attack of theeveryday environment that people experience,while the most corrosion-resistant grades caneven withstand boiling seawater.

If these alloys were to have any relativeweaknesses, they would be:

1. Austenitic stainless steels are less resistantto cyclic oxidation than are ferritic grades

because their greater thermal expansion co-efficient tends to cause the protective oxidecoating to spall.

2. They can experience stress corrosion crack-ing (SCC) if used in an environment to whichthey have insufficient corrosion resistance.

3. The fatigue endurance limit is only about30% of the tensile strength (vs. ~50 to 60%for ferritic stainless steels). This, combinedwith their high thermal expansion coeffi-cients, makes them especially susceptible tothermal fatigue.

However, the risks of these limitations can beavoidable by taking proper precautions.

Alloy Families in Perspective

The fundamental criterion in the selection ofa stainless steel is generally that it can survivewith virtually no corrosion in the environmentin which it is to be used. Good engineeringpractice sometimes requires that materials beselected for sufficient, but finite, service life.This is especially true for high-temperatureservice, for which creep and oxidation lead tolimited life for all materials. The choice amongthe stainless steels that can be used in that envi-ronment is then based on the alloy from whichthe component can be produced at the lowestcost, including maintenance, over the intendedservice life. The ferritic stainless steels are lessexpensive for the same corrosion resistance butsometimes are found lacking because of:

• Lack of toughness, as is the case at subambi-ent temperatures or in thicknesses greaterthan about 1.5 mm

• Lack of great ductility, specifically if morethan about 30% elongation is needed




• Susceptibility to high-temperature embrit-tling phases when moderately alloyed

The less-expensive martensitic grades areused instead of austenitic when high strengthand hardness are better achieved by heat treat-ing rather than by cold work, and mechanicalproperties are more important than corrosion re-sistance. This is also the case for the more ex-pensive PH grades, which can achieve corrosionresistance only matching the least corrosion re-sistant of the austenitic alloys.

Duplex grades match austenitic grades in cor-rosion resistance and have higher strength in theannealed condition but present the designerwith challenges with regard to embrittlingphases that can form with prolonged exposureto elevated temperatures and only moderateductility like the ferritic alloys.

So, the austenitic grades are the most com-monly used grades of stainless mainly because,in many instances, they provide very predictablelevels of corrosion resistance with excellent me-chanical properties. Using them wisely can savethe design engineer significant costs in his or herproduct. They are a user-friendly metal alloywith life-cycle cost of fully manufactured prod-ucts lower than many other materials.

The austenitic alloys can have compositionsanywhere in the portion of the Delong diagramlabeled austenite shown in Fig. 1 (Ref 1). Thisdiagram was designed to show which phases arepresent in alloys in the as-solidified condition,such as found in welds. Thus it also applies tocastings and continuously cast products. As apractical matter of castability, the compositionof most commercial alloys falls along the zoneof several percent ferrite as cast. The salient fea-ture of austenitic alloys is that as chromium andmolybdenum are increased to increase specificproperties, usually corrosion resistance, nickelor other austenite stabilizers must be added ifthe austenitic structure is to be preserved.

The traditional way of displaying theaustenitic stainless steels is to present 302 as abase. Figure 2 shows one such diagram. Dia-grams such as these treat alloys as an evolution-ary family tree and subtly mislead. Many alloyswere pushed toward obsolescence because ofadvances in processing. For instance, 321 wasdeveloped as an alloy in which the detrimentaleffects of carbon were negated by addition of ti-tanium. The widespread adoption of the argonoxygen decarburization (AOD) in the 1970smade this alloy unnecessary, except for specialcircumstances, since carbon could be cheaply

Fig. 1 Schaeffler-Delong stainless steels constitution diagram. Adapted from Ref 1, 2

Chapter 6: Austenitic Stainless Steels / 71

removed routinely. Likewise, 302 gave way tothe lower-carbon 304, for which the even lower-carbon 304L is commonly substituted and du-ally certified to qualify as either grade. Whilelow carbon prevents sensitization, stabilizedgrades may still be preferred for special applica-tions such as type 321 in aerospace and type347 in refinery service. Similar inertia keeps thehigher-nickel 300 series as the de facto standardwhen the more cost-efficient high-manganese200 series is the logical basic grade. The rele-vant types of austenitic alloys can nonethelessbe rationalized with this diagram.

As chromium is added, oxidation resistanceand corrosion resistance increase. Because

nickel equivalents (manganese, nitrogen, carbon,etc.) must also be added in matching amounts,austenite stability is also increased. If molybde-num, a chromium equivalent, is added, corro-sion resistance but not oxidation resistance isenhanced. And, if nitrogen is the austenite stabi-lizer added to balance increases chromium ormolybdenum, then corrosion resistance is alsoincreased. With small exceptions, that is therationale of austenitic grade design. Siliconis used as an alloy to promote oxidation resist-ance and resistance to corrosion by oxidizingacids. Copper is used to promote resistance tosulfuric acid. Rare earths make a more stablyoxidation-resisting scale. Niobium increases

Fig. 2 The austenitic stainless family. Source: Ref 3


creep resistance. Sulfur and selenium increasemachinability.

In this chapter, austenitic alloys are classifiedinto three groups:

• Lean alloys, such as 201 and 301, are gener-ally used when high strength or high forma-bility is the main objective since the lower,yet tailorable, austenite stability of these al-loys gives a great range of work-hardeningrates and great ductility. Richer alloys, suchas 305, with minimal work hardening are thehigh-alloy, lowest work-hardening rategrade for this purpose. The general-purposealloy 304 is within this group.

• Chromium nickel alloys when the objectiveis high temperature oxidation resistance.This can be enhanced by silicon and rareearths. If the application requires high-tem-perature strength, carbon, nitrogen, niobium,and molybdenum can be added. 302B, 309,310, 347, and various proprietary alloys arefound in this group.

• Chromium, molybdenum, nickel, and nitro-gen alloys when corrosion resistance is themain objective. Alloys such as silicon andcopper are added for resistance to specificenvironments. This group includes 316L,317L, 904L, and many proprietary grades.

Wrought alloys generally have cast counter-parts that differ primarily in silicon content.Versions that require enhanced machinabilityhave a high content of controlled inclusions,sulfides, or oxysulfides, which improve machin-ability at the expense of corrosion resistance.Carbon is kept below 0.03% and designated anL grade when prolonged heating due to multi-pass welding of heavy section (greater thanabout 2 mm) or when welds requiring a post-weld stress relief are anticipated.

Lean Alloys

Lean austenitic alloys constitute the largestportion of all stainless steel produced. These areprincipally 201, 301, and 304. Alloys with lessthan 20% chromium and 14% nickel fall intothis unofficial category. Since they are stainless,it is generally taken for granted that these alloyswill not corrode, and these alloys have suffi-cient corrosion resistance to be used in any in-door or outdoor environment, excluding coastal.These grades are easily weldable and formableand can be given many attractive and useful sur-face finishes, so they are very much general-purpose alloys. Table 1 lists some typical com-positions of the most commonly used leanaustenitic alloys. These typical compositionsvary with end use, raw material cost factors, andthe preference of a given manufacturer. Thecompositions of standard alloys are often fine-tuned to the intended end use. In this table, theword drawing indicates higher nickel for lowerwork hardening, while tubing indicates alloyswith higher sulfur to facilitate gas tungsten arcwelding (GTAW) penetration. Tensile indicateslower alloy levels to increase the work-harden-ing rate for material that is intended to be usedin the cold-worked, high-strength condition.316L is included in its most common tubing enduse chemistry even though it is a corrosion-re-sisting alloy because it is so pervasively used asa service center sheet item.

The main difference among the leanaustenitic alloys lies in their work-hardeningrate: the leaner the alloy, the lower the austenitestability. As unstable alloys are deformed, theytransform from austenite to the much hardermartensite. This increases the work-hardeningrate and enhances ductility since it delays theonset 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 Cu201 drawing S220100 0.08 0.07 16.9 5.4 0.02 7.1 0.5 0.001 S 0.30 P 0.6 Cu201LN S20153 0.02 0.13 16.3 4.5 0.2 7.1 0.45 0.001 S 0.03 P 0.5 Cu301 tensile S30100 0.08 0.4 16.6 6.8 0.2 1.0 0.45 0.001 S 0.03 P 0.3 Cu301 drawing S30100 0.08 0.04 17.4 7.4 0.02 1.7 0.45 0.007 S 0.03 P 0.6 Cu303 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 Cu304 drawing S30400 0.05 0.04 18.4 8.6 0.3 1.8 0.45 0.001 S 0.03 P 0.3 Cu304 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 Cu304L tubing S30403 0.02 0.09 18.3 8.1 0.3 1.8 0.45 0.013 S 0.030 P 0.4 Ci305 S30500 0.05 0.02 18.8 12.1 0.2 0.8 0.60 0.001 S 0.02 P 0.2 Cu321 S32100 0.05 0.01 17.7 9.1 0.03 1.0 0.45 0.001 S 0.03 P 0.4 Ti316L S31603 0.02 0.0 16.4 10.5 2.1 1.8 0.50 0.010 S 0.03 P 0.4 Cu


deformation is more than offset by greater lo-calized strain hardening.

These grades are best viewed as a continuumwith a lower boundary at 16%Cr-6%Ni and anupper boundary at 19%Cr-12%Ni. This repre-sents the range from minimum to maximumaustenite stability. Since that is the main distinc-tion within this grade family, let us examine itsbasis.

Martensite and Austenite. Stability. Theformation of martensite at room temperaturemay be thermodynamically possible, but thedriving force for its formation may be insuffi-cient for it to form spontaneously. However,since martensite forms from unstable austeniteby a diffusionless shear mechanism, it can occurif that shear is provided mechanically by exter-nal forces. This happens during deformation,and the degree to which it occurs varies withcomposition according to (Ref 4):

Md30 (°C) = 551 – 462(%C + %N)

– 9.2(%Si) – 8.1(%Mn) – 13.7(%Cr)

– 29(%Ni + Cu) – 18.5(%Mo)

– 68(%Nb) – 1.42 (GS – 8) (Eq 1)

This is the temperature at which 50% of theaustenite transforms to martensite with 30%true strain (Ref 5). It should be noted that evenelements that are chromium equivalents in pro-moting ferrite are austenite stabilizers in thatthey impede martensite formation. This temper-ature is the common index of austenite stability.This regression analysis was generated for ho-mogeneous alloys. If alloys are inhomogeneous,

such as occurs when they are sensitized or whensolute segregation occurs, as from welding, thenthe equation applies on a microscopic scale.Sensitized zones (i.e., the regions near grainboundaries where chromium carbides have pre-cipitated) will have a much higher tendency totransform to martensite. Figures 3(a) and (b)show the changes in phase structure as a func-tion of composition over ranges that encompassthese alloys.

Martensite can be present in two differentforms. The α′-form is the bcc magnetic form,while ε is a nonmagnetic, hcp (hexagonal close-packed) version. The formation of ε versus α′ isrelated to the stacking fault energy of the alloy,which is given by (Ref 6):

Y300SF (mJ m-2) = Y0SF + 1.59Ni – 1.34Mn

+ 0.06Mn2 – 1.75Cr + 0.01Cr2

+ 15.21Mo – 5.59Si

–60.69(C + 1.2N)1/2

+ 26.27(C + 1.2N)

(Cr + Mn + Mo)1/2

+ 0.61[Ni•(Cr + Mn)]1/2 (Eq 2)

Epsilon martensite formation is favored inalloys of lower stacking fault energy. The fccstructures deform by slip between (111)planes. Viewed from these planes, the structureis a series of ABCABC atom arrangements.Slip between planes can result in anABCA/CAB structure. This so-called stackingfault generates an hcp structure. With lowerstacking fault energies, these are more readily

Fig. 3 (a) Iron-chromium phase diagram at 8% nickel; (b) iron-nickel phase diagram at 18% chromium


formed, and ε predominates. The stacking faultcan also be viewed as two partial dislocationswith the material between them faulted. Thesepartial dislocations, when generated in abun-dance, cannot readily slip past one anotherand thus pile up, increasing work-hardeningrates.

As in carbon and alloy steels, the martensitetransformation can take place simply by cool-ing, but in the lean austenitic alloys the temper-atures are well below ambient. The more stablealloys do not transform even with cryogenictreatment. Figure 4 shows the variation ofmartensite formation with temperature and truestrain for 304. Martensite formed in these alloysis quite stable and does not revert until heatedwell above the temperatures (Fig. 5) at which itwas formed. The carbon levels of austeniticstainless steels are always relatively low, sostrain-induced martensite is self-tempering andnot brittle.

Martensite has been found to form in unstableaustenite due to the electrochemically inducedsupersaturation by hydrogen (Ref 9). Underconditions of cathodic charging, superficial lay-ers were found to transform to ε under condi-tions of intense hydrostatic compression. Dur-ing subsequent outgassing, α′ was found toform due to reversals in the stress state. Marten-site thus formed is, of course, susceptible to hydrogen embrittlement.

Mechanical Properties. The tensile proper-ties in the annealed state not surprisingly relatewell to composition. The 0.2% yield strength

and tensile strength, respectively, are reported(Ref 10) to follow the equations:

(Eq 3)

In each case, d is the grain diameter in mil-limeters.

Another researcher (Ref 11) gave the rela-tionships as:

(Eq 6)

Again, d is grain diameter in millimeters, andδ is percent ferrite. The claimed accuracy forthe latter set of equations is 20 MPa and is said

TS N

Mo

= + ++ + + −

470 600 0 2

14 1 5 8 1 2

( . )

. /δ d

YS a( ) .

( .

MP N Mn Cr

Mo Cu

= + + + ++ + +120 210 0 02 2 2

14 10 6 115 0 054

7 35 0 2 1 2

−+ + + −

. )

( ( . )) /

δ δN d

TS MP C N

Si

( ) . [ (% ) (% )

. (% ) .

a = + ++ +15 4 29 35 55

2 4 0 11((% ) . (% )

. (% ) . (% ) . (% )

Ni Mo

Nb Ti Al

++ + ++

1 2

5 0 3 0 1 2

00 14 0 82 1 2. (% ) . ( )/Ferrite + −d

YS MP C N

Cr

( ) . [ . (% ) (% )

. (% ) .

a = + ++ +15 4 4 4 23 32

0 24 0 994

1 3 1 2

0 29 2 6

(% )

. (% ) . (% )

. (% ) . (% )

Mo

Si V

W Nb

+ ++ +++ +++ −

1 7 0 82

0 16

0 46

. (% ) . (% )

. (% )

. (

Ti Al

Ferrite

d 11 1 2/ / )

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

(Eq 4)

(Eq 5)


to apply to both austenitics and duplex stainlesssteels, but clearly the tensile strength relation-ship must break down for leaner alloys, such as301, in which tensile strength increases with de-creasing alloy content because of the effect ofincreasing alloying causing less transformationto martensite, which inarguably produces highertensile strengths in austenitic stainless steels.Equation 3 must also be favored over Eq 5 inthat it accounts for carbon explicitly.

One other hardening mechanism is possiblein austenitic stainless steels, and that is precipi-tation hardening. Most precipitation-hardeningstainless steels are unstable austenite, which istransformed to martensite before the precipita-tion hardening takes place. One commercialalloy, A-286, is entirely austenitic and employsthe precipitation within the austenite matrix ofNi3 (titanium, aluminum) for strengthening.This is dealt with in a separate section.

Austenitic stainless steels do not have a clearyield point but can begin to deform at as little as40% of the yield strength. As a rule of thumb,behavior at less than half the yield strength isconsidered fully elastic and stresses below two-thirds of the yield strength produce negligibleplastic deformation. This quasi-elastic behavioris a consequence of the many active slip sys-tems in the fcc structure. Even highly cold-worked material exhibits this phenomenon, al-though stress-relieving cold-worked materialwill cause dislocations to “lock in place” and

form more stable dislocation arrays that breakloose at a higher and distinct yield point.

The tensile properties of austenitic stainlesssteels with unstable austenite, that is, those withMd30 temperatures (Eq 1) near room tempera-ture, are very strain rate dependent. This is sim-ply due to the influence of adiabatic heatingduring testing increasing the stability of theaustenite. Tests run under constant temperatureconditions, either by slow strain rates or use ofheat sinks, produce lower tensile strengths.Thus, reported tensile strengths should not betaken as an absolute value but a result that canbe significantly changed by changes in testingprocedure, even with accepted norms andstandards.

Highly cold-worked austenitic stainlesssteels are often used for their robust mechanicalproperties. Few metallic materials can matchthe very high strengths they can achieve. Verylean 301 can be cold worked to yield strengthson the order of 2000 MPa because of its unsta-ble austenite transforming to martensite. Whencold worked to lower degrees, it can providevery high strength while keeping impressiveductility.

Austenitic stainless steels have exceptionaltoughness. The ambient temperature impactstrength of austenitic stainless steels is quitehigh. This is not surprising in view of their hightensile strengths and high elongations. What ismost remarkable is the absence of a transitiontemperature, which characterizes ferritic andmartensitic materials. Figure 6 shows impactstrength of the various stainless steel types ver-sus temperature. This again is due to the multi-plicity of slip systems in the fcc structure andthe fact that they do not require thermal activa-tion. This makes the austenitic stainless steels,especially the 200 series, the optimal cryogenic

Fig. 5 Reversion of martensite formed by cold work. Source:Ref 8

Fig. 6 Variation of impact strength with temperature for (a)austenitic, (b) duplex, and (c) ferritic stainless steels


material, surpassing the 9% nickel martensiticsteels in cost, toughness, and, of course, corro-sion resistance.

Precipitation of Carbides and Nitrides.Carbon is normally considered as an undesir-able impurity in austenitic stainless steel. Whileit stabilizes the austenite structure, it has a greatthermodynamic affinity for chromium. Becauseof this affinity, chromium carbides, M23C6, formwhenever carbon reaches levels of supersatura-

tion in austenite, and diffusion rates are suffi-cient for carbon and chromium to segregate intoprecipitates. The solubility of carbon in austen-ite is over 0.4% at solidification but decreasesgreatly with decreasing temperature. The solu-bility is given by (Ref 12):

(Eq 7)

The equilibrium diagram for carbon in a basic18%Cr10%Ni alloy is shown in Fig. 7. At roomtemperature, very little carbon is soluble inaustenite; even the 0.03% of L grades is mostlyin a supersaturated solution. The absence of car-bides in austenitic stainless is due to the slowdiffusion of carbon and the even slower diffu-sion of chromium in austenite. At a carbon levelof 0.06%, which is found in most 304, supersat-uration is reached below about 850 °C. Belowthis temperature, supersaturation increases ex-ponentially, while diffusion decreases exponen-tially. This results in precipitation rates that varywith temperature and carbon level as shown inFig. 8. At these temperatures, grain boundarydiffusion is much more rapid than bulk diffu-sion, and grain boundaries provide excellent nu-cleation sites, so precipitation occurs alonggrain boundaries. And, because carbon diffusesseveral orders of magnitude more rapidly thanchromium, carbon diffuses to and combineswith chromium essentially in situ, depleting thegrain boundaries of chromium in solution.

log ( )( )

C ppmK

= −77716272

T °

Fig. 7 Carbon solubility in 18–10 austenitic stainless. Source:Ref 13

Fig. 8 The precipitation rates for Cr23C6 as a function of carbon content


Figure 9 shows that the local chromium de-pletion is such that the chromium level can be-come low enough that it has not even enough tobe stainless and certainly much lower corrosionresistance than the surrounding area. This zone,because it is lower in chromium, also has veryunstable austenite and is quite prone to marten-site formation. Figure 10 shows how the locusof precipitation changes with time and tempera-ture. Carbon relatively far from grain bound-aries in the interior of grains remains in super-saturation until much longer times and muchgreater supersaturation since bulk diffusion isrequired for the nucleation and growth of theseprecipitates.

The key observation is that any solid-stateprecipitation of a chromium-rich precipitatewill necessarily cause local chromium depletionand a resulting loss of corrosion resistance.

Much longer term heat treatment is required toeliminate these depleted zones by rehomoge-nization of slowly diffusing chromium than theshort time required to form them. This is veryevident for carbides, but also true for oxides.Underneath chromium-rich oxide scales is alayer depleted in chromium and lower in corro-sion resistance. This is why not only scale fromwelding must be removed, but also the underly-ing chromium-depleted zone.

Other precipitation processes that give rise tochromium depletion are α and χ and the solid-state precipitation of oxides, nitrides, and sul-fides. Chromium precipitates that form in theliquid alloy do not cause depletion of chromiumlocally because no chromium gradients are setup around them during precipitation as diffu-sion in the liquid is very rapid. Thus, primarycarbides, oxides, and sulfides are not per seharmful to corrosion resistance. But, if the samecompounds form and grow in the solid state,chromium depletion occurs (Ref 15).

Alloying elements can have a major influenceon carbide precipitation by their influence on thesolubility of carbon in austenite. Molybdenumand nickel accelerate the precipitation by dimin-ishing the solubility of carbon. Chromium andnitrogen increase the solubility of carbon andthus retard and diminish precipitation. Nitrogenis especially useful in this regard (Fig. 11).

Increasing austenite grain size acceleratesprecipitation, as does cold work, especially inthe interior of grains, where diffusion is en-hanced by increased defect density.

Nitrogen is much more soluble than carbonand does not give rise to sensitization phenom-ena as does carbon even though Cr2N can be a

Fig. 9 Depletion of chromium from the austenite near grainboundaries due to chromium carbide precipitation.

Source: Ref 14

Fig. 10 Variation of carbide precipitation locus with time.Source: Ref 16


stable phase when the solubility limit is ex-ceeded. The solubility is over 0.15% in austen-ite, so its precipitation seldom has the possibil-ity of occurring, but it does become an issue inferritic stainless steels in this regard, for whichsolubility is much lower. Manganese andchromium increase the solubility of nitrogen inaustenite.

Stabilization. Before carbon was easily low-ered to harmless levels, it was found that addingmore powerful carbide formers than chromiumcould preclude the precipitation of chromiumcarbides. Titanium and niobium are the mostuseful elements in this regard. They form car-bides with solubility that follows the followingequation type:

(Eq 8)

For titanium carbide and niobium carbide, therespective solubilities are:

(Eq 9)

(Eq 10)

Oxides and sulfides are more energetically fa-vorable than are carbides and nitrides of thesemetals. Thus, any additions made to form car-bides must be sufficient to account for the priorformation of these compounds. Nitrogen alsocompetes with carbon for available titanium orniobium. Thus, for successful gettering of allcarbon, there must be sufficient titanium or nio-bium to combine stoichiometrically with allthese species present.

This requires in rough terms that titanium ex-ceed four times the carbon plus nitrogen, or thatniobium exceed eight times, after accountingfor the oxygen and sulfur. It would be a mistake

to ignore the titanium-consuming capacity ofoxygen and sulfur unless they have been mini-mized by refining, which can be done quitereadily.

Even if sufficient titanium or niobium is pres-ent to combine with all carbon, kinetic consid-erations may result in that not occurring. Hightemperatures, such as encountered in welding,dissociate carbides. If quenched from this state,carbon can be free to form Cr23C6 if it is re-heated to temperatures above 500 °C.

Carbon has always been considered totallyundesirable from a corrosion point of view be-cause of its tendency to form chromium car-bides. Recently, however, new processes havebeen developed to supersaturate carbon inaustenite below the temperatures at which it hassufficient mobility to form carbides. This so-called colossally supersaturated austenite re-sults in very high hardness (Fig. 12) and corro-sion resistance over limited depths. From this,however, we can see that carbon, like nitrogen,is actually beneficial to corrosion resistance insolid solution, although this is not observed atnormal concentrations. It is possible to see thatif it could be kept in solution it would be appro-priate to give it a factor of around 10 in the pit-ting resistance equivalent number (PREN)equation:

(Eq 11)

This is consistent with the similar thermody-namic interaction coefficients that carbon andnitrogen share with regard to chromium.

PREN Cr Mo

N C

= ++ +% . (% )

(% ) (% )

3 3

30 10

log [ ] [ ] .Nb C = −4 559350

T

log [ ] [ ] .Ti C = −2 976780

T

log [ ] [ ]M X A HRT

= + −

Fig. 11 Delay in carbide precipitation induced by nitrogenlevel. Source: Ref 17

Fig. 12 Variation of hardness with depth and therefore car-bon content in colossal supersaturation


High-Temperature Alloys

The austenitic stainless steels can have anexceptional combination of strength and corro-sion resistance at temperatures above 500 °C.They are often called on to resist attack byoxygen, sulfur, carburizing, nitriding, halogens,and molten salts. Austenitic stainless steels arethe most creep resistant of the stainless steels.Alloying with carbon, nitrogen, and niobiumproduces the greatest strength at elevated tem-peratures. Refer to the properties database forcomparisons among the grades. This discus-sion concentrates on their resistance to high-temperature environments, which is their salientcharacteristic.

Oxidation Resistance. Resistance to oxida-tion comes from the protective Cr2O3 scale thatforms on the surface of the material. Aboveabout 18% chromium, a continuous scale forms.The scale acts as a barrier to oxygen and greatlyslows further oxidation of metal below thescale. Below the composition at which com-plete Cr2O3 coverage occurs, the film will alsocontain the less-protective spinel FeCr2O4. TheCr2O3 scale is more protective because it betterrestricts the diffusion of oxygen to the interfacebetween the scale and the base metal, which iswhere the oxidation reaction occurs. As theoxide grows, the path to the interface lengthens,and the rate of oxidation slows. This generatesthe parabolic-type oxide growth rate that char-acterizes these alloys. The rate of oxide growthis expressed simply as:

(Eq 12)

in which the rate is in units of mass gained perunit of surface area and time. This rate is astrong function primarily of chromium level, ascan be seen in Fig. 13.

The rate increases exponentially with temper-ature since diffusion governs this phenomenon.The rate drops dramatically as chromiumreaches the concentration necessary to generatethe protective Cr2O3 layer. Above this suffi-ciency level, further increases in chromium arenot as beneficial; they mainly provide a reser-voir of chromium to re-form the Cr2O3.

As long as chromium can diffuse to the inter-face at a sufficient rate to satisfy the incomingflux of oxygen, the parabolic rate is maintained.If there is insufficient chromium flux, then theoxygen penetrates beyond the interface andforms Cr2O3 in situ. The oxide will change to

the less-protective FeCr2O4, and the scalegrowth rate will increase beyond the parabolicrelationship, leading to breakaway oxidation.The breakaway temperature increases with in-creasing chromium level.

The austenitic alloys benefit over the ferriticalloys from the presence of nickel. For a givenchromium level, oxidation rates decrease withincreasing nickel content. Figures 14, 15, and16 display this relationship. The optimal rangefor the iron base stainless steels, shown in Fig.14, is reached by the commercial alloy 310,with 25Cr-20Ni composition.

Alloying can alter the oxidation-resisting per-formance of the austenitic stainless steels. Someelements form more protective oxide layersthan Cr2O3. Aluminum and silicon are most use-ful in this regard. Aluminum forms a layer ofAl2O3 that is more restrictive to oxygen diffu-sion than is Cr2O3, as does silicon through theformation of SiO2. The alloys 302B, 153MA,and 253MA all use elevated silicon levels. Alu-minum’s strong ferrite-promoting tendency re-stricts its utility in austenitic grades, however.

While the gains from using under 3% siliconare impressive, rare earths can yield evengreater benefits from mere trace additions.153MA (UNS S30415) is a variation on 304using silicon and cerium. Cerium appears to actat the metal-scale interface such that the oxides

R kt=

Fig. 13 Variation of parabolic oxidation rate with chromiumlevel and temperature. Source: Ref 18


formed are thinner, tougher, more adherent, andmore protective, although there is no consensuson the mechanism. Figure 17 shows the im-provement quantitatively (Ref 21).

Because austenitic stainless steels have agreater thermal expansion coefficient than fer-

ritic alloys, they stress their scale more duringthermal cycling. This can fracture the scale,causing spalling and rapid subsequent oxidationof the underlying metal. This serious perform-ance flaw also is remedied by rare earths, asshown in Fig. 18 (Ref 21).

Fig. 14 Influence of nickel on oxidation of iron-chromiumalloys. Source: Ref 19

Fig. 15 Isooxidation curves. Source: Ref 20

Fig. 16 Corrosion rates for various stainless steels and nickelbase alloys. Source: Ref 20


While the mechanism by which rare earthsmake the scale tougher and more adherent arevague, their effect in making austenitic alloysmuch better at resisting high-temperature oxida-tion, especially cyclic oxidation, are undeniable.

Alloying elements can also be detrimental tooxidation resistance. Manganese, an even morepotent oxide former than chromium, forms amanganese-chromium spinel that is less protec-tive than the Cr2O3. Molybdenum and tungsten,which are refractory metals and are beneficial toaqueous corrosion resistance, form volatile, low-melting oxides (MoO3 and WO3) that promotecatastrophic oxidation (Ref 22, 23). Vanadiumalso forms an oxide, V2O5, which melts at 660°C and can also cause catastrophic oxidation.

The formation of an oxide on stainless steelshould be understood to imply de facto the de-pletion of chromium from the underlyingmetal surface. Whether the scale is formed inservice, during heat treating, or during weldingthe surface, once the oxide is removed, therewill be less chromium than the bulk alloy,often by a very significant amount, and there-fore the corrosion resistant will be less. This iswhy welds must have not only their heat tintremoved, but also the underlying metal whichis depleted in chromium, to a depth on theorder of 10 μ (Ref 24).

Microstructure can also affect oxidation re-sistance. As a generalization, it can be said thatchanges that promote the diffusion of chromiumassist in the formation of a protective scale andimprove oxidation resistance. Thus, cold workand finer grain size are positive factors via theirenhancement of chromium diffusion.

At the very highest temperatures, Cr2O3 can befurther oxidized to CrO3, which has significantvapor pressure above about 1100 °C. The com-positions of some of the main high-temperatureaustenitic alloys mentioned here are given inTable 2.

Other Environments. The most common ad-dition species that aggravates high-temperatureoxidation is water vapor. At 10%, water vaporwill increase oxidation by a factor of ten. It actsby increasing the porosity of the oxide scale andby promoting formation of the volatile CrO2

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

Fig. 18 310S (4845) versus rare earth-modified 253MA incyclic oxidation. Source: Ref 21


(OH)2 species. As a rule of thumb, maximumservice temperatures should be reduced by 50 to100 °C in the presence of steam.

Halogens can attack oxide scales and causetheir degradation or volatilization.

Carburization and nitriding are best preventedby an oxide layer that forms at very low oxygenpartial pressures as chromium and silicon con-tents are increased. Austenitic alloys have noadvantage over ferritic in this regard.

Intermetallic Phases. Transition elementsmay combine to form intermetallic phases inwhich the formula can vary from B4A to A4B.Table 3 lists the most common secondaryphases encountered in austenitic stainless steels,i.e., apart from austenite and ferrite. Sigma for-mation is retarded by nitrogen, so alloys such as153MA are less prone to it. Lower chromiumand higher nickel are beneficial. Silicon andaluminum are detrimental, as is molybdenum.

The most relevant is σ. It can contain as littleas four iron to one chromium or molybdenum ina tetragonal structure. Thus, it can exist in manyconventional austenitic alloys. Other relevantphases are χ and Laves. The greatest risk fromthese phases is the loss of room temperaturetoughness, followed by some loss of corrosionresistance.

In lean austenitic alloys used in high-temper-ature, 600 to 1000 °C service, formation timesare relatively long, on the order of 100 h ormore. In richer alloys, such as 310, times can beas short as several hours. If the temperature atwhich the alloy is to be used is one in this tem-perature range, then some σ is a foregone con-clusion, and while σ will have little detrimentaleffect on short-term properties at these tempera-tures, long-term properties such as creep, stressrupture, and especially rupture ductility are de-graded by σ. For alloys, σ is an even greaterconcern as these are prone to its formation andcan inadvertently form some during processing.If such alloys are intended for use near roomtemperature, then their toughness will be seri-ously reduced by the brittle σ, which will formfirst at triple points and then throughout grainboundaries. Because of this morphology, just afew percent intermetallic phase can causetoughness to decrease by an order of magnitude.

High-Temperature Mechanical Properties.Above about 500 °C, yield strength is less ap-propriate than creep strength in assessing theadequacy of an austenitic stainless steel forstructural purposes. The resistance of a materialto creep is generally measured by the creep rup-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 . . . 950304H S30409 0.08 0.08 18.8 8.1 . . . 1.8 0.50 . . . 820321H S32109 0.06 0.03 17.8 9.1 . . . 1.8 0.50 0.6 Ti 820153MA S30415 0.05 0.15 18.5 9.5 . . . 0.6 1.3 0.05 Ce 1000309S S30909 0.08 0.07 23.0 12.2 . . . 1.7 0.50 . . . 1040309Si DIN 1.4828 0.08 0.07 19.8 11.1 . . . 1.8 2.0 . . . 1040253MA S30815 0.08 0.17 21.0 10.5 . . . 0.6 1.5 1.0 Al, 0.05 Ce 1100310S S31008 0.05 0.03 24.6 19.2 . . . 1.6 0.60 . . . 1090353MA 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 . . . 1200332Mo 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 = 10.57–10.68 Cr16Fe5 Mo2C (e.g.)

M6C Diamond cubic a = 10.62–11.28 (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 = 8.807–8.878 Fe36Cr12 Mo10

G-phase fcc a = 11.2 Ni16Nb6 Si7, Ni16Ti6 Si7

(a) fcc, face-centered cubic; (b) body-centered cubic


rupture after 10,000 or 100,000 h. If deforma-tion is a greater concern, however, the creep de-formation strength, that is, the stress that resultsin a strain of 1% after 10,000 or 100,000 h, canbe used as a basis for design.

Cold work and precipitates tend to be ineffec-tive strengtheners at temperatures that producesolution annealing and precipitate coarsening(overaging). Thus, solid solution hardening is

preferred. Substitutional elements have limitedeffect, but interstitial solid solution elements,such as carbon and nitrogen, are quite useful.Nitrogen is the better addition in this regard,plus it has the collateral benefit of stronglyretarding intermetallic phase precipitation.Figures 19 to 22 compare mechanical propertiesof the major high-temperature austenitic alloys(Ref 25).

Fig. 19 Charpy V toughness after 200 hr aging

Fig. 20 Relative 100,000-h creep strength


Corrosion-Resistant Austenitic Alloys

Stainless steels are almost always chosen atleast in part for their corrosion resistance. Innormal atmospheric conditions, alloys withmore than 10.5% chromium do not rust.Austenitic alloys require higher levels thanthis to stabilize the austenitic structure at room

temperature, thus giving the common perceptionthat they are superior in corrosion resistance.

The main advantage austenitic alloys have istheir ability to utilize the powerful and inexpen-sive alloying element nitrogen. That is the keyaspect of the more modern austenitic stainlesssteels that have come into use in the last twodecades.

Fig. 21 100,000-h creep rupture strength

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


The ion that is most aggressive against stain-less steel is one of the most pervasive in ourenvironment. The chloride ion is found inabundance over the entire earth. It is, ofcourse, in seawater, but also in the rain, onroads, in food, and even in our bodies. Chlo-rides destabilize the passive film. If the condi-tions of chloride concentration, temperature,and acidity are sufficiently aggressive to breakdown the film, then active corrosion ensues. Ifthis is highly localized because of a localweakness in the passive film, then pitting oc-curs. Such a pit may be unstable and repassi-vate, or it may grow without limit. Otherhalides have the same effect, but they are lessubiquitous.

Because of the specific virulence of the chlo-ride ion and because of its universal presence,corrosion-resistant austenitic stainless steels alllook like they were designed to resist chloride-pitting attack. Pitting in stainless steels is inmost instances the threshold level of corrosion.Crevice corrosion is, however, more severe andusually design limiting vis-à-vis corrosion.Crevices can exist not only in deliberate jointsbut also under environmental deposits, paintfilms, weld splatter, etc. There are other envi-ronments in which the resistance follows differ-ent rules, such as oxidizing acids, bases, and or-ganic acids, but these are best regarded asexceptions.

The main factors in the resistance of austeniticalloys to pitting attack are generally given by:

PREN = %Cr + 3.3(%Mo) + 30(%N) (Eq 13)

Pitting resistance is measured by ASTM G 48(practice C) in which the lowest temperature atwhich pitting occurs in a 6% FeCl3 is measured.This is the critical pitting temperature, CPT. Therelationship between PREN and CPT is shownin Fig. 23 (Ref 26).

The function of chromium in the passive filmis intuitively clear. As the chromium content ofan alloy increases, the ready reservoir ofchromium to form the chromium-rich layer is fa-cilitated. The roles of molybdenum and nitrogenare subtler and are still subject to controversy. Itis the subject of much research, which has beensummarized in reviews. The obvious paradox ishow can elements that are not active in the pas-sive film be so effective in maintaining its in-tegrity. We do know that the essential chromiumin the matrix of stainless steels is quite reactiveand will form compounds with carbon, oxygen,sulfur, and other transition elements. When itdoes, it is no longer effective as a passive filmformer. The regions from which the chromiumdiffused to form the chromium-rich phase willbe poor in chromium unless subjected to alengthy homogenizing anneal. Most theories ofpitting 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


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 Nb254SMO 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 . . .

a homogeneous passive film, which is an impos-sible goal in reality. The chapter on corrosiondeals with this topic in more depth.

The passive layer is extremely thin comparedto oxide layers. It is on the order of 1 to 10 nmthick. Its formation does not cause chromiumdepletion beneath it, as oxide layers do. As thealloy content of chromium and molybdenum in-crease, the film is thinner, and the current den-sity required to form the film is correspondinglyreduced.

The corrosion-resistant austenitic stainlesssteel grades range from 316 to the various high-molybdenum, high-nitrogen alloys commercial-ized in the last ten years, the most notable ofwhich are listed in Table 4 with their typicalanalyses.

Early grades were based on alloying withchromium and molybdenum with sufficientnickel added to preserve the austenitic structure.Each of these elements facilitates the formationof the passive film and reduces the corrosionrate in the active state. Further experimentationshowed that molybdenum was not beneficialunder highly oxidizing conditions, but that sili-con was helpful under such conditions. Copperwas beneficial against sulfuric acid.

Alloy development came in stages. First, 317was the most significant corrosion-resistantalloy. Then, more chromium and molybdenumwere added, and the class of alloys known col-lectively as the 6%Mo alloys was commercial-ized. Allegheny’s AL-6X™ is representative ofthis group. With PRENs of around 40, these al-loys were resistant to seawater at ambient butnot at elevated temperatures. This left a greatdeal wanting in corrosion resistance. These al-loys were also very difficult to process, at leastin part because they rapidly formed brittle grainboundary σ-phases in as little as several minutesat some temperatures. This limited chromiumand molybdenum levels to a total of about 30%.

The discovery that nitrogen was beneficialagainst corrosion permitted a breakthrough inalloy development by the 1980s. Nitrogen wasincreased to around 0.20% from nominal levelsof 0.05%. This was found to increase PREN byanother five units, but more importantly, alsosuppressed σ formation to times that permittedthicker sections to be welded without embrittle-ment. Research into the thermodynamics ofnitrogen in austenite showed that manganese in-creased the solubility of nitrogen appreciably.This permitted even higher levels of total alloy-ing to be achieved. This was exploited in thealloys UNS S34565 and S32654, which contain3 to 6% manganese and about 0.50% nitrogen.The PRENs of these alloys are over 50, whichgives them a critical pitting temperature around100 °C. Table 5 lists the performance of the var-ious popular corrosion-resistant austenitic stain-less steels.

The advances are quite appreciable and madestainless steel a viable material for many appli-cations for which previously nickel base or tita-nium alloys had been required.

The obvious question in view of the successof the use of high manganese levels in conjunc-tion with high nitrogen levels in the most highlyalloyed austenitic stainless steels is when thisapproach will be used for the medium-alloyedaustenitics to make alloys superior to 316, 317,and 904 without the high levels of nickel andmolybdenum that render these alloys so expen-sive. It does not take much imagination to envi-sion alloys equal to 904L in PREN with lessnickel and molybdenum than 316L that wouldbe almost totally resistant to intermetallic phaseprecipitation and have much greater resistanceto SCC because of higher austenite stability.The same case could be made for a 317-equiva-lent alloy in corrosion resistance with less than7% nickel and 0.5% molybdenum, essentially a301 in alloy cost. In view of these trends in


alloy development, the future use of 316 and317 should be numbered. There is no justifica-tion for the use of scarce and expensive re-sources such as nickel and molybdenum whencheap, abundant replacements like manganeseand nitrogen are available. The use of 316 as astandard alloy should in the future be eroded bymore cost-effective alloys such as the lean du-plex alloys like 2003.

The same environments that cause pittingcorrosion also cause crevice corrosion. Acrevice is a volume in and out of which diffu-sion is restricted to a degree that corrosion prod-ucts accumulate and cause the contained envi-ronment to become increasingly aggressive inpH and [Cl–]. Conditions that are below thethreshold for pitting can cause crevice corro-sion. The critical temperature for crevice corro-sion is also measured in 6% FeCl3 (ASTM G-48B or D). It is the lowest temperature at whichcrevice corrosion occurs. This temperature, theCCT (critical crevice temperature), is lowerthan the CPT. GTAW a wrought alloy also low-ers the CPT to about the level of the CCT. Thereason for this lowering of resistance to local-ized attack has been thought to be related toalloy depletion caused either by dendrite coringduring solidification or chromium depletionaround inclusions. The relation to creviceswould thus seem to be that surfaces contain nu-merous flaws with respect to corrosion resist-ance, which, while not capable of sustaining pit-ting, can in a crevice dissolve and alter theenvironment sufficiently that the new harsherenvironment can generally destabilize the pas-sive film and proceed autocatalytically.

Stress corrosion cracking is the bane ofaustenitic stainless steels. SCC occurs whenthere is both a tensile stress of a sufficient mag-nitude and a sufficiently aggressive environ-ment. The threshold stress varies with alloy and

thermomechanical history. As a rule of thumb,the environment to initiate SCC must be suffi-ciently severe to cause localized corrosive at-tack. The most dangerous situation is one inwhich the expectation of pitting is marginal.

The mechanism of SCC has been a subject ofintense academic controversy for many years.Because of this, much of the literature has fo-cused on arguing a case rather than clarifyingthe phenomenon. What can be said about SCCin austenitic stainless steels with consensus?

• Risk of SCC is low at room temperature andincreases exponentially with temperature.

• SCC is preceded by localized corrosive at-tack, which has an incubation time, and thenproceeds in a discontinuous manner.

• Fracture may be transgranular, intergranular,or both. It is almost entirely lacking in plas-tic deformation with little, if any, metal loss.

• Alloying or treatments that delay localizedattack or stabilize austenite can delay SCCup to the point of virtual immunity.

• SCC is aggravated by increased chlorideconcentration and acidity, but also exists incaustic environments.

• Stress must exceed a threshold for a givenset of conditions for SCC to occur.

• Anodic or cathodic polarization may preventSCC under conditions at which it would oth-erwise be expected, or it may cause SCC inenvironments in which it would not other-wise occur.

Austenitic stainless steels are not alone in theirsusceptibility to SCC. All stainless steels sufferfrom SCC under some set of conditions of envi-ronment and material thermomechanical history.The key is to choose an alloy that is resistantunder the conditions of use. To a first approxima-tion, this means using an alloy that will not pitunder 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 –3316Ti S31635 23 15 –3317L S31703 30 25 0317LMN S31726 34 30 4904L N80904 35 40 15JS700 N08700 36 43 15254SMO S31254 46 73 384565 S34565 53 90 50654SMO S32654 63 105 75AL6-XN N08367 50 78 43AL-6XN Plus 50 min 95 60

(a) PREN, pitting resistance equivalent number.


use to be below the threshold stress completes asound design approach if residual stresses can beaccurately known. Otherwise, assuming that themetal will have residual stresses equal to 100%of the yield strength is the prudent engineeringapproach. Figure 24 shows threshold stress for anumber of alloys.

Special Corrosive Environments. Knowl-edge of the ability of the various stainless steelsto resist specific environments is essential to thedesign process. This information is extensivesince it must correlate many environments andtemperatures for many materials. Hence, referto the publications of organizations such as theNational Association of Corrosion Engineers(NACE) or to the Web sites of companies suchas Allegheny Ludlum or Outukumpu, wheresuch information is available freely. The morereputable producers will give assistance on spe-cific questions. Engineering forums on the In-ternet, such as www.eng-tips.com, should alsobe considered a resource. The following discus-sion presents just the principles of the resistanceof austenitic stainless steels to specific, morecommon environments.

Sulfuric acid is common, aggressive, and mustbe contained. Figure 25 shows the isocorrosioncurves for several alloys in pure sulfuric acid.

Alloy 20, 904L, and alloy 825 were devel-oped specifically for sulfuric acid service. Eachcontains several percent copper that, while notbeneficial against pitting, concentrates in thepassive film and diminishes general corrosion.Molybdenum and tungsten are also very benefi-cial for resistance to sulfuric acid. Phosphoricacid is similar to sulfuric acid in its effect onaustenitics but somewhat less aggressive.

Nitric acid is not particularly aggressiveagainst stainless steels. Resistance to it is pro-portional to chromium content. So, attack,should it take place, is preferably at grainboundaries, where segregation of elements suchas carbon, phosphorus, and silicon can lowerchromium locally. These elements are kept lowfor nitric acid service. Pitting is not a risk. Stan-dard usage is:

• Below 50% concentrations and below 100°C, 304L and 17% Cr ferritics are used.

• Around the 65% aziotrope, 310 is most re-sistant, especially a low carbon version, but304 NAG with low carbon, phosphorus, andsilicon is more often used.

• For 98% solutions or for lower concentra-tions that contain other stronger oxidizers,alloys with 4% Si, 18% Cr, and 15% Ni or5% Si, 17% Cr, and 17% Ni (UNS S30600and S30601) have been developed.

Hydrochloric acid, not surprisingly, is quiteaggressive against stainless steel. It is very ef-fective in destabilizing the passive film. Thus,resistance to hydrochloric acid is simply an ex-treme case of resistance to pitting in chlorideswith resistance given by Eq 12. Only the mosthighly alloyed austenitic alloys, such as AL-6XN®, should be considered and then underconditions that are tolerable, such as thoseshown in Fig. 26.

Strong bases such as NaOH and KOH are notespecially aggressive against stainless. The 17%chromium alloys can be used up to 50 °C, while304L can be used to 90 °C. As is the case withnitric acid, chromium and nickel are beneficial,while molybdenum is counterproductive. The25% 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


chromium duplex can be used to 150 °C, abovewhich temperature nickel base alloys are re-quired. High-chromium ferritic stainless steelsare also very good choices.

Organic acids are generally less aggressiveagainst stainless than are mineral acids sincethey are less dissociated in solution. They be-come hazardous when they contain chlorideions, at high temperatures, or when they disso-ciate strongly, such as with formic acid.

Because of the large number of organic com-pounds that may be considered, refer to the var-ious corrosion tables.

Surface Finish. The corrosion resistance ofaustenitic stainless steels is quite dependent onsurface condition, as are other stainless steels.Treatments that enhance surface concentrationsof beneficial elements or remove detrimentalconstituents can greatly alter performance.Oxide formation depletes surface chromium, sostrong pickling or electropolishing of thedescaled surface is especially important. Studieshave shown chromium depletion of a maximumof 6% extending 10 μ before reaching bulkchromium levels. This is equivalent to the de-pletion seen in sensitization. The increase in at-tack rate from this depletion is huge. A 1000-fold increase in weight loss in the ASTM G 48B test has been seen by a superficial loss of 6%chromium.

Likewise, surface abrasion, especially coarseabrasion, has a major detrimental effect. The120 grit #3 finish often seen on stainless reducespitting resistance by as much as the equivalentof 5 PREN, that is, equal to a reduction inchromium content of 5%. Rolled finishes aremuch preferred. The mechanism for this has notbeen clearly established; exposure of MnS in-clusions, the microcrevices abrasion produces,and residual stress have been cited as possible

contributing causes. Powder injection-moldedstainless components often have porosity that isgenerally spherical. When exposed to the sur-face, such pores act as crevices and lower thepitting potential also. All of these factors are op-erative and can act in unison.

Very fine abrasive polishing causes littleresidual stress and has very minimal crevicecreation. Thus, mirror-type polished finishes donot degrade corrosion resistance, but they donot enhance corrosion resistance as does elec-tropolished mirror finishes, which removechromium-depleted sites, which can initiatepitting.

REFERENCES

1. A.L. Schaeffler, Constitution Diagram forStainless Steel Weld Metal, Met. Prog.,Vol 56, Nov 1949, p 680–688

2. W.T. Delong, A Modified Phases Diagramfor Stainless Steel Weld Metals, Met. Prog.,Vol 77, Feb 1960, p 98

3. Design Guidelines for Selection and Use ofStainless Steel, SSINA,1998, p 3

4. K.-J. Blom, “Press Formability ofStainless Steels,” paper presented at Stain-less steels ‘77

5. F.B. Pickering, “Physical Metallurgical De-velopments in Stainless Steel,” paper pre-sented at Stainless Steel ‘84, Goteborg

6. Q.-X. Dai et al., Chin. Phys., Vol 11, 2002, p596–600, doi:10.1088/1009-1963/11/6/315

7. Aciers Inoxidables, Les Editions dePhysique, Les Ulis, Paris, 1993, p 564


9. P. Marshall, Austenitic Stainless Steels, Mi-crostructure and Mechanical Properties,Elsevier, 1984


11. H. Nordberg, Mechanical Properties ofAustenitic and Duplex Stainless Steels, In-novation in Stainless Steels ‘93 (Firenze),1993, p 2.217


13. S.J. Rosenberg and C.R. Irish, J Res. Nat.Bar. Stand., Vol 48, 1952, p 40


15. M. McGuire, “A Diffusion Model for theInfluence of Oxygen and Sulfur on the Non-Equilibrium Distribution of Chromium in

Fig. 26 Resistance to hydrochloric acid. Source: Ref 28


Austenitic Stainless Steel Welds and Slabs,”paper presented at Proceedings MS&T‘04,2004

16. R. Stickler and A. Vinckier, Trans. ASM,Vol 54, 1961, p 362

17. Aciers Inoxidables, Les Editions de Phy-sique, Les Ulis, Paris, 1993, p 568


19. Aciers Inoxidables, Les Editions de Phy-sique, Les Ulis, Paris, 1993., p 453


21. www.outukumpu.com22. W.C. Leslie, Mechanism of Rapid Oxida-

tion at High Temperature, Trans. ASM, Vol41, 1958, p 1213–1219

23. N. J. Grant, Accelerated Oxidation of Met-als at High Temperature, Trans. ASM, Vol44, 1961, p 128–137

24. J.F. Grubb, paper 04291 presented at Corro-sion 2004, NACE, 2004, p 1–15

25. ACOM Files, High Temperature StainlessSteels, www.outukumpu.com

26. J. Okamoto et al., A Super-Austenitic Stain-less Steel for Tubing and Piping Applica-tions, Nippon Steel Technical Report 90,July 2004

27. ACOM Files, High Temperature StainlessSteels, www.outukumpu.com

28. Allegheny Technologies, “AL6-XN®

Alloy”

CHAPTER 7

Duplex Stainless Steels

Summary

THE NEWEST FAMILY of stainless steels isthe duplex alloys. The mixture of ferrite andaustenite in their structure gives them higherstrength than either phase by itself. Duplex al-loys have at least 20% chromium, so they areconsidered as highly corrosion-resistant alloysbut not high-temperature alloys because of em-brittling phases. Their low nickel content makesthem more economical than austenitic alloys ofthe same level of corrosion resistance, espe-cially when their greater strength can be utilizedto reduce the amount of material required. Theyshould largely replace alloys such as 316Land317L in the future.

Introduction

Duplex stainless steels are the newest andfastest-growing alloy group in the stainless steelfamily. They are called duplex because at roomtemperature they consist of two phases, ferriteand austenite. Discovered in the 1920s, theylanguished in a suboptimized and underutilizedstate until recently. They possess excellentstrength, toughness, and corrosion resistance.They also display exceptional resistance tostress corrosion cracking (SCC) and corrosionfatigue. The leaner grades, such as 2304, corre-spond to 316L in corrosion resistance but havedouble the yield strength, while the higher alloygrades like 2507 compete with the 6% molyb-denum superaustenitics in corrosion resistancewhile still possessing much greater strength.Their limitations lie in their lack of cryogenictoughness and their inability to withstand tem-peratures much above 300 °C without formingembrittling phases. But between –100 and 300 °C

they are exceptional materials. Whether theseduplex alloys will grow to the full extent oftheir potential depends on several factors:

• Will high nickel and molybdenum prices besufficient motivation to drive designers toexplore alternatives to traditional austeniticgrades?

• Will producers overcome their inhibition toaggressively market these grades throughtheir cost-saving potential?

• Will producers perfect the techniques to pro-duce these grades reliably so that their avail-ability is unquestioned?

• Will design codes change to correctly reflectthe duplex materials’ higher ratio of yieldstrength to tensile strength?

Why are there such issues with a family of al-loys that has been successfully used for 20 years?The concept of duplex stainless steels is simple:islands of austenite in a continuous matrix ofhighly alloyed ferrite, as seen in Fig. 1. This

Fig. 1 Wrought 2205 duplex microstructure




combination in principle offers high strength be-cause of the possibility of refining the dual-phasegrain structure and thereby raising yield strengthaccording to the Hall-Petch relationship as wellas by solid solution hardening, especially withnitrogen. In addition, the absence of a continuousaustenite phase provides relief from SCC by hav-ing any propagation of cracks in austenite ar-rested by the ferrite phase.

The optimization of the alloy system had towait for two events, both related to nitrogen.First, the control of nitrogen in the refining bythe argon oxygen decarburization (AOD) processallowed the control nitrogen content up to thesolubility limit. Second, the understanding ofthe thermodynamics of the alloy system becameunderstood and reduced to a computer model.At this point, the alloys developed over the first 50 years of development became obsolete, andnew grades with higher nitrogen vastly improvedthe performance and user friendliness. Why thiswas so important can be seen by studying thestructure of these alloys.

Structure and Alloy Design

The ideal structure of a duplex grade wouldbe a stable 50-to-50 ratio of austenite to ferriteat all temperatures at which it is to be used with-out other phases. The austenite would be islands

in the ferrite matrix, and each phase would haveequal corrosion resistance despite having differ-ent compositions. It took a long time for that tobe accomplished.

Figure 2 shows a simple Fe-Cr-Ni constitu-tional diagram (Ref 1). The salient points arethat the typical successful alloys nearly bisectthe two-phase field for austenite and ferrite. It isalso obvious that the composition of the austen-ite and the ferrite must be quite different.

Ferrite contains a great deal more chromiumthan austenite; hence, its pitting corrosion re-sistance contribution from chromium is muchgreater than the resistance of the austenite be-cause in duplex grades:

PREN = %Cr + 3.3 × %Mo + 16 × %N (Eq 1)

If one were to add molybdenum to increasepitting resistance, it would preferentially parti-tion to the ferrite, further exacerbating the dif-ferential between the two phases.

This is where nitrogen saves the day. Addi-tions of nitrogen concentrate nearly entirely inthe austenite. This lowers the activity ofchromium and thereby effectively attracts morechromium to the austenite phase than wouldotherwise be present. This stabilizes the austen-ite, keeping the ratio of ferrite to austenite morenearly constant with temperature. The pittingresistance of the austenite increases signifi-

Fig. 2 The Fe-Cr-Ni phase diagrams. The shaded area results from nitrogen additions

cantly to approximately that of the ferrite. In ad-dition, the nitrogen solid solution strengthensthe austenite and retards the formation of inter-metallic phases, which is not bad for an elementthat costs nothing.

ThermoCalc, developed by the Swedish RoyalAcademy, has been an especially valuable tool inhelping us understand and design better duplexstainless steels. Without being able to computermodel the thermodynamics of the system, itwould be impossible to project the partitioning ofthe various potential alloying elements. Figure 3shows isopleth diagrams for a basic 2205 compo-sition in which nickel is varied.

The 2205 is the workhorse grade of duplex. Ithas a pitting resistance equivalent number

(PREN) of about 35 and fills a niche in corrosionresistance where austenitics and ferritics arelacking, greater than 317L stainless, PREN = 30,and below the 6% molybdenum grades, such asAL-6XN alloy, with PRENs of around 45. Fer-ritics have a gap between 442 (18Cr-2Mo) andthe super ferritics (28Cr-4Mo). By varying thechromium, nickel, and molybdenum, leaner al-loys can be devised that save cost based on re-duced molybdenum and nickel. Conversely,more corrosion-resistant alloys with higherPRENs can also be mapped, such as Fig. 4, withthe same diagrams varying nickel for a highermolybdenum level. This composition includesthe important 2507 alloy.

Partitioning of elements (Fig. 5) betweenaustenite and ferrite is an important issue. Thepartitioning tendency is a strong function oftemperature. Figure 6 shows that as temperature

Fig. 3 The iron-nickel diagram for 22% Cr, 3% Mo, 0.15% N

Chapter 7: Duplex Stainless Steels / 93

Fig. 4 The iron-nickel diagram for 25% Cr, 4% Mo, and0.25% N: N is a nitride, χ is chi, σ is sigma, α is fer-

rite, and γ is austenite

Fig. 5 Partitioning tendencies of various elements betweenferrite and austenite. Source: Ref 2

Fig. 6 Variation of partitioning ratio with temperature.Source: Ref 2


increases, the partitioning diminishes until atjust above 1300 °C it approaches unity for allnormal substitution-alloying elements (Ref 2).For nitrogen, however, the tendency is to in-creasingly segregate to austenite as temperatureincreases (Ref 2).

A danger in these alloys is that austeniteformed from ferrite on heating, such as duringwelding or annealing, will contain only the lowamount of nitrogen that was in the ferrite fromwhich it was formed, until diffusion can restoreequilibrium. If the heating time does not permitthis, this so-called secondary austenite willhave low nitrogen and therefore low pittingcorrosion resistance, as shown in Fig. 7(e). Ni-trogen alters the phase stability, making austen-ite stable to higher temperatures. This helpskeep welds from becoming excessively ferriticand disturbing the desirable 50-to-50 ratio ofaustenite to ferrite. Secondary austenite withlow nitrogen is remedied by diffusion if thephase forms at higher temperatures at whichdiffusion of nitrogen can rehomogenize the ni-trogen level.

A crucial aspect of alloy design in the duplexalloys involves the avoidance of unwantedphases. The duplex stainless steels have all thepotential problems with embrittling phases ofthe ferritic and austenitic stainless steels com-bined since they contain both as phases. Ferriteforms two main embrittling phases, α′ and σ.The α′ is generally believed to be a result ofthe miscibility gap that exists in the iron-chromium system, by which ferrite undergoesspinodal decomposition into the iron-rich α,normal ferrite, and the chromium-rich α′,which is a brittle ordered alloy. Higher levelsof chromium or the presence of copper ormolybdenum exacerbate this reaction, whichhas a formation that follows an Arrhenius-typecurve with a maximum at around 400 °C. Fig-ure 8 shows the α′ formation kinetics for fiveduplex alloys (Ref 2). While duplex gradeshave good oxidation resistance and high-tem-perature strength, the α′ problem restricts theiruse to below about 315 °C. Ferrite and austen-ite both form intermetallic phases, of whichthe most prominent and dangerous is σ, atetragonal phase richer in chromium andmolybdenum than the ferrite from which itforms. It is brittle and forms at grain bound-aries, so its precipitation has the immediate ef-fect of lowering toughness. Cold work acceler-ates the precipitation process by up to an orderof magnitude by virtue of its dual effect on nu-

cleation and diffusion. The areas around thenewly formed σ are naturally somewhat di-minished in chromium and molybdenum, sothe alloy’s resistance to localized corrosion iscompromised also.

Figure 9 shows the TTT (time-temperature-transformation) diagram for various high-alloystainless steel, including austenitic, ferritic, andduplex. Alloys of all structures, ferritic,austenitic, and duplex, with high chromium andmolybdenum encounter the σ problem fairlyequally and in proportion to their alloy content(Ref 2). This is the reason that the use of nitro-gen instead of molybdenum is so beneficial tothe leaner alloys, not just in cost, but for themajor reduction in rate of formation of sigma.Figure 10 shows the large reduction in σ forma-tion enjoyed by the lean alloy AL 2003™ mate-rial compared to the higher molybdenum 2205alloy (Allegheny Ludlum).

There are other intermetallic phases in addi-tion to σ. They include χ, R, π, and τ. These areof more research than practical interest becauseσ, with its bad consequences, forms sooner andin greater quantity under the same conditionscompared to the others.

Carbides and nitrides can also form in duplexalloys. The nitride Cr2N can form when satu-rated ferrite is quenched from a high tempera-ture, as can occur in the welding process. It ispossible that this would result in nearbychromium depletion and a decrease in corrosionresistance. Carbide formation does not as easilycause chromium depletion in duplex alloys be-cause the precipitation at the ferrite-austenitegrain boundary does not deplete the austenite asgreatly in chromium locally because of theneighboring ferrite having a much higher diffu-sivity for chromium. The point is generallymoot since all modern duplex grades containless than 0.030% carbon.

Table 1 lists the duplex grades currentlyavailable commercially. Figure 7 shows a seriesof duplex photomicrographs.

Mechanical Properties

In many ways, the duplex stainless alloysrepresent a best of both worlds in combiningtraits from the austenitic and ferritic alloys.They offer high as-annealed strength withgood toughness and ductility. Table 1 lists the major grades of duplex stainless steels;


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 propertiesfor the major duplex alloys and those of somecomparable ferritic and austenitic alloys forcomparison.

The most striking and unexpected character-istic of the duplex grades is their high yieldstrength, more than double that of comparableaustenitic grades.


Fig. 10 Delay in ( precipitation in lean duplex 2003

Fig. 8 Kinetics of ( formation

Fig. 9 Sigma formation kinetics at various alloy levels

The strength of the duplex grades is driven bythe strength of the continuous ferrite phase. Itowes its strengthening primarily to:

• Solid solution hardening by nickel, molyb-denum, chromium, copper, and manganese

• Interstitial solid solution hardening by car-bon and nitrogen

• Strengthening by grain refinement

These components have been related to themechanical properties by the following equa-tions (Ref 3):

(Eq 2)

Rp1.0 = Rp0.2 + 40 ± 9 (Eq 3)

(Eq 4)

where δ is the ferrite content in percent, d is thelamellar spacing, and results are in megapas-cals.

The influence of nitrogen is interesting in thatat lower levels (e.g., below 0.1% nitrogen)austenite is the weaker phase, but additional ni-

R

dm

/

( . )

.

= + +

+ + + −

470 600 0 02

14 1 5 8 1 2

N

Mo δ

R Np0 2 120 210 0 02

2 14 10

6 1

. .

( )

( .

= + +

+ + + ++

Mn Cr Mo Cu

55 0 54

7 35 0 02 1 2

−+ + + −

. )

( ( . )) /

δ δN d


trogen strengthens the austenite so that above0.2% nitrogen, the austenite becomes thestronger phase.

The two phases are elongated parallel to themajor strain axis of working such as from hot orcold rolling. As working increases, the mi-crostructure and properties become increasinglyanisotropic, with the austenite taking on a (110)[223] texture and the ferrite (100) [011] to (211)[011] (Ref 2).

Because the ferrite phase controls mechani-cal properties, the dependence of these proper-ties on temperature is significant since flow inbody-centered cubic (bcc) structures is ther-mally activated. Figure 11 shows the variationof yield and tensile strengths of various gradesalong with that of austenite and ferrite of simi-lar composition. Because use of these alloys

above 300 °C is not recommended, no higher-temperature properties are shown. Since theyhave a ductile-to-brittle transition, they also arenot well suited to cryogenic use.

Impact Strength. Toughness is a significantconsideration when using duplex alloys to re-place the extremely tough austenitic alloys.Duplex alloy low-temperature toughness is in-termediate to that of ferritic and austenitic al-loys. This having been said, it should be notedthat the duplex alloys can have excellent tough-ness levels, such as 100 J at –100 °C in the so-lution-annealed condition. As would be ex-pected, toughness improves with decreasinggrain size and deteriorates with cold work. Themost deleterious effect on toughness comesfrom the precipitation of intermetallic phases,such as α′ and σ, which cause a sharp decrease

Table 1 Duplex compositions

UNS Name C N Cr Ni Mo Mn Si Cu W P S

S32900 329 0.08 . . . 23.0–28.0 2.5–5.0 1.0–2.0 1.0 0.75 . . . . . . 0.040 0.030S31200 44LN 0.03 0.14–0.20 24.0–26.0 5.5–6.0 1.2–2.0 2.0 1.0 . . . . . . 0.040 0.030S31260 DP3 0.03 0.10–0.30 24.0–26.0 5.5–7.5 2.5–3.5 1.0 0.75 0.2–0.8 0.1–0.5 0.030 0.030S31500 3RE60 0.30 0.05–0.10 18.0–19.0 4.25–5.25 2.5–3.0 1.2–2.0 1.4–2.0 . . . . . . 0.030 0.030S31830 2205(old) 0.03 0.08–0.20 21.0–23.0 2.5–3.5 2.5–3.5 2.0 1.0 . . . . . . 0.030 0.020S32001 19 D 0.03 0.05–0.17 19.5–21.5 1.0–3.0 4.0–6.0 1.0 . . . . . . 0.040 0.030S32003 2003 0.03 0.14–0.20 19.5–21.0 3.0–4.0 1.5–2.0 2.0 1.0 . . . . . . 0.040 0.030S32101 2101 0.04 0.20–0.25 21.0–22.0 1.35–1.70 0.1–0.8 4.0–6.0 1.0 0.1–0.8 . . . 0.040 0.030S32205 2205 0.03 0.14–0.20 22.0–23.0 4.5–6.5 3.0–3.5 1.0 2.0 . . . . . . 0.030 0.020S32304 2304 0.03 0.05–0.20 21.5–23.5 3.0–5.0 2.5 1.0 0.05–0.6 . . . 0.040 0.040S32520 Uranus

52N+0.03 0.20–0.35 24.0–26.0 5.5–8.0 3.0–5.0 1.5 0.8 0.5–3.0 . . . 0.035 0.020

S32550 255 0.04 0.10–0.25 24.0–27.0 6.0–8.0 2.9–3.9 1.5 1.0 1.5–3.0 . . . 0.040 0.030S32750 2507 0.03 0.20–0.30 24.0–26.0 6.0–8.0 3.0–5.0 1.2 0.8 0.5 . . . 0.035 0.020S32760 Zeron 100 0.03 0.20–0.30 24.0–26.0 6.0–8.0 3.0–5.0 1.0 1.0 0.5–1.0 0.5–1.0 0.030 0.010S32906 2906 0.03 0.30–0.40 28.0–30.0 5.8–7.5 1.5–2.6 0.8–1.5 0.5 0.8 . . . 0.030 0.030S32950 7–Mo Plus 0.03 0.15–0.35 26.0–29.0 3.5–5.2 1.0–2.5 2.0 0.6 . . . . . . 0.035 0.010S39274 DP3W 0.03 0.24–0.32 24.0–26.0 6.0–8.0– 2.5–3.5 1.0 0.8 0.2–0.8 1.5–2.5 0.030 0.020S39277 AF 918 0.025 0.23–0.33 24.0–26.0 6.5–8.0 3.0–4.0 0.8 0.8 1.2–2.0 0.8–1.2 0.030 0.020

Table 2 Duplex mechanical properties

Grade Name Rp0.2 Rm A5 HB RC

Charpy-V–40 °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 . . .


in toughness level and a concurrent increase intransition temperature. The combined effect ofcold work and α′ can be seen in Fig. 12. Leanalloys such as 2001, 2003, and 2101 have amuch slower rate of formation of α′ and aremuch less at risk for loss of toughness from ex-posure in the 300 to 600 °C range, as wasshown in Fig. 11.

Fatigue. Fatigue tests on duplex stainlesssteels indicate that they possess a fatigue limitof about 50% of the yield strength when testedin air (Ref 4). The ratio of the fatigue strength ina hostile environment to that in air is a usefulmeasure of the complementary strong points ofthe duplex grades (i.e., strength and corrosionresistance). Figure 13 shows that ratio for vari-ous alloys plotted versus their PREN. As analloy’s resistance to corrosive attack increases,its fatigue limit in a given environment ap-proaches that in air, indicating simply that cor-rosion plays an increasingly small role in fa-tigue crack propagation as corrosion resistanceincreases. While this is intuitively reasonable, itis not diminished because the duplex reward theuser with a higher level of yield strength and fa-tigue strength in air, so the net useful strengthunder cyclic loading is much greater than that ofequivalent-PREN austenitic alloys.

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

Fig. 12 Increase in transition temperature with α′ formationwith aging for (a) annealed 2705 and (b) cold-

worked 2205. Source: Ref 4


Forming and Machining

The higher strength and lower ductility of theduplex grades compared to austenitics givesthem correspondingly less ability to be coldformed. Duplex alloys have sufficient ductilityto be cold drawn; they behave like ferritics oraustenitics of similar alloy level. This, however,is an alloy level at which excellent formabilityis seldom expected. Nevertheless, duplex alloyscan be cold formed like austenitic alloys. Oper-ations such as bending, drawing, and pressingcan readily be performed. Bend radii should beat least twice sheet thickness. Tubing can be ex-panded into tube sheets, but care must be takento produce tight roller-expanded joints. Tubingbend radii should be at least twice tubing out-side diameter (OD).

Heavily formed sections should be fully annealed, not just stress relieved, wheneverthere is a potential for SCC in the service environment.

Corrosion Resistance

Because duplex alloys are made up of twophases, ferrite and austenite, each must carry itsown weight in resisting corrosion. Early alloysthat were lacking in nitrogen generally had a fer-rite phase that, because of the greater partition-ing of the chromium and molybdenum to the fer-rite, had higher corrosion resistance than theaustenite. As nitrogen is added, it enriches theaustenite phase preferentially until the corrosionresistance of the austenite phase reaches that of

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


the ferrite. This approach is common to all morerecently developed alloys starting with the revi-sion of 2205 from UNS S31803 to S32205,which has primarily higher nitrogen. The net re-sult is a type of alloy that has most of the highlydesirable corrosion resistance characteristics ofsuperferritic grades without their limiting lack ofmechanical properties, mainly toughness.

General Corrosion

The duplex alloys offer important advantagesin performance over the austenitic grades in anumber of significant aggressive media, includ-ing sulfuric acid, hydrochloric acid, sodium hy-droxide, phosphoric acid, and organic acids.This performance extends to situations in which

the aggressiveness of these media is enhancedby contamination.

Sulfuric Acid. Figure 14 shows the behaviorof S32304 compared to 304 and 316 in sulfuricacid. Figure 15 shows additional, more highlyalloyed duplex grades. The use of copper as analloying element in S32550 (1.5%) and S32760(0.5%) gives them much better performancethan the otherwise similar S32750.

In real-life situations, such as seen in flue gasdesulfurization, sulfuric acid can be contami-nated with chlorides. While this contaminationis deadly to 316 and 317, it has only a minor ef-fect on the copper-alloyed duplexes (Fig. 16).

Hydrochloric Acid. Historically stainlesssteels have had their poorest performance whenconfronted by hydrochloric acid. Here again, the

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


copper/tungsten-alloyed duplexes show excep-tionally good performance, as seen in Fig. 17.This extends the usefulness of stainless steels toan environment that had previously been offlimits. Indeed, the duplex stainless steels in gen-eral can be said to be relatively indifferent to thepH of chloride solutions and are affected ratherby the chloride concentration and temperature.

Nitric Acid. It is fairly well known and ac-cepted that resistance to nitric acid, which wasone of the first uses of stainless steel, dependsalmost entirely on the chromium content.Molybdenum, in all other instances a very bene-ficial alloying element, has a strongly negativeinfluence on resistance to this highly oxidizingacid. Consequently, only the leanest-molybde-num duplex alloys, such as S32304, should be

considered for use with nitric acid, and eventhen no advantage can be claimed.

Sodium Hydroxide. Much of the older pub-lished data on the behavior of stainless steel hasseemed to promote the notion that higher nickellevels were beneficial in strong bases. Thereseems now to be little to support that notion.Figures 18 and 19 clearly indicate, respectively,that the duplex alloys with their relatively lownickel levels significantly outperform the highernickel 304L and 316L, with performance im-proving with increasing chromium content. Theadvantage is magnified when the environment iscontaminated with chlorides, as is the case ofthe white liquors of kraft digesters.

Phosphoric Acid. While pure phosphoricacid 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. 18 Corrosion rates in boiling NaOH. Source: Ref 7Fig. 19 Corrosion rates in white liquors plus chlorides.

Source: Ref 8


steel, contaminants again can render it so.Halides are particularly common and aggressivecontaminants. Figure 20 shows the substantialimprovement in performance of the duplex al-loys over 316L when contaminants are present.Performance again improves with increasingchromium, molybdenum, and nitrogen levels.

Organic Acids. Duplex alloys perform partic-ularly well in organic acids and have an excellentrecord in industrial plants. In acetic acid, 304Lhandles lower temperatures and concentrations.

Alloys such as S32205 perform well. Informic acid, the most aggressive organic acid,S32750 is resistant at all concentrations almostto the boiling point, outperforming even tita-nium (see Fig. 21).

In combinations of acetic and formic acid, thesuperiority of duplex alloys is quite evident, asseen in Fig. 22. S32750 shows virtual immu-nity, while in mixtures contaminated withhalides its performance ranks very closely toexpensive nickel-based superalloys such asN06625 and N06455. Even the lower alloyedS32205 can offer an order of magnitude im-provement over S31703 in hot contaminatedacetic acid.

Pitting Corrosion

The different analysis of the two main phasesin duplex alloys means that each has its own pit-ting resistance equivalent number, PREN. Theferrite phase has the relationship common toferritic grades:

PREN + %Cr + 3.3%Mo (Eq 5)

while the austenite obeys the more familiar:

PREN + %Cr + 3.3%Mo + 30%N (Eq 6)

The duplex grades partition these critical ele-ments in such a way that the overall PREN ofmost alloys comes out to be approximately Eq 1.

If one has the actual analysis of each phase,then the proper relationship to use is Eq 2. Theserelationships are incomplete in that they only ad-dress the major alloying elements. Tungsten hashalf the value of molybdenum and is frequentlyincluded:

PREN = %Cr + 3.3(%Mo + 0.5 × %W) + 16%N(Eq 7)

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

Fig. 21 Isocorrosion (0.1 mm/yr) performances of variousalloys. Source: Ref 9


If nonwrought material is involved, as in as-cast and welded alloys, these relationshipsgreatly overstate PREN. This is because non-equilibrium-diminished concentrations ofchromium are often found around precipitates,especially (manganese, chromium) S inclusions(Ref 11, 12) and because of lower alloy contentlocally due to solidification segregation, princi-pally of molybdenum. This is most significantin welded tubing, which can have higher sulfurlevels to increase weld penetration. Tube weldscan be reequilibrated by high-temperature an-nealing, but field girth welds will show dimin-ished corrosion resistance if unannealed. So,untreated welds can have PREN’s 5 to 15 lowerthan the parent alloy, which equates to the local-ized lowering of chromium levels. The criticalpitting temperature (CPT) of welds often de-creases to near the critical crevice corrosiontemperature (CCT) of the parent metal.

The precipitation of chromium- or molybde-num-rich second (third, in this case) phases,such as σ or α′ inevitably results in diminish-ment of these key alloying elements in the re-gion surrounding the precipitate, which willmake it more prone to localized corrosion.

This can also occur when secondary austen-ite forms during the heating of alloys to high

temperatures. This austenite, which forms fromferrite, has very little nitrogen, which clearlylowers its pitting corrosion resistance.

The duplex alloys stand up very well in com-parison to corresponding superaustenitic alloys.

Figure 23 shows how CPT varies with PREN.This ranking is not always linear, as Fig. 24shows, with pitting potential dropping fairlyrapidly with temperature and at different ratesfor different alloys. Figures 25 and 26 show theinfluence of pH and chloride concentration, re-spectively. In 3% NaCl (Fig. 26), the rankingsshow a minor variation with pH and a rationalrelationship to alloy content. The influence ofchloride concentration is strong over a widerange of concentrations.

These tests are best for judging relative per-formance of alloys and must be used cautiouslywhen extrapolating lab results to service per-formance. The degree to which short-term tests,whether potentiostatic or strictly immersion, re-flect long-term performance has not been wellestablished.

Crevice Corrosion

Crevices exist both by design and inadver-tently. Crevices are occluded volumes of liquid

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


in which oxygen and corrosion products reachlevels quite different from the exterior environ-ment and become highly corrosive. Thus, thetighter the crevice is, the greater the restrictionof diffusion between the crevice and the bulkand therefore the greater the chance of crevicecorrosion occurring. An alloy’s susceptibility tocrevice corrosion is proportional to its resist-ance to pitting corrosion under the same condi-tions. The CCT is lower than the CPT by about10 to 30 °C.

The difference increases with total alloy con-tent, as can be seen in Fig. 27. Interestingly, thedifference is approximately the same as is thedifference in CPT between the wrought alloyand the welded alloy.


Stress corrosion cracking (SCC) has longbeen the Achilles’ heel of stainless steels. Onlysoft ferritic stainless steels are immune to it. It

Fig. 24 Variation of pitting potential with temperature. SCE, saturated calomel electrode. Source: Ref 14

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


occurs at temperatures and in environmentswhere stainless would be a perfect material ifonly it did not stress corrosion crack. The arrival of duplex stainless steels has to a verylarge degree ameliorated, if not solved, thatproblem.

SCC is unfortunately poorly understood. Likepitting, whose initiation mechanism has notbeen identified, SCC has both its initiation andpropagation mechanisms still open to debate.But the duplex alloys have good strengthmainly through fine grain size and solid solutionhardening, which seems to avoid the hydrogen-trapping dislocation types that seem to be asso-ciated with hydrogen failures. So, while we can-not state the mechanism for SCC, we can mapout the conditions under which duplex alloysare susceptible to SCC. The major environmen-tal factors that affect SCC are chloride concen-tration and temperature. Figure 28 shows the remarkable advantage the duplex alloys haveover the comparable austenitic alloys with re-gard to the temperatures at which they may beused without SCC. The duplex alloys in this re-gard are governed in their behavior by their fer-rite matrix, through which cracks must propa-gate (Ref 16). Ferritic stainless steels are knownfor their resistance to SCC in the annealed con-dition. The advantage of the duplex lies in theircomposite-type microstructure with the crack-arresting austenite phase and the tougheningfine grain structure. The duplex alloys show ahigher threshold stress for SCC as a percentageof their yield strength (Fig. 29) than austeniticalloys. This is in spite of their higher yieldstrength, again giving these alloys more usablestrength.

In ferrite, SCC susceptibility is a maximumbelow 100 °C, while in austenite susceptibilityappears to begin around 50 °C and increasemonotonically with temperature. The tempera-ture at which SCC occurs at the fastest rate in-creases with nickel content. This is also char-acteristic of ferritic and martensitic materialsand mirrors their hydrogen embrittlement be-havior. H2S also accelerates failure in chlorideenvironments (Fig. 30), and cold work acceler-ates failure and lowers threshold stress values.While duplex alloys behave in many regardslike ferritic alloys in their SCC or hydrogenembrittlement response, they do not have thesame relationship between susceptibility andbulk hardness. Other ferritic and martensiticalloys display pronounced susceptibility tothese failures modes when their hardness

Fig. 25 Variation of critical pitting temperature (CPT) withpH. Source: Ref 14

Fig. 26 Critical pitting temperature (CPT) as a function of NaCl concentration. SCE, saturated calomel

electrode. Source: Ref 5

Fig. 27 Critical crevice temperature (CCT) and critical pit-ting temperature (CPT). Source: Ref 15


exceeds Rc 22. The duplex alloys have an-nealed hardness over Rc 30 without being indanger. This probably simply indicates thathardness as a measure of susceptibility is validonly insofar as it reflects a certain yieldstrength threshold as it does in temperedmartensite and is not valid for ferrite/austenitecomposite structures. Thus, it is very importantto understand duplex SCC behavior as a sepa-rate study and not interpret it in terms ofaustenitic or martensitic SCC.

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

Fig. 29 Constant load stress corrosion cracking (SCC) tests inaerated MgCl2 at 150 °C. Source: Ref 5

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


REFERENCES

1. P. Lacombe, B. Baroux, and G. Beranger,Stainless Steels, Les Editions de Physique,2003

2. R.N. Gunn, Duplex Stainless Steels, Abing-ton Publishing, 1997, p 28

3. H. Nordberg H, Innovation of StainlessSteel, Conf. Proc., AIM, Florence, 1994, p2.217–2.229

4. Charles, Duplex Stainless Steels ’91, Vol 1,Beaune, Les Editions de Physique, 1991, p3–48

5. S. Bernhardsson, Duplex Stainless Steels’91, Vol 1, Beaune, Les Editions dePhysique, 1991, p 137–150

6. J. Nichols J, 12th International CorrosionCongress, Houston, NACE, p 1237

7. E.-M. Horn, Werkstoffe und Korrosion, Vol42, 1991, p 511–519

8. J.P. Audouard, Stainless Steel Europe, April1992, p 45

9. Avesta Sheffield, Corrosion Handbook forStainless Steels, 1994

10. B. Walden et al., Stainless Steel ’93,Florence, AIM, 1993, p 3.47

11. M.F. McGuire, MS&T Conf. Proc., 2004, p831–846

12. M. Ryan, D. Williams, R. Chater, B.Hutton, and D. McPhail, Why StainlessSteel Corrodes, Nature, Vol 412, p 770

13. C.V. Roscoe et al., Duplex Stainless Steels’86, The Hague, Nederlands Instituut voorLasteckniek, 1986, p 126–135

14. J.M. Drugli et al., Paper 270 presented atCorrosion ’90, Las Vegas, NACE, 1990

15. S. Bernhardsson, Paper 164 presented atCorrosion ’90, Las Vegas, NACE, 1990

16. T. Kudo, H. Tsuge, and A. Seki, StainlessSteel ’87, The Institute of Metals, 1988, p168–175

17. R. Francis, Duplex Stainless Steels ’94, Vol3, Glasgow, TWI, 1994, paper KIV

CHAPTER 8

Ferritic Stainless Steels

Summary

THE FERRITIC STAINLESS STEELS arethe lowest-cost highly corrosion- and oxidation-resisting alloys in existence. They are usefulmainly as light-gauge sheet since their tough-ness drops off rapidly for heavier sections. Evenas they have grown in use more than any othertype of stainless, they could still economicallydisplace the popular but expensive 304 formany routine applications.

Introduction

Ferritic stainless steels are simplest, lowest-cost stainless steels. In their minimal form, theycontain simply enough chromium to overcometheir inherent level of carbon impurity and hitthe 11% chromium in solution required for“stainlessness.” Early in the 20th century, 430came into being, and the attainable levels ofcarbon removal required 16% chromium forthis to occur. So much extra chromium was re-quired because during annealing, to develop thefully ferritic structure, carbon combines withchromium, rendering it useless as a corrosionfighter. In October 1967, the first commercialuse of argon oxygen decarburization (AOD)changed the world for ferritic stainless steel.This process, in which argon and oxygen areblown through the molten metal to selectivelyremove carbon without removing chromium(described in detail elsewhere in this book), re-duced the carbon plus nitrogen levels suffi-ciently that their effect could be nearly negatedby small additions of titanium or niobium,which combine strongly with carbon and nitro-gen and effectively remove them from solution.This process is called stabilization, and the

technology was documented long before AODwas invented (Ref 1). It was not until carbonand nitrogen levels were brought down to AODlevels that it became truly practical for ferriticalloys. The level of carbon plus nitrogen waslowered from around 0.10% to around 0.04%,and less-expensive high-carbon ferrochromiumcould be used instead of expensive low-carbonversions. Thus, there exist two types of ferritics:the early high-carbon types such as 430, 434,436, and 446 and the more modern stabilized al-loys led by 409 and 439.

The older, unstabilized grades are not alwaysfully ferritic. Their carbon levels cause them toform some high-temperature austenite, whichtransforms to martensite if quenched. Thismakes their welds brittle. To be used, they arenormally in the annealed condition, which re-quires a lengthy subcritical anneal to avoidmartensite and to evenly distribute chromiumafter all carbides have stably formed. The newerstabilized alloys behave as if they are interstitialfree. They are ferritic at all temperatures (ex-cluding for the moment the possibility of extra-neous phases such as (α' and σ) and can be eas-ily welded without fear of unwanted phases.Stabilization does not preclude excessive graingrowth in the fusion or heat-affected zone(HAZ) of welds, which can render them brittle.

The mechanical properties of ferritic stainlesssteels appear similar to austenitics strengthwise,but they lack the ductility of austenitics, andthey are limited at low temperatures by brittle-ness and at high temperatures by softness.

The lower thermal expansion coefficient offerritics makes their scale more compatible withthe base alloy and provides them with a lessertendency to spall. This makes them excellentfor high-temperature applications with thermalcycles, provided their strength is adequate.




The corrosion resistance of ferritics is ham-pered by their inability to utilize nitrogen. Theabsence of nickel, which characterizes these al-loys, is not a problem since nickel adds little tocorrosion resistance. The titanium stabilizationof the modern alloys has quite a beneficial effectsince titanium is a powerful deoxidizer anddesulfurizer, both of which can cause localchromium depletion and pitting. Ferritics, more-over, are essentially free from stress-corrosioncracking (SCC) since they are below the thresh-old hardness for hydrogen embrittlement inbody-centered cubic (bcc) ferrous alloys. Thereare a few exceptions.

The main attraction of ferritic stainless steelsover austenitics is their cost. The old compari-son of 430 versus 304 is a bit unfair since 304 isricher in chromium. A fair comparison might bebetween 439 and 304. The corrosion resistanceof these two alloys is barely distinguishableunder normal ambient conditions. They are bothvery formable and weldable. The vast majorityof the objects made commercially from 304could be switched to 439 with no adverse con-sequence. But, if nickel is selling for $7 perpound, then the total cost of 304 versus 439 isdoubled by its presence. No design engineer canafford to ignore this level of incentive to learnto use ferritic stainless steels.

Ferritic Stainless Alloys

The ferritic stainless alloys generally group inlow (10.5 to 12.0%), medium (16 to 19%), andhigh (greater than 25%) chromium. They can bestabilized or not. These distinctions are some-what imposed after the fact. Rather than givingthem an order that they truly do not possess, themost significant alloys are all listed in Table 1with their compositions.

The low-chromium ferritic stainless steelsbegan with the development of MF-1, the prede-cessor of 409, in the 1960s. Its excellent corro-sion resistance, compared to carbon steel; rela-tively low cost; good welding; and formabilitypermitted it to replace aluminized carbon steeland cast iron in automotive exhaust systems,opening up what eventually became the largestsingle market for stainless steel. It was madepossible by the very low carbon plus nitrogenlevels the AOD process provided and the use ofstabilization. Thus, 409 was an improvement on405 in which aluminum performed a quasi stabi-lization, and low carbon suppressed martensite.

A similar predecessor was 410S, a low-carbonversion of 410 to which some understabilizingamount of titanium is added but that still re-quires annealing for full ferritic properties. Thekey issue of the 11% chromium ferritics is howto deal with carbon and nitrogen. The 405 and410S take the approach of minimizing it and livewith annealing. The 409 uses full titanium stabi-lization. The hidden problem with using only ti-tanium is that unless nitrogen levels are madevery low, the amount of titanium required tocombine with it can reach levels at which thefirst TiN precipitates in the molten metal. Thisslaggy precipitate agglomerates, causing castingproblems and surface defects. This gave 409 areputation as a grade unsuitable for applicationsthat required good appearance because the tita-nium streaks were difficult to avoid and greatlyhighlighted by polishing. This has largely beenovercome by better refining techniques to reducecarbon plus nitrogen to levels below 0.02% andthe use of dual stabilization by titanium and nio-bium; 468 (UNS S40930) is such an alloy.

The historical archetype of ferritic stainlesssteels was 430, which has existed since the1920s and is still widely used. Its drawbacks arelack of weldability, relatively poor corrosion re-sistance because so much of its chromium istied up as carbides, and modest formability. Thenew archetype for this medium-chromium levelis 439. With 17% chromium and single (439) ordual stabilization (468), this alloy overcomesthe problems of 430 and can readily replace 304in most applications with significant cost sav-ings. In North America, 439 is mainly used as ahigher-temperature automotive exhaust alloy,but in Europe 430Ti is used extensively in morevisually challenging applications, such as appli-ances. There, it is generally used instead of 439whenever the part can be designed to be formedfrom it.

Now, 434 and 436 are little used as their his-torical application in automotive trim finds littleplace in today’s automotive styling. A modernoffshoot of these alloys, which are basicallymolybdenum enhanced 430, is 444. This alloyhas roughly the corrosion resistance of 316L butis fully resistant to SCC in the welded or an-nealed condition. This makes it especially usefulfor applications such as hot water heaters, heatexchangers, and food- and beverage-processingequipment.

Both the nominally 11 and 18% chromiumalloys are sometimes modified to enhancetheir high-temperature strength or oxidation

Chapter 8: Ferritic Stainless Steels / 111

Tabl

e 1

Ferr

itic

sta

inle

ss c

ompo

siti

ons

Allo

yD

esig

nati

onC

NC

rN

iM

nSi

Mo

PS

Ti

Nb

Oth

er

405

S405

000.

08. .

. 11

.5–1

4.5

0.6

11

. . .

0.04

0.03

. . .

. . .

0.10

–0.3

0 A

l40

0A

K a

lloy

0.05

. . .

12.0

–13.

0. .

.1

1. .

.0.

030.

03. .

.. .

.0.

25 A

l

409

S409

000.

08. .

.10

.5–1

1.75

0.5

11

. . .

0.45

0.45

6x(C

+N

) to

0.7

5. .

.. .

.

409

S409

100.

030.

0310

.5–1

1.7

0.5

11

. . .

0.04

0.02

6x(C

+N

) to

0.5

0.17

. . .

409

S409

200.

030.

0310

.5–1

1.75

0.5

11

. . .

0.04

0.02

8x(C

+N

) to

0.1

5–0.

50. .

.. .

.

409

ultr

afor

mA

K a

lloy

0.02

0.02

10.5

–11.

70.

50.

751

. . .

0.04

0.01

8x(C

+N

). .

.. .

.

466

S409

300.

020.

0210

.5–1

1.75

0.05

11

. . .

0.04

0.01

0.8

+8x

(C+

N)

Ti+

Nb

. . .

. . .

409C

bS4

0940

0.06

. . .

10.5

–11.

70.

51

1. .

. 0.

040.

0410

xC to

0.7

5 N

b. .

.. .

.

409N

iS4

0975

0.03

0.03

10.5

–11.

70.

5–1

11

. . .

0.04

0.03

. . .

. . .

. . .

11 C

r-C

bA

K a

lloy

typi

cal

0.01

0.01

511

.35

0.2

0.25

1.3

. . .

. . .

. . .

. . .

0.35

. . .

12 S

RA

K a

lloy

typi

cal

0.02

0.01

512

. . .

. . .

. . .

. . .

. . .

. . .

0.3

0.6

1.2

Al

Alf

a I

AT

I al

loy

typi

cal

0.02

5. .

.13

. . .

0.03

50.

03. .

.. .

.. .

.0.

4. .

.3

Al

Alf

a II

AT

I al

loy

typi

cal

0.02

5. .

.13

. . .

0.03

50.

03. .

.. .

.. .

.0.

4. .

.4

Al

4724

Out

ukum

pu

typi

cal

0.08

. . .

13.5

. . .

0.7

1. .

.. .

.. .

.. .

.. .

.1

Al

429

S429

000.

12. .

.14

.0–1

6.0

0.75

11

. . .

0.04

0.03

. . .

. . .

. . .

430

S430

000.

12. .

.16

.0–1

8.0

0.75

11

. . .

0.04

0.3

. . .

. . .

. . .

430F

S430

200.

12. .

.16

.0–1

8.0

. . .

1.25

1. .

.0.

060.

15. .

.. .

.. .

.m

in43

0Se

S430

230.

12. .

. 16

.0–1

8.0

. . .

1.25

1. .

.0.

060.

06. .

.. .

.0.

15 S

e

430T

iS4

3036

0.1

0.04

16.0

–19.

51

11

. . .

0.04

0.03

0.20

+ 4

x(C

+N

) to

1.1

0. .

.0.

15 A

l

439

S430

350.

070.

0417

.0–1

9.0

0.5

11

. . .

0.04

0.03

0.20

+ 4

x(C

+N

) to

1.1

0. .

.. .

.

439L

TS4

3932

0.03

0.03

17.0

–19

0.5

11

. . .

0.04

0.03

0.20

+ 4

x(C

+N

) to

0.7

5 T

i+N

b. .

.0.

15 A

l

439

HP

439

ultr

afor

mA

TI,

AK

al

loys

0.01

0.01

17.5

0.2

0.35

0.45

. . .

0.02

0.00

10.

35. .

.. .

.

468

S468

000.

03. .

.18

.0–2

0.0

0.5

11

. . .

0.04

0.03

Ti+

Nb:

0.2

0+

4x(C

+N

) to

1.1

0. .

.. .

.

18 C

r-C

bA

K a

lloy

typi

cal

0.02

. . .

18. .

.0.

30.

45. .

.. .

.. .

.0.

250.

55. .

.

18SR

AK

allo

yty

pica

l0.

015

. . .

17.3

0.25

0.3

. . .

. . .

. . .

. . .

0.25

. . .

1.7

Al

4742

Out

ukum

puty

pica

l0.

08. .

. 18

. . .

0.7

1.3

. . .

. . .

. . .

. . .

. . .

1 Al

(con

tinue

d)


Tabl

e 1

Ferr

itic

sta

inle

ss c

ompo

siti

ons

Allo

yD

esig

nati

onC

NC

rN

iM

nSi

Mo

PS

Ti

Nb

Oth

er

434

S434

000.

12. .

.16

.0–1

8.0

. . .

11

0.75

–1.2

50.

040.

03. .

.. .

.. .

.43

6S4

3600

0.12

. . .

16.0

–18.

8. .

.1

10.

75–1

.25

0.04

0.03

. . .

Nb

+Ta

5xC

0.7

. . .

441,

4509

,43

0J1L

S441

000.

03. .

.17

.5–1

8.5

. . .

11

. . .

0.04

0.01

50.

1-0.

69x

C 0

.3–1

.0. .

.

442

S442

000.

2. .

. 18

.0–2

3.0

0.6

11

. . .

0.04

0.04

. . .

. . .

. . .

436S

AT

I al

loy

typi

cal

0.01

0.01

517

.30.

30.

20.

41.

20.

020.

001

8x(C

+N

) m

in. .

.. .

.

444,

YU

S 19

0-E

MS4

4400

0.02

50.

035

17.5

–19.

51

11

0.75

–1.2

50.

040.

03T

i+N

b: 0

.20

+4x

(C+

N)

to 0

.80

. . .

. . .

433

AT

I al

loy

typi

cal

0.01

. . .

200.

250.

30.

4. .

.0.

020.

001

. . .

10x(

C+

N)

. . .

4762

Out

ukum

puty

pica

l0.

08. .

.24

. . .

0.7

1.4

. . .

. . .

. . .

. . .

. . .

1.5

Al

453

AT

I al

loy

typi

cal

0.03

. . .

220.

30.

30.

3. .

.0.

020.

030.

02. .

.0.

60 A

l0.

1R

EM

E-B

rite

,26

-1S4

4627

0.01

0.01

525

.0–2

7.5

0.5

0.4

0.4

0.75

–1.2

50.

020.

02. .

.0.

5–0.

200.

2 C

u0.

5C

u+

Ni

Mon

itS4

4635

0.02

50.

035

24.5

–26.

03.

5–4.

51

0.75

3.5–

4.5

0.04

0.03

Ti+

Nb:

0.2

0+

4x(C

+N

) to

0.8

0. .

.. .

.Se

a-cu

reS4

4660

0.02

50.

035

25.0

–27.

01.

5–3.

51

12.

5–3.

50.

040.

03T

i+N

b: 0

.20

+4x

(C+

N)

to 0

.80

. . .

. . .

29-4

CS4

4735

0.02

5. .

.28

.0–3

0 0.

51

0.75

3.5–

4.5

0.04

0.03

Ti+

Nb:

0.2

0+

4x(C

+N

) to

0.8

0. .

.. .

.44

6S4

4600

0.2

0.25

23.0

–27.

00.

61.

51

. . .

0.04

0.03

. . .

. . .

. . .

CC

-50

Cas

t allo

y0.

5. .

.26

.0–3

04

11.

5. .

.. .

.. .

.. .

.. .

.. .

.

(con

tinu

ed)


resistance. Again, the driving force has beenthe requirements of the hot end of exhaustsystems (e.g., exhaust manifolds). Alloyingwith niobium and molybdenum adds to thehigh-temperature strength, while additions ofchromium, silicon, and aluminum increase ox-idation resistance. There exists an array ofproprietary alloys as shown in Table 1; theseare usually developed for specific automotiveneeds and employ all or some of these alloy-ing variations. The use of silicon and alu-minum decreases formability and can acc-elerate (formation, so their use involvestrade-offs.

Alloys with more than 20% chromium areused specifically for high-oxidation or corrosionresistance. Despite the relative lack of high-temperature strength, these alloys are particu-larly useful because of their high-oxidation re-sistance, which they derive from the tightadherence of their oxide scale. The close matchbetween the thermal expansion coefficient ofthe scale and the alloy prevents spallation of theoxide, which would lead to breakaway oxida-tion. This was the purpose of the earliest high-chromium ferritic stainless, 446. The perform-ance of 446 has been exceeded by lower alloyedgrades, such as the aluminum-alloyed ferritics.A prime example of the state of the art is 453,which has not only 22% chromium and 0.6%aluminum but also rare earths in trace amounts(i.e., 0.1%).

As in austenitic alloys, rare earths act as verypowerful oxide and sulfide formers that concen-trate at the metal-oxide interface and stabilize it,again preventing spallation. This type of alloysfinds use in high-temperature applications suchas planar oxide fuel cells.

The high-chromium alloys, when used forcorrosion resistance, are usually called super-ferritics. In the 1960s, E-Brite® was developed.To obtain high toughness, it was vacuum refinedto very low carbon plus nitrogen levels. It wasfollowed by the more capable 29-4®. Later, thisalloy was stabilized and became the still-popu-lar AL 29-4C®. (E-Brite now has a new life as afuel cell material based on its oxidation resist-ance and very low thermal expansion coeffi-cient.) These alloys saw success as replace-ments for 316L when SCC was a problem. Thisalloy and its close neighbor SeaCure® are usedprimarily in tubing where corrosion resistanceis most important. It was developed for weldedcondenser tubing where seawater or brackishwater is involved. It is also used in heat ex-

changers and extensively in condensing por-tions of high-efficiency residential furnaces.The lower-alloyed Seacure had a slight tough-ness advantage that permitted it to be used atwall thicknesses of 1/16 in. when AL29-4C®was too brittle. As with other ferritics, these al-loys are generally only suitably tough whenused in thin section size (i.e., less than severalmillimeters).

It is difficult to say ferritic stainless steels areunderutilized since they account for about halfthe world’s production of stainless, but there aremany applications in which more expensiveaustenitic stainless steels are used needlessly.Ferritic stainless steels are a viable alternativeto nickel-bearing austenitics when thickness is 2mm or less and drawing and bending instead ofstretch forming is permitted. There are manyapplications where the longer corrosion life oflow-chromium ferritics should economically re-place carbon steel, as they have in automotiveexhaust systems. There are no technical barriersto obtaining these savings; design engineersneed to learn how to use these alloys.

Metallurgy of Ferritic Stainless Steels

Chromium stabilizes the ferritic structure athigh temperatures. Thus, above about 11%chromium, austenite does not exist at any tem-perature in pure iron chromium alloys, as seenin Fig. 1.

However, iron-chromium alloys devoid ofcarbon are not practical, so early metallurgistssaw the diagram shown in Fig. 2 with the level

Fig. 1 Iron-chromium phase diagram from Thermocalc


of carbon at 0.20%, which represented the pu-rity level attainable in arc furnace refining. Car-bon is essentially insoluble in ferrite at ambienttemperatures, and carbides of chromium andiron will form to the extent carbon is available.Since carbon diffuses interstitially much morerapidly than chromium can substitutionally,chromium is combined in situ, especially alonggrain boundaries, which are fast-diffusionpaths. This locally depletes chromium, and thealloy is sensitized. This can be eliminated by a

sufficiently long homogenization anneal at alow enough temperature that carbon and nitro-gen have very little solubility. This is standardin the processing of unstabilized ferritic stain-less steels, such as 430. Rapid cooling of unsta-bilized alloys causes carbon and nitrogen to pre-cipitate within grains. This severely embrittlesthe material and does not avoid sensitization.This is called high-temperature embrittlementbecause it comes from putting carbon and nitro-gen into solution at a high temperature and thencausing it to precipitate in a harmful manner.

These alloys were only ferritic at room tem-perature if they were given a subcritical annealto transform austenite to ferrite. Otherwise, atroom temperature they would be ferrite plusmartensite. There are alloys that are intended touse a mixed ferrite/martensite structure, butthey are treated later as a variation from the nor-mal ferritic alloys.

The introduction of AOD refining permittedmuch lower levels of carbon, as seen in Fig. 3,opening the door for fully ferritic stainless steels.

Carbon and nitrogen added together produceabout the same effect as carbon alone. So, un-stabilized fully ferritic alloys are not feasiblebelow 20% chromium without extreme refiningtechniques, such as electron beam refining,which are not commercially viable for low-costalloys. Thus, nearly all modern ferritic alloysFig. 2 Iron-chromium phase diagram at 0.20% carbon

Fig. 3 Iron-chromium diagram at low carbon levels Source: Ref 2


are “stabilized.” This means that a strong car-bide former such as titanium or niobium isadded in sufficient quantity to combine with allthe carbon plus nitrogen, removing them fromsolution.

These reactions are simply:

Ti + C = TiC (Eq 1)

Ti + N + TiN (Eq 2)

Nb + C = NbC (Eq 3)

Nb + N = NbN (Eq 4)

Titanium is the stronger getter for carbon andnitrogen. The thermodynamic driving force forcarbide and nitride formation is given by

(Eq 5)

(Eq 6)

It must be noted that titanium has an evenhigher affinity for oxygen and sulfur than forcarbon, so that the removal of carbon from so-lution is preceded by the removal of oxygen,nitrogen, and sulfur in that order. This will beseen to have a major influence on corrosion re-sistance as the MnS inclusions generally asso-ciated with the initiation of pitting are notfound in titanium-stabilized grades of nor-mally low sulfur. In practice, the removal ofoxygen begins in the molten state with the for-mation of titanium sulfide and nitride and nextin the molten or solid state, depending on con-centrations. It is desirable to keep sulfur andnitrogen low enough that precipitation is in thesolid state so that precipitates do not agglom-erate and cause large primary inclusions thatbecome unsightly surface defects. TiCS formsin the solid state if sulfur is present; if not, TiCforms. Essentially all carbon is removed fromsolution below 1250 °C if carbon and nitrogenare kept as low as possible and a stoichiomet-ric amount of titanium is available (i.e., greaterthan about four times the carbon plus nitrogen).

The stabilization formula in various specifi-cations is more than four times the carbon plusnitrogen because experimentally it has beenfound that sometimes understabilization oc-curs. This is due to the influence of oxygen andsulfur having prior compound formation withthe titanium and less importantly that kinetic

factors prevented TiC formation. The latter ef-fect was real in early austenitic alloys, such as321, leading to knife-line corrosion attack afterwelding, but does not exist in low interstitialferritic alloys, which have much greater diffu-sion rates than austenitic alloys. But, since car-bon mobility is quite high, it is not practical toquench alloys quickly enough to prevent car-bide precipitation as is possible in austenitics(detailed in the Chapter 6, “Austenitic StainlessSteels”). Figure 4 shows the time-temperature-transformation (TTT) curve for an unstabilized430-type alloy with carbon plus nitrogen of0.08% (Ref 3).

Stabilization causes nonchromium carbides toform at high temperatures, precluding chromiumcarbide precipitation. The net effect is thatmodern stabilized ferritic alloys behave as inter-stitial free and can be mapped using the pureiron-chromium diagram shown in Fig. 1.

The rate of diffusion of carbon in ferrite isaround 100 times greater than that of carbon inaustenite. The solubility of carbon in ferrite isvastly lower than it is in austenite. Because ofthese factors, the heat treatments to avoid sensi-tization are essentially reversed. Carbon inaustenite can be retained in supersaturation forextended periods of time. This is why austeniticL grades do not sensitize even though they areslightly supersaturated. Sensitization occursat higher levels of carbon by prolonged heatingat 600 to 850 °C. In ferritics, carbon cannotbe kept in supersaturation even by the mostrapid quenching, and sensitization is alleviatedby prolonged heating in the 600 to 850 °Crange to allow chromium to equalize wherecarbide precipitation has previously made itinhomogeneous.

( )( ) .Ti N = − +157905 40

T

( )( ) .Ti C = − +77002 75

T

Fig. 4 430 time-temperature-transformation (TTT) curve. K,carbideSource: Ref 3


Ferritic alloys, like austenitic alloys, can formintermetallic phases. The most prominent is σ,which can be seen to form in higher-chromiumstainless steels (i.e., those with chromium plusmolybdenum of 20% or more). Formation of σoccurs when such alloys are held between 500and 800 °C; it is a hard, brittle tetragonal phasewith equal parts iron and chromium. Thus, itsformation causes chromium depletion of the ad-joining ferrite. Formation requires substitutionaldiffusion of chromium so is slower to form thancarbides, minutes rather than seconds. Sincecold work enhances substitutional diffusion, itaccelerates σ formation. The σ forms preferen-tially along grain boundaries for diffusion rea-sons, and this causes it to have a major embrit-tling effect. The σ may be redissolved bysolution annealing, but regaining full homo-geneity is not immediate.

Another embrittling phenomenon is the for-mation of α'. This was named 885 °F or 475 °Cembrittlement before its cause was understood.Before the nature of α' was known, it was con-fused with temper embrittlement, which occursin martensitic alloys at the same temperature.Temper embrittlement is the segregation ofphosphorus to prior austenitic grain boundariesand does not occur in fully ferritic alloys. The α'is the ordered equiatomic chromium iron phasethat forms by spinodal decomposition; it has thesame composition as σ but exists at a lower

temperature with the same structure as ferritebut with the chromium and iron atoms in an or-dered bcc matrix in which iron and chromiumoccupy sites equivalent to two interlocking sim-ple cubic matrices. Because the lattice soclosely matches that of ferrite, the precipitate iscoherent and causes hardening. The α' embrit-tlement causes an extreme loss of toughness aswell as hardening. It also causes a loss in corro-sion resistance via the chromium depletion ofthat part of the matrix that surrenders chromiumto the α'-phase.

Figures 5 and 6 show the hardening effect ofα' and the resulting loss of toughness, respec-tively (Ref 6).

Mechanical Behavior

Ferritic stainless steels are quite similar intheir mechanical behavior to carbon steel. Themain influence of chromium is to produce somesolid solution hardening. Let us review thestrengthening mechanisms of bcc iron. Pureiron is an extremely soft material with a yieldstrength well under 10,000 psi. This softness isnot seen in practice because steel is never pure.Carbon has an extremely powerful effect onhardening, as does nitrogen.

The influence of substitutional alloying ele-ments is also quite significant. According to

Fig. 5 Influence of α' formation on hardnessSource: Ref 4


Paxton (Ref 7), the misfit of solute atoms causeslattice strains proportional to the amount dis-solved and provides strengthening through thelattice friction term. This mechanism also in-creases the impact transition temperature unfa-vorably. Elements that produce a refining ofgrain size are the exception to this general rule incarbon steel, but the lack of an austenite-to-fer-rite transformation in stabilized ferritic stainlesssteels negates this benefit for them. Figure 7shows that fairly common ingredients and impu-rities have strong hardening effects (Ref 6). Man-ganese and silicon are normally deoxidizers, butin titanium-stabilized alloys, titanium takes overthe deoxidizing role so their presence can be lim-ited. Phosphorus is virtually impossible to refinefrom stainless steel, so its presence at around0.02% is normally a given unless low-phospho-rus raw materials are used as a starting point.

The worst toughness-inhibiting effects comefrom interstitial elements to grain boundaries:oxygen, carbon, and nitrogen. The effect of car-bon plus nitrogen on transition temperature isprofound, as seen in Fig. 8 (Ref 8).

Stabilizing removes the interstitial carbon andnitrogen, along with oxygen and sulfur, fromsolution. This does not produce a major soften-ing, however, because the precipitate itself has ahardening effect.

The softest ferritic stainless alloys are the409 variations made for highly formed exhaustsystem components. They contain as littlemanganese, silicon, nickel, and other substitu-tional elements as possible and have a mini-mum of carbon plus nitrogen, so that the re-sulting precipitate fraction after titaniumaddition is as low as possible. To maximizesoftness and formability, titanium and niobiumin excess of that required for stabilization mustalso be minimized as they will cause solid so-lution hardening.

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

Fig. 7 Influence of substitutional elements on hardness ofiron alloys


Stabilization

Stabilization is essential to ferritic stainlesssteels to avoid the precipitation of grain bound-ary carbides. Combined carbon plus nitrogenlevels below 100 ppm are necessary to avoidboth sensitization and embrittlement, but with-out proper heat treatment even alloys of this pu-rity can incur debilitating loss of toughness dueto carbide and nitride precipitates (Ref 9). Theselevels are not economically attainable for com-mercial alloys, so stabilization is the correct en-gineering answer.

Stabilization is generally considered as thesimple gettering of carbon and nitrogen by asuitable carbide and nitride former. It was notknown until about 1980 just what the mecha-nisms of embrittlement were in the ferriticstainless steels, however. The distinguishing ofα' from those related to interstitials and theirstabilizers (Ref 10) permitted stabilizing ele-ments to be optimized.

Titanium combines with carbon and nitrogenstoichiometrically by:

(Eq 7)

Niobium requires a greater weight percentage:

(Eq 8)

As titanium and niobium are added to alloys,their corrosion resistance is improved (Figs. 9and 10) (Ref 11). Maximum improvement incorrosion resistance levels off once full stabi-lization is reached. Excess amounts of the stabi-lizing elements have negligible effect, but tita-nium-stabilized alloys have a lower rate ofcorrosion than niobium-stabilized alloys. This isprobably due to titanium’s ability to eliminatesulfur and oxygen from solution.

Toughness improves for niobium-stabilizedalloys up through full stabilization and then be-gins to decline. This is a result of excess stabi-lizing alloy acting as a solid solution hardenerand therefore a toughness reducer. This tough-ness reduction is more pronounced with tita-nium, which is a stronger solid solution hard-ener (Figs. 11 and 12) (Ref 11).

The upshot of this understanding was the in-troduction of dual stabilization, through whichboth weld and base metal toughness and corro-sion resistance are optimized. The same studyrecommended that dual stabilization follow thefollowing formula:Nb C N= × + ×7 7 6 6. .

Ti C N= × + ×4 3 4.

Fig. 8 Influence of interstitial carbon and nitrogen on tough-ness transition temperature Source: Ref 8

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

Fig. 10 Corrosion of niobium-stabilized 29% Cr plus 4% Moalloys in ASTM A 763 Y test. Source: Ref 11


(Eq 9)

The toughness of these alloys has a broad op-timum that takes advantage of the corrosion-re-sisting benefits of titanium (Fig. 13) (AlleghenyLudlum).

Other strong carbide formers such as zirco-nium and vanadium are ineffective stabilizersbecause their mobility at the temperatures atwhich they are thermodynamically capable offorming sufficiently large percentages of car-bides and nitrides is too low to rid the matrix ofthese elements. They also have too great a ten-dency to form intermetallic compounds.

Toughness in ferritic stainless steels is amajor consideration. If ferritic alloys enjoyedthe same toughness as austenitic alloys, therewould be few instances when the use of themuch more expensive nickel-bearing gradeswould be justified. Because stabilized alloysare ferritic at all temperatures, there is noautomatic grain-refining transformation as ex-ists in carbon steel. If grains grow large fromannealing at high temperatures or welding, thenthe transition temperature increases. Sectionsize also has an effect. Stabilized ferritic stain-less steels are seldom used in thicknesses ofover several millimeters because of decreasingtoughness. Figure 14 shows how transition

( ) ( )Ti Nb C N+ ≥ × +6

Fig. 11 Charpy V-notch impact ductile to brittle transitiontemperature (DBTT) of titanium-stabilized 29%Cr

plus 4%Mo alloys test. Source: Ref 11

Fig. 12 Charpy V-notch impact ductile to brittle transitiontemperature (DBTT) of niobium-stabilized 29%Cr

plus 4%Mo alloys test. Source: Ref 11

Fig. 13 Toughness of dual-stabilized low-alloy ferritic stainless. AL 466 is recognized as S40930


temperature can increase with thickness (Ref12). This effect is due simply to stress statestransitioning from biaxial to the more embrit-tling triaxial with increasing thickness.

Texture and Anisotropy

The deformation of ferritic bcc materials ischaracterized by limited slip systems, highstacking fault energy, and lattice anisotropy.So, when ferritic stainless are deformed, dis-locations tend not to dissociate as they do inaustenitic stainless steels. The lack of dissoci-ation of dislocations encourages cross slip.This minimizes dislocation tangles and workhardening.

When ferritic stainless steels are deformed,certain crystallographic slip systems predomi-nate, so that large deformations mechanicallybring different grains via rotation into closercrystallographic alignment. This preferred de-formation along easier slip planes resultsmacroscopically in overall mechanical proper-ties varying with direction with respect to theprior deformation. Thus, ferritic stainlesssteels, like low-carbon steels, have pronouncedmechanical anisotropy. This is manifest intheir deep drawing characteristics. Heavilycold-rolled and annealed ferritic stainlesssteels draw quite well. They resist thinning.When elongated, they contract in the width di-rection while keeping virtually the same thick-ness. This same phenomenon means that theycannot be stretch formed since plain strainquickly results in fracture because of the re-sistance to deformation in the thickness direc-

tion. But, the anisotropy does result in remark-able drawing characteristics, with ferriticstainless steels with elongations in tensile testsin the mid-30% range being nearly equal toaustenitic stainless steels with over 50% elon-gation.

The measure of anisotropy is the Lankfordratio. It is expressed as:

(Eq 10)

When this expression equals 1, then a mate-rial is isotropic. As the value increases from 1,the drawability increases, as measured by thelimiting drawing ratio (LDR), the ratio of thediameter of a disk to that of the deepest cylinderinto which it can be drawn. The ferritic stainlesssteels in sheet form have LDRs of around 2.2compared to 2.0 for 304.

The good formability of ferritic stainlesssteels has some drawbacks. They are subject toridging, which is the formation of visible ridgesparallel to the direction of elongation. This is anartifact of texture in the material. A combinationof careful chemistry design and thermomechan-ical processing is required to keep it under con-trol. The approach centers on variables that in-crease stored energy from deformation topromote recrystallization over recovery duringannealing.

The ferritic stainless steels even carry for-ward some of the preferred grain orientationthat come from initial solidification whengrowth of dendrites is along preferred crystallo-graphic directions. Hot working merely reori-ents these similarly oriented grains en masse.Without phase changes or enough stored energyto provoke full recrystallization, randomness ofgrain orientation is never achieved.

Titanium-stabilized steels show more textureand recovery versus recrystallization than doniobium-stabilized alloys. This is because tita-nium carbides and nitrides form at higher tem-perature and are therefore coarser. They thuspresent less obstruction to dislocation motionthan finer niobium precipitates. Furthermore,niobium precipitates tend to dissociate to agreater degree than those of titanium. This putsniobium in solution during hot working where itcan interact with dislocations. Thus, alloys atleast partially stabilized with niobium canachieve greater recrystallization, which cantranslate to finer grain size and less anisotropy.

Rr r r

=+ +0 45 902

4

Fig. 14 Change in transition temperature with thickness for29Cr-4Mo-2Ni alloy. Source: Ref 12


Boron additions to ferritic stainless steelsresult in the formation of grain boundary car-bides, M23(C, B)6. If added to titanium-stabi-lized steels, the carbides form on preexistingTiN particles and result in coarser overall pre-cipitate arrays since finer, lower-temperatureprecipitating TiC or TiCS precipitates are atleast partially precluded. The net result iscoarser grain size and no major improvementin mechanical properties over the use of tita-nium alone (Ref 13). Additions of boron toniobium-stabilized steels does cause finer pre-cipitates and grain size than would niobiumalone (Ref 14).

High-Temperature Properties

High-temperature mechanical properties offerritic stainless steels are often important totheir successful use because their oxidation re-sistance is excellent and better than austenitics,but their high-temperature strength is lowerthan that of austenitics. This has led to consider-able development of high-temperature proper-ties, primarily for the automotive market. Re-search has determined that high-temperaturestrength and creep resistance are best served bystabilizing grain size and having niobium insolid solution. Adding titanium to niobium-sta-bilized steels stabilizes the type of carbide, es-pecially preventing the formation of the coarseM6C, whose growth decreases strength. Therelatively high insolubility of TiC causes this.Niobium is concurrently made available forhigh-temperature solid solution strengthening.

Corrosion and Oxidation Resistance

Corrosion resistance is chemistry dependentrather than structure dependent, so ferritic stain-less steels behave just as do other stainlesssteels of the same crucial alloy content.

The main alloying elements that provide re-sistance to localized corrosion, general corro-sion, and crevice corrosion are chromium,molybdenum, and nitrogen. Since nitrogen isessentially insoluble in ferrite, it cannot con-tribute to the corrosion resistance of ferriticstainless steels as it can in austenite. Other al-loying elements, such as copper and nickel, canadd to corrosion resistance in special cases, butthey are of secondary importance compared to

chromium and molybdenum. Likewise, otherelements can have a negative effect. Any ele-ment that can combine with chromium ormolybdenum can detract from corrosion resist-ance by their removal of these essential ele-ments from solution. The most notorious ofthese is carbon, whose tendency to formchromium carbides causes areas around suchcarbides to be partially depleted of chromium.However, nitrogen, oxygen, and sulfur can alsoform chromium compounds and cause localizedloss of corrosion resistance. Manganese sul-fides, for instance, are almost always seen to bethe locus of pitting corrosion (Ref 15). Morecareful examination has shown that such sul-fides grow in the solid state as chromium/man-ganese sulfides and deplete their very close sur-roundings of chromium, inviting corrosion tobegin at the inclusion-matrix interface, wherechromium levels in solution are reduced (Ref16).

Other factors that lead to loss of localized cor-rosion resistance are the formation ofchromium-rich phases such as α' and σ. Eitherof these with about 50% chromium will causeadjoining ferrite to have lower chromium levels.

Because ferrite has a non-close-packed struc-ture, diffusion rates, both substitutional and in-terstitial, are about two orders of magnitudehigher than in austenite. That means that anydeleterious chromium-depleting reaction canhappen more rapidly. Alloys cannot bequenched rapidly enough to forestall sensitiza-tion, the precipitation of chromium carbidesthat depletes grain boundary regions ofchromium. Instead, carbon must be neutralizedby stabilization, or the chromium depletionmust be removed by homogenization in long-box anneals. Note that the latter technique isalso possible in austenitics but would requireannealing for excessively long times, 102 h orso.

The ferritic stainless steels are valued fortheir resistance to SCC. Even in environmentsthat cause pitting, the normal initiation step forSCC, annealed ferritic stainless steels do notundergo SCC as long as alloying elements suchas nickel, copper, and cobalt are kept below0.5% in aggregate. Cold work sufficient to raisetheir hardness above Rc 20 to 22 can make themsusceptible to both SCC and its cousin, hydro-gen embrittlement. The more highly alloyed su-perferritic alloys are even susceptible to hydro-gen embrittlement in the annealed condition(Ref 17). As with martensitic stainless steels,


this susceptibility is a maximum near room tem-perature and declines with increasing tempera-ture, as opposed to austenitics, which see theirmaximum susceptibility above room tempera-ture. This limits these alloys’ ability to employcathodic protection safely to –0.80 Vsce, atwhich point corrosion in seawater is, if noteliminated, reduced to very low levels (Ref 18).

REFERENCES

1. F.M. Beckett and R. Franks, Trans AIME,Vol 113, 1934, p 126–143

2. Stainless Steel, Les Editions de Physiques,1992, p 483

3. Stainless Steels, Les Editions de Physique,2003

4. H.D. Newell, High Chromium Irons, Met.Prog., April 1947, p 617–626

5. P.J. Grobner, The 885 °C (475 °C) Embrit-tlement of Ferritic Stainless Steels, Metall.Trans., Vol 4, 1973, p 251–260

6. Handbook of Stainless Steels, Peckner andBernstein, McGraw Hill, 1977, p 5–9, 5–12

7. H.W. Paxton, Alloying, ASM, 1998, p 2138. H. Abo et al., Stainless Steel ‘779. J. Grubb and R. Wright, The Role of C and

N in the Brittle Fracture of Fe-26Cr, Met.Trans. A, Vol 10A, Sept 1979, p 1247–1255

10. J. Grubb, R. Wright, and P. Farrar, “Micro-mechanisms of Brittle Fracture in Titanium-Stabilized Stainless Steels,” Special Publi-cation 706, ASTM, 1980

11. J. Grubb, Stabilization of High-ChromiumFerritic Stainless Steels, Proc. Int. Conf.Stainless Steels, ISIJ, Chiba, 1991

12. M.A. Streicher, Stainless Steel ‘77, p 2713. E. El-Kashif, K. Asakura, T. Koseki, and

K. Shibata, ISIJ Int., Vol 44, 2004, p1568–1575

14. N. Fujita, K. Ohmura, E. Sato, and A. Yamamoto, Nippon Technical Report 71,Oct 1996

15. T. Suter, E. Webb, H. Bohni, and R. Alkire,Pit Initiation in I M NaCl With and WithoutMechanical Stress, J. Electrochem. Soc.,Vol 148 (No. 5), 2001, B174

16. M. Ryan, D. Williams, R. Chater, B. Hutton,and D. McPhail, Why Stainless Steel Cor-rodes, Nature, Vol 412, 2002, p 770

17. J. Grubb, “Hydrogen Embrittlement of Su-perferritic Stainless Steels,” paper presentedat 1984 ASM Int’l Conference on New De-velopments in Stainless Steel Technology,Detroit, September 1984

18. J. Grubb and J. Maurer, “Use of CathodicProtection With Superferritic StainlessSteels in Seawater,” paper presented at Cor-rosion 84, New Orleans, April 1984

CHAPTER 9

Martensitic Stainless Steels

Summary

THE SMALLEST CATEGORY of stainlesssteels in usage volume is the martensitic stain-less steels. This is mainly because these alloysare limited in corrosion resistance because ofthe necessity of keeping alloy levels low to pro-duce the martensite structure. Even so, they fillan important niche as a strong, hard, and toughalloy of fairly good corrosion resistance and asa strong, stable, high-temperature alloy.

Introduction

Nearly 100 years ago cutlery was first sold inGreat Britain with a composition of 13%chromium and 0.25% carbon. This was the firstcommercial use of stainless steel and cutlerywith the same basic analysis is still sold today.The useful alloys of martensitic stainless steelcontain from roughly 11 to 18% chromium andup to 1.0% carbon. Relatively small amounts ofnickel, molybdenum, tungsten, vanadium, andniobium are also added at times for specific pur-poses explained in this chapter. Those marten-sitic stainless steels in which elements such ascopper and titanium are added to produce addi-tional hardening through precipitation are dis-cussed in Chapter 4, “Corrosion Types.”

The designers and engineers already familiarwith martensitic carbon and alloy steels willfind nothing confusing about martensitic stain-less steels. There is no aspect of martensiticsteels that does not apply directly to stainlessmartensitic steels. The additional concerns onemust have with stainless martensite relatemainly to those that are due to the strong ferri-tizing influence of chromium. Chromiumstrongly promotes the formation of ferrite,

which restricts the temperature and compositionranges over which it is possible to obtain a fullyaustenitic structure from which to form marten-site. The presence of ferrite in a martensiticstructure is detrimental to strength, hardness,and toughness. Ferrite can appear in the as-caststructure and be formed during austenitizing ortempering. All the usual concerns inherent inany martensitic alloys are still present; temperembrittlement, retained austenite, etc.

Martensitic stainless steels are the most mar-ginally corrosion resistant of all the stainless al-loys. The requirement that they be fully austeni-tizable limits the amount of corrosion-resistingchromium and molybdenum they can contain.Much of the carbon in them detracts from the ef-fective chromium content by forming chromiumcarbides. In addition they are always susceptibleto stress corrosion cracking (SCC) when theirhardness exceeds about Rc 22. These limitationscombine to make their excellent properties us-able in only mild environments compared toother stainless steels. Their high strength andhardness for their relatively low cost ensure theirplace as a very useful engineering material.

Table 1 lists the most significant of themartensitic stainless steel alloys. The readershould be aware that some alloys which arequite similar are discussed primarily in otherchapters dealing with specifically PH stainlesssteels or primarily ferritic stainless steels. Thedistinction between martensitic stainless steelsand some other stainless alloy families is some-times vague. Nearly all the precipitation-hard-ening stainless steels are used in the martensiticstate, but their special hardening mechanism ofprecipitation within a martensitic matrix causesthem to be categorized separately somewhat ar-bitrarily. By this conventional logic, some of themartensitic alloys containing molybdenum or




Tabl

e 1

Com

posi

tion

s (w

t%)

of m

arte

nsit

ic s

tain

less

ste

els

Allo

yU

NS

For

mC

M

nS

SiC

rM

oN

iO

ther

Oth

er

403

S403

00W

roug

ht0.

15 m

ax1

0.03

0.5

11.5

-13.

5. .

.. .

.. .

.. .

.41

0S4

1000

Wro

ught

0.15

max

10.

031

11.5

-13.

5. .

.. .

.. .

.. .

.41

0SS4

1003

Wro

ught

0.03

10.

03. .

.10

.5-

. . .

1.5

0.03

. . .

12.5

N41

0S4

1008

Wro

ught

0.08

1.5

0.03

111

.5-

0.6

. . .

. . .

. . .

13.5

410C

bS4

1040

Wro

ught

0.18

max

10.

031

11.5

-. .

.. .

.0.

05-0

.30

. . .

13.5

Nb

412

S410

03W

roug

ht0.

030

max

1.5

0.03

110

.5-

. . .

1.5

0.03

0 m

ax. .

.12

.5N

414

S414

00W

roug

ht0.

15 m

ax1

0.03

111

.5-1

3.5

. . .

1.25

-2.5

0. .

.. .

.41

4 m

odS4

1425

Wro

ught

0.05

0.5-

1.0

0.00

50.

612

.0-

1.5-

4.0-

0.06

-0.1

20.

30 C

u15

27

N41

5S4

1500

Wro

ught

0.05

max

0.50

-1.0

00.

030.

611

.5-1

4.0

0.50

-1.0

03.

50-5

.50

. . .

. . .

416

S416

00W

roug

ht0.

15 m

ax1.

250.

15-0

.30

112

.0-1

4.0

. . .

. . .

. . .

. . .

416S

eS4

1623

Wro

ught

0.15

max

1.25

0.06

112

.0-

. . .

. . .

0.15

min

. . .

14Se

418

S418

00W

roug

ht0.

15-0

.20

0.5

0.03

0.5

12.0

-14.

0. .

.1.

80-2

.20

2.50

-3.5

0 W

. . .

420

S420

00W

roug

ht0.

15 m

in1

0.03

112

.0-1

4.0

. . .

. . .

. . .

. . .

4116

DIN

W

roug

ht0.

5. .

.. .

.. .

.14

.50.

65. .

.. .

.0.

15 V

1.41

16 N

omin

al42

0FS4

2020

Wro

ught

0.15

min

1.25

0.15

min

112

.0-1

4.0

0.6

. . .

. . .

. . .

420F

SeS4

2023

Wro

ught

0.15

min

1.25

0.06

112

.0-1

4.0

0.6

. . .

0.15

min

Se

. . .

422

S422

00W

roug

ht0.

20-0

.25

10.

030.

7511

.0-1

3.5

0.75

-1.2

50.

50-1

.00

0.75

-1.2

5 W

0.15

-0.3

0V42

4S4

2400

Wro

ught

0.06

max

0.50

-1.0

00.

030.

30-0

.60

12.0

-14.

00.

30-0

.70

3.50

-4.5

0. .

.. .

.42

5S4

2500

Wro

ught

0.08

-0.2

01

0.01

114

.0-1

6.0

0.30

-0.7

01.

00-2

.00

. . .

. . .

425m

od. .

.W

roug

ht0.

50-0

.55

10.

031

13.0

-14.

00.

80-1

.20

0.5

. . .

. . .

Tri

nam

et. .

.W

roug

ht0.

30 m

ax1

0.03

112

.01.

00. .

.2.

00. .

.-1

4-3

-3.0

0 C

uH

P13C

r-1

JFE

Wro

ught

0.02

50.

45. .

.. .

.13

14

. . .

. . .

Nom

inal

HP1

3Cr-

2JF

EW

roug

ht0.

025

0.45

. . .

. . .

132

5. .

.. .

.N

omin

alN

T-C

RS

Nip

pon

Wro

ught

0.03

1.45

. . .

. . .

12.7

1.4

4.5

1.5

Cu

0.04

0 N

Nom

inal

(con

tinue

d)

Chapter 9: Martensitic Stainless Steels / 125

Tabl

e 1

Com

posi

tion

s (w

t%)

of m

arte

nsit

ic s

tain

less

ste

els

Allo

yU

NS

For

mC

M

nS

SiC

rM

oN

iO

ther

Oth

er

NT-

CR

SSN

ippo

nW

roug

ht0.

022

. . .

. . .

12.3

25.

81.

5 C

u0.

015

NN

omin

alK

L-1

2Cr

JFE

Wro

ught

0.01

. . .

. . .

. . .

11. .

.2.

40.

5 C

u0.

010

NN

omin

alK

L-H

PJF

EW

roug

ht0.

01. .

.. .

.. .

.12

25.

5. .

.0.

010

N12

Cr

Nom

inal

431

S431

00W

roug

ht0.

20 m

ax1

0.03

115

.0-1

7.0

. . .

1.25

-2.5

0. .

.. .

.44

0AS4

4002

Wro

ught

0.60

-0.7

51

0.03

116

.0-1

8.0

0.75

. . .

. . .

. . .

440B

S440

03W

roug

ht0.

75-0

.95

10.

031

16.0

-18.

00.

75. .

.. .

.. .

.44

0CS4

4004

Wro

ught

0.95

-1.2

01

0.03

116

.0-1

8.0

0.75

. . .

. . .

. . .

440F

S440

20W

roug

ht0.

95-1

.20

1.25

0.10

-0.3

51

16.0

-18.

00.

40-0

.60

0.75

. . .

. . .

440F

SeS4

4023

Wro

ught

0.95

-1.2

01.

250.

031

16.0

-18.

00.

60.

750.

15 m

in S

eB

G-4

2N

omin

alW

roug

ht1.

15. .

.. .

.0.

314

.54

. . .

1.2

V. .

.A

TS-

34N

omin

alW

roug

ht1.

050.

4. .

.0.

3514

4. .

.. .

.. .

.14

-4N

omin

alW

roug

ht1.

050.

5. .

.0.

314

4. .

.. .

.. .

.C

rMo

154

CM

Nom

inal

Wro

ught

1.05

0.45

. . .

0.3

144

. . .

. . .

. . .

CPM

Nom

inal

PM1.

45. .

.. .

.. .

.14

2. .

.4.

0 V

. . .

S30V

CPM

Nom

inal

PM2.

150.

4. .

.. .

.17

0.4

. . .

5.5

V. .

.S6

0VC

PMN

omin

alPM

2.2

. . .

. . .

. . .

131

. . .

9.0

V. .

.S9

0VC

A-1

5J9

1150

Cas

t0.

15 m

ax1

0.04

1.5

11.5

-14.

00.

51

. . .

. . .

CA

15M

J911

51C

ast

0.15

max

10.

040.

6511

.5-1

4.0

0.15

-1.0

01

. . .

. . .

CA

-40

J911

53C

ast

0.20

-0.4

01

0.04

1.5

11.5

-14.

00.

51

. . .

. . .

CA

-40F

J911

54C

ast

0.20

-0.4

01

0.02

0-0.

040

1.5

11.5

-14.

00.

51

. . .

. . .

CB

-6N

J916

50C

ast

0.06

max

0.5

0.02

110

.50-

12.5

01

. . .

. . .

CB

-6M

NJ9

1540

Cas

t0.

06 m

ax1

0.03

111

.50-

14.0

00.

40-1

.00

3.50

-4.5

0. .

.. .

.C

A-2

8MV

WJ9

1422

Cas

t0.

20-0

.28

0.50

-1.0

00.

031

11.0

0-12

.50

0.90

-1.2

50.

50-1

.00

0.90

-1.2

5 W

0.20

-0.3

0 V

(con

tinu

ed)


tungsten should also be considered precipitation-hardening alloys, but they customarily are notand will not be in this work.

The ferritic alloys often have compositionsthat allow them to be partially martensitic undersome conditions. 430 (UNS S43000) and3CR12 (UNS S41003) can contain somemartensite if their heat treatment is such thataustenite is allowed to form and is followed byrapid cooling. Even 409 (UNS S409XX) canform some austenite if chromium is at the highend of its possible range and nickel and man-ganese residual levels are high. The martensiticalloys themselves can be made to be partiallyferritic by forcing their carbon contents to lowlevels as is customarily done with 410S (UNSS41003). Not understanding these alloys canlead to unexpected consequences in mechanicalproperties or corrosion performance.

Martensite Formation

Martensite as a phenomenon deserves a briefreview. Martensite forms as result of the diffu-sionless transformation of austenite. Theaustenite may be supersaturated with carbon ornitrogen, but that is not necessary for the trans-formation. The driving force for the transforma-tion is simply the much lower free energy of theferrite phase over the austenite phase, whichcan be attributed largely to large mutual repul-sion between iron atoms that possess unpairedouter electrons with the same quantum numberand magnetic polarity. This free-energy differ-ential increases with decreasing temperature. Ata certain temperature, the martensite start tem-perature Ms, the transformation occurs sponta-neously via the coordinated movement of atomsin a shearing-type mode at very high speeds ap-proaching the speed of sound in the material.The composition of the martensite is identical tothat of the parent austenite.

There is regularity to the relationship betweenthe parent austenite and the martensite.Greninger and Troiano determined that theclose-packed planes of the austenite {111} var-ied from the {011} of the martensite by only0.2°. Further, the direction of the <101 bar> ofthe austenite was only 2.7° from the <1 bar 11bar> of the martensite. These relationships de-fine the habit plane that constitutes the austenitemartensite boundary.

Martensite forms essentially independent oftime and the fraction transformed depends only

on temperature. The amount is given by theKoistinen and Marburger equation (Ref 1):

(Eq 1)

The martensite is coherent with the parentaustenite and resembles the passage of slip dis-locations through the crystal. The sum of manysuch dislocations is shear, and this can bemacroscopically visible as in Fig. 1.

The formation of martensite is essentially me-chanical (i.e., via deformation, not diffusion).

The shear and volume expansion, about 4%,which accompanies the transformation, in-volves a great deal of strain energy that must betaken into account. This is shown diagrammati-cally in Fig. 2 (Ref 2).

1− = −′V Ms Tα βexp{ ( )}

Fig. 1 Martensite platelets emerging from the surface.Source: Ref 2

Fig. 2 The martensite reaction ab contrasted to the nucleationand growth-type transformation of austenite to ferrite, ac


This energy differential between ferrite andmartensite is stored in the high-strain energymatrix. Applied strains affect the transforma-tion. Indeed, metastable austenite can readily betransformed to martensite by deformation.However, the untransformed austenite is hin-dered from transforming by the compression itreceives from the already-formed martensite.Thus, some residual austenite is commonlyfound between lathes of martensite.

At the Ms temperature, the body-centeredcubic (bcc) phase becomes preferable energeti-cally, but this temperature is too low for diffu-sion transformation, and a slight shear in theaustenite lattice causes a rearrangement of theatoms from a face-centered cubic (fcc) to a dis-torted bcc structure. The amount of distortion isproportional to the amount of carbon in the in-terstices of the structure. These interstices areconsiderably smaller in the bcc structure eventhough it is expanded from the fcc. The octahe-dral sites change from 2.86 by 3.56 A to 2.86 by2.86 A, as shown in Fig. 3 (Ref 3). The distor-tion is accommodated by accommodation fromsite to site at low carbon levels, but above about0.018% carbon this can no longer be accommo-dated and a tetragonal distortion occurs (Ref 3).The carbon is in a state of supersaturation in theas-formed martensite. When the martensite istempered, the carbon diffuses from these inter-

stitial sites and forms various carbides, leavingthe parent martensite less strained, softer, andtougher. Figure 4 shows that the large strain en-ergy in martensite varies with the carbon con-tent, and Fig. 5 (Ref 4) shows how hardnessvaries with carbon content.

Nitrogen behaves similarly to carbon in bothaustenite and martensite, but its solubility islower, and it is less significant as an alloying el-ement accordingly. Hydrogen and boron, as in-terstitials, also raise hardness.

Phase Structure

Figure 6(a) to (h) shows a series of photomi-crographs of various martensitic alloys (Ref 5).A stainless martensitic alloy should have thefollowing characteristics:

• It must have at least 10.5% chromium toqualify as stainless and even more for bettercorrosion resistance.

• It should be fully austenitic at some temper-ature.

• The temperature at which austenite forms onheating should be sufficiently high to permittempering above the temper embrittlementrange.

These criteria are somewhat challenging. Fig-ure 7(a) shows that at low-carbon (0.05%) lev-els austenite is stable up to about 12%chromium, above which some δ-ferrite tends tobe stable at all temperatures below the meltingpoint. Increasing carbon slightly expands thechromium level at which full austenitization canoccur (Fig. 7b) (Ref 3).

Fig. 3 Change in size of the octahedral interstitial site withthe change from face-centered cubic (fcc) to body-

centered cubic (bcc). Source: Ref 2Fig. 4 Strain energy of martensite dependence on carbon

content. Source: Ref 2


The interplay between chromium and carbonis further explained in Fig. 8(a) and (b), inwhich it becomes clear that for higher-chromium alloys the range over which fullaustenitization can occur is further restricted.

The variety of martensitic stainless steelswould be very limited if only chromium andcarbon were available as alloying elements, butfortunately nickel again can make an importantcontribution. Nickel greatly expands the

Fig. 5 Variation in martensite hardness with carbon content


Fig. 6 (a) Annealed 410 showing carbides within an equiaxed ferrite matrix. (b) 410 quenched and tem-pered. (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 showingsurface decarburization. (f) 440A annealed displaying primary and smaller secondary carbides in a ferrite ma-trix. (g) 440B quenched and tempered displaying both primary and secondary carbides. (h) 440C quenched andtempered displaying significant primary carbides plus finer secondary carbides in a martensite matrix


chromium levels and temperatures at whichaustenite is stable as is shown in Fig. 9 (Ref 3).

Table 2 quantifies the influences of the vari-ous possible alloying elements on the key prop-erties of martensitic stainless steels.

It can be seen that the elements that promoteaustenite, with the exception of cobalt, all de-press the Ms temperature. This puts a limit on the

amount of total alloy that can be used and in theend puts an upper limit on the ability of marten-sitic stainless steels to achieve high corrosion re-sistance. This is because as the main corrosionfighters, chromium and molybdenum, which areferritizers, are increased, so must austenitizerssuch as nickel. The coordinated increase in theseelements lowers the Ms to such a degree that the

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

1800

1600

1400

1200

1000

800

6000 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Mass, %C

Tem

pera

ture

, °C

1800

1600

1400

1200

1000

800

6000 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Mass, %C

Tem

pera

ture

, °C

LiquidLiquid

L + α

α + γ

α + γ γγ

γ + carbide

γ + carbide

α + carbideα + carbide

σ

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


alloys become stably austenitic before muchhigher corrosion resistance is obtained.

The ability to temper without austenite rever-sion is an important trait. Obviously, if trans-forming martensite to austenite during temper-ing caused subsequent untempered martensiteor other undesirable phases, this would limitone’s ability to temper at a high enough temper-ature to achieve desired toughness. This limitsthe use of nickel while encouraging the use ofelements like molybdenum.

Copper has become an important alloying el-ement in martensitic stainless steels because itgreatly improves corrosion resistance in certainenvironments without diminishing an alloy’sability to be tempered.

Thermal Processing

The main concerns with processing marten-sitic stainless steels are austenitizing, quench-ing, tempering/stress relieving, and annealing.

Austenitizing is complicated in martensiticstainless steels because many grades containcarbon at levels intended to produce carbidesfor wear resistance purposes. Since carbon solu-bility varies strongly with temperature ataustenitizing temperatures (Fig. 7a and b), con-trol of temperature is vital to have the correctbalance of carbon in solution versus carbon ascarbide since carbon in solution has such astrong influence on ferrite content, Ms, and me-chanical properties.

Austenitizing temperature also determinesaustenite grain size. This affects Ms, but moreimportantly it influences subsequent toughness.Phosphorus precipitates at prior austenite grainboundaries during tempering with a maximumeffect at 475 °C. This is the infamous temperembrittlement. Figure 10 (Ref 3) shows the sig-nificant toughness change that occurs as in-creasing austenitizing temperature increasesaustenite grain size and permits greater phos-phorus concentrations at grain boundaries.

Refining phosphorus from any chromium-containing steel is quite challenging thermody-namically, so achieving low phosphorus levelsdepends mainly on restrictions on raw materialsfor melting. Because this is difficult or costly,grain size control is the main tool for control-ling temper embrittlement.

The higher-carbon grades, those above 0.20%carbon, should be heated gradually throughstage heating to avoid cracking due to thermalstresses. Soaking at 800 °C until uniform tem-perature is achieved minimizes this risk.

Another concern during austenitizing is su-perficial carbon loss, an example of which isshown in Fig. 6(e). Heating in air to 1050 °Ccan cause surface carbon to decrease by approx-imately 0.10% per hour, resulting in muchlower surface hardness. This loss increases withbase carbon level and austenitizing temperature.Carbon or nitrogen pickup could also occur ifthe atmosphere was rich in these elements. Thecarbon potential of the furnace atmosphere mustbe controlled to avoid potentially serious prob-lems. If hydrogen atmospheres are used theFig. 9 The expansion of the range of austenite stability with

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 % ferrite per % element

–220 –210 –20 –7 –7 –6 6 5 14 18 54

Lowering of MSper % element

–475 –475 –17 0 to 10 –17 –30 –11 –21 –17 –46 . . .

Change of ACper % element

0 to 280 0 to 250 –30 to –115 0 0 –25 to –66 25 to 73 25 to 60 0 to 35 50 to 290 30 to 750


danger of embrittlement after quenching mustbe recognized. Stress relief without delay wouldbe mandatory.

The high chromium content of these alloysrenders them very deep hardening. Air harden-ing is generally sufficient. Oil quenching whichis faster may be slowed by heating the oil.Avoiding quench cracking and excessivewarpage is almost always a greater concern thandepth of hardening so air quenching is standard.

Because the quenching and the transforma-tion it causes are inevitably accompanied byresidual stresses in a brittle material, stress re-lieving should be immediate to avoid cracking.Higher-carbon grades should not even be al-lowed below room temperature before stressrelief. Pickling should never be done on as-quenched material because this could easily re-sult in hydrogen uptake and delayed crackingby hydrogen embrittlement.

Heating as-quenched material to between 150and 400 °C produces stress relieving. Besidesthe normal flow on a microscopic scale, whichwe understand as stress relieving, there is aslight growth in the number of fine cementiteparticles and a corresponding decrease in theamount of carbon in solid solution. This resultsin a slight decrease in hardness. At 400 °C, afurther precipitation of M2X and M7C3 as well

as the transformation of M3C into M7C3 can re-sult in a secondary hardening, a true precipita-tion-hardening effect. In the presence of strongcarbide-forming alloying elements such asmolybdenum, vanadium, and tungsten, the M2Xcarbide can become the more stable species andbe responsible for the secondary hardening. At500 °C, coarser M23C6 and M7C3 begin to growat grain boundaries. This is accompanied by apronounced softening. The hardening reductionwith stress relief and tempering for a 12% Cralloy is shown in Fig. 11 (Ref 6).

Separately at the 475 °C range, the previouslymentioned phosphorus segregation to prioraustenitic grain boundaries occurs. This effectbegins to disappear above 550 °C. Thus truetempering is conducted above this temperature.The microstructural changes at these tempera-tures are the above-mentioned loss of carbonfrom solid solution, carbide precipitation andcoarsening, and, of course, stress relief. The re-sult is a pronounced softening and toughening.If the material contains retained austenite, itmay decompose to ferrite and carbide with anegative effect on toughness.

The molybdenum, vanadium, and tungsten-alloyed grades will resist softening during tem-pering because of the strength of the secondaryhardening they undergo due to precipitation

Tran

sitio

n te

mpe

ratu

re, °

C+120

+80

+40

0

−40

−80

[ ]

[ ]

[ ]

[ ]

2001005020105

Austenite grain size, μm

[ ]

[ ]

P: 0.003-0.004%

P: 0.021%

P: 0.035%

P: 0.047%

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


hardening of carbides and nitrides. Nickelseems to amplify this action by its influence ondiminishing the solubility of carbon in the ma-trix. Thus, the tempering of the higher-alloymartensitic stainless steels can truly be consid-ered a precipitation-hardening reaction.

The higher-carbon, higher-chromium gradesare typically only stress relieved because the re-moval of chromium from solution by carbideformation at higher temperatures causes an un-acceptable loss of corrosion resistance.

Applications

High-Temperature Use. The basic 12 % Crmartensitic alloy has been the basis of alloyingimprovements that were done to produce betterhigh-temperature performance, especially forturbines.

The addition of vanadium and niobium, bothof which form much more stable carbides thanchromium, results in alloys that have vastly im-proved creep resistance in the 550 °C range, asshown in Fig. 12 (Ref 3).

Tool and Cutlery Alloys. A high-profile useof martensitic stainless steels is in cutlery. Hunt-ing knives, sport knives, and chefs’ cutting toolsare highly valued items and contain some of the

most sophisticated martensitic stainless alloys.While 420 is the common alloy and is quiteserviceable, much more wear- and corrosion-re-sistant alloys exist. At one time, 440C was themaximum step up from 420; however, furtheralloying with molybdenum for corrosion resist-ance and vanadium for hardness of the carbidephase has led to improvements. The wear resist-ance of a blade is largely determined by thehardness and amount of carbides while thetoughness is governed by the matrix properties.

These alloys are used at very high hardnesslevels, so cleanliness is very important to tough-ness, which measures the ability to withstandchipping in use. Electroslag remelting (ESR) orvacuum induction melting-vacuum arc remelt-ing (VIM-VAR) provides the cleanliness re-quired, while powder metallurgy is optimal forobtaining very fine carbide size and uniformity.The nominal analyses of some prominentgrades are shown in Table 3.

The martensitic alloys have a tendency to-ward centerline segregation during solidifica-tion as well as toward the formation of primarycarbides. This has produced limitations in theamount of highly wear-resistant constituentssuch as vanadium carbide (hardness Rc 75),which can be introduced into the matrix in con-ventional production. Powder metal techniques

Har

dnes

s500

450

400

350

300

250

200

150

300°C350°C400°C450°C500°C550°C600°C650°C700°C750°C

Initial hardness

750°C

700°C650°C

600°C

550°C500°C

450°C400°C

350°C300°C

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


are not subject to the same limitations as contin-uous casters and have alloys the production ofalloys with high volume content of VC. Onesuch alloy is Crucible CPM 90V with 14% Cr,9% V, 1% Mo, and 2.3% C. This alloy has equalor better toughness and corrosion resistance as440C but has ten times the wear resistance atthe same macrohardness.

Oil Country Tubular Good and Line Pipe.The need for corrosion resistance in oil produc-tion tubulars has grown as the quality of petro-leum deposits has become less optimal. Use ofstainless can eliminate for corrosion inhibitorsin H2S and CO2 environments. This has led tothe use of low-carbon martensitic stainlesssteels. Low carbon and nitrogen levels give

good toughness without tempering and mini-mize the loss of chromium to carbides, main-taining it in solution for corrosion resistance.The addition of nickel and molybdenum yieldsfull austenite and martensite transformation andimproves corrosion resistance. Table 4 lists sev-eral such alloys by JFE: the first two can bemade to meet L80 specifications and producedas seamless. JFE reports production of over100,000 tons per year of this product (Ref 7).The third alloy is a near match for the precipita-tion-hardening stainless Custom 450 (UNSS45000) (Ref 8). Like other precipitation-hard-enable steels, it shows excellent resistance toSCC at high strength levels. Figures 13 and 14show the improvements in corrosion resistance

App

lied

stre

ss, k

g/m

m2

50

45

40

35

30

25

20

16

3 10 30 100 300 1000 3000 10,000

Rupture life, h

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

1

3

2

4

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

Nominal440A 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 . . . . . . . . .CrMo154 CM Nominal Wrought 1.05 0.45 . . . 0.3 14 4 . . . . . . . . .CPM Nominal PM 1.45 . . . . . . . . . 14 2 . . . 4.0 V . . .S30VCPM Nominal PM 2.15 0.4 . . . . . . 17 0.4 . . . 5.5 V . . .S60VCPM Nominal PM 2.2 . . . . . . . . . 13 1 . . . 9.0 V . . .S90V


over carbon steel L80 oil country tubular goods(OCTG) under test conditions representative ofdifficult real-use environments (Ref 7).

The improvements in martensitic steels forthese applications are hardly more than a thor-ough revisiting of the developments of the1950s and 1960s. This does not diminish theirimportance. Carbon steels are limited in sour envi-

ronments to about Rc 22 to avoid SCC by the hy-drogen embrittlement mechanism. The stainlesscan resist this failure mode at higher strengths. Asspecifying bodies such as the American PetroleumInstitute (API) approve the use of stainless tubu-lars at higher strength levels than carbon steeltubulars are safely capable of handling, then thestrength improvement, coupled with the orders ofmagnitude improvement in corrosion resistance,will cause a great increase in their use.

Lower carbon levels permit the use of fieldwelds without tempering so that similar alloyscan be used for line pipe. These are corrosionresistant and yet meet X70 and X80 class speci-fications. These alloys are the last two in Table2. These uses of martensitic stainless steels foroil production represent possibly the greatestgrowth area for any kind of stainless steel in thefirst decade of the 21st century.

REFERENCES

1. D.R. Koistinen, R.E. Marburger, “A GeneralEquation Prescribing the Extent of theAustenite/Martensite Transformation inPure 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 dePhysique les Ulis, Paris, 1993, p 481

4. ASM Handbook Desk Edition, 1985, p 28–95. http://products.asminternational.org/mgo/6. K.J. Irvine et al., JISI, Vol 195, ISIJ Interna-

tional 1960, p 386–4057. S. Deshimaru et al., “Steels for Production,

Transportation and Storage of Energy, JFETechnical Report (No. 2), March 2004 p 55–67

8. M. Kimura et al., “High CR StainlessOCTG with High Strength and SuperiorCorrosion Resistance,” JFE Technical Re-port (No. 7), Jan 2006, p 7–13

Fig. 13 Corrosion rates of stainless versus carbon steel

Fig. 14 Corrosion rates for stainless oil country tubular goods(OCTG) alloys under severe operating conditions

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 NNominal

NT-CRSS Nippon Wrought 0.02 2 . . . . . . 12.3 2 5.8 1.5 Cu 0.015 NNominal

KL-12Cr JFE Wrought 0.01 . . . . . . . . . 11 . . . 2.4 0.5 Cu 0.010 NNominal

KL-HP JFE Wrought 0.01 . . . . . . . . . 12 2 5.5 . . . 0.010 N12Cr Nominal

CHAPTER 10

Precipitation-Hardening Stainless Steels

Summary

THE PRECIPITATION-HARDENABLE (PH)grades are a highly specialized family of stain-less steels whose existence derives from theneed for very high-strength materials with goodcorrosion resistance. The workhorse alloys arethe martensitic PH grades, which are used inmany forms. Primarily used as forgings, bar,and other hot-worked forms, they can also beobtained in cold-rolled sheet and strip, althoughnot with the flatness expected from non-PHstainless. The semiaustenitic alloys are moreamenable to production as sheet, strip, and wireand are designed for applications that requireextensive forming before hardening. The fullyaustenitic PH alloys fill a small niche wherehigh mechanical properties are required at tem-peratures above or below which the other PHgrades are found lacking, when a nonmagneticmaterial is required, or when the higher thermalexpansive coefficient of an austenitic material isdesired. In no case is corrosion resistance betterthan that of normal 304 found in PH stainlesssteels. If enhanced strength and very high corro-sion resistance are required, then the designershould look to duplex stainless steels for the op-timal material. If cost is a greater concern thancorrosion resistance or toughness, then marten-sitic stainless steels should be considered forapplications where strength and hardness overthat of annealed ferritic and austenitic stainlessis required.

The increased use of titanium alloys and ad-vanced composite materials may occur at theexpense of the stainless PH alloys and at thesame time may create some new niche applica-tions for them.

Introduction

The PH stainless steels exploit the low austen-ite stability possible in the chromium/nickelstainless steels by making the alloys so lean incomposition that they can be made to trans-form nearly entirely to martensite by thermalor mechanical treatment. This martensite canthen be further hardened by the coherent pre-cipitation of intermetallic compounds, ele-mental copper, nitrides, or even phosphides.This precipitation hardening can also be madeto occur in a fully austenitic matrix, and thisalso provides a commercial PH alloy. But, themartensitic PH grades are by far the morecommon. The border between the more highlyalloyed martensitic stainless steels, which un-dergo secondary hardening during tempering,and the PH alloys is indeed vague. Some au-thors have astutely treated them as a singlegroup. Here, we treat them separately becausethey are traditionally considered as separatealloys.

The advantage of the PH alloys over thestrictly martensitic stainless steels is that theyattain great strength with higher toughness andcorrosion resistance than can be obtainedthrough the hardening of martensite throughcarbon. In addition, they can be fabricated in arelatively soft state and then hardened with verylittle dimensional change.

The PH grades were developed at the begin-ning of World War II, with Stainless W (UNSS17600) by U.S. Steel generally acknowledgedas the first. The later-developed grades are dis-tinguished from the first by their more uniform,and therefore tougher, microstructure throughthe elimination of residual δ-ferrite and retained




austenite and by more astute alloy design andchemistry control.

The mechanism of precipitation hardening isparallel to that used to strengthen aluminum al-loys in which the precipitation of a coherent sec-ond phase from a supersaturated solid solution isproduced by an aging heat treatment. The coher-ent precipitate strains the lattice and impedes themotion of dislocations, producing strengthening.Overaging causes the precipitates to lose co-herency, and softening follows. The precipitatethat causes the hardening is normally nickel(aluminum/titanium) (Ref 1). Figure 1 shows thecompounds that can form from the precipitationof supersaturated aluminum and titanium in aniron alloy matrix.

It is also possible to produce a hardening reac-tion by the precipitation of elemental copper(Ref 2). In nitrogen-bearing alloys, a hardeningmay be produced by the precipitation of Cr2N(Ref 3). The precipitation begins with the diffu-sion of the precipitating species to sites on theexisting matrix. These enriched zones are calledGuinier-Preston (GP) zones. Close dimensionalmatchup between the precipitating species andthe parent matrix is required. The differentialshould be on the order of a percent. This allowsnot only coherency but also strain. The coherentprecipitate is a effective barrier to dislocation

movement. As time and temperature of precipi-tation increase, the zones can grow to sizes thatcannot accommodate the small size differential;coherency is lost, and with it the hardening ef-fect diminishes. The precipitation has the dualfunction of stress relieving the martensite whilefurther hardening the matrix through the precipi-tation of the coherent precipitate. The mechani-cal properties of the final microstructure dependon the initial strength of the matrix before aging,the amount of precipitate, and the coherency ofthe precipitate. The ideal microstructure for theinitial matrix is 100% martensite. To the extentthere is δ-ferrite or retained austenite, properties,especially yield strength and toughness in thetransverse direction, are compromised. Theaging temperatures can also be high enough thatreversion of martensite to austenite occurs,which also lowers subsequent tensile properties.

While the presence of persistent, large bandsof either δ-ferrite or γ-austenite is undesirable,but both also have benefits. The presence ofsome fine bands of δ-ferrite promotes easier andmore reproducible precipitation of chrome car-bides at the δ/γ interface during the “austeniteconditioning” or “trigger anneal” heat treatmentstep for semiaustenitic alloys (17-7, AM350,etc.). Although bands of stable austenite are undesirable, it is the presence of residual interlath

Alu

min

um, w

t%

4

3

2

1

0 1 2 3 4

Titanium, wt%

Cellularprecipitation

Limit of austeniteferrite region insolution treatedconditions

γ + Ni3(AlTi)

γ + Ni3Ti + Ni3(AlTi)

γ + Ni2(AlTi) + Ni3AlTi

γ + Ni3Ti

γ

γ + Ni(AlTi)γ + Ni2AlTi

γ + Ni(AlTi) + Ni3AlTi Ni(AlTi)

Ni(AlTi) + Ni2AlTi

Ni2AlTi

Ni2(AlTi) + Ni3AlTi

Ni2AlTi

Ni3Ti

σ

Fig. 1 Possible aluminum/titanium precipitates

Chapter 10: Precipitation-Hardening Stainless Steels / 139

austenite that provides the work-hardening abil-ity in many of these PH alloys. It is this workhardening that gives the PH alloys, especially thesemiaustenitic ones, their unusual combinationof high strength plus ductility and toughness inthe fully hardened state.

The complexity of PH steels comes from theprocessing involved in producing the martensiticstructure in which the precipitation will occur.The most straightforward alloys are the marten-sitic, also called the martensitic PH alloys. Thesesteels are supplied in the fully martensitic condi-tion with hardness in the low Rc 30s. This is con-fusingly called the annealed condition, or condi-tion A, even though the matrix is untemperedmartensite. After the material is fabricated, it issubjected to an aging treatment designated bythe aging temperature in Fahrenheit (e.g., H-950). These aging temperatures range from950 °F (510 °C) to 1150 °F (620 °C).

A second major group of PH grades is thesemiaustenitic. These grades in the normallyfurnished condition A are fully austenitic. Thisis accomplished by adding elements that lowerthe martensite start temperature, such as morechromium, molybdenum, and nickel. Theaustenite is more formable than martensite, andit has the possibility of superior corrosion resist-ance because of higher chromium content. Thisis balanced by the need to use either cold work,cryogenic treatment, or a destabilizing anneal tocause the matrix to become martensitic beforeits precipitation aging treatment.

Last, if the austenite is made very stable byfurther alloying additions, a precipitation reac-tion can still be made to occur by the same typeof aging treatment without martensite ever form-ing. The precipitation takes place in austeniteand therefore results in lower room temperature

strength than that of which the martensitic orsemiaustenitic alloys are capable. The austeniticPH strength is better above 750 °F.

Martensitic Precipitation-HardenableStainless Steels

The martensitic PH alloys are, as stated, fullymartensitic at room temperature. Their marten-site is a relatively soft, low-carbon (less than0.05%) martensite as opposed to the higher car-bon found in the martensitic stainless steels. Theearly alloys of this type, 17–7 PH and 17-4 PH,contained up to 10% δ-ferrite stringers, whichcaused poor through-thickness toughness. Thiswould be expected from the Schaeffler-Delongdiagram, but this is asking too much of the Scha-effler-Delong diagram, which was developed forwelds, to predict the phase composition of alloysthat have been homogenized by hot working.The inaccuracy of the diagram for more com-plex systems was overcome, and alloys were de-signed that had minimal δ-ferrite and still trans-formed entirely to martensite, if not at roomtemperature, at least at a reasonably attainablesubzero temperature. This was done first by trialand error and more recently by using thermody-namic computer models, such as ThermoCalc, topredict equilibrium phase composition. This de-velopment was very significant for making thealloy family useful as a high-strength/high-toughness material for demanding applicationsrequiring high mechanical properties and corro-sion resistance. The most advanced PH alloysare martensitic PH grades by Cartech, Custom465 and 475.

Table 1 shows the more significant of these al-loys compared on a strength basis; Table 2 shows

Table 1 Mechanical properties of martensitic precipitation-hardenable alloys

Alloy UNS Condition Yield, MPa Tensile, MPa Elongation, % HRCToughness,CVN ft-lb

Stainless W S17600 H-950 ( 510) 1240 1340 14 42 . . .17-4 PH S17400 H-925 (495) 1210 1310 14 41 4015-5 PH S15500 H-925 (495) 1210 1300 15 41 20

H1100 (595) 930 1025 17.50 34 7013-8 PH S13800 H-950 (510) 1450 1550 12 47 25

H1050 (565) 70Custom 450 S45000 H-900 (480) 1270 1350 14 42 60

H1100 (595) 460 970 23 . . . 180Custom 455 S45500 H-950 (510) 1515 1585 10 48 8

H1050 (565) 1205 1310 14 40 25Custom 465 S46500 H-950 (510) 1650 1765 11 49 13

H1000 (535) 1500 1600 13 48 28Custom 475 . . . H-975 (525) 1855 2005 5 54 . . .

H 1100 (595) 1315 1572 13 48Ferrium S53 . . . . . . 1565 1985 14–16 54 18


their nominal compositions. Figure 2 shows a se-ries of photomicrographs of PH alloys.

The main advancement metallurgically inthese alloys from the top, and earliest, in Stain-

less W to the latest in Custom 475 besides theelimination of δ-ferrite is in the volume fractionof the precipitating phase and the elimination ofretained austenite. Alloy designers found that to

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 0.317-4 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 NbCustom 455 S45500 0 0.3 0.3 12 8.5 . . . . . . 2.5 1 0.3 NbCustom 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 NbCustom 475 . . . 0 0.4 0.4 11 8 5 1.2 . . . . . . 8.0 CoFerrium 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 so-lution 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


reduce δ-ferrite they also tend to stabilizeaustenite, which also happens to reduce thetemperature at which martensite forms, MS.Last, the higher levels of molybdenum reducethe tendency to form secondary austenite duringaging. To minimize δ-ferrite requires reducingchromium or molybdenum, which also reducescorrosion resistance. As a result, as alloystrength increases, corrosion resistance is com-promised. The alloys with the greatest strengthpotential, Custom 465 and 475, barely qualifyas stainless with around 11% chromium. But,from a utility point of view, these alloys are de-signed to have maximum mechanical propertieswith adequate corrosion resistance, so this isviewed as an acceptable compromise.

The newest alloy is Ferrium S53, one of therecent alloys designed by computer-assistedthermodynamic calculations. It was designed toreplace 300M, 4340, and AerMet 100 on anequal mechanical properties basis but also pro-vide the corrosion resistance necessary to beused in aircraft components without cadmiumplating. Its composition superficially seems de-ficient in chromium to provide “stainlessness,”but the cobalt level raises the thermodynamicactivity of chromium sufficiently that the equiv-alent of 12% chromium in a non-cobalt-contain-ing alloy is achieved. The precipitation harden-ing mechanism is the precipitation of Mo2C. Ithas been established that this hardening mecha-nism optimizes resistance to stress corrosioncracking (SCC) for a given strength level.

The alloying characteristics of these gradesare:

• Low carbon, nitrogen, silicon, and man-ganese because these elements lower MSwithout contributing to age hardening

• Low chromium to suppress δ-ferrite• Sufficient nickel to suppress δ-ferrite and

provide for precipitates without excessivelydepressing MS

• Molybdenum to offset loss of corrosion re-sistance by minimization of chromium, toincrease the temperature at which austeniteforms, and to form another hardening pre-cipitate in the presence of cobalt

• Cobalt to stabilize austenite while raisingMS

• Aluminum or titanium to form intermetallicprecipitates with nickel or copper to precipi-tate as elemental copper

It is possible to quantify these various influ-ences on phases. This is summarized in Table 3in terms of the influence of the element on dif-ferent factors measured in degrees Centigradefor a 12% chromium alloy.

Rapid quenching of these alloys is not re-quired. They are air hardenable. But, the cool-ing of these alloys must be completed expedi-tiously through the final stages of martensiteformation with minimal delay. During delaysafter the start of martensite transformation hasoccurred, the remaining austenite tends to stabi-lize, and full transformation to martensite doesnot occur. When this happens, the higher levelsof austenite reduce subsequent mechanicalproperties after aging.

As with any alloy used at such high strengthlevels, microstructural cleanliness is essential,but air melting and argon oxygen decarburiza-tion (AOD) refining are quite adequate.

Corrosion Resistance. The martensitic PHstainless steels obey the same rules as otherstainless steels with regard to corrosion resist-ance. The martensite carries no nitrogen in solu-tion, so the resistance to pitting is given by:

PREN = %Cr + 3.3%Mo (Eq 1)

In the martensitic PH alloys, no chromium isrendered ineffective by the formation of Cr23C6since carbon is either held low or stabilized bytitanium 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

Loweringof % ferriteper % element

–220 –210 –20 –7 –7 –6 6 5 14 18 54

Lowering of MSper % element

–475 –475 –17 0 to 10 –17 –30 –11 –21 –17 –46

Change of ACper % element

0 to 280 0 to 250 –30 to –115 0 0 –25 to –66 25 to 73 25 to 60 0 to 35 50 to 290 30 to 750


resistance will equal that of stabilized ferritic al-loys of the same pitting resistance equivalentnumber (PREN) for which voluminous data areavailable.

A greater concern is the risk of SCC in thesealloys. While the mechanism of SCC inaustenitic alloys is still debatable, it has longbeen clear that, for martensitic alloys, SCC issimply a manifestation of hydrogen embrittle-ment in which the hydrogen is provided by localcorrosion. The existence of pitting is a suffi-cient, if not necessary, condition for SCC tooccur if the temperature is within the range ofsusceptibility and the material is inherently sus-ceptible. The material susceptibility is largely afunction of resistance to crack propagation inany given alloy that is measured by fracturetoughness. The martensitic PH grades have ex-cellent toughness and low rates of crack propa-gation, but none should be considered immuneto SCC since their hardness is never below theRc 22 level, which is considered to be thethreshold hardness for susceptibility to SCC inbody-centered cubic (bcc) ferrous alloys.

The suitability of high-strength alloys for usein potential SCC-provoking environments con-taining H2S is regulated in many locales by Na-tional Association of Corrosion Engineers(NACE) International Standard MR01-75. In it,the use of S17400 is permitted if it is doubletempered at 620 °C and its hardness is 33 HRCor less, while S45000 can be used if it has beenaged at 620 °C for 4 h and its hardness is 31HRC or less. These permissible hardness levelsare significantly higher than allowed in non-PHmartensitic alloys, 22 HRC, which reflects thefact that the martensitic matrix has the tough-ness of a lower-hardness martensite.

In marine environments, the PH alloys aresusceptible to SCC if used at a high strengthlevel. S17400 aged at 480 °C with a yieldstrength of 1240 MPa is susceptible to SCC,while higher aging temperatures (above 540 °C)producing lower strengths renders the materialsimmune at stresses near the yield strength, ap-proximately 1170 MPa. The threshold strengthfor ordinary martensitic stainless steels wouldbe 1030 MPa (Ref 4).

Semiaustenitic Precipitation-Hardenable Stainless Steels

If a martensitic stainless steel were alloyedmore strongly with austenite-stabilizing ele-

ments, the austenite could be made stable atroom temperature. This would make the alloysofter and more fabricable and, most impor-tantly, permit them to be manufactured as cold-rolled sheet and strip. If the austenite could thenbe transformed to martensite by cryogenic treat-ment, cold work, or special heat treatment, thenit could be age hardened just like the martensiticPH grades. This has been accomplished for agroup of alloys called the semiaustenitic PHgrades. The “semi” signifies that the austenitein these alloys is metastable rather than stable atambient temperatures. Also, it should be notedthat these semiaustenitic alloys usually containsome δ-ferrite in their predominantly austeniticmicrostructure after annealing.

These alloys are complex metallurgically be-cause of the technique used to make the austenitestable at room temperature after a full solutionanneal. The austenite is rendered stable by fairlyhigh levels of carbon, a powerful austenite stabi-lizer, in solution. The amount of carbon that canbe held in solution is a function of annealing tem-perature. The 1050 °C anneal of the condition Amill anneal puts all the carbon in solution, givingthe austenite the stability of a normal 301-typealloy. This permits extensive forming. The key isto apply a subsequent lower-temperature annealso that less carbon goes into solution. Some of itwill thus form M23C6. This is in a sense deliber-ately sensitizing the alloy, but the sensitizationtakes place at such a high temperature thatchromium deficits around precipitated carbidesare minimized by diffusion. This causes a higherMs temperature because of the lower amount ofcarbon, and chromium, in solution in the austen-ite. Depending on the temperature at which theanneal is done, the Ms temperature can be con-trolled so that a transformation to martensite canbe raised to either room temperature or some at-tainable cryogenic temperature. Figure 3 shows achart of these heat treatment options.

The lower strength levels achieved by thecondition T route in Fig. 3 reflect the lower car-bon content of the martensite, while the higheststrength of the condition C route reflects thecompound influence of cold work and marten-site hardness with similar subsequent contribu-tion from age hardening.

The main alloys of this group are listed inTable 4. Examination of the chemistries in thistable shows that the first two alloys rely on theprecipitation of Ni3Al for the hardening, whilethe last two have no apparent precipitatingcomponents. Their hardening is a more subtle


secondary hardening from the tempering ofmartensite rather than the classic precipitationhardening via precipitation of intermetalliccompounds. In 17-7 and 15-7, aluminum ratherthan titanium is the precipitating agent because

titanium would preferentially deplete the alloyof carbon and nitrogen, precluding the action ofthe conditioning heat treatment, which relies onmanipulating the amount of carbon in solution.In AM350 and AM355, it is the precipitation of

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 . . . 1 015-7 PH S15700 0.1 0.5 0.3 15 7.1 2.2 1 0AM-350 S35000 0.1 0.8 0.4 17 4.3 2.8 . . . 0.1AM-355 S35500 0.1 0.9 0.4 16 4.3 2.8 . . . 0.1


(Cr,Fe)2N within the martensite phase that isresponsible for age hardening. In addition,molybdenum produces a secondary hardeningin carbon-bearing martensite. These two earlyalloys did not possess the hardening potentialthat alloys employing copper-, titanium-, oraluminum-based precipitates enjoy.

The mechanical properties of these alloys are not greatly different from the martensitic al-loys, as can be seen in Table 5. Their separateexistence is due to the need for alloys that aremore fabricable at room temperature than are thealloys that are martensitic at room temperature.This benefit is offset by the necessity to condi-tion anneal before age hardening. If corrosionresistance is a high concern, then the alloy andheat treatment that yields the greatest amount ofchromium in solution should be chosen. Thuscondition CH is better than RH, which is betterthan TH, the order of ascending solution anneal-ing temperature and ascending Ms.

The martensitic PH grades have somewhatbetter strength because they have a more uni-formly martensitic structure, and they employ atougher, lower-carbon martensite. The fact thatthe semiaustenitic alloys are typically sheetproducts generally makes their service tough-ness requirements less onerous, so that their re-tained δ-ferrite is not a crippling drawback be-cause of its detrimental effect on throughthickness toughness.

Corrosion Resistance. The semiausteniticPH alloys tend to higher values of PREN thanthe martensitic alloys inherently since they arealloyed to have lower Ms temperatures. Thethermal processing of these alloys causes a sig-nificant portion of the chromium to be removedfrom solution as chromium carbide. This lowersthe corrosion resistance from what would be ex-pected based on the bulk composition. From anengineering point of view, it is best to assume

that all carbon is present as chromium carbide,and that the chromium content is diminished bythat amount before applying Eq 1. In addition,carbon can remove some molybdenum in theform of carbides, and nitrogen can removechromium as a nitride. The effective corrosionresistance of these alloys thus is similar to fer-ritic 430. The designer is thus advised to consultwith producers about corrosion resistance de-pending on the thermal processing that will beused, especially if double aging is performed,which can cause some degree of chromium de-pletion at grain boundaries.

While the general and pitting corrosion re-sistance of the semiaustenitic PH alloys arenever quite as good as most austenitic stainless,they have very good resistance to SCC com-pared to ordinary martensitic stainless steels.The δ-ferrite and the generally well-temperedmartensitic matrix provide a crack-arrestingfeature and good inherent toughness that resistSCC at higher strength levels than in straightmartensitic stainless steels. AM-355 in the SCC(850 °F) condition can withstand stresses of75% of 0.2% offset yield strength in salt spraywithout SCC failure.

Austenitic Precipitation-HardenableStainless Steels

The austinitic PH class consists of just oneimportant commercial alloy, A-286. The impor-tance of the alloy is that it is entirely stableaustenite in both the solution-annealed and theage-hardened condition. This means it is veryformable and nonmagnetic. And, because theprecipitation takes place in an austenite matrix,the precipitation takes place at a higher temper-ature, around 700 °C. This gives the alloy the

Table 5 Mechanical properties of semiaustenitic precipitation-hardenable alloys

Alloy UNS ConditionYield,MPa

Tensile,MPa Elongation, % HRC

17–7 PH . . . TH 1050 (565) 1100 1310 10 42RH 950 (510) 1380 1520 9 46CH 900 (480) 1585 1655 2 49

15–7 PH . . . TH 1050 (565) 1380 1450 7 45RH 950 (510) 1550 1650 6 48CH 900 (510) 1720 1790 2 50

AM-350 . . . SCT 850 (450) 1210 1420 12 46SCT1000 (540) 1020 1165 15 40

AM-355 . . . SCT 850 (450) 1250 1510 13 48SCT 1000(540) 1035 1124 22 38


potential to be used to near 700 °C without fearof overaging. Thus, austenitic PH stainless rep-resents a way to strengthen the austenite matrix,which has the following advantages:

• High ductility and therefore high formabilityin the soft, unaged condition

• High toughness at all temperatures andstrength levels

• Excellent creep and stress rupture properties• Excellent oxidation, corrosion, and SCC re-

sistance

This alloy is rightly considered an iron-based superalloy and is the root of the groupthat succeeds it in properties, the nickel- andcobalt-based superalloys. These last alloys canattain greater strength and creep resistancethan A-286, effectively ending further devel-opment of austenitic PH alloys. The lower costof the A-286 alloy, compared to nickel-base

PH alloys, makes it attractive for a variety ofaerospace and nonaerospace uses.

The hardening mechanism of A-286 is theprecipitation of Ni3 (aluminum, titanium). Dif-fusion, even at the higher temperatures, isslower in austenite, so aging treatments are typ-ically 16 h. Table 6 gives the typical composi-tion of A-286, and Fig. 4 shows some typicalproperties as a function of temperature for astandard 980 °C solution treatment followed bya 720 °C, 16 h aging.

The mechanical properties can be greatly en-hanced by cold working prior to aging, as isshown in Fig. 5.

The toughness of the austenitic matrix isabundant and quite temperature insensitive.Charpy V-notch values of over 60 J are typicalfrom –200 to 800 °C.

The corrosion resistance of A-286 is compa-rable to that of 304 and 316. It has slightly

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 2.15Discalloy 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


better resistance to SCC despite its higherstrength level.

REFERENCES

1. F.B. Pickering, Physical Metallurgical De-velopments of Stainless Steel, Stainless ’84,Goteborg, Sept 3–4, 1984, p 2–28.

2. M. Murayama, Y. Katayama, and K. Hono,Microstructural Evolution in a 17-4 PHStainless Steel After Aging at 400 °C, Met-

allurgical and Materials Transactions A,Vol 30A, Feb 1999.

3. G. Aggen, Ph.D. thesis, Carnegie MellonUniversity

4. E.E. Denhard, “Stress Corrosion Crackingof High Strength Stainless Steels in Atmos-pheric Environments”, paper presented atthe Twenty-fourth Meeting of the AGARDStructures and Materials Panel (Turin,Italy), April 17–20, 1967.

5. Allegheny Technology Blue Sheets

Fig. 5 The influence of cold work on aging response in A-286. DPH, diamond pyramid hardness. Source: Ref 5

CHAPTER 11

Casting Alloys

Summary

WITH TYPICAL ALLOY SYSTEMS, cast-ing is often the most convenient method bywhich to produce components. This is true forstainless steels—both for corrosion-resistingand for heat-resisting applications. This chapterdiscusses primarily the alloys used for stainlesssteel castings and their metallurgy. Foundrymethods are discussed to the degree they arespecific to the stainless alloys.

Stainless Steel Casting Alloys

Essentially any wrought stainless alloy compo-sition can be modified to be made as a cast alloy.The systemic difference between cast alloys andtheir wrought equivalents is that cast alloys gen-erally contain between 1.0 and 2.5% silicon. Aswith other ferrous alloys, this is done to increasethe fluidity of the melt to make it cast more effec-tively. Silicon has strong metallurgical effects,both beneficial and detrimental, which should beunderstood by the user of cast stainless steels.These are explained. A second general observa-tion is that the stabilized ferritic stainless steel al-loys, which constitute almost half the tonnage ofall stainless steel used, are notably absent fromthe cast alloys. This is because these alloys aresingle phase at all temperatures in the solid stateand because they have large as-cast grain sizesthat can only be refined by heavy cold work fol-lowed by annealing. This makes them quite lack-ing in toughness as cast. Since heavy cold workdefeats the purpose of casting to achieve a near-net shape, stabilized ferritic stainless steels areseldom used as castings. Also, the standard stabi-lizing alloy, titanium, is too readily oxidized fornormal foundry practice to avoid the loss of this

essential element. Thus, the casting alloys listedin Tables 1 and 2 (Ref 1) are recognizable as ap-proximate counterparts of the co-listed wroughtalloys (AISI grade). This cross reference towrought equivalents is helpful when looking fordata about an alloy that may be more easilyfound for wrought alloys than for cast.

The High Alloy Product Group of the SteelFounder’s Society of America employs a nam-ing system (ACI, the Alloy Casting Institute)for cast alloys that is significant; these designa-tions are currently assigned by ASTM as gradesand are added to ASTM specifications. The firstletter, “C” or “H,” indicates corrosion resisting.The second letter indicates the relative amountof nickel, from a minimum of 0 to 1% for “A”up to 30% nickel for “N” alloys. The numberfollowing the hyphen for “C” alloys designatesthe maximum carbon in hundredths of a percent.The suffix letters designate additional alloyingelements, such as Cu for copper, M for molyb-denum, N for nickel or nitrogen, F for free ma-chining, and C for columbium (niobium). Theheat-resisting, “H,” alloys have generally only asecond letter designating relative nickel level onthe same scale as “C” alloys but going past stain-less steels all the way to nickel-based alloys. Theinclusion of a number after the first two lettersindicates the center of the carbon range ex-pressed in hundredths of a percent by weight.

To learn more about the influence of alloyingelements, refer to the chapters on the individualalloy families; see Section 3. Here, we brieflysummarize:

• Pitting and crevice corrosion resistance, aswell as general corrosion resistance, are en-hanced by chromium, molybdenum, tung-sten, and nitrogen and carbon in solution.

• Localized corrosion is caused by chromiumdepletion, which occurs when precipitates




Table 1 Compositions of cast stainless corrosion resisting alloysComposition(a), wt%—maximum or range

Nearest ACI designation AISI grade UNS %C %Mn %Si %Cr %Ni %Mo %Other

Chromium alloysCA-15 410 J91150 0.15 1.00 1.50 11.5–14.0 1.0 0.50CA-15M J91151 0.15 1.00 0.65 11.5–14.0 1.0 0.15–1.00CA-40 420 J91153 0.40 1.00 1.50 11.5–14.0 1.0 0.50CA-40F 420F J91154 0.2–0.4 1.00 1.50 11.5–14.0 1.0 0.20–0.40 SsCB-30 431,442 J91803 0.30 1.00 1.50 18.0–22.0 2.0CC-50 446 J92613 0.30 1.00 1.50 26.0–30.0 4.0Chromium-nickel alloysCA-6N J91650 0.06 0.50 1.00 10.5–12.5 6.0–8.0CA-6NM S41500 J91540 0.06 1.00 1.00 11.5–14.0 3.5–4.5 0.4–1.0CA-28MWV 422 J91422 0.20–0.28 0.50–1.00 1.00 11.0–12.5 0.5–1.0 0.9–1.25 0.9–1.25 W,

0.2–0.3 VCB-7Cu-1 17-4PH (AISI 630) J92180 0.07 0.70 1.00 15.5–17.7 3.6–4.6 2.5–3.2 Cu,

0.2–0.35 Nb,0.05 N max

CB-7Cu-2 15-5 PH (XM-12) J92110 0.07 0.70 1.00 14.0–15.5 4.5–5.5 2.5–3.2 Cu, 0.2–0.35 Nb,0.05 N max

CD-3MN 2205 (S32205) J92205 0.03 1.50 1.00 21.0–23.5 4.5–6.5 2.5–3.5 1.0 max Cu, 0.10–0.30 N

CD-3MCuN 255 (S32550) J93373 0.03 1.20 1.10 24.0-26.7 5.6–6.7 2.9–3.8 1.4–1.9 Cu, 0.22–0.33 N

CD-3MWCuN (S32760) J93380 0.03 1.00 1.00 24.0–26.0 6.5–8.5 3.0–4.0 0.5–1.0 Cu, 0.5–1.0 W, 0.20–0.30 N

CD-4MCu J93370 0.04 1.00 1.00 24.5–26.5 4.75–6.0 1.75–2.25 2.75–3.25 CuCD-4MCuN J93372 0.04 1.00 1.00 24.5–26.5 4.7–6.0 1.75–2.25 2.75–3.25

Cu, 0.10–0.25 NCD-6MN J93371 0.06 1.00 1.00 24.0–27.0 4.0–6.0 1.75–2.25 1.75–2.5 Cu,

0.15–0.25 NCE-3MN 2507 (S32750) J93404 0.03 1.50 1.00 24.0–26.0 6.0–8.0 4.0–5.0 0.10–0.30 NCE-8MN J93345 0.08 1.00 1.50 22.5–25.5 8.0–11.0 3.0–4.5 0.10–0.30 NCE-30 312 J93423 0.30 1.50 2.00 26.0–30.0 8.0–11.0CF-3 304L J92500 0.03 1.50 2.00 17.0–21.0 8.0–12.0CF-3M 316L J92800 0.03 1.50 2.00 17.0–21.0 8.0–12.0 2.0–3.0CF-3MN 316LN J92700 0.03 1.50 1.50 17.0–21.0 9.0–13.0 2.0–3.0 0.10–0.20 NCF-8 304 J92600 0.08 1.50 2.00 18.0–21.0 8.0–11.0CF-8C 347 J92710 0.08 1.50 2.00 18.0–21.0 9.0–12.0 NbCF-8M 316 J92900 0.08 1.50 2.00 18.0–21.0 9.0–12.0 2.0–3.0CF-10 304H J92590 0.04–0.10 1.50 2.00 18.0–21.0 8.0–11.0CF-10M 316H J92901 0.04–0.10 1.50 1.50 18.0–21.0 9.0–12.0 2.0–3.0CF-10MC 316H J92971 0.10 1.50 1.50 15.0–18.0 13.0–16.0 1.75–2.25 (10xC)–1.2 NbCF-10SMnN NITRONIC™ 60 J92972 0.10 7.00–9.00 3.50–4.50 16.0–18.0 8.0–9.0 0.08–0.18 NCF-12M 316 0.12 1.50 2.00 18.0–21.0 9.0–12.0 2.0–3.0CF-16F 303 J92701 0.16 1.50 2.00 18.0–21.0 9.0–12.0 1.5 max 0.2–0.35 SeCF-20 302 J92602 0.20 1.50 2.00 18.0–21.0 8.0–11.0CG-6MMN NITRONIC™ 50 J93790 0.06 4.00–6.00 1.00 20.5–23.5 11.5–13.5 1.5–3.0 0.1–0.3 Nb,

0.1–0.3 V, 0.2–0.40 N

CG-8M 317 J93000 0.08 1.50 1.50 18.0–21.0 9.0–13.0CG-12 308 J93001 0.12 1.50 2.00 20.0–23.0 10.0–13.0CH-8 309S J93400 0.08 1.50 1.50 22.0–26.0 12.0–15.0CH-10 309H J93401 0.04–0.10 1.50 2.00 22.0–26.0 12.0–15.0CH-20 309 J93402 0.20 1.50 2.00 22.0–26.0 12.0–15.0CK-3MCuN 254SMO™ J94653 0.025 1.20 1.00 19.5–20.5 17.5–19.5 6.0–7.0 0.5–1.0 Cu,

0.18–0.24 NCK-20 310 J94202 0.20 2.00 2.00 23.0-27.0 19.0–22.0CN-3M 904L J94652 0.03 2.00 1.00 20.0–22.0 23.0–27.0 4.5–5.5CN-3MN AL-6XN® J94651 0.03 2.00 1.00 20.0–22.0 23.0–27.0 6.0–7.0 0.18–0.24 NCN-7M 320 N08007 0.07 1.50 1.50 19.0-22.0 27.5–30.0 2.0–3.0 3.0–4.0 CuCN-7MS J94650 0.07 1.50 3.50 18.0–20.0 22.0–25.0 2.5–3.0 1.5–2.0 CuCT-15C N08151 0.05–0.15 0.15–1.50 0.50–1.50 19.0–21.0 31.0–34.0 0.5–1.5 Nb

(a) Balance Fe for all compositions. Source: Ref 1

Chapter 11: Casting Alloys / 149

form in the solid state. These precipitates arecarbides, oxides, and sulfides as well as inter-metallic phases richer in chromium, molyb-denum, or nitrogen than the matrix.

• General corrosion resistance follows theabove guidelines but is also helped by cop-per and nickel, which do not assist in pittingresistance.

• High-temperature oxidation resistance is en-hanced by increasing chromium and silicon.Wrought alloys employ aluminum and rareearths to help oxidation resistance, but thedifficulty of keeping these elements frombeing oxidized requires special techniquessuch as vacuum induction melting and inertrefractories for molds.

• Iron-chromium (ferritic) alloys have betterthermal fatigue resistance but poorer creepresistance than iron-chromium-nickel(austenitic) alloys.

The alloy designation system largely ignoresthe wrought alloy distinctions by microstructure(i.e., ferritic, austenitic, duplex, PH [precipita-tion hardening], and martensitic). One reason isthat the most widely used wrought-stabilizedferritics (e.g., 409, 439) do not exist as commoncasting alloys, and nominally austenitic alloysin the cast form contain enough ferrite to be sig-nificantly magnetic. Thus, the distinctions basedon phase are not as well defined for casting al-loys. The high ferrite content in the nominallyaustenitic casting alloys is to avoid or at least

minimize solidification hot cracking or to allowweld repair of cracks that do form. It has beenshown that the existence of ferrite can increasethe resistance to stress corrosion cracking.

Metallurgy of “C” Alloys

The corrosion-resistant “C” series havewrought counterparts from which they differ es-sentially only in silicon content. This silicon hasno significant influence on corrosion resistanceor mechanical or physical properties, so an un-derstanding of these alloys by approximatingthem to their wrought counterparts is justified.The main difference between the cast andwrought product forms of these alloys is thegrain structure. In wrought grades, the grainstructure can be manipulated by deformationand heat treatment. The use of deformation isnot an option in cast alloys; consequently, theopportunities for grain refinement in cast alloysare limited. To counteract the problems of lowercorrosion resistance of cast grades, a homoge-nizing solution heat treatment is necessary tocounteract the chromium depletion that occursdue to solidification segregation and precipitatingphases. Representative mechanical propertiesfor “C” alloys are listed in Table 3 (Ref 1).

Martensitic. CA alloys are martensitic. Themetallurgy is straightforward and equivalent totheir wrought counterparts. The mechanical

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 8–10 1.00HC 446 J92605 0.50 max 26–30 4–max 2.00HD 327 J93005 0.50 max 26–30 4–7 2.00HE 312 J93403 0.20–0.50 26–30 8–11 2.00HF 302B J92603 0.20–0.40 19–23 9–12 2.00HH 309 J93505 0.20–0.50 24–28 11–14 2.00HI J94003 0.20–0.50 26–30 14–18 2.00HK 310 J94224 0.20–0.60 24–38 18–22 2.00HK-30 J94203 0.25–0.35 23–27 19–22 1.75HK-40 J94204 0.35–0.45 23–27 19–22 1.75HL N08604 0.20–0.60 28–32 18–22 2.00HN J94213 0.20–0.50 19–23 23–27 2.00HP N08705 0.35–0.75 24–28 33–37 2.00HP-50WZ 0.45–0.55 24–28 33–37 2.50HT 330 N08605 0.35–0.75 13–17 33–37 2.50HT-30 N08603 0.25–0.35 13–17 33–37 2.50HU N08005 0.35–0.75 17–21 37–41 2.50HW N08006 0.35–0.75 10–14 58–62 2.50HX N06050 0.35–0.75 15–19 64–68 2.50

(a) Balance Fe for all compositions. Manganese content: 0.35–0.65% 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: 0.90–1.2%. Maximum molybdenum for other alloys is 0.5%. HH contains 0.2% N(max). HP-50WZ also contains 4–6% W, 0.1–1.0% Zr, and 0.035% S (max) and P (max). Source: Ref 1


Tabl

e 3

Roo

m t

empe

ratu

re m

echa

nica

l pro

pert

ies

of c

orro

sion

res

isti

ng c

ast

stai

nles

s al

loys

Tens

ile s

tren

gth

Yie

ld s

tren

gth,

0.2

% o

ffse

tE

long

atio

n %

in 5

0 m

m, 2

in.

Red

ucti

onin

are

a, %

Har

dnes

s,H

B

Cha

rpy

toug

hnes

s

Allo

yH

eat

trea

tmen

t(a)

MP

aks

iM

Pa

ksi

Jft

-lb

Spec

imen

CA

-6N

M

>95

5 o C

(17

50 o F

), A

C, T

827

120

689

100

24

60

269

94.9

70

V

-not

ch

CA

-15

980

o C (

1800

o F),

AC

, T79

3 11

5 68

9 10

0 22

55

22

5 27

.1

20

Key

hole

not

ch

CA

-40

980

o C (

1800

o F),

AC

, T10

34

150

862

125

10

30

310

2.7

2 K

eyho

le n

otch

C

B-7

Cu

1040

o C (

1900

o F),

OQ

, A13

10

190

1172

17

0 14

54

40

0 33

.9

25

V-n

otch

C

B-3

0 79

0 o C

(14

50 o F

), A

C

655

95

414

60

15

. . .

195

2.7

2 K

eyho

le n

otch

C

C-5

0 10

40 o C

(19

00 o F

), A

C

669

97

448

65

18

. . .

210

. . .

. . .

. . .

CD

-4M

Cu

1120

o C (

2050

o F),

FC

to

1040

o C (

1900

o F),

WQ

74

5 10

8 55

8 81

25

. .

. 25

3 74

.6

55

V-n

otch

1120

o C (

2050

o F),

FC

to

1040

o C (

1900

o F),

A89

6 13

0 63

4 92

20

. .

. 30

5 35

.3

26

V-n

otch

CE

-30

1095

o C (

2000

o F),

WQ

66

9 97

43

4 63

18

. .

. 19

0 9.

5 7

Key

hole

not

ch

CF-

3 >

1040

o C (

1900

o F),

WQ

53

1 77

24

8 36

60

. .

. 14

0 14

9.2

110

V-n

otch

C

F-3A

>10

40 o C

(19

00 o F

), W

Q

600

87

290

42

50

. . .

160

135.

6 10

0 V

-not

ch

CF-

8 >

1040

o C (

1900

o F),

WQ

53

1 77

25

5 37

55

. .

. 14

0 10

0.3

74

Key

hole

not

ch

CF-

8A>

1040

o C (

1900

o F),

WQ

58

6 85

31

0 45

50

. .

. 15

6 94

.9

70

Key

hole

not

ch

CF-

20

>10

95 o C

(20

00 o F

), W

Q

531

77

248

36

50

. . .

163

81.4

60

K

eyho

le n

otch

C

F-3M

>

1040

o C (

1900

o F),

WQ

55

2 80

26

2 38

55

. .

. 15

0 16

2.7

120

V-n

otch

C

F-3M

A>

1040

o C (

1900

o F),

WQ

62

1 90

31

0 45

45

. .

. 17

0 13

5.6

100

V-n

otch

C

F-8M

>

1065

o C (

1950

o F),

WQ

55

2 80

29

0 42

50

. .

. 17

0 94

.9

70

Key

hole

not

ch

CF-

8C

>10

65 o C

(19

50 o F

), W

Q

531

77

262

38

39

. . .

149

40.7

30

K

eyho

le n

otch

C

F-16

F >

1095

o C (

2000

o F),

WQ

53

1 77

27

6 40

52

. .

. 15

0 10

1.7

75

Key

hole

not

ch

CG

-8M

>

1040

o C (

1900

o F),

WQ

56

5 82

30

3 44

45

. .

. 17

6 10

8.5

80

V-n

otch

C

H-2

0 >

1095

o C (

2000

o F),

WQ

60

7 88

34

5 50

38

. .

. 19

0 40

.7

30

Key

hole

not

ch

CK

-20

1150

o C (

2100

o F),

WQ

52

4 76

26

2 38

37

. .

. 14

4 67

.8

50

Izod

V-n

otch

C

N-3

MN

1150

o C (

2100

o F),

WQ

770

112

365

5350

185

190

140

V-n

otch

CN

-7M

11

20 o C

(20

50 o F

), W

Q

476

69

214

31

48

. . .

130

94.9

70

K

eyho

le n

otch

(a) A

C, a

ir c

ool;

FC, f

urna

ce c

ool;

OQ

, oil

quen

ch; W

Q, w

ater

que

nch;

T, t

empe

r; A

, age

. Sou

rce:

Ref

1


properties are governed by the thermal pro-cessing, and strength, hardness, and toughnesscan be varied over a wide range. The CB 30and CC 50 alloys are ferritic and, as such, havenegligible toughness but effectively delivercorrosion resistance. The toughness of CB 30can be improved by balancing the chromiumand silicon to a lower part of the range and thecarbon and nickel to the higher end to renderthe microstructure partly martensitic.

Precipitation Hardening. The cast PH alloysinclude CB-7Cu-1, which behaves in a similarway to 17-4PH, which has an overlapping com-position range. Note that most other majorwrought PH grades rely on titanium and alu-minum to form coherent strengthening precipi-tates and so do not have cast counterparts. Cop-per, which can harden ferrite but not austenite,is thus the only strengthener available. There isone cast PH alloy that has no wrought counter-part. It is CD-4MCu; however, it is rarely usedin the precipitation-hardened condition and ismost commonly classified as a duplex stainlesssteel in which the nitrogen level is closely con-trolled. This is a highly alloyed duplex gradethat contains copper to precipitation harden theferrite phase. Oil field CO2 corrosion is resistedby alloys that resemble the martensitic PHgrades. These alloys are discussed in Chapter22, “Petroleum Industry Applications” and canbe considered castable alloys.

Duplex. The cast equivalents of alloys 2205and 2507, J 92205, and J 93380 have similarproperties and corrosion resistance. Modernwrought duplex alloys rely on nitrogen to parti-tion the alloy with uniform corrosion resistancein each phase and to suppress intermetallic phaseformation. Cast alloys are effectively limited to0.25% nitrogen before gas porosity becomes ex-cessive. Porosity can be reduced by replacingsome nickel with manganese, which increases ni-trogen solubility. Doing so would expand themost promising area of stainless steel develop-ment, lean duplex alloys such as 2101 and 2003,to the cast grades. Alloy 2101 with 4 to 6% man-ganese provides the corrosion resistance of CF-8M or 316L with total nickel plus molybdenumof only 2% versus the 12% required for theaustenitic alloy. The duplex alloys also havegreater strength and are nearly immune to stresscorrosion cracking. These alloys represent signif-icant cost-savings potential for the foundry andfor its customers. CE-30 is duplex steel, which isfairly simple metallurgically and uses onlychromium for corrosion resistance. However, its

high level of nickel negates much of the potentialcost savings duplex alloys offer. All castings aresolution annealed and quenched to eliminate em-brittling intermetallic phases.

The cast duplex alloys may offer a better engi-neering approach than the equivalent austeniticcast alloys because they have greater strengthand lower alloy cost for the same level of corro-sion resistance. They do not have the same prob-lems of hot cracking that make casting austeniticsteels difficult. The poor hot workability of duplex steel is not an issue for castings. It is im-portant for designers to understand that cast du-plex steels are totally compatible galvanicallywith wrought or cast austenitic alloys of thesame corrosion resistance. Mixing componentswith different microstructures does not create agalvanic differential when corrosion resistancelevels are similar. Reluctance to mix alloys forgalvanic reasons can be an expensive error whentheir similar corrosion resistances makes themcompatible, even if they are quite differentmicrostructurally.

Austenitic-Ferritic. The typical CF alloys,which make up about two-thirds of U.S. stain-less steel castings, are nominally austenitic butalways contain ferrite. This is not detrimentaland improves resistance to stress corrosioncracking and sensitization. Homogenization an-nealing can reduce the amount of ferrite and re-sult in lower yield and tensile strength andhigher elongation and toughness. The composi-tion balance is the main determinant of ferritelevel. Increasing the nickel, nitrogen, manganese,or carbon content decreases ferrite. Increasingchromium, silicon, or molybdenum content in-creases ferrite. Increasing the solidification ratewill increase the ratio of austenite to ferrite induplex or austenitic-ferritic alloys. The predom-inantly austenitic matrix has a very high tough-ness even at cryogenic temperatures. Ferrite, ifcontinuous, decreases toughness. Fortunately, itis seldom present as a continuous phase. Theloss of toughness associated with high ferritecontent can be aggravated by heating the ferriteabove 475 oC (885 oF) for a sufficient time forthe ferrite to decompose to the brittle α and α'.At higher temperatures, development of the σphase would have a similar embrittling effect.These phases thus formed are quickly redis-solved and removed by annealing. Note thatsometimes copper is added to austenitic alloysto improve corrosion resistance in sulfuric acidenvironments. It has no precipitation hardeningeffect in austenite, as it does in ferrite. When


used as a precipitating hardening agent, copperdoes not increase corrosion resistance.

Virtually any non-titanium-bearing, corro-sion-resistant, austenitic, wrought alloy canhave a cast counterpart. Curiously, the 2xx low-nickel alloys are not found in most cast alloyslists. If a specific wrought alloy cannot be foundto have a published cast counterpart, the de-signer should not avoid requesting a producerto supply a version that the foundry is confidentof making. The designer must thoroughly un-derstand the design of the alloy desired so thatany alterations to its composition necessary toallow castability will not compromise expectedperformance.

Metallurgy of “H” Alloys

The heat-resisting “H” alloys are principallyaustenitic. Alloying elements and impuritiesdiffuse more slowly through the face-centeredcubic (fcc) austenitic structure than the bccferrite structure, making the austenite more re-sistant to diffusion-controlled creep. Austenite

has higher thermal expansion and lower thermalconductivity than ferrite, which aggravates ther-mal fatigue and oxide spalling. Nevertheless,the better high-temperature strength of austenitegenerally is the predominant consideration, andmost “H” alloys are austenitic. Tables 4 and 5list properties of “H” alloys (Ref 1 to 3).

Ferritic HA, HC, HD. Of the ferritic alloysHA, HC, and HD, HA with less than 10%chromium is not quite stainless but is useful to650 oC (1200 oF) for petroleum refinery applica-tions. HC and HD are very high chromiumferritics that have very low toughness and creepresistance but are quite oxidation and sulfida-tion resistant. They can be cost-effective materi-als when high-temperature strength is not anoverriding concern.

Austenitic HE-HP. The predominant high-temperature grades are the austenitic HEthrough HP, after which come the nickel alloys,which are not generally classified as stainlesssteels because they contain less than 50% iron.The high material cost of the nickel base alloysrestricts their use to those specific environmentswhere maximum carburization or nitriding

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

Alloy Condition

Tensile strength Yield strength

Elongation, % Hardness, HB MPa ksi MPa ksi

Standard gradesHA 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) Agingtreatment: 48 h at 980 °C (1800 °F), furnace cool. Source: Ref 1


resistance is mandatory. For oxidation and sulfi-dation resistance, the iron base alloys are pre-ferred. These cast stainless steels derive theiroxidation resistance from their chromium level.The chromium near the surface acts as a reser-voir to replenish the protective iron/chromiumoxide scale as explained in Chapter 6 in the sec-tion on oxidation. Silicon, another stable oxideformer, assists in forming this protective scaleand resistance to carburization. Other typicalalloying elements do not aid in oxidation resist-ance. If it were possible to cast these alloys withaluminum or rare earth additions without thembeing lost to oxidation before solidification,there could be some impressive benefits. Suchalloys exist in wrought form, for example,153MA and 253MA. The metallurgical basis ofthe benefits from aluminum and rare earth

alloying also are discussed in the oxidation sec-tion of Chapter 6.

The major problem that all producers of stain-less steels face is that of transferring moltenmetal from the furnace to the mold cavity. Thisproblem is heightened when the foundry makescomplex shapes. Methods developed to protectthe molten stream from exposure to air to preventreoxidation have shown great promise and havebeen demonstrated by the wrought alloy produc-ers who tend to produce much simpler shapesthan the foundry. Protection of the molten streamcould result in castings with much better high-temperature performance that could be used in-stead of some use of higher nickel alloys.

High-temperature strength is modestly im-proved by higher levels of chromium andnickel. Molybdenum improves high-temperature

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

Alloy Temp Yield strength

ksi, MPa

Tensilestrength ksi, MPa

Creep rate0.0001%/h psi, MPa

1% in 100,000h psi, MPa

Stress to rupture

in 1000 h

Stress to rupture

in 10,000 h

HA 1400 oF 16 27

1800 oFHC 1400 oF 8.7 10.5 1.3 1.3

1800 oF 2.1 2.5 3.6 0.6HD 1400 oF 36 3.5 7.0

1800 oF 15 1.0 2.5HE 1400 oF 3.5 11.0

1800 oF 1.0 2.5HF 1400 oF 35 6.0 4.4 9.1

1800 oF 1.6 (est)HHTYPE 1

1400 oF 17 33 3.0

1800 oF 6.3 9 1.1HHTYPE 2

1400 oF 18 35 7.0 2.0 8.0

1800 oF 7 11 2.1 1.6 0.9HI 1400 oF 6.6 8.5

1800 oF 1.9 2.6HK 1400 oF 6.8 6.3 12.0

1800 oF 2.7 0.9 2.8 1.7HL 1400 oF 50 7.0 15

1800 oF 18.7 2.8 (est) 5.2HN 1400 oF

1800 oF 2.4 2.1HP 1400 oF

1800 oF 2.1HPNb(a) 800 oC . . . . . . 51 MPa . . . 55 MPaHPNbTi(a) 800 oC . . . . . . 54 MPa . . . 64 MPaHT 1400 oF 26 35 8.0 12

1800 oF 8 11 2.0 2.7 1.7HU 1400 oF 40 8.5

1800 oF 6.2 10 2.2 2.9 1.8HW 1400 oF 23 32 6.0 7.8

1800 oF 8 10 1.4 2.6HX 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


strength, but its detrimental effect on oxidationresistance and its promotion of intermetallicprecipitation limits its use. Carbon is very effec-tive for promoting high-temperature strengthand suppression of intermetallic phase forma-tion. All “H” alloys, therefore, employ muchhigher carbon levels than the “C” alloys. Thisdoes, however, directly imply that the corrosionresistance of “H” alloys, should it be an issue, issignificantly degraded over otherwise similar“C” alloys.

The HP grades have undergone significantdevelopment over the last 30 years. This devel-opment has come about through the addition ofniobium to increase creep and rupture proper-ties. The use of microalloying additions has de-livered creep and rupture properties some 30%higher than the HP grade without niobium mi-croalloying. It is unfortunate that these HPgrades have not at this time found their way intoASTM standards; however, work is under wayto remedy this omission. Currently, the most up-to-date collection of these grades can be foundin ISO 13583-2.

Foundry Practice

While the scope of this book does not extendto the production of castings, certain aspects areimportant to the user of castings. For the last 50years, virtually all stainless has been refined inargon oxidation decarburization (AOD) vesselsor versions thereof. For “C” alloys, this refiningmethod should be considered a basic require-ment for good quality where the carbon levelsare restricted to low levels (e.g., CF3M). “H”alloys are less refined inherently and can besimply arc or induction melted; however it maybe necessary to use refining techniques for the

higher-performance grades. It is also possible touse AOD-refined master melt stock to achievethe same benefits as AOD refining while usinginduction melting.

Welding of cast stainless alloys is a commonpractice and does not present problems whenusing approved weld procedures and qualifiedwelders. Chapter 17 describes joining methodsin detail. The same precautions about sensitiza-tion apply to castings. Welding of non-niobium-stabilized “C” alloys with carbon levels above0.03% will require postweld annealing to redis-solve chromium carbides, which will otherwisemake the alloy susceptible to corrosive attack inthe chromium-depleted regions of the heat-affected zone.

Iron and nickel base “H” alloys that are fullyaustenitic can suffer from hot shortness due tosulfide films that precipitate along grain bound-aries even at low bulk sulfur levels. This makesthem susceptible to hot cracking of welds. Al-loys with some ferrite are less susceptible to hotcracking, so most “C” alloys are highly resistantto this problem.

REFERENCES

1. Cast Stainless Steels, Metals Handbook,desk ed., J.R. Davis, Ed., ASM Interna-tional, 1988, p 386–390

2. International Organization for Standardiza-tion, www.iso.org, ISO 13583-2

3. Steel Founders Society of America, onlinedocuments: http://www.sfsa.org/sfsa/pubs/index.html

SELECTED REFERENCE

• http://www.sfsa.org

CHAPTER 12

Melting, Casting, and Hot Processing

Summary

THE PRIMARY PRODUCTION PROC-ESSES of melting, casting, and hot processingare invisible to the end user. The vast majority ofstainless steel is made by arc furnace melting fol-lowed by argon oxygen decarburization (AOD)refining and continuous casting. It is not normal,and it is seldom beneficial for the end user tospecify processing paths. The end user should,however, be knowledgeable and require the pro-ducer to document the process and the producer’scontrol of it.

Introduction

The manner in which stainless steel is madeat the producing mill can have a great impact onits final properties. These production methodshave undergone a major evolution over the last50 years and are mainly responsible for stain-less steels becoming the practical, widespreadengineering materials they are today. Traditionalcarbon and alloy steel-making methods are notsuitable for stainless steels. The fundamentaldifference is that the basic decarburization step,which is common to all steel making, is thermo-dynamically very difficult in stainless steel be-cause the essential element, chromium, reactsmore strongly with the purifying agent, oxygen,than does carbon. Thus, early stainless steel mak-ing, done in an arc furnace, was a lengthy processthat necessarily involved high chromium lossesto the slag as carbon was removed. This processwas not only very expensive, the carbon levelsthat could be achieved were not much below0.10%, making most of today’s stainless steels,whose carbon levels range from 0.010% in sta-bilized ferritic alloys to about 0.07% in normal

austenitic alloys, impossible to produce. Theadvent of AOD, continuous casting, ladle metal-lurgy, and powerful hot rolling mills has led tostainless steels of much higher quality producedat lower cost. Ironically, the low processing costof stainless steel has spurred demand and madesome of its ingredients, such as molybdenumand nickel, which are relatively scarce and ex-pensive commodities, even more costly, forcingthe cost of many alloys to spike even higherthan in earlier years.

Melting and Refining

The arc furnace is nearly universally used forthe first step in the production of stainless steel.The arc furnace is quite flexible in the types ofcharge materials it can accept. Since the chargematerials for stainless steel are typically carbonsteel and stainless steel scrap, this flexibility al-lows scrap of all types to be used. The necessarychromium is added as ferrochromium, whosecost is inversely related to its carbon content.The carbon content of the heat of steel is roughly1.5 to 2.5% when it is melted and ready tocharge into the separate refining vessel.

It is this carbon whose removal is the primaryfocus of refining. In the 1960s, Union Carbideengineers perfected a method, the previouslymentioned AOD process, of removing nearly allthe carbon from molten stainless steel withoutsignificant loss of chromium. This process isbased on the following chemical reaction:

Cr3O4 (Solid) + yC = yCO (gas) + Cr (Eq 1)

The equilibrium for this reaction is:

(Eq 2)Ln (KG

T)

4575= −Δ




where K is the equilibrium constant, and G isthe Gibbs free energy.

Working through the thermodynamics yieldsthe relationship that summarizes the importantrelationship among carbon, chromium, and CO(Ref 1):

(Eq 3)

Thus, increasing the temperature works to in-crease the elimination of carbon as CO, whichevolves from the melt. This is similar in principleto the basic oxygen furnace (BOF) process forcarbon steel in which oxygen is injected intomolten steel to remove carbon by oxidizing it.The key to the AOD process, though, is the in-jection of oxygen and argon into the bath to keepthe partial pressure of CO (pCO) very low. This isdone at a temperature consistent with economicrefractory life. The injection is done throughtubes called tuyeres in the bottom of the barrel-shaped vessel. The injection and the reactioncause extremely thorough mixing, which wouldnever happen in the flat, stagnant, arc furnacebath. This mixing not only allows the CO-pro-ducing reaction to reach equilibrium, but alsothe mixing of the slag and metal also permitsdesulfurization. By increasing the ratio of argonto oxygen in the injected gas as the refining pro-ceeds, the carbon is selectively oxidized with-out concurrent chromium oxidation. A typicalstarting ratio is 3 to 1 oxygen to argon/nitrogenby volume. The ending ratio can be as low as 1to 9, oxygen to argon/nitrogen. The choice ofwhich inert gas to use, argon or nitrogen, is basedon cost and final nitrogen content desired. Stabi-lized stainless steels require low carbon and ni-trogen levels, for instance, so the more expensiveargon must be used.

It is possible to use a vacuum system to keepthe partial pressure of CO low when refiningwith injected oxygen. This is the vacuum oxy-gen decarburization (VOD) process. The VODprocess can achieve slightly lower carbon levelsbut does not achieve cleaner steel as somebelieve.

In both processes, after final carbon contenthas been achieved ferrosilicon is added to reducethe chromium in the slag and have it return to themolten steel. The excellent mixing of the slagand metal in the AOD permits this to be doneefficiently. The silicon plus the manganese in the

steel combine to reduce the oxygen content ofthe steel to around 100 ppm. This could be fur-ther reduced by aluminum, but aluminum-basedinclusions are generally undesirable. The ther-modynamic activity of aluminum is consider-ably reduced in iron as chromium levels in-crease, so its role as a deoxidizer is less valuablein stainless steels. Titanium, on the other hand, isenhanced as a deoxidizer in chromium-iron al-loys, and consequently small amounts of it aresometimes used as a supplementary deoxidant inalloys even though an alloy specification maynot call for any. Titanium is believed to reducehot working defects. More active deoxidants,such as calcium and magnesium, can be usedwhen required. Also note that even if no inten-tional addition of metallic calcium is made,strong deoxidation with aluminum or titaniumcan reduce small amounts of calcium from theCaO in the slag, producing measurable calciumcontent in the metal.

Besides carbon and oxygen, other impuritiescan be removed from the molten stainless. Oncethe steel has been deoxidized, sulfur can bereadily removed by contact with a basic slag.Sulfur can be reduced to less than 0.001% inthe AOD, and this excellent purity level is com-mercially furnished without additional pricepremium. Sulfur, although a harmful impurityfrom a corrosion standpoint, is often deliber-ately kept at moderate levels (0.008 to 0.015%)for tungsten inert gas (TIG) welding penetration(see Chapter 17) and at high levels (0.15%+) formachinability (see Chapter 15). These trade-offs,which are beneficial to processors, should beviewed with skepticism by end users, whoseproduct integrity is compromised. There areprocessing methods for which higher levels ofsulfur are not necessary that are preferable tothe end user while not compromising weldingor machining costs. For example, machinabilitycan be improved by calcium additions that pro-duce malleable oxides to replace the deleterioussulfides (see Chapter 15), and welding methods,such as laser welding, can be used in manycases to eliminate the need for the weld penetra-tion enhancement of sulfur while increasingwelding speeds.

Phosphorus is an impurity for which no prac-tical removal technology exists in stainlesssteel. Any known process to remove it first re-moves chromium. Thus, it exists in almost allstainless steel at levels close to its normal speci-fication limit, about 0.030% in austenitic alloysand 0.020% or less in ferritic alloys, which are

LogCr

%C

13,800CO= = − + −%

. . logT

p8 76 0 925

Chapter 12: Melting, Casting, and Hot Processing / 157

made from a higher percentage of low-phospho-rus carbon steel scrap. The deleterious effects ofphosphorus on corrosion are not avoided unlessmuch lower levels are achieved. Consequently,its presence is tolerated since it has no differen-tial effect over the range in which it is found.

Heavy metals are eliminated by high-temper-ature AOD blowing, as is hydrogen. Care mustbe taken not to reintroduce such impurities afterrefining, which is a risk when using damp orcontaminated scrap for coolant.

Alloy adjustment can be done in the AOD orpreferably in a treatment-and-transfer ladle. Thetapped molten steel generally has excess heatfrom the highly exothermic refining process.This allows the composition to be measured andadjusted before it must be cast. This can be donevery precisely by wire feeding of alloying ele-ments through the slag into the heat, which canbe stirred by argon bubbling via porous plugs.This technique is very effective for the fine-tuning of reactive elements such as titanium.

The refining treatments used for carbon steeland stainless steel are very similar, but there aresubtle differences because of the difference inthe thermodynamics of dilute solutions like car-bon steel and highly alloyed, nondilute solutionslike stainless steel. Table 1 shows the factors bywhich additions of various elements to stainlesssteel (j) alter the thermodynamic activity ofother alloying elements (i).

Equation 4 is used to calculate the activity ofelements in steel. The activity coefficient γvaries with the concentration of alloying ele-ment x by:

(Eq 4)

This calculation is best left to computer pro-grams such as Thermo-Calc that have been per-fected for these lengthy procedures. It shouldbe noted that chromium, which is always pres-ent in nondilute quantities, has a powerful effect

on interstitial solubility. The higher solubilityof carbon, nitrogen, and oxygen in stainlesssteels is significant. A manganese/silicon deox-idized stainless steel will still have about 100ppm of dissolved oxygen at the freezing tem-perature as opposed to the less than 10 ppm ofoxygen found in aluminum-killed carbon steel.This oxygen precipitates as oxides in the solidstate.

Vacuum induction melting (VIM) is anothermethod of melting stainless steels. This is anearly slag-free process, and little refining ispossible. Melt purity is largely controlled by thepurity of the starting material, and use of AODmaster melt stock for VIM remelting is com-mon. Limited decarburization is possible via in-jection of oxides such as Fe3O4 or SiO2 to createCO evolution inside the vessel. Using this tech-nique, very low carbon levels (less than 50ppm) are achievable commercially. Use of VIMis generally limited to high-value, high-purity,or low-tonnage melts.

Remelting

Some stainless steels and related alloys areremelted to refine composition or ingot struc-ture. There are two principal remelt processes:vacuum arc remelting (VAR) and electroslagremelting (ESR).

In VAR, the material to be remelted is castinto a cylindrical electrode and placed inside acylindrical water-cooled vacuum chamber. Ahigh-current direct current (dc) arc is estab-lished between the electrode and a starter plateat the bottom of the chamber. The end of theelectrode is melted, and the molten drops fallthrough the intervening vacuum. Volatile con-stituents escape from the molten drops, and thepurified drops collect to form a molten pool ontop of the starter plate. VAR parameters are ad-justed to maintain a shallow pool, which solidi-fies in a bottom-up fashion. The shallowness ofthe molten pool produces a refined grain

RT RT RTxi i

j n

ni

j

ln lnγ γδ γδ

= +=∑0

1...

ln

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 . . . –.005N –.03 .13 –.05 –.02 –.01 0.0 .01 .05 .01 .05 –.53 –.001S . . . .11 –.01 –.03 .003 .01 0.0 –.27 –.03 .06 –.07 .010 –.39 –.45 –.04 –.02 .003 .06 .006 –.20 –.13 –.13 –.6 –.01


structure with less solidification segregationthan found in typical cast product.

In ESR, the material to be remelted is castinto an electrode of similar shape, but slightlysmaller than the water-cooled mold. A gap be-tween the electrode and a starter plate at the bot-tom of the mold is filled with a prepared slag.Typically, this slag is calcium fluoride-basedwith high lime (CaO) content. Additional ingre-dients control the basicity, fluidity, oxidizingpotential, and other properties of the slag. Ahigh current is used to melt the slag, which inturn melts the end of the electrode, and themolten drops fall through the slag. Reaction ofthe molten drops with the slag removes sulfurand some other impurities, and the purifieddrops collect to form a molten pool on top of thestarter plate. ESR melting typically is done at ahigher rate than VAR, and the molten pool isdeeper. This deeper pool produces a grain struc-ture between that of VAR and typical cast prod-uct, with commensurate intermediate segrega-tion patterns.

Casting

Continuous slab, billet, and bloom castinghave become the standard methods of makingstainless steel primary products, replacing theobsolete ingot method. There are some alloysthat cannot be continuously cast, but these repre-sent a miniscule percentage of stainless produc-tion. Continuous casting produces slabs directly,thus removing the costly soaking and slab-rolling processes. In a well-executed continuouscasting operation, slabs are of sufficient qualitythat they require no surface conditioning beforebeing hot rolled. Slabs range in thickness from13 to 63 cm (5 to 15 in.). The segregation in con-tinuous casters is less than in ingots because ofthe smaller section size. It is not eliminated,however, and certain alloying elements concen-trate at the centerline, where they defy homoge-nization. Carbon and molybdenum are examplesof alloying elements with this tendency.

In properly executed continuous casting, theladle feeds by a slide gate, or preferably a stop-per rod gate, into a ceramic tube into the largetundish situated over the caster mold. The metalin the tundish is covered with a protective slagcover, and flow patterns within the tundish aredesigned to minimize dead spots and encourageremoval of inclusions by impingement with theslag cover. The metal feeds through another

ceramic tube, called the submerged entrynozzle, into the mold, which is covered with aconsumable protective and lubricating slagcover, called a mold powder. The mold powder,which melts in the mold as it is added, containsceramics, fluxes, and carbon. The level of themolten metal should be carefully controlled byultrasonic measurement, or other methods, toprevent fluctuations in level that may entrapslag in the slab surface. The entire water-cooled,copper alloy mold oscillates in a precise patternas the solidifying strand of steel is withdrawnfrom the mold bottom by pinch rolls andsprayed with water to cool it. The pinch rollsapply enough pressure to slightly deform theslab. This deformation has a crucial, seldom-rec-ognized effect. It causes a beneficial recrystal-lization that improves hot working characteris-tics of austenitic and duplex alloys. In ferriticalloys, it can cause excessive grain growth,which detracts from hot workability. The initialportion of slab cast in a sequence is seldom of adequate quality to be used because of exogenous inclusions, entrapped mold powder,and non-steady-state solidification structure.The defective portion must be identified andscrapped or diverted to low-quality requirementend uses.

The strand is bent from an initial slightlycurved shape to flat and cut into slabs. Morethan one heat of steel may be cast sequentiallywithout restarting the process. This is ideal eco-nomically and for quality reasons since initialand final segments of a casting can containmore inclusions and aberrant structure. Someend users stipulate that no first slabs be appliedto their orders. Producers generally apply firstslabs to less-critical uses or discard suspect sec-tions of them. If casting conditions are not opti-mal, the result can be slabs with poor surfacequality that must be surface ground.

Slabs are sometimes quenched to avoid pre-cipitation of phases; however, they may be heldat high enough temperatures prior to hot rollingto stay above the temperature range in whichembrittlement can occur or to stay above thetemperature at which an embrittled slab canfracture. Ferritic and martensitic alloys are es-pecially prone to these problems.

There has been great interest for decades inproducing stainless steel coils directly fromthe melt in so-called strip casters. Eliminationof hot rolling could be quite valuable in stain-less steel, whose hot rolling from slab can beboth expensive and problematic. There are a

Chapter 12: Melting, Casting, and Hot Processing / 159

number of such machines in pilot or limitedproduction. They have not had sufficient com-mercial or technical success to have become afactor in the industry. Since their developmentis only being undertaken by those large stain-less steel producers who already have the hotrolling assets that strip casting would replace,it seems unlikely that strip casting will soonbecome a major factor even if it is perfectedtechnically.

Another method of shortcutting the casting/ingot step has been perfected: the powder met-allurgy approach. In powder metallurgy, the re-fined molten metal is atomized by gas or liquidand made to freeze into small particles. Theseparticles, having been quenched extremelyrapidly, are quite homogeneous. Powder tech-nology methods allow for the design of alloysthat would otherwise freeze with too muchsegregation and too coarse a structure withconventional production methods. Traditionalpowder metallurgy production methods areused to make small near-net shape compo-nents, avoiding most of the costly machiningsteps. More impressively, powder technologyis also used to produce massive components.For example, very high carbon/vanadiumstainless tool steel components can be made byencapsulating powder in an evacuated canisterin which it can be sintered and hot worked to100% density and virtually complete homo-geneity. Chapter 9 on martensitic alloys dis-cusses these materials.

Hot Rolling

Hot rolling remains an essential process forthe vast majority of stainless steel used. Hotrolling characteristics of stainless steels varygreatly. Ferritic stainless steels are extremelyeasy to hot roll since they have a soft, single-phase structure at hot rolling temperatures.Martensitic stainless steels roll like their carbonand alloy steel counterparts since their mi-crostructure during hot rolling is a moderatelyalloyed austenite similar to alloy steels. The mi-crostructure during hot rolling is the crucial fac-tor. Austenitic stainless steels have high strengthat hot rolling temperatures. Furthermore, thelow diffusion rates in austenite slow recrystal-lization so that the steel does not always softenbetween stands in tandem mills. This increasesmill loads, and lower reductions must be takenthan for alloy steels. Powerful hot strip tandem

mills that routinely roll carbon steel to 1.5 mm(0.06 in.) can struggle to attain 4.5-mm (0.18-in.)thickness for 316 stainless.

The high separating forces on the hot rollingmill stands also cause greater roll deflection andcompression, which if not countered by rollbending or roll shifting schemes can lead to sig-nificant variation in thickness across the sheet,as much as 0.25 mm (0.01 in.). This variation asa percentage of thickness is not reduced by coldrolling and is a major cause of tolerance loss insheet and strip. Hot-rolled bands vary in thick-ness along the length of the coil because the tailend of the slab is colder and harder to roll. Coilboxes (on reversing mills) address this problemto a degree by permitting the semirolled coil toequalize in temperature.

Hot strip tandem mills powerful enough tosuccessfully roll high-quality stainless steel hot-rolled bands are massively expensive and areseldom justified for the tonnage of stainless steelrolling a given melt shop produces, althoughrolling stainless on hot tandem mills used pri-marily for carbon steel can be an excellent pro-duction method.

Hot Steckel mills have become the favoredmethod of hot rolling stainless steel becausetheir throughput better matches stainless steelmelt shop production outputs. This permits themelt shop and caster to be adjacent to the hotmill, which permits energy-saving hot chargingof slabs. In hot Steckel mills, typically a four-high reversing rougher rolls slabs to about 3-cm(1.2-in.) thick. Then, the transfer band is rolledto final gauge on a separate reversing four-highfinishing mill with coil boxes to preserve tem-perature. The economy of having only two millstands makes these mills ideal for typical stain-less production quantities and permits the costof sophisticated mill capabilities, such as rollshifting, roll crossing, or roll bending, not tohave to be duplicated among many stands. Thisis the same justification for using Sendzimirmills to cold roll stainless. In both cases, thelogic applies more to austenitic alloys than tothe easily rolled ferritic stainless alloys.

In either case, the hot-rolled band carries aheavy, embedded scale that must be removedfrom the surface before further processing inmost cases. Some alloys can be cold rolled in the“black band” state at a cost of coarser surfacefinish and greater rolling mill roll wear. If nor-mal cold rolling or use as hot-rolled coil is fore-seen, the hot-rolled band must then be annealedand pickled since the as-rolled hot-rolled band


has poor corrosion resistance, poor mechanicalproperties, residual cold work and hardness vari-ations, as well as a heavy oxide layer.

Defects

Stainless steel hot-rolled bands can containmany types of defects. These are seldom seenby the end user because they are removedwhen they are not prevented. They do haverepercussions on delivery. The major cate-gories are:

• Hot mill defects• Inclusion-related defects• Hot ductility-related defects

Stainless steel is less forgiving of hot millfaults because its surface is not removed by oxi-dation to the degree carbon steel’s surface is.Thus, a skid mark from a slab-heating furnacewill remain through the hot rolling, annealing,and cold rolling processes. This is true of all hotmill scratches, gouges, digs, etc. Rolling stain-less requires a different mindset than rollingcarbon steel, which argues against the benefitsof rolling stainless on a mill built and primarilyused for carbon steel.

Inclusion-related defects are all essentiallyavoidable by using state-of-the-art technology.Protecting metal from reoxidation and keepingprecise mold-level control in the continuouscaster prevents all inclusions of a size that canproduce a defect.

Hot ductility defects are more subtle. Theyarise from many causes and are manifest prima-

rily as edge cracks and slivers. Edge cracks aresimply a lack of ductility at the colder strip edge.Stainless hot ductility often has a narrow tem-perature window, and many factors can affectthe size of that window depending on alloy type.The most inherently challenging alloys for hotworking are the duplex alloys and the alloys thatsolidify in the fully austenitic state. The formerhas a mixed-phase structure, and the phases canexhibit mechanical incompatibility at certaintemperatures. The latter alloys reject sulfur andoxygen during solidification and slab reheatingto the grain boundaries, where they form veryweak films. But, even alloys such as 304 and316 can have very poor hot ductility if they con-tain much sulfur and oxygen or if they are re-heated for long times or at temperatures above1250 °C (2280 °F), which facilitates diffusion ofsulfur and oxygen to the grain boundaries andalso encourages very large grains. This poor hotductility manifests itself as “slivers,” which canrequire grinding of the entire hot band surface.These tendencies are fought by low oxygen andsulfur levels and minimal slab-reheating temper-atures and times, as well as slab surface workingin the caster pinch rolls. Sometimes, very poorhot working alloys are given a single hot reduc-tion pass on a hot mill to produce a full recrystal-lization that disperses grain boundary-weakeningelements on subsequent reheat.

REFERENCE

1. D. Peckner and I.M. Bernstein, Handbook ofStainless Steels, McGraw-Hill, 1977, p 3–13

Summary

THE THERMAL PROCESSING of stainlesssteel is a topic the end user should approachwith great respect. It is not simple in concept orin practice. Before attempting to carry out anythermal processing on stainless steel, the practi-tioner must understand the alloy design, compo-sition, and processing history of the material inquestion. The thermal processing then must bedesigned and executed in a planned, controlledmanner. The consequences of failure in thermalprocessing can become catastrophic to mechan-ical properties and corrosion resistance.

Introduction

The thermal processing of stainless steels canhave many purposes. Normally, the objectivesare simple: heating for hot working, annealing tosoften after cold working, solution annealing to homogenize, heating to temper martensite, orto stress relieve. However, even if the objectiveis simple, the processes that occur are anythingbut simple. Variations in temperature, times attemperature, heating and cooling rates, and at-mosphere can have complex and easily unin-tended consequences. There is no substitute forunderstanding the processes that are occurringwhen stainless steels are heated for successfulheat treating to be achieved.

Stainless steels have many alloying elementsin large amounts. Many of these elements arehighly reactive thermodynamically. The practi-cal consequence of this is that many phases arethermodynamically possible at different temper-atures. Stainless also reacts with its environmentat high temperatures, causing changes in surfacealloy content. Some of the resulting phases are

desirable, and some are potentially very detri-mental. Readers are encouraged to review theearlier chapters on phases in stainless steel(Chap. 6–10) to familiarize themselves withthese phases.

Each of the alloy groups of stainless steels hasradically different thermal processing objectivesand requirements; therefore, each is discussedseparately.

Austenitic Stainless Steels

Thermal processes applied to austenitic stain-less steels include:

• Soaking for homogenization and preparationfor hot working

• Annealing to remove the effects of coldwork and to put alloying elements into solidsolution (solution annealing)

• Stress relieving

The temperatures at which these processesare carried out are shown in Table 1 for typicalaustenite compositions.

Soaking

Because virtually all stainless steel is continu-ously cast, the older soaking process of holdingingots in soaking pits for many hours is rarelyused. The soaking had two functions. The obvi-ous one was to equilibrate at the right tempera-ture for hot working. The second, less-obvious,one was to achieve greater chemical homogene-ity. The lack of homogeneity comes from thesolute segregation that occurs as solute ele-ments are rejected from the material that wasfirst to freeze. Solute segregation was exagger-ated by the slow solidification of ingots, andcontinuous casting helped make the stainless

CHAPTER 13

Thermal Processing




steel more homogeneous. Nevertheless, castslabs and blooms must be soaked to eliminateas-cast segregations. This process, to the extentit is done, occurs as they are reheated to the ap-propriate temperature for hot working. Soakingdissolves the few percent of residual delta fer-rite that are present on slab solidification. It isimportant to soak at the highest temperature atwhich delta ferrite is not a stable phase so that itwill dissolve, about 1250 °C (2280 °F) for mostaustenitic stainless steels.

Soaking at higher temperatures causes ferritelevels to increase, negating the homogenizationand causing very poor hot workability. Longertimes at temperature than the minimum requiredfor thermal uniformity also cause problems asany sulfur and oxygen impurities are rejectedfrom austenite and can diffuse to grain bound-aries, where they form weak, plastic films thatalso degrade hot workability. Grain growth, byreducing grain boundary area, exacerbates thiseffect. Thus, soaking times are best minimizedand closely controlled. Alloys are therefore de-signed to have only a slight amount of delta fer-rite to be redissolved during soaking. Ferrite isuseful because it has a high solubility for oxygenand sulfur. Having none would result in impurityrejection of these elements to grain boundariesduring solidification, which results in the worst-possible hot working characteristics. The oxygenand sulfur trapped in the ferrite during solidifica-tion precipitates in the solid state as inclusions,which also must be equilibrated with the sur-

rounding matrix by sufficient soaking. Welds thatare unannealed have such precipitated inclusionsin an unequilibrated state, and the result is dimin-ished chromium concentration and poorer pittingresistance.

Annealing

Annealing serves two main functions instainless steel: It removes the effect of coldwork by replacing strained microstructure withnew strain-free grains, that is, recrystallization.New grains nucleate and grow. If stored strainenergy is insufficient, as happens often withferritic stainless steels, true recrystallization isdifficult to achieve, and the annealing processmay only produce recovery without recrystal-lization, leaving the same grains relieved ofstress. This leaves the surviving grains with thesame crystallographic orientation that deforma-tion produced and may or may not be the de-sired outcome. Second, annealing returns intosolution solute that has been precipitated as un-wanted phases, principally carbides, but alsointermetallic phases. Annealing also may helpto reduce solute segregation remaining fromthe casting process, making the compositionmore homogeneous. The homogenizationprocess is accelerated by the reduction in di-mensions from hot and cold working. A reduc-tion in dimension by a factor of two reduces thetime to achieve a given degree of homogeniza-tion by a factor of four.

Table 1 Recommended thermal processing temperatures for austenitic alloysAnnealing Annealing ASTM A480(a) Stress Stress

Alloy temperature, °C temperature, °F 2006, °F relieving, °F relieving, oC

Standard alloys . . . . . . . . . 1500–1600 non-L, 815–870 non-L, 1000–1600 L grades 540–870 L grades

201, 202, 201LN 1010–1120 1850–2050 1900 min . . . . . .301, 301LN, all versions 1010–1120 1850–2050 1900 min . . . . . .304, 304L, 305, all versions 1010–1120 1850–2050 1900 min . . . . . .316, 316L, 316N, 317, 317L 1040–1175 1900–2150 1900 min . . . . . .308, 309, 309S, 310, 310S 1040–1175 1900–2150 1900 min . . . . . .Stabilized alloys . . . . . . . . . 1000–1600 540–870321 955–1065 1750–1950 1900 min . . . . . .347, 348 980–1025 1800–1950 1900 min . . . . . .20Cb-3 925–955 1700–1750 . . . 925–1010 . . .Moderately alloyed, Creq<26 1120–1175 2050–2150 Various 1500–1600 815–870S31725, N08028, JS700Highly alloyed, lower sigma 1120–1175 2050–2250 Various Not recommended . . .

alloys, Creq>30, high NAL6-XN, 4565, 654SMO,

254SMOHighly alloyed, sigma-prone 1205–1230 2200–2250 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.

Chapter 13: Thermal Processing / 163

Annealing to recrystallize is fairly rapid. To afirst approximation, it is instantaneous, and theresults are merely a function of the maximumtemperature attained. This may not be the casefor continuous annealing lines, in which transittime can be short enough, less than a minute attemperature, to limit the grain size attained. Thedriving force for recrystallization is the strainenergy stored in the lattice from deformation.The strain energy in a given material is propor-tional to the square of the material’s flow stress.As the material is heated, recovery occurs first.This is the change in physical and mechanicalproperties associated with dislocation annihila-tion and polygonalization that occurs before thenucleation and growth of new grains.

The nucleation of new grains occurs at high-angle grain boundaries and proceeds by themovement of roughly hemispherical growthfronts into strained areas. The percentage re-crystallized, once a sufficient temperature isreached, grows sigmoidally. It is normal forthe time to fully recrystallize to be rather lessthan the time to attain that temperature. Evenat the lower range of annealing temperatures,times are generally less than 1 min. Recrystal-lization will not occur if the lattice contains in-sufficient strain energy. Thirty percent coldwork should be used as a rough threshold forthe required amount. Annealing after loweramounts of cold work is characterized byscarce nucleation sites and can result in verylarge and nonuniform grain size. Hot-workedmaterial often has a composite structure thatmay have already had some recrystallizationdepending on the final reduction temperature.Annealing may not produce a clear recrystal-lized structure in this case.

The relative rapidity of recrystallization an-nealing is due to the fact that it is rate con-trolled by short-range diffusion. Solution an-nealing requires longer-range diffusion andthus can require much longer times. Some stud-ies have shown that welds, for instance, do notrecover completely from their loss of corrosionproperties that arise from local alloy depletionunless they have been annealed for times on theorder of 1 h (Ref 1). Others have seen homoge-nization in as little as 10 min (Ref 2). Wroughtmaterials can require much shorter times be-cause reductions during hot working have re-duced diffusion distances. It should be notedthat precipitates can be redissolved and not ap-parent in the annealed microstructure withoutfull homogeneity being achieved. For example,

carbides can be redissolved and carbon dif-fused away from the carbide, but this does notmean that all composition gradients have beenreduced to zero. This means that precipitatesmay re-form more rapidly in such a materialthan they would in a completely homogeneousalloy.

The annealing temperature for a given alloy ischosen based on the temperature required to putall alloying elements into solution. Higher car-bon levels, for instance, require higher tempera-tures to dissolve all the carbon. This is one ofthe principle values of accurate phase diagrams.If it were simply a consideration of recrystal-lization, all alloys could be annealed at similartemperatures at the low end of the recom-mended range. Within the recommended range,the temperature selected should be determinedby the desired grain size. End use determineswhether a fine or coarse grain size is preferable.Table 1 lists recommended annealing tempera-tures for austenitic stainless steels.

The overall interplay between prior cold workand annealing temperature on mechanical prop-erties of annealed material can be summarizedas (Ref 3):

• Grain size of a given alloy is the most im-portant parameter in characterizing mechan-ical properties.

• Yield and tensile strength are essentiallyconstant for a given grain size regardless ofthe amount of prior cold work; however, theelongation depends on the prior reduction.

• Yield strength, tensile strength, and hardnessare essentially linear functions of grain size.

• Elongation decreases with finer grain sizeand at an increasing rate as grain size be-comes finer as long as cross-section size ofthe test specimen is not extremely small.This is not true for very coarse-grained ma-terial.

• Maximizing elongation comes from maxi-mum prior cold work and medium annealingtemperatures

• Anisotropy coefficients, or plastic strain ra-tios r are constant up to about 40% reductionafter, which r45 and rn increase sharply, whilert decreases. This leads to earing duringdrawing.

• The increase in properties for a one ASTMgrain size increment is:a. 13 MPa (2 ksi) for tensile strength b. 20 MPa (3 ksi) for yield strengthc. 2 HRB for hardness


Atmospheres for annealing are important.Austenitic stainless steels heated in air, ofcourse, form oxide scales. Beneath this oxide,the metal matrix becomes significantly depletedof chromium (Ref 4), often more than 5% lowerin chromium and to a depth of as much as 10 µ(395 µin.). So, not only must any oxide be re-moved, so must the chromium-depleted layer.This requires aggressive pickling, which whiledone commonly, may not be practical for manystainless users. The chromium-depleted zone,however, does pickle rapidly precisely becauseit does have less chromium. To avoid oxidescale formation, vacuum, hydrogen, or inert gasatmospheres may be used.

If vacuum is used, it should be less than 2 × 10–3 torr (0.3 Pa). If an inert gas or hydrogenis used, the key consideration is moisture con-tent. The dew point must be –40 °C (–40 °F) orlower. More stringent levels may be required ifmirror finishes are desired. Cool down must berapid as oxidation potential increases as tempera-ture decreases. Vacuum or inert gas is preferableto hydrogen for alloys containing stable oxideformers such as aluminum or titanium or for al-loys containing boron.

Austenitic alloys that are subject to sensitiza-tion must be cooled rapidly enough from anneal-ing temperatures to avoid carbide precipitationduring cooling. If forced air or water quenchingare impractical or if section size prohibits rapidcooling, then using stabilized or low-carbongrades is indicated.

Superaustenitic stainless steels, and even al-loys like 317, present a special problem becausethese alloys have significant sigma-forming ten-dencies. Sigma forms initially because solidifi-cation segregation causes local enrichment ofsigma-promoting elements, such as molybde-num. It can also form from slow cooling of slabsor hot bands. This latter sigma forms at grainboundaries and will cause embrittlement and re-duced corrosion resistance, so it must not onlybe redissolved, but also the alloy must be ho-mogenized to remove the residual concentrationgradients from the sigma. If this is not done,chromium- and molybdenum-depleted regionswill still exist, and sigma will re-form muchmore rapidly during subsequent exposure tohigh temperatures. For this reason, the higherends of the annealing ranges are recommended,and annealing times should be generous. Neweralloys have higher nitrogen contents to suppressformation of sigma and other deleterious inter-

metallic phases. Use of high-chromium and-molybdenum alloys without enhanced nitrogenis no longer recommended, and the use of lower-nitrogen alloys should be reexamined and ques-tioned if specified.

Last, stainless surfaces should be scrupu-lously clean before annealing. Even hard waterdeposits can cause differential oxide growth,which can cause etched spots on the surface,where the postanneal pickling attacks the differ-ent oxide more strongly. Carbonaceous materi-als left on the surface are even more objection-able because they can cause carburization andsubsequent loss of corrosion resistance.

Stabilizing anneals are sometimes conductedon stabilized alloys such as 321 and 347. This isuseful when carbon levels are sufficiently highthat significant dissociation of carbides occursat annealing temperatures. A second anneal atlower temperature, about 900 °C (1650 °F),then is done to permit the carbon to combinewith the stabilizing element rather than leavingit available to form chromium carbides. Currentpreferred practice for these alloys is to maintaincarbon and nitrogen below 0.03% for corrosion-resistant service, which renders this stabilizingunnecessary. Alloys used for high-temperatureservice benefit from the creep-resisting contri-butions of higher carbon levels.

Stress Relieving

Austenitic stainless steel weldments often con-tain residual stresses, which can cause distortionor lead to stress corrosion cracking in service.They are commonly stress relieved at tempera-tures slightly below the annealing temperature,so that residual stresses may be relieved bycreep. One hour at 900 °C (1650 °F) reducesresidual stress by about 85%. Lower tempera-tures require exponentially longer times for thesame stress relief, with times doubling for each100 °C (180 °F) decrement as decreasing diffu-sion rates, which govern creep, are encountered.

Cold-worked austenitic stainless steels have amarkedly diminished proportional limit, partic-ularly in compression. This Bauschinger effect,which arises from the easy mobility of disloca-tions, can be eliminated by stress relieving ataround 350 °C (660 °F) for 2 h, which providesthe thermal energy for dislocation interactionsto lock into place. This produces a sharp yieldpoint without premature nonproportional elasticdeformation.


Ferritic Stainless Steels

Ferritic stainless steels, from an annealingpoint of view, must be discussed in two cate-gories. First are the modern, stabilized alloys,which are ferritic at all temperatures. These al-loys behave as interstitial-free (IF) alloys be-cause the interstitial carbon and nitrogen are re-moved from solution as a stable precipitate. Inthe second category are the older ferritic steels,which have enough austenitizing elements, usu-ally carbon, in solution to cause them to formaustenite at what would otherwise be a good an-nealing temperature. This makes them trulyquasi-martensitic alloys, and they must betreated accordingly. Table 2, which lists heat-treating temperatures for ferritic stainless alloys,also shows which grades fit into which category.

Soaking

Heating of ferritic stainless for hot working isstraightforward. Whether stabilized or not,these alloys are heated to the 1000 to 1100 °C(1830 to 2010 °F) range for hot working. Thesuperferritics can be heated to up to 1300 °C(2370 °F). At this temperature, no debilitatingphases occur, and ductility is good. The highdiffusion rate inherent to the ferritic structuremakes homogenization easy. As long as hotworking is completed at temperatures abovethat at which austenite forms, good hot ductilityis expected. This is not a concern with IF alloys,which do not form austenite.

Annealing

The IF ferritics do not undergo any phasechange with temperature during the course ofproperly executed heat treatment. The objectiveof annealing is generally simply to remove theeffects of cold work. This is because they do notneed to have carbon put into solution and, ex-cept in rare cases, do not have intermetallic

phases that require dissolution. Alloys with highchromium and molybdenum contents can formσ and/or α' , the brittle, ordered body-centeredcubic (bcc) phase, at temperatures below an-nealing temperatures, so rapid cooling is pru-dent when chromium plus molybdenum ex-ceeds 20%.

The driving force for recrystallization in thesealloys is limited by the lower stored energyfrom deformation inherent in the bcc structure.In addition, the pronounced deformation textureleads to annealing responses that are more accu-rately characterized as recovery and graingrowth with diminished recrystallization. Thesealloys retain this texture after annealing, andthis characteristic anisotropy is exploited forgood drawability. The major concern is to avoidexcessive annealed grain size, which greatly re-duces toughness. Anneal at the higher end of therange only if the loss of toughness associatedwith large grain size is not a concern. Stabiliz-ing anneals are normally unnecessary for stabi-lized ferritics as their high diffusion rates ensurefreedom from knife-edge attack due to sensiti-zation from free unbonded carbon combiningwith chromium at grain boundaries. The stabi-lizing additions of titanium and/or niobium tieup the carbon as stable TiC or NbC, which doesnot redissolve during annealing.

The interstitial-bearing ferritic stainlesssteels must be annealed subcritically, or the for-mation of austenite at higher temperatureswould make martensite formation on coolingvirtually unavoidable. Thus, a typical primaryanneal cycle for a typical alloy such as 430would be nearly 24 h at 750 °C (1380 °F), themajority of which is thermal equilibration ofthe large coil mass. The actual time at tempera-ture required is less than 1 h. Continuous an-nealing is not practical because the diffusion ofcarbon is too slow to occur in the dwell time attemperature typical in continuous annealinglines. This cycle also precipitates essentially allthe carbon and nitrogen as mixed Cr/Fe car-bides and nitrides and homogenizes chromiumcontent. This necessarily slow process permitssubsequent subcritical annealing for mechani-cal properties (to alleviate the effects of coldwork) to be done in a few minutes since carbonhas been eliminated from solution by the for-mation of fairly stable carbides. Since the ma-terial is generally purchased in the annealedcondition, the user need never be concernedwith such lengthy anneals.

Table 2 Recommended annealing temperaturesfor ferritic alloys

Annealing AnnealingAlloy temperature, oC temperature, oF

Stabilized, Cr+Mo<20 409, 439,18 SR 870–925 1600–1700

Unstabilized, Cr+Mo<20405, 430, 434, 436 705–790 1300–1450

Stabilized, Cr+Mo>2029-4C, Monit, Seacure, 444 1010–1065 1850–1950

Unstabilized, Cr+Mo>20,446 760–830 1400–1525


Stress relieving is rarely a concern for anytype of ferritic stainless. Unstabilized gradesshould not be welded, and if they are, full sub-critical annealing is required. Stabilized gradeshave no need for postweld heat treatment. Low-temperature heat treatment runs the risk of α'formation and is best avoided.

Martensitic Stainless Steels

The martensitic stainless steels resemble theunstabilized ferritic stainless steels described.The martensitic stainless steels form essentially100% austenite on heating and have very highhardenability, so their ability to be softened byannealing is limited. The traditional martensiticstainless steels are iron/chromium/carbon al-loys, sometimes with a small amount of nickeland/or molybdenum. More recently, alloys havebeen developed for petroleum applications thatcontain high copper, nickel, and/or molybde-num and low carbon. The principles of heattreatment of the two alloy categories are thesame. The more highly alloyed newer alloysare, in fact, simpler to heat treat because theirlow carbon and nitrogen levels alleviate theneed to temper.

Soaking

Hot working should be carried out in theaustenitic range. Temperatures for this arelisted in Table 3. Forging and hot workingshould always be followed by annealing toavoid stress cracking due to the deep hardeningof these alloys.

Annealing

Martensitic stainless steels can be annealedby subcritical anneal and sometimes by full an-neal depending on alloy level. If the alloy levelis such, as in the nickel-containing grades, that

martensite cannot be avoided by furnace cool-ing from austenitic temperatures, then only sub-critical annealing is feasible. But, even fornickel-free alloys the hardenability is so greatthat annealing by slow cooling is quite difficult.Martensitic alloys are put into the annealed con-dition for processing before they are quenchedand tempered for their final use. Thus, the moreeconomic subcritical anneal is the predominantannealing heat treatment.

The nickel-bearing alloys have such highhardenability that annealing in the critical rangecannot produce softening by any practical cool-ing rate, so subcritical annealing is always rec-ommended for these alloys. Nickel reduces thetemperature at which austenite is stable asshown in Chapter 9, Fig. 9. Other additions likevanadium, molybdenum, and tungsten promotesecondary hardening and tempering resistance,and subcritical annealing of these alloys be-comes a slow, difficult process. This is a charac-teristic of the so-called super 12Cr alloys.

Martensitic alloys have lower corrosion re-sistance in the annealed condition than in thehardened condition because in this state theyhave the maximum amount of chromium tied upas chromium carbide.

Austenitizing

Table 3 lists the austenitizing and temperingranges for martensitic stainless steels. Fullaustenitizing is crucial to producing a micro-structure that is fully martensitic. Only austenitetransforms to martensite. If other constituents,such as δ ferrite or carbides, exist during theaustenitizing heat treatment before quenching,they will not transform to martensite. Some alloys, such as the 440 group, have enough car-bon that the austenitizing temperature deter-mines how much carbon is put into solution.The carbon in solution in the austenite will be-come the carbon level in the martensite, whichdirectly determines strength and corrosion

Table 3 Recommended annealing, austenitizing, and tempering temperatures for martensitic alloysAlloy Subcritical anneal, Full anneal, Austenitizing, Tempering, low Tempering, high

oC (oF) oC (oF) oC (oF) range, oC (oF) range, oC (oF)

Straight 650–760 830–885 925–1010 205–370 400–700Cr, C<0.20, 410, (1200–1400) (1525–1625) (1700–1850) (565–605) (1050–1125)

416,403Ni/Mo, C<0.20, 620–705 Not recommended 980–1065 205–370 400–700

414, 431, 415, (1150–1300) (1800–1950) (565–605) (1050–1125)425, C>0.20

440A,B,C F, 420 675–760 845–900 1010–1065 150–300 Not recommended(1245–1400) (1555–1650) (1850–1950) (300–700)


resistance because undissolved carbides containchromium, which diminishes that available forcorrosion resistance. Austenitizing temperatureand holding time become most significant whencarbon exceeds 0.20%, where its solubility is asteep function of temperature.

The � ferrite is a generally undesirable phasethat can be produced by temperature excursionsor composition variations. Excessive austenitiz-ing temperatures can cause its formation, as canlow levels of carbon, which may be originallypresent in the alloy or arise from decarburiza-tion. It will cause lower hardness and toughnessif present.

Heating rates should be such that a uniformtemperature is attained before the allotropictransformation from bcc to face-centered cubic(Fcc), which involves a more than 1% linear di-mension change and can cause distortion orcracking. Oxidation during austenitizing cancause serious carbon loss on the surface, whichwill result in serious loss of surface hardness.Heating 410 in air for 10 min at 1100 °C (2010°F) can cause surface carbon to decrease byone-half, lowering hardness from HRC 45 tounder 20.

Quenching rate is not a significant issue forthe martensitic stainless steels since they havesuch high hardenability, but some, especiallythose with higher carbon levels, may have re-tained austenite, which can lower hardness andcause problems with dimensional stability. Thequenching rate must be sufficient, however, toavoid precipitation of carbides in the austeniteduring quenching since the sensitization wouldpersist in the final microstructure. If this oc-curs, a subzero treatment at below –75 °C(–100 °F) should be undertaken immediatelyafter quenching.

Tempering

Untempered martensite has insufficient tough-ness to be a useful engineering material. Duringtempering, carbon is precipitated from the super-saturated state it is in when it is quenched intothe bcc martensite structure during the diffusion-less transformation. The strain energy associatedwith the lattice strain of the bcc martensitecaused by the poor fit of the carbon in the tetrag-onal interstices is very large. Heating to evenlow temperatures allows carbon enough mobil-ity to diffuse and precipitate as carbide. Sincecarbon diffuses 106 times as fast as iron,chromium, or other carbide formers, it tends to

precipitate with the abundant iron atoms first. Athigher temperatures and longer times, more ther-modynamically stable carbides, such as Cr23C6,form. Carbide formation is a complex functionof temperature, time, and composition. Thegrowth of carbides reduces strain and hardness.There are exceptions, such as the precipitation ofMo2C, whose morphology produces a precipita-tion hardening (PH), called secondary harden-ing. Niobium and vanadium also form carbidesthat result in higher hardness at all temperingtemperatures. Had this been understood earlyon, these steels could have been correctly in-cluded in the PH stainless group. The PH steels,AM-350 and AM-355, both derive their PHfrom the precipitation of Mo2C and Mo2N. In allother cases, higher tempering temperatures leadto lower hardness.

The nickel-bearing alloys have a restrictedupper tempering temperature because of thedanger of re-forming austenite, which wouldthen transform to untempered martensite duringcooling, requiring a second tempering operation.

Intermediate temperatures can lead to thephenomenon of temper embrittlement, which iscaused by the precipitation of phosphorus andother species, such as, but not limited to, car-bides, at prior austenitic grain boundaries. Thisphenomenon is distinct from the precipitation ofα' , which causes the so-called 475 °C embrittle-ment, which occurs more severely in alloyswith higher chromium levels. Because the for-mation of martensite is diffusionless, theaustenite boundaries maintain the microcompo-sitions they have at high temperatures. Auste-nite has low solubility for impurities such asphosphorous, so phosphorous is more highlyconcentrated in the grain boundary regions.This compositional inhomogeneity can be elim-inated by higher tempering temperatures or bythe addition of molybdenum, which combineswith the phosphorus and prevents the embrittle-ment. The existence of temper embrittlement isthe reason that Table 3 recommends avoidingcertain temperature ranges for tempering.

Lower tempering temperatures and higheraustenitizing temperatures are best for corrosionresistance because both minimize the amount ofchromium tied up as carbide. Quenching fromhigher temperatures also enhances toughness.

Stress Relieving

If quenched martensitic steels are not imme-diately tempered, then they should be promptly


stress relieved. Otherwise, the residual stressfrom quenching could result in stress corrosioncracking even in seemingly benign environ-ments. Stress relieving is simply a low-tempera-ture tempering operation, but some eliminationof residual stress does occur. Higher tempera-tures and longer times produce greater stress relief and maximize elastic properties, but opti-mal toughness is obtained at intermediate timesand temperatures.

Precipitation Hardening Stainless Steels

There are three classes of PH steels, and eachrequires totally different heat treatment. Theclasses are martensitic, semiaustenitic, andaustenitic. The most straightforward alloys arethe martensitic PH grades. Like the plain marten-sitic alloys, the martensitic PH alloys are care-fully designed to produce a nearly fully austeniticstructure at high temperature that quenches to anearly fully martensitic structure on cooling. Themartensite is low in carbon, so it is relatively softand not prone to brittleness. So, the hardness andstrength of these alloys is derived from a subse-quent tempering-type heat treatment duringwhich various constituent elements form ex-tremely fine coherent precipitates that greatlystrain and therefore strengthen the matrix. Thereare numerous precipitates that can provoke thiseffect, and they are described in detail in the PHchapter (see Chapter 10). All require the short-range diffusion of substitution elements to formthese optically invisible precipitates.

Martensitic PH Grades

Solution treatment of these alloys is con-ducted to achieve a fully austenitic structure.The constituent elements are easily dissolved,so excessive temperature or time is unnecessaryand could be counterproductive if it were to re-sult in ferrite formation or surface oxidation,which would be detrimental to final mechanicalproperties. The presence of retained ferrite ismainly a function of alloy and compositionwithin the allowed range. Earlier grades such as17–4 and the obsolete stainless W intrinsicallycontained some ferrite. The subsequent alloysare substantially ferrite free. Most alloys mayretain some austenite after quenching to roomtemperature, in which case subzero treatmentshould be done within 24 h to avoid further sta-bilizing the austenite. Subzero treatment, by

eliminating retained austenite, enhances dimen-sional stability but diminishes toughness. Table 4lists the solution treatments for all PH grades.

The as-quenched state is called condition A.This is the normal condition in which the mate-rial is supplied from the mill and is intended tobe soft enough for machining and some forming.If softer material is required, the H-1150M con-dition exists in which the material is first highlyoveraged at 760 °C (1400 °F), allowing someaustenite to re-form. The subsequent aging thenoverages that martensite while retaining somestable austenite. The result is a very tough microstructure.

Aging. The time and temperatures requiredto produce this precipitation are also given inTable 4. The condition code itself tells the tem-perature at which the aging is conducted inthat the code numbers are based on thermalprocessing temperatures expressed in degreesFahrenheit (oF), for example, TH 900 meanstransformed to martensite (T) and aged at 900degrees Fahrenheit. The final properties are afunction of both aging time and temperature.Lower temperatures result in higher possiblehardness but lower toughness. The precipi-tates, as mentioned, are optically invisible andcause very little dimensional change. Contrac-tion on the order of 0.0005 in./in. from aging istypical, often permitting machining to final di-mensions in condition A. All aging treatmentsare above the temper embrittlement range towhich these alloys are susceptible. Servicetemperatures in this range would result in em-brittlement, so use above 350 °C (660 °F)should be avoided. Molybdenum-bearinggrades should be selected to minimize thisphenomenon if high-temperature use is con-templated.

Solution Heat Treatment and Conditioning.The semiaustenitic grades are more complicatedthan the martensitic PH alloys. These alloys aredesigned so that they are austenitic whenquenched from the solution heat treatment tem-perature. This is also called condition A, and itpermits them to be highly formable. This stabi-lization of the austenite comes mainly fromhigher chromium and carbon levels. These al-loys essentially resemble a lean 301 austeniticalloy, many with some molybdenum substitut-ing for part of the chromium. The key to thesegrades is making them behave as a martensiticalloy. This is done by precipitating some of thecarbon as chromium carbide at a temperature atthe high end of what would normally be consid-


ered sensitization, 760 °C (1400 °F). The car-bide precipitation occurs at the interfaces of thesmall amount of residual ferrite these alloyshave and also at grain boundaries. The deletionof carbon and chromium from the matrixchanges the matrix composition sufficiently thatthe temperature for the start of the martensitictransformation (Ms) of the depleted austenite in-creases from below zero to about 65 to 100 °C(18 to 212 °F). The martensitic transformation

finishes (Mf) near room temperature. Thisprocess is called austenite conditioning. Theheat treatment scheme just described would becondition A-1400. This material, after quench-ing to room temperature, would be said to becondition T.

The higher the temperature of the condition-ing, the less carbon is precipitated and the lowerthe resulting Ms. The highest conditioning tem-peratures of 955 °C (1750 °F) cause a sufficiently

Table 4 Recommended annealing and stress-relieving temperatures for martensitic PH gradesAlloy Condition code Solution anneal Conditioning Aging

Martensitics13-8 A 925 °C 15 min, oil or air cool below 15 °C –75 °C 8 h

RHxxx 925 °C 15 min, oil or air cool below 15 °C –75 °C 8 h xxx °F for 4 hHxxx 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 –75 °C 8 hHxxx 980 °C for 1 h, cool rapidly –75 °C 8 h xxx °F for 4 h

475 A 925 °C 1 h, air cool –75 °C 8 hHxxx 925 °C 1 h, air cool –75 °C 8 h xxx °F for 4 h

Semiaustenitics17-7, 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 –75 °C for 8 hTHxxx 1065 °C for 30 min, air cool 760 °C (1400 °F) xxx °F for 4 h

90 min, air cool to RT for 30 minCHxxx 1065 °C for 30 min, air cool Cold reduce xxx °F for 4 hRHxxx 1065 °C for 30 min, air cool 955 °C (1750 °F) 10 min, air xxx °F for 4 h

cool, chill to –75 °C for 8 hAM-350 A 1010–1065 °C . . .

L (equivalent to T) 1010–1065 °C 930 °C for 90 min, air cool . . .SC (equivalent to R) 1010–1065 °C 930 °C for 90 min, air cool, . . .

180 min at –75 °CSCTxxx 850 °F or 1010–1065 °C 930 °C for 90 min, air cool, xxx °F for 180 min1000 °F 180 min at –75 °C

DA (double aged) . . . 930 °C for 90 min, air cool, 450–540 °C 180 min730–760 °C 180 min

Am-355 A 1025–1040 °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 –75 °CSCTxxx 850 °F or 930 °C for 90 min, air cool, xxx °F for 180 min1000 °F 180 min at –75 °CDA (double aged) 930 °C for 90 min, air cool. 440–470 °C

730–760 °C 180 min for 180 minEqualized and 930 °C for 90 min, 540–590 °C for overtempered air cool, 730–760 °C 180 min 180 min

AusteniticA-286 ST1650 900 °C 120 min, oil/water quench . . . . . .

ST1650A 900 °C 120 min, oil/water quench 730 °C 16 hST1650DA 900 °C 120 min, oil/water quench 730 °C 16 h,

650 °C 8 hST1800 980 °C 120 min, oil/water quenchST1800A 980 °C 120 min, oil/water quench 730 °C 16 h

PH, precipitation-hardenable; RT, room temperature.


low Ms that subzero treatment is required to obtain the fully martensitic structure required forage hardening. This would be called A-1750.After subzero treatment at –73 °C (–100 °F), itwould be called condition R-100.

It is even possible to obtain the fully marten-sitic structure by cold work from full conditionA, which is quenched from 1065 °C (1950 °F).This is called condition C and requires heavycold rolling to accomplish. Since there is no heattreatment to precipitate carbon from the austen-ite matrix, the resulting martensite is the hardest.

The aging treatments of the semiausteniticalloys are identical to those for the martensiticalloys because the treatments are standardized.The resulting mechanical properties vary in acomplex fashion with the alloy composition andthermomechanical treatment history beforeaging. The principles at work that determine themechanical properties, besides the strain in-duced by the precipitate phase, are:

• Carbon level of the quenched martensite• Amount of cold work of austenite prior to

aging

These factors do not come into play with themartensitic PH grades because all of the carbonis in solution, and they are not significantly coldworked prior to aging. The semiaustenitic PHalloys, because of the variable amount of car-bide precipitation, have an equally variablechromium content, so corrosion resistance willvary accordingly. The highest-temperature solu-tion and carbide precipitation treatments pro-vide the best corrosion resistance, as do thelowest aging temperatures.

Austenitic PH Alloys

In contrast to the martensitic and semi-austenitic PH alloys, the austenitic PH alloysare austenitic under all conditions. Theirstrengthening reaction comes from precipitationof titanium- and aluminum-nickel intermetallicphases within the austenite matrix as occurs insuperalloys. Solution treatment is very typical

for a normal solution anneal of non-PH austen-ite. The subsequent precipitation aging requireshigher temperatures and longer times becausediffusion is much slower in austenite.

Duplex Stainless Steels

Duplex stainless steels are both ferritic andaustenitic, so their heat treatment combines thesame elements and principles as their principlephases.

Soaking

Duplex alloys are multiphase at all usefulworking temperatures, making their hot worka-bility quite poor. It is extremely important todrive sulfur to the lowest possible levels, lessthan 0.001%, to achieve satisfactory hot ductil-ity. Otherwise, soaking should proceed the sameas for ferritic stainless alloys since ferrite con-stitutes the continuous phase to be worked.

Because these alloys always contain highchromium and generally high molybdenum,they should be cooled as rapidly as possiblefrom high temperatures to avoid formation ofsigma or other intermetallic phases.

Annealing

The function of annealing in the duplex alloysis generally to:

• Remove the effects of cold work• Restore the balance between the volume

fraction of ferrite and austenite• Achieve equilibrium composition within

both the austenite and ferrite• Dissolve unwanted intermetallic phases

The annealing range of duplex alloys is some-what restricted, approximating the overlap be-tween what each of the two constituent phaseswould be annealed at separately. Table 5 liststhe normal annealing temperatures for these al-loys. The use of nitrogen as a key alloying

Table 5 Recommended annealing and stress-relieving temperatures for duplex alloysAnnealing Annealing

Alloy temperature, °C temperature, °F ASTM A480 2006 Stress relieving, °F Stress relieving, °C

Lean duplex, Cr+Mo<23, 1010–1100 1850–2010 Various Not recommended2003-2101, 19-D-2304

Medium alloy, Cr+Mo<26, 1040–1100 1900–2010 1040 min Not recommended2205

Cr+Mo>26, 2507, 52N+, 1050–1150 1925–2100 Various Not recommendedZeron 100, 255


element has improved the annealing behavior ofthese alloys since its diffusion is quite rapid,causing ferrite-austenite equilibrium to be at-tained very rapidly. Nitrogen also hinders (for-mation and facilitates the dissolution of second-ary austenite, which can form after quenchingfrom welding temperatures and cause regions ofpoor corrosion resistance.

These alloys are not very susceptible to car-bide sensitization and normally have very lowcarbon content. Thus, the guiding principle inannealing is simply to achieve phase balanceand avoid cooling so slowly that intermetallicphases may form.

The strengthening of duplex is normallyachieved by the strong grain refinement andsolid solution hardening. No strengthening heattreatments are used. Stress relief would have tooccur at temperatures at which embrittling reac-

tions, either from �' beginning at 350 °C (660°F) or from �/�, which takes over at 600 °C(1110 °F), would occur and so is not indicatedfor these alloys.

REFERENCES

1. A. Garner, The Effects of AutogenousWelding on Pitting Corrosion in AusteniticStainless Steel, Corrosion, Vol 35 (No. 3),1979, p. 108–114

2. J.F. Grubb, personal communication, June4, 2006

3. Data courtesy J&L Specialty Steel, October2, 1998

4. J.F. Grubb, Proceedings of the InternationalConference on Stainless Steels, 1991,Chiba, ISIJ

CHAPTER 14

Forming

Summary

STAINLESS STEELS RANGE in formabilityfrom the extremely formable austenitic alloys tothe hard-to-form martensitic alloys. For steelswith equivalent corrosion resistances, formabil-ity increases with the level of austenitizing al-loying elements. The ferritic alloys are the leastalloyed, least expensive, and least formable;duplex steels are next, and austenitic steels arethe most formable but most expensive. How-ever, if the specific structural anisotropy offerritic alloys, which gives particularly gooddeep-drawing characteristics for a given level ofductility, can be exploited, the best formingeconomies can be gained.

Introduction

The technology for forming stainless steel isquite similar to that for forming carbon steels.The primary difference is the higher strength ofall types of stainless steels compared to drawing-type carbon steels. This higher strength requiresgreater sophistication in tooling and lubricants,and it requires more powerful forming equip-ment. The higher initial strength is alsoaccompanied by a higher work hardening rate inaustenitic stainless steels, which further distin-guishes them from carbon steel. Galling alsomust be recognized as a danger and prevented.Stainless steel has lower thermal conductivitythan carbon steels, which can cause it to retainheat from deformation and friction, therebydecreasing lubricity. Last, stabilized stainless al-loys contain abrasive carbide microconstituents.

The various types of stainless steel have verydifferent deformation characteristics in terms ofstrain hardening and anisotropy. It will be

shown that it is important to understand andexploit these characteristics to optimize formingof stainless steels.

Flat, Rolled Stainless Steel

The vast majority of carbon steel and, espe-cially, stainless products are flat products. Thesesteels are formed by bending, roll forming, spin-ning, hydroforming, and deep drawing. Rollforming is most commonly used to producewelded pipe and tubing and is simply bendingdone in a continuous manner. Bending is a sim-ple operation, and there is no meaningful changein thickness of the sheet during the operation.The higher yield strength and work-hardeningrates of most stainless steels will result in greaterspringback than would be experienced in carbonsteel. Tooling must be adjusted to compensatefor this. In neither of these forming methods isthere large motion of the formed material acrossthe die surface, so lubricant is not normallyused. The reader is referred to the Forming andForging in Volume 14 of the ASM Handbook fordetailed charts on bend radii and springback re-lated to bending.

Other forming techniques employ more com-plex deformation processes. Deep drawing isthe foremost of these. Figure 1 (Ref 1) showsschematically what occurs during drawing. Around blank is held between dies over a cavity,and a punch pushes the material into the cavityto produce the part. If the dies pinch the blankto only a small degree, the process is normal, orordinary, drawing. If the dies significantlyrestrain the periphery of the blank so it cannotmove, stretch forming occurs. Material proper-ties determine whether a material is most suc-cessfully formed by stretch forming or drawing.




Stretch Forming. In stretch forming, thematerial is constrained from moving whollyinto the die. Thus, the section that enters the dieis stretched more and must become thinner.Whether the material becomes so thin locallythat it fails is governed by its work-hardening

rate. If it work hardens faster than it becomesthinner, the strain is distributed, and local fail-ure is prevented. Austenitic materials have theface-centered cubic (fcc) crystal structure withmany slip systems and low stacking fault ener-gies. This means that they can generate manycomplex arrays of tangled dislocations, whichcause strain hardening. They can also transformduring stretching to the much harder martensite,for which the deformation is greatest, againredistributing the deformation away from thepotentially thinning area. This makes austeniticstainless steels particularly suitable for stretchforming.

Deep Drawing. Deep drawing, referred tosimply as drawing, without stretching requires adifferent material characteristic. For drawing, alow work-hardening rate is desirable so that thematerial can compress in the circumferentialdirection while elongating in the radial direction.Obviously, a high ability to elongate is alwaysuseful regardless of any other characteristic.But, ferritic material has one other advantage:Body-centered cubic (bcc) alloys have moreslip systems than fcc alloys. When bcc alloysare rolled to become flat stock, they may retaina preferred crystallographic orientation, calledtexture, as a result of the deformation. This non-random crystal structure can cause the materialto have higher strength in the through-thicknessdirection. This directional variation in proper-ties is called anisotropy.

When a material with desirable texture isstretched, it flows in the stretching direction andcontracts laterally at lower stresses than arerequired to initiate plastic flow in the through-thickness direction. As long as the work-hard-ening rate keeps the flow stress below thethrough-thickness yield strength, there will beno thinning. The geometry of deep drawingwith constraint fits such materials’ capabilities.This is why carbon steels and ferritic stainlesssteels deep draw well. If the material were con-strained from contracting while being stretched,the tensile strength would be exceeded beforethrough-thickness flow occurred and the mate-rial would fracture with little deformation. So,these materials cannot be stretch formed. Evenwithout stretching, some hold-down pressure isrequired to prevent wrinkling of the blankbefore it is pulled into the die. This is more pro-nounced for thinner blanks and for materialwith higher work-hardening rates. In manyforming situations, adjustments to the drawingprocess (i.e., hold-down force adjustment, draw

Fig. 1 Deep drawing schematic. Source: Ref 1

Chapter 14: Forming / 175

bead contour, blank size, lubrication, die radii,etc.) may be more important than material prop-erties in determining whether the desired partcan be made successfully.

The material properties that are important toformability are ductility, as measured by tensiletest elongation; the work hardening rate, whichis the instantaneous slope of the true stress, truestrain curve and is called n; and the anisotropy.The measure of anisotropy is the Lankford ratio(Ref 2), expressed as:

(Eq 1)

where R is the average strain ratio, r0 is the strainratio in the longitudinal direction, r45 is the strainratio measured at 45° to the rolling direction (ofthe sheet metal-forming operation), and r90 is thestrain ratio in the transverse direction. R deter-mines the average depth (that is, the wall height)of the deepest draw possible. When this expres-sion equals 1, then a material may be consideredisotropic, that is, the material properties are thesame for all crystallographic orientations. As afirst approximation, the Lankford ratio equalsthe ratio of the lateral strain to the through-thick-ness strain during the tensile test of a sheet spec-imen. As the value increases from 1, the drawa-bility increases because the material tends tomaintain a constant thickness while changingshape from a flat blank to a cup shape. The abil-ity to be deep drawn is measured by the limitingdrawing ratio (LDR), the ratio of the diameter ofa disc to that of the deepest cylinder into whichit can be drawn.

The ferritic stainless steels in sheet form haveLDRs of around 2.2 compared to 2.0 for 304. Foraustenitic steel, the ratio is about 1.0, while forflat-rolled carbon steel and ferritic stainless steel,it can be greater than 2.0, but values between 1.5

and 2.0 are more common. Figure 2 shows theLDR as a function of the Lankford ratio. Itshould noted that the very ductile 304 fares nobetter than carbon steel because of the advantageof a beneficial anisotropy, which the ferrous bccstructure has. Indeed, the best-performing deep-drawing stainless steels are low interstitial ferriticsteels with boron added. A number of stainlesssteels are compared in Table 1 (Ref 4).

It shows that when comparing different typesof materials, some tests are not good predictorsof deep-drawing performance. Optimizing thematerial/drawn component combination is farfrom simple, especially when other considera-tions, such as cost and material performance inservice must be factored in. End users areencouraged to deal directly with the producingmill early in the design stage of any new high-production, deep-drawn component. The pro-ducing mills, while not necessarily exhaustivesources of information, are certainly reservoirsof knowledge of current practice.

The most widely used summary of a mater-ial’s formability is contained in its forming limitdiagram (FLD), developed by Keeler and Back-ofen (Ref 5). This diagram shows the locus offailure under varying strain states. Figure 3shows a comparison of the FLDs for austeniticstainless steel and carbon steel. The FLD tells

Rr r

=+ 2r +

40 45 90

Table 1 Deep-drawing materials comparison0.2% proof Tensile Erichsen Conical cup

stress, strength, Elongation, Hardness Lankford value value, value, Steel N/mm1 N/mm2 % HV n value ro r45 r90 r̄ mm mm

YUS 190 343 497 33.8 173 0.20 1.60 1.47 2.10 1.66 9.5 26.7YUS 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 26.7SUS 430 308 472 31.8 159 0.21 0.94 0.92 l.50 1.07 8.9 28.5SUS304 281 705 64.0 172 0.44 0.91 1.19 0.83 1.03 12.5 27.0

Source: Ref 2

Fig. 2 Limiting drawing ratio variations with Lankford ratio.Source: Ref 3


the point of failure for a given sheet materialwith a given thermomechanical history over afull range of combined strain states. These dia-grams are generated by examining circle gridsprinted on material that is deformed to failure invarious modes. The single most important valueon the curve is the intersection of the curve withthe major strain axis at zero minor strain. Thiscan be used as an index of formability and is theelongation possible for plane strain conditions.This value within a given class of materials isproportional to the strain-hardening exponentbecause a higher work-hardening rate causeshigher localized resistance to thinning, which isthe precursor of failure.

There is much more variety within the fami-lies of stainless steel than within carbon steel.Figure 4 shows generalized FLDs for austenitic,ferritic, high-strength ferritic, high-strengthaustenitic, and duplex stainless steels.

Deep drawing of components is seen as a wayto obtain near-net shape. Since tooling is costly, itis necessarily a high-volume application. Quiteoften, designers push component design to thelimit of a material’s ability to be formed. Thereare various drivers that cause this. One is to elim-inate extra operations or components by consoli-dating them into one more complex deep-drawnpart. Another is to use the least-expensive alloy.In some industries, such as the household appli-ance industry, as many components as possibleare deep drawn from ferritic stainless alloys andthe more costly austenitic alloys are used onlywhen the part cannot be made from a ferritic.

There has been much research to develop fer-ritic stainless alloys with improved formability.This has been accomplished by reducing the

total interstitial content (i.e. carbon plus nitro-gen) and by thermomechanical working to givea fine-grained, fully recrystallized, yet benefi-cially anisotropic, microstructure. Figure 5shows how the FLD of an enhanced 409 ferriticstainless steel, 409 Ultra Form, compares to thealready highly evolved 409.

For austenitic stainless steel, maximumdrawability is obtained by low work-hardeningrates coupled with maximum elongation, asexemplified by 305 and high-nickel 304. Thiscomes with a cost penalty as the easiest way toimprove formability is to increase the nickellevel, although replacing the expensive nickelwith copper or manganese, as in 204Cu, hasbeen shown to be effective. Many austeniticcomponents are made by stretched deep draw-ing. In this case, the preferable alloys are theleaner austenitics, 201 and 301. These alloysform martensite more rapidly than do 304 and305 during stretching. Martensite has a 4%greater volume than the austenite from which itforms and a much greater strength. This gives201 and 301 the ability to redistribute deforma-tion from thinning areas elsewhere and stretchextensively, making them an optimal materialfor objects such as sinks.

The specific alloy composition is often finelytuned for a given part and tooling design, andsmall deviati ons can dramatically increasebreakage rates. Even such minor processchanges as blank temperature variations due toambient temperature can alter work-hardeningrates enough to cause breakage problems. Cer-tainly, this can occur when designs push the

Fig. 3 Forming limit diagram of carbon steel compared toaustenitic stainless steel. Source: Ref 3

Fig. 4 Forming limit diagrams for categories of stainlesssteels. A, austenitic stainless steel; F, ferritic stainless

steel; HAS, high-strength austenitic stainless steel; HSF, high-strength ferritic stainless steel; FA(50), ferritic-austenitic stainlesssteel. Source: Ref 3


envelope of a material’s capability, but costpressures generally drive designers to thisextreme.

On very severe forming, intermediateannealing may be required to either enhanceductility or reduce required pressure. Stainlesssteel in the as-drawn condition will have resid-ual stress and may have sufficient hardness tobe susceptible to delayed failure if placed in acorrosive environment. Bright-annealed alloyswith high martensite levels from forming canfail by hydrogen embrittlement with just theresidual hydrogen from annealing. Therefore,the use of bright-annealed lean alloys such as301 is not recommended.

A last key variable in which practice and ma-terial interact is strain rate. Ferritic steels flowmore easily at lower strain rates and are thusmore formable. Austenitic steels experience theopposite effect if they are susceptible to marten-site formation. Adiabatic heating can retard themartensitic transformation and reduce the work-hardening rate, changing their forming charac-teristics. When drawing, this is good, but forstretching it may not be.

Hydroforming, a variation on deep drawingin which hydrostatic pressure forces a blankinto the die cavity, can improve the degree towhich stainless steels can be deep drawn. Thehydroforming process avoids friction between

Fig. 5 Optimized 409 for forming versus normal 409. Source: Ref 5


the blank and the tool. Deformation is spreadmore evenly across the blank, and the materialforms close to its theoretical best. Productivityusing this technique is relatively low, so its useis justified mainly when it is not otherwise pos-sible to make a certain design in one drawncomponent. An example of this may be the pro-duction of a complex exhaust manifold that re-quires a higher-alloyed ferritic stainless withrelatively low formability.

Besides failure by breakage, there are otherless-severe defects found on deeply drawnparts. Austenitic steels can develop a surfacecondition known as orange peel, the result ofslip planes within a grain disrupting the sur-face. Orange peel is prevented by keepinggrain size fine so that the surface relief is toosmall to be seen. Austenitic stainless steels canalso develop anisotropy, which while lesssevere than ferritic steels, can cause “earing,”in which round blanks deform nonuniformly ina four-, six-, or eight-fold symmetry, causingexcess ear-shaped material to extend beyondthe intended dimensions of the component.Material is wasted because larger blanks haveto be used. The steel producer can minimizethe phenomenon by keeping cold roll reduc-tions above about 60%. One measure of theearing tendency is derived from the Lankford rmeasurements:

(Eq 2)

The left side of Eq 2, �r, is a measure of thevariation of plastic strain ratio r with directionin the plane of a sheet. Values of �r near zerogenerally indicate minimal tendency towardearing, while � values significantly above orbelow zero indicate increased tendency towardearing. A combination of a high R value fromEq 1 and a low �r value provides optimaldrawability.

It should be noted that deformation alwaysproduces some surface relief, so highly reflec-tive surfaces become spectrally diffuse, orcloudy, after plastic deformation. This has beenan issue for items such as automotive brighttrim. Mechanical buffing can restore the luster,but the time and expense of buffing increasedramatically if orange peel or roping (a similarsurface defect) is excessive.

Duplex stainless steel flat products exhibitsignificant in-plane anisotropy that can cause

forming difficulties and lack of compliance withmechanical property requirements. They alsohave significant rolling anisotropy, which causesthe yield strength transverse to the rolling direc-tion to be consistently higher than it is in therolling direction. This behavior is contrary to thegeneral behavior of single-phase alloys. The dif-ference in yield strength is sufficient, reportedlyup to 15% lower in the longitudinal directionthan in the transverse direction required for ten-sile tests (Ref 7), that it is both serious designand forming considerations.

Ferritics undergo a more specialized surfacerelief because their anisotropic grain structurecan yield in a more concerted fashion and giveeven greater surface relief, called ridging androping. This is minimized by refining grainsize, achieving full recrystallization versus justrecovery during annealing, and temper passing(i.e., elongation of about 1% on a cold-rollingmill) to suppress the yield point phenomenonthat is characteristic of ferrous bcc materials.End users should always make their use of thematerial known to the producing mill so thatthe correct thermal processing path can beemployed for the manufacturing process thematerial will undergo.

One of the most important material consider-ations for deep drawing is surface finish. Flat-rolled stainless should be fully annealed andpickled so that the surface holds lubricant welland yields as readily as possible. Temper pass-ing will drastically reduce the drawability ofstainless by smoothing the surface (rather thanincreasing the yield strength). Temper passingwith roughened rolls does not significantlyharm drawability. The surface finishes that areproduced by temper rolls with special finishes,such as Koolline, retain lubricant well and canbe drawn with minimal distortion.

Tooling for stainless must be strong and wearresistant. Traditional tooling materials are D2tool steel and high-strength aluminum bronze.D2 tool steel must be hardened to HRC 60 to62 and must have smooth surfaces. The use ofpowder metal techniques to produce tool steelsfor dies has permitted much higher volume frac-tions of ultrahard microconstituents such asvanadium carbide to be introduced, therebyvastly improving wear resistance without harm-ing toughness or even raising overall hardness.The benefits of cast aluminum bronzes are lowfriction, high thermal conductivity, and low ten-dency to gall. They are preferred when finishedpart surface appearance is more important than

Δrr r r

=+ + 2

20 90 45


absolute die life and forming pressures aremoderate.

Lubricants for stainless steel forming must beable to prevent metal-to-metal contact underhigher pressures than those seen with carbonsteel. The ASM Metals Handbook, Desk Edi-tion, lists common lubricants as shown in Table2 (Ref 1). Not listed in the table are the newerthermoplastic acrylic polymers that, when ap-plied to the surface at a density of around 1 g/m2

(0.004 oz/ft2), provide a dry film with lubricat-ing properties surpassing any of those listed inTable 2.

Stainless Long Products

Cold heading, one of the most importantforming operations conducted on stainless longproducts, is a forming process that increases thecross-sectional area of a room temperatureblank at one or more points along its length.Cold heading is typically a high-speed processin which the blank is progressively movedthrough a multistation machine. The process iswidely used to produce a variety of small- and

medium-sized hardware items, such as screws,bolts, nuts, rivets, and specialized fasteners.

As with flat-rolled forming operations, theprimary difference between carbon/alloy steelsand stainless steels comes from the higher yieldstrength and higher work-hardening rates ofstainless. Anisotropy is not a significant consid-eration for long products. The ferritic stainlesssteels are the most easily cold headed. The useof the most formable stainless alloys for flatproducts, the stabilized ferritic alloys, is limitedfor long products because of the severe lack oftoughness these alloys show for cross sectionsgreater than about 2 mm (0.08 in.). But, thenonstabilized ferritics, the martensitic, precipi-tation-hardenable (PH), austenitic, and duplexgrades are all cold formable.

In cold-heading terminology, the maximumpossible deformation an alloy can tolerate is ex-pressed in terms of the length of long productexposed beyond the die that can be successfullyforged into the upset. This is measured in thenumber of diameters of initial stock. So, an op-timal ferritic such as 430 can tolerate upsets upto about 2.25 diameters, while a very low work-hardening austenitic, such as 384, can tolerate3.0. The martensitic, PH, and richer duplex

Table 2 Suitability of various lubricants for use in forming of stainless steelBlanking Press- Drop Contour

and brake Press Multiple-slide Deep hammer rollLubricant piercing forming forming forming drawing Spinning forming forming Embossing

Fatty oils and C B C 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 B B C B Apastes(b)

Heavy-duty B NR B A B B NR A B emulsions(c)

Dry film (wax B B B NR B A B NR Aor soap plus borax)

Pigmented B NR A B A C NR NR NRpastes(b)(d)

Sulfurized or A A B+ A C NR A B Asulfochlorinated oils(e)

Chlorinated oils or A(h) NR A NR A NR A(i) A NRwaxes(f) high- viscosity types(g)

Chlorinated oils or B+ A A A B NR A(i) A Awaxes(f) low- viscosity types(j)

Graphite or NR (l) (l) NR (l) NR (l) NR NRmolybdenum disulfide(k)

A, excellent; B, good; C, acceptable; NR, not recommended; SUS, Say-bolt universal second. Ratings consider effectiveness, cleanliness, ease of removal, and othersuitability 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 con-centration of extreme pressure (EP) sulfur or chlorine compounds. (d) Chalk (whiting) is most common pigment; others sometimes used. (e) EP types; may containsome 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 20,000SUS. (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.


alloys are in the 1.5 to 2.0 range of formability.The lean duplex, when they find their way intowider use, should resemble 430 with 2.25 diam-eters maximum.

Hot forming of stainless steel is done as an in-tegral part of their production; therefore, allstainless alloys can be forged. The main issue isthat the high hot strength of stainless requiresmuch more force than would be required forcarbon steels. Martensitic stainless steels re-quire 10 to 100% more force than 4340 alloysteel, while the austenitics require much morebecause of their high hot hardness and because

their work hardening is not instantly removedby annealing. For example, the initial pressurerequirements for a given deformation range for304 can be three times what is required to de-form carbon steel and up to five times for highermolybdenum alloys. For greater deformation,this multiple increases (see Fig. 6a and b).

Austenitic stainless steel loses ductility whenheated above 1280 °C (2335 (F) because of lowmelting phases in the grain boundaries. Asmuch as possible, all alloys should be hotworked in a single phase field of the phase dia-gram to avoid mixtures of ferrite and austenitesince the great difference in their great strengthscan cause failures. Duplex stainless steels andother alloys (e.g., 17-7 PH) that have high lev-els of ferrite in austenite or austenite in ferrite atthe hot-working temperature exhibit reducedhot ductility compared to either fully austeniticor fully ferritic stainless steels and are more dif-ficult to hot work.

REFERENCES

1. Forming of Sheet, Strip, and Plate: DeepDrawing, in Metals Handbook, desk ed.,J.R. Davis, Ed., ASM International, 1998, p782–829

2. U.F. Kocks, C. Tomé, H.-R. Wenk, Ed., Tex-ture and Anisotropy, Cambridge UniversityPress, Cambridge, UK, 1998

3. E. Schedin, “Forming Stainless Steel,”ACOM Technical Paper, www.outokumpu.com

4. H. Sumitomo and T. Tanoue, Nippon SteelTechnical Report 71, October 1996

5. S.P. Keeler and W.A.A. Backofen, ASMTrans. Q, Vol. 56 (No. 163), 1963, p 25–48

6. “409 Ultra Form Stainless Steel,” ProductData Bulletin, www.AKSteel.com

7. R. Cordewener et al., “Duplex StainlessSteels,” Paper 109, TWI conference, Glas-gow, 1994

SELECTED REFERENCE

• ASM Handbook, Vol. 14, Forming andForging, ASM International, 1988

Fig. 6 Forces required for hot working. Source: Ref 1

CHAPTER 15

Machining

Summary

MACHINING STAINLESS STEELS is acomplex operation. Not only does a shop needthe correct supporting equipment and supplies,a better understanding of the metal itself is ad-vantageous. Technology in the production of amore machinable stainless steel is advancing.The incorporation of complex oxides has led tothe development of materials that allow highermachining speeds and increased productivities,both of which are reducing machining costs andkeeping shops competitive.

Introduction

Stainless steel forgings, castings, plate, andlong products all are frequently machined. Thisfundamentally involves the removal of a layerof material from the workpiece with a cuttingtool one or multiple times until a finished orsemifinished part is produced. Machining, in it-self, is a complex topic with many variables.Rather than attempt to understand all aspects ofmachining, it is helpful to consider a material’smachinability, that is, its ability to be machinedand the factors that affect its ability to be ma-chined. In Fig. 1, a macroview shows how themachinability of a material is influenced by theinteraction of humans, machine, methods, mate-rial, and management. Some of the variablescan affect the appearance of the material, whileothers affect the performance of the piece, mak-ing machining an art as well as a science. Opti-mum machinability is obtained when each ofthese sectors come together, providing the bestpossible conditions for efficient machining. Anychange in one of these sectors can change thebehavior or efficiency of a machining job.

From a more focused viewpoint, the machin-ability of a material is further described by:

1. Consistency: Does the material machinabil-ity stay the same when bundles are changed?

2. Tool life/wear: How long does the tool lastin the machining operation? This could beminutes, hours, shifts, or days.

3. Productivity: How many parts were madein an hour, shift, or day?

4. Cost per part: What is the cost of the finalgeometry?

5. Cycle time: How fast can a part be com-pleted?

6. Surface finish: How smooth or shiny is thepart?

7. Chip control: Are the chips manageable?8. Maximum cutting speed: How fast can the

part be cut without affecting tool life?9. Maintaining tolerances: How long can the

machining operation continue before ad-justments are made?

10. Minimal operator intervention: Does theoperator need to constantly adjust setup?

Fig. 1 The 5 M’s of machinability




This list is somewhat empirical or job re-lated, but it provides guidelines for defining ajob since cutting conditions can be very differ-ent for each material and part. For example, ifthe surface finish of a part is very important, itmay be necessary for chip control and tool lifeto be sacrificed. Clearly, machining involvesmuch more than simply cutting a piece ofmetal.

Machining is a very empirically mature sub-ject. The recommended feed rate, depth of cut,tool material, and cutting fluid for a given mate-rial/material condition (thermomechanical his-tory) can be found in readily available pub-lished tables. Books such as the ASMHandbooks; Machinery’s Handbook, publishedby Industrial Press; Marks’ Standard Hand-book for Mechanical Engineers, published byMcGraw-Hill Book Company; or the Machin-ing Data Handbook, 3rd edition, by theMachinability Data center at the Institute of Ad-vanced Manufacturing Services (IAMS; for-merly known as Metcut Research AssociatesInc.) in Cincinnati, OH; include much of thedata used in industry today. Material manufac-turers are also a source of valuable machiningdata. A typical guide from ASM is shown asTable 1 (Ref 1).

Rather than simply reproducing data, thefocus of this chapter is the metallurgical factorsgoverning the machinability of stainless steels.Most of the information regards machiningstainless bar products; however, many of theconcepts could be applied to forgings as well ascastings.

Physical and Mechanical Properties

The machinability of stainless steels is verydifficult to characterize in definitive terms be-cause of the broad nature of these materials. Aferritic stainless steel, such as type 430, willmachine very differently from the martensitic.In some sense, this is like comparing brass tocarbon steel. Both type 410 and type 430 arestainless steels, but the chemistry and structuraldifferences create diversity in machining char-acteristics.

The machinability of stainless steels can bethought of as a function of the steel’s chemistry,cleanliness, structure, processing history, andthe cross-section size of the stock, with no onefactor more important than another:

Each variable contributes uniquely to machin-ability. Machine shops and users generally havevery little influence on these material variables.Because no two mills are exactly identical, therewill be differences in machinability of a steelgrade provided by different mill suppliers. How-ever, having an understanding of how these vari-ables contribute to machinability is invaluable.Armed with an understanding of the material andhow it is made, one can determine the tooling,coolant, and setup of the machining job.

Let us take a closer look at these variables.

Chemistry

The role of chemistry is to define not only the different grades of stainless steel (ferritic,martensitic, etc.), but also how the grade is chemically balanced within the specificgrade; for example, the amount of carbon in amartensitic stainless can change tool wear char-acteristics, or a change in nickel content withinspecification limits can alter the stringiness of achip. Combined, both will be the basis the mate-rial’s machinability.

Each of the elements used to produce stain-less steels will contribute some general machin-ing attributes. The effects of the elements as de-scribed next are general, and slight deviationsmay be encountered depending on the stainlessgrade. However, for the more common stainlessgrades used today, these effects of these alloy-ing elements are fairly accurate.

Iron is the base element in a stainless steel. Itis a soft, gummy material that has high work-hardening characteristics. Iron is characterizedby surface finishes that are difficult to obtainand chips that are stringy, and it has a high ten-dency toward tool built-up edge (BUE).

Chromium strengthens and reduces ductilityof stainless steel. Machine and tool setup re-quire more rigidity. Chromium allows chips tobegin breaking.

Carbon content increases strengthen stainlesssteels and promote carbide formation. Low car-bon levels, typical in ferritic stainless steels, donot help machinability much. Increasingamounts of carbon to greater than 0.08% will aidin chip breakability and reduced BUE in thesegrades. However, as carbon content increases,

Machinability of =ƒ(chemistry, cleanlinesss,

stainless steels structure, processing,ccross section)

Chapter 15: Machining / 183

the amount of carbide increases, the structurechanges to martensitic, and the wear on tools increases.

Nickel increases the toughness and ductilityof stainless and reduces the work hardeningrate. Nickel also increases elevated temperaturemechanical properties. This causes chips to bemore difficult to break. Nickel will have a ten-dency toward increased BUE; however, bettertool life will generally result.

Sulfur reduces mechanical and corrosionproperties and can be a cause of hot cracking inthe resulfurized grades. It is best known as afree-machining contributor that promotes bettertool life and greater machining speeds.

Manganese is generally added to combinewith sulfur to form manganese sulfide (MnS),which acts as a self-lubricant and improvesmachinability. In high-manganese grades, such

as duplex and 200 series alloys, manganese hasthe same relative effect as nickel when used ingreater amounts, as, for instance, in the 200 se-ries stainless steels.

Molybdenum increases the strength and ele-vated temperature mechanical properties. Thisincrease in hot hardness and strength meansmore energy will be needed to cut the material,thus creating hotter cutting conditions. Whilethe molybdenum helps in chip breakability, itwill require more rigid setups and will reducetool life.

Copper improves ductility and reduces thestrain-hardening or work-hardening rate (withthe exception of participation-hardening alloys,for which copper is used as the precipitant).Chips can be more difficult to break, which in-creases the tendency of BUE and promotes bet-ter tool life.

Material Hardness,

HB ConditionDepth of cut(a), in.

Speed, fpm

Feed, ipr

Tool materialAISI Brazed Indexable

Feed, ipr

Tool material

gradeSpeed,

fpm Feed,ipr

Toolmaterial

grade

Ferritic steels 405, 409, 429,

430, 434,436, 442,446(c)

135-185 Annealed 0.040 150(235) 0.007 M2, M3 575 650 0.007 C-7 850 0.007 CC-7 0.150 120(190) 0.015 M2, M3 450 500 0.015 C-6 650 0.015 CC-6 0.300 95(150) 0.020 M2, M3 350 400 0.030 C-6 525 0.020 CC-6 0.625 75(115) 0.030 M2, M3 275 310 0.040 C-6 . . . . . . . . .

Austenitic and duplex steels 201, 202, 301,

302, 302B,304, 304L,305, 308,309, 309S,310, 310S,314, 316,316L, 317,321, 330,347, 348,384, 385(c)

135-185 Annealed 0.040 95 0.007 M2, M3 325 375 0.007 C-3 500 0.007 CC-3 0.150 75 0.015 M2, M3 300 325 0.015 C-3 425 0.015 CC-3 0.300 60 0.020 M2, M3 225 250 0.020 C-2 325 0.020 CC-2 0.625 45 0.030 M2, M3 175 200 0.030 C-2 . . . . . . . . .

225-275 Cold drawn or duplex annealed

0.040 80 0.007 T15, M42(b) 300 325 0.007 C-3 425 0.007 CC-3 0.150 65 0.015 T15, M42(b) 250 275 0.015 C-3 350 0.015 CC-3 0.300 50 0.020 T15, M42(b) 290 215 0.020 C-2 275 0.020 CC-2 0.625 40 0.030 T15, M42(b) 140 165 0.030 C-2 . . . . . . . . .

2205, 2507 295-310 Annealed

Martensitic and PH steels 403, 410, 420,

422, 501,502(c)



275-325 Quenched and tempered

0.040 95 0.007 T15, M42(b) 360 465 0.007 C-7 700 0.007 CC-7 0.150 75 0.015 T15, M42(b) 280 360 0.015 C-6 450 0.015 CC-6 0.300 60 0.020 T15, M42(b) 225 280 0.020 C-6 375 0.020 CC-6

375-425 Quenched and tempered

0.040 65 0.007 T15, M42(b) 290 320 0.007 C-7 475 0.007 CC-7 0.150 50 0.015 T15, M42(b) 225 250 0.015 C-6 300 0.015 CC-6 0.300 40 0.020 T15, M42(b) 180 200 0.020 C-6 250 0.020 CC-6

High-speed steel toolSpeed, fpm

CoatedUncoated

Table 1 Machining setup recommendations for turning wrought stainless steels

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.


Nitrogen strengthens stainless steels. It aidsin chip breakability and reduces BUE but in-creases tool wear.

Titanium promotes carbide formation and in-creases tool wear.

Niobium promotes carbide formation and in-creases tool wear.

The production of stainless steels is identi-fied by industry specifications, such as AISI,UNS, EN, JIS, etc. These specifications are alldefined with fairly broad elemental chemicalcompositions. For example, an AISI 304 has a2 wt% window for the nickel content; that is,this grade can have a nickel level of 8 to 10%.A type 304 with 8% nickel can have differentmachinability characteristics from a type 304with 10% nickel. This 2% difference alters chipmorphology and surface finish capability. Sincetoday’s mill technology can meet very tight ele-mental targets within the grade specification,how the mill balances the grade’s chemistrywill provide the foundation of its machinabilitycharacteristics.

Cleanliness

The cleanliness of steel is determined by theamount and type of inclusions it contains. Vac-uum and argon oxygen decarburization (AOD)melting and refining along with proper steel-making techniques can reduce the inclusions tonegligible levels. It is beneficial to machinabil-ity to avoid hard inclusions. However, certaininclusions are plastic and act as solid-state lu-bricants and chip breakers and prevent adhesionof the material to the tool. The beneficial effectof controlled inclusions is discussed in thischapter.

Structure

Material structure consists of both the phasesthat are present and the microstructure of thosephases. Each type of stainless steel belongs to alarger family, which is characterized by a singlepredominant phase or a combination of two.These are ferritic, austenitic, martensitic, pre-cipitation hardening, and duplex (see the chap-ters on stainless steels, Chapters 6 to 10, in thisVolume). Their machining characteristics aredescribed in the next section. The microstruc-ture of a given alloy is independent of the gradetype and composition and is mainly influencedby grain size. Grain size is not normally speci-fied or reported on certifications; however, mills

measure and control it to varying degrees. Thematerial’s grain size results from the thermaland mechanical history during manufacturingand from the mill’s equipment capability andpractices.

The grain size of a particular product can dra-matically change its machinability. It is entirelypossible for the grain size difference betweentwo lots of material to be large enough to pre-vent both lots from being effectively machinedwith the same setup, requiring adjustments inthe machining setup to remedy the situation.Finer grain sizes strengthen the stainless steel,cause hotter cutting conditions, and have ahigher tendency of BUE. On the brighter side,finer grain sizes yield better surface finishes andsmoother roll thread crests.

Process

The type of equipment used by the stainlessmanufacturer, the manufacturing sequence, andthe practices employed by the mill can affectmachinability as well as mechanical properties,but more important, processing affects howconsistently the material can be machined. Themelt type, hot rolling parameters, thermal treat-ments, cold-finishing parameters, and sequenceof these operations can affect how consistentlya material machines. Many times, the culprit isequipment operational procedures or practicesthat can vary one day to the next. Equipmenttypes can also play a role in machinability. Forexample, machinability can vary when thesame-size material is drawn across two differentdraw benches using different pulling mecha-nisms and two different straightening mecha-nisms. Whether the material is continuously an-nealed or batch annealed can cause differentstrain distributions across the material crosssection as well as material strength differences.Various annealing lines vary in time/tempera-ture profiles and therefore result in differentgrain size and mechanical properties.

With all this in mind, manufacturing consis-tency can be a great asset in machinability. Amachine shop can adjust when a material is con-sistently bad, but it is very difficult when one lotis easy to machine followed by a lot that istough to machine, while a third bundle performsdifferently from the first two. Mills that promotemachining consistency pride themselves bypracticing manufacturing consistency. Toler-ance variation will be tighter and machiningcosts will be lower with their products.


Cross-Section Size

Mill processing equipment dictates a manu-facturing route based on size. Smaller diametersare cold drawn, while larger diameters arestraightened/cut/turned, yielding a softer prod-uct. This can have an impact on machining per-formance despite all other factors being thesame. Cold finishing of stainless steels can beaccomplished via a couple of general manufac-turing routes.

The first is by cold drawing to bar, and the sec-ond is simply a straightening-turning operation.The mechanical properties of the straightened-turned bars will be softer than the bars made bycold drawing. Typically, sizes greater than 1 in.(25 mm) are annealed/turned and straightened,with virtually no strain in the product.

Machinability of the Stainless Steel Families

Comparing the machinability of stainlesssteels with other materials such as carbon steels,brass, or aluminum, there are some striking dif-ferences. In general, stainless steels have:

1. Low thermal conductivity2. High work-hardening rates3. High tensile strengths4. High toughness5. High ductility6. Large spreads between the yield and tensile

strengths

Each stainless steel family (ferritic, marten-sitic, etc.) brings its own general set of machin-ing rules. This is mainly due to the chemistry ofthese families and its resultant effect on thephysical and mechanical properties. A generaldescription of the machining behavior is pro-vided next. One must keep in mind that these aregeneral characteristics. Further alloying of thesefamilies, such as with a sulfur addition, can re-sult in a radical change in machining behavior.

Ferritic

Ferritic stainless steels are the most basicstainless steels and are part of the 400 seriesgrades. Their basic chemical composition isiron and chromium. These grades generally ex-hibit lower strengths, more ductility and soft-ness, and a close yield-to-tensile ratio. Thesegrades will have a high tendency to BUE, chips

will tend be stringy but can be broken throughaggressive chip breaking, and surface finisheswill be somewhat of a challenge. Ferritics arethe easiest of the stainless steels to machine, donot require much horsepower, have a low work-hardening rate and better tool wear, and willgenerally have higher speed and feed capabili-ties than other stainless families.

Martensitic

Martensitic stainless are also very basicstraight chromium stainless steels, 400 seriesstainless grades, and are similar to the ferriticgrades. The difference is that the martensiticgrades have much higher carbon levels, whichfurther strengthen the materials and allow thesematerials to be hardenable by heat treatment.These grades will have higher carbide levels,which will lead to higher tool wear. This is es-pecially true if the material is being machined inthe hardened condition. The higher strengthswill require more horsepower to cut and willneed more rigid setup than ferritic steels. Thework-hardening rate of martensitic stainless islower than for ferritic stainless. Martensiticstainless also has a small yield-to-tensile ratio,making chips easier to break.

Austenitic

The austenitic grades, the 300 series stainlessgrades, are more difficult to machine than theferritic and martensitic families. Austeniticstainless steels are more highly alloyed and aremore prone to higher work-hardening rates. Thisleads to the need for higher horsepower andmore rigid setups. These grades are very proneto BUE and hence are prone to poorer surfacefinishes and tend to tear. The yield-to-tensile ra-tios of austenitic stainless steel is very large,making chips hard to break. Chips in this familyof alloys tend to be long and stringy. The higherstrength and higher ductility of these grades alsotend to increase cutting temperatures, necessitat-ing tooling with higher heat resistance.

Precipitation Hardening

Precipitation hardening stainless steels arecharacterized by higher strength and toughness.The solution-annealed hardness of AISI 630, forinstance, is HRC 36 versus HRC 23 for a 304.Higher horsepower requirements, high tendencyto BUE, higher tool wear, and difficulty break-ing chips are familiar scenarios for this class of


stainless. Except for alloy A-286, the precipita-tion-hardenable (PH) grades are all martensiticalloys and can be treated as such for machiningpurposes.

Duplex

The duplex is a unique class of stainless char-acterized by a dual-phase structure. Duplex al-loys have a structure that is roughly a 50% mixof austenite and ferrite; thus, two hardness ma-terials with different hardnesses coexist side byside. The tool will alternate cutting between softand hard grains of the duplex structure, leadingto an automatic tendency to initiate chatter inthe cutting system. Strength levels of duplex al-loys are quite a bit higher than austenitic grades.Between the duplex structure and high-strengthlevels, high horsepower is necessary, and highlyrigid setups are required. Some work at Ugitechfound that to effectively machine these grades,highly alloyed carbide tooling with high hard-ness and high heat resistance, such as theC7/C8-type carbides, should be used.

Super Stainless Steels

Super stainless steels are today’s highly spe-cialized stainless grades. These grades, like theduplex alloys, are being developed to increasecorrosion performance parameters to meet someof today’s increasing performance require-ments. These alloys are more highly alloyedthan the duplex materials. Strength levels arehigher and toughness is greater, driving machin-ability downward.

Role of Inclusions

Metallurgists have long known that the pres-ence of a soft second phase dispersed in the ma-trix of a parent metal can improve its machin-ability. These particles provide a solid-statelubricant between the chip and tool or a discon-tinuity in the material to aid in chip breaking.The challenge to the alloy designer has been todevelop second phases that produce these bene-ficial effects with a minimal of degradation tothe material itself.

Lead, Selenium, Tellurium

The range of additions possible to stainlesssteel is the same as for carbon steel. Lead addi-tions probably are the best source for improving

machinability of a material. It was once a lead-ing addition to carbon steels, but its reported en-vironmental toxicity has diminished its role.Many carbon mills are looking at other machin-ing agents to replace lead. Lead, however, hasnot been a large factor for stainless steels be-cause of the extremely negative effect it has onhot workability, always a serious considerationin stainless steel design.

Selenium and tellurium have similar charac-teristics to lead as additives but also are non-competitive due to cost, toxicity, and incompati-bility with stainless.

Sulfur

It became obvious very early to metallurgiststhat higher sulfur levels correlated to bettermachinability, and sulfur remains the popularadditive choice. Sulfur is cost-effective as amachinability additive and can be easily re-moved with modern refining methods. Sulfur isa natural impurity and has negative effects onmechanical and corrosion properties, discussedseparately here. The role of sulfur as a machin-ability agent in stainless steels is very complexand not necessarily straightforward, but in gen-eral sulfur has been extremely beneficial in in-creasing the machinability of stainless steels.

Generally, as the molten metal cools to solidform, sulfur combines with manganese to formmanganese sulfide inclusions. Manganese be-comes a very important variable during this re-action. Two basic sulfide forms are found instainless steels—manganese sulfides andchromium sulfides—and the form the sulfidewill take depends on the manganese content.When manganese levels are less than 0.4%,chromium sulfides and chromium-rich sulfideswill be present. As manganese levels reach 0.4to 1.8%, chromium-rich manganese sulfides arepresent. For manganese levels beyond 1.8%,comparatively pure manganese sulfide will befound. The manganese-to-sulfur ratio is also im-portant. For highest machinability levels, a highmanganese-to-sulfur ratio is desired. However,if corrosion resistance is desired, a low man-ganese level is preferred to encourage the for-mation of chromium sulfides or chromium-richmanganese sulfides since these sulfide formshave superior corrosion resistance.

Sulfides form initially as spherical inclusionswithin the cast structure. Hot working, as well ascold working, elongates these inclusions asshown in Fig. 2. This elongation increases the


surface area of the sulfides and will form weakplanes that mechanically weaken the steel, per-mitting chip breakage. The deformation causedby the severe bending of the chip during cuttingcauses the chip to break at the sulfide striations,shown in Fig. 3. An additional benefit of thesemanganese sulfide stringers is that when thesestringers touch and pass the cutting tool, a smallamount of manganese sulfide is deposited on thetool surface, providing a lubricating layer be-tween the chip and tool. This reduces friction,thus reducing heat to the tool. Consequently,

machining speeds can be increased, improvingmachining productivity.

Much has been written about the benefits ofsulfides in a machining operation. A metallurgi-cal perspective shows the more complex natureof sulfides. And, there is more to this than justadding sulfur. The discussion of the combina-tion of manganese and sulfur revealed that thereis a particular balance of manganese to sulfur toachieve desired needs.

In addition to chemistry (manganese-sulfurbalance), the size and shape (relative elongationor globular nature) of sulfides contributes to themachinability of stainless steels. Sulfides aredefined into four categories based on morphol-ogy: type I to IV sulfides. Type I sulfides formfrom the melt as large globular sulfides and areassumed to be best for machinability. Type IIsulfides generally form as a eutectic-like distri-bution of finer rod-like sulfides in interdendriticregions. Type III sulfides form as angular-shaped particles. Type IV sulfides form from themelt as plate-like sulfides in a ribbon-shapedpattern. It has been shown that the larger, moreglobular (less-elongated) type I, homoge-neously dispersed sulfides are better than theother types for enhancing machinability in bothcarbon and stainless steels. However, it seemsthat sulfides that are too large or too small canbe detrimental to machinability. The coarsesttype I sulfides, once thought to be best formachinability, are difficult to attain with today’sstainless manufacturing equipment. Ingot cast-ing, with the slow cooling and solidificationrates, is beneficial to the creation of the coarsertype I sulfides. With the transition to continu-ously cast blooms, solidification rates are muchfaster, effectively creating a finer type I sulfide.

The effect of sulfur on the machinability ofstainless steels is more effective in smalleramounts than in larger amounts. Figure 4 showsa graph of drill penetration results in varyingamounts of sulfur in an 18-9 stainless. It is eas-ily seen that small additions of sulfur have thegreatest effect at sulfur levels <0.10%. This alsocan be proven with a 304 or 316 stainless steel,for which sulfur levels are inherently lower than0.030%. Machinists readily can see differencesin tool wear between a 316 with 0.023% andone with 0.028% sulfur. That small addition ofsulfur has a dramatic effect. On the oppositeside of the scale, with sulfur levels greater than0.20%, the curve flattens out.

From a machining perspective, stainlesssteels can be classified into three groups based

Fig. 2 Typical AISI 303. Source: Ref 2

Fig. 3 AISI 303 chip breaking at the sulfides. Courtesy ofUgitech


on the sulfur additions. In the first group, sulfurcontent is limited to 0.010%. For these stainlesssteels, corrosion resistance, weldability, or me-chanical properties are more important thanmachinability. Grades in this group are316LVM and other remelted stainless grades,many duplex and super stainless grades. Thesecond group has sulfur contents ranging from0.010 to 0.030%. This group of stainless steelsstill holds corrosion resistance, welding, andmechanical properties as critical but has the ad-vantage of increased machinability. Grades inthis group include 316/316L, 304/304L, 321,347, 410, and 430, among others. The thirdgroup with sulfur levels �0.15% are consideredresulfurized grades. Sulfur levels are generally0.25 to 0.35%, with some grades reaching0.45%. The grades in the third group include303, 420F, 430F, 430FR, 1.4570, and others. Afourth group of grades can be included in thisclassification, but the amount of sulfur is not thegrading criterion. This is where mills will rebal-ance chemistries and processing to enhancemachinability beyond what the first three groupsoffer to create enhanced machining grades.Product offerings include Ugitech’s Ugima andUgima XL (Ugima 2) materials, Sandvik’s San-mac materials, Outokumpu’s Prodec materials,and Carpenter’s Project 70+ materials.

Sulfides in stainless steels have a dark side,especially when sulfur levels exceed 0.15%. Itis well documented that sulfides negatively af-fect corrosion resistance and mechanical prop-erties of stainless steels. Sulfur tends to formsegregated films with low melting points duringsolidification. And, since these films are lowstrength, they may induce the formation of mi-crocracks brought on by solidification shrinkagestresses. Further processing of these materialscan induce cracking even further, leading topoor processing yields.

Sulfur hurts corrosion resistance by locally de-pleting chromium from the matrix to precipitateas a sulfide. These manganese sulfide inclusionsbecome exposed on the surface of the bar and be-come initiation sites for pitting corrosion. Passi-vation of the components will help, but the holeleft behind by the sulfide becomes a collectionsite for contaminants. Other sulfide inclusionsthat form on the grain boundaries contribute tointergranular corrosion. Hot workability is alsohurt by sulfides. Sulfur increases the hot short-ness of materials during hot-forming operations.

Manganese sulfides form stress risers withinthe material, which lead to reduced mechanicalproperties in notch-sensitive alloys, especially inthe transverse direction. As the amount of sul-fides increases in stainless steels, a susceptibilityto longitudinal cracking can become an issue.With these stress risers in place, any cold de-formation can lead to cracking along the sul-fide stringers. As a general rule, the smaller thebar, the higher the sulfur, and the higher thestrain produce a high probability for a crack toinitiate.

Despite these deficiencies, the sulfides foundin stainless steels are very effective in improvingmachinability, especially in austenitics, wherecontact forces are very high. The benefits of sul-fur to improve machinability outweigh thelosses due to defective parts, at least from themachinist’s viewpoint. The question of perform-ance of the finished component is another issue,which concerns the end user; for the end user,sulfur is a major negative factor for corrosion re-sistance. This has led to the development of al-ternative methods of improving machinability.

Oxides

The basic machinability-enhancing agent dis-cussed thus far has been sulfur. The beneficialeffects of sulfur are undeniable, but the detri-mental effect is equally evident. This has ledsteel producers to look at other inclusion sys-tems for a viscoplastic inclusion without thenegative effects of sulfur. Since the 1900s,steelmakers have known that injecting calciuminto the melt converts refractory inclusions intosoft, malleable, complex oxides that act as free-machining agents with high-temperature lubri-cating capabilities. The oxide inclusion chem-istry is based on the CaO-Al2O3-SiO2 system.Figure 5 shows that small amounts of calciumcan greatly increase tool life. These calcium-based oxide formulations have been commer-

Fig. 4 Effect of sulfur on stainless machinability. Source: Ref 2


cialized for stainless steels, but the machinabil-ity agent is, at this point, not standardized asthis process is difficult to reproduce consis-tently from heat to heat. However, Ugitech SA(formerly Ugine Savoie and now part ofSchmolz and Bittenbach) developed a propri-etary and patented process sold under the tradename Ugima and Ugima XL (Ugima and Ugima2 in Europe and Asia). Figure 6 shows theUgima oxide coexisting with sulfur in AISI.

The Ugima oxide works similarly to sulfur bycoating the cutting tool and acting as a lubricant.Figure 7 shows EDAX (energy dispersive analy-sis by x-ray) spectra of the surface of a carbidetool, proving the existence of coatings of man-ganese sulfide and Ugima oxide. The Ugimaoxide performs synergistically with sulfur. Sincethe oxide alone has limited lubricating abilitiesat slow machining speeds, the manganese sul-fides in the stainless are the dominant machin-

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

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

Fig. 7 X-ray examination showing manganese sulfides and Ugima oxides coating the tool surface. EDAX, energy dispersive analysisby x-ray. Courtesy of Ugitech


ing agent. As machining speeds increase, man-ganese sulfide eventually loses its lubricity, andthe oxide acts as a high-temperature lubricantthat will allow faster machining speeds withoutthe need to use specialized tooling. Data gener-ated at Ugitech as well as data generated byother producers have shown that once sulfurcontent reaches 0.30%, the contribution towardmachinability flattens and possibly decreases,which seems to contradict the experience ofmany machine shops. However, the combina-tion of sulfur and the Ugima oxide extends themachinability range for sulfur levels beyond0.30%, as seen in Fig. 8.

Tool lubrication is only one function the com-plex oxide performs. Like sulfur, the Ugimaoxide is a discontinuity and will aid in chipbreakage. Figure 9 exhibits cross-sectional pho-tomicrographs of 304L chips with a 0.025% sul-fur level. As shown, the chip is shearing alongcomplex oxide stringers, helping the chip break.

Ugitech has seen synergistic effects with copper additions as well. Machining tests atUgitech and many field experiences have shownincreases in machining performance of 30%with the addition of 1.4 to 1.8% copper to a303 stainless steel with the Ugima oxide. Thegrade chemistry meets EN 1.4570 and is soldunder the trade name 303 Ugima UX (4570Ugima in Europe). Further additions of copperup to 4% have exhibited even better machiningperformance.

Tooling and Coolants

The machining of stainless steels can bemuch more complicated than for other materi-als. Machine and tool rigidity, machine powerrequirements, sharp cutting tools, and optimumlubrication type and amounts are very impor-tant. Feed rates need to be sufficiently more ag-gressive than with carbon steels to prevent workhardening and glazing of the material as well asto avoid reducing tool life.

High-Speed Tool Steels. The cutting toolsthemselves are the main variable other than the

Fig. 9 Comparison of 304L chips with and without the Ugima oxide. Courtesy of Ugitech

Fig. 8 Comparison of machinability of AISI 303 at differentsulfur levels with and without the Ugima oxide.

The vertical axis, VB30/0.3, represents 0.3 mm of tool wear in30 min.


material in determining the quality of the ma-chined part. The earliest cutting tools werethose made of the high-speed steels (HSSs). Thewrought HSS tooling was very versatile, al-lowed resharpening of the tools many times,and is still a preferred choice in many shops.With time, mills perfected their ability to pro-duce more highly alloyed tool steels to meet theincreasing demands of the machining industry.Today, tools are being made with a powder met-allurgy process, by which ingots of compactedhigh-speed tool steel powders have more struc-tural homogeneity and thus better wear and heatresistance. Powder metallurgy techniques pro-duce even more highly alloyed high-speed toolsteels with properties approaching those of car-bide tooling, allowing better machining per-formance. Although high-speed tool technologyhas improved, its limiting factors are heat andwear resistance. Use of HSS tooling still limitsmachining speeds and performance.

Carbides. Carbide tooling is the next genera-tion of tool materials after HSS, although ittends to be hard and brittle. In the past, unless itwas possible to feed hard, maintain fast speeds,and have uninterrupted cuts, carbide toolingwas a good choice. However, carbide tool tech-nology has come a long way in grades andtechnology. Micrograin and nanograin carbidesare providing carbide tooling with increasedtoughness. Single-point tooling is now able towithstand the punishment of an interrupted cut,and drills are able to withstand some flexing.

Generally, the carbide grades to be used whenmachining stainless steels are the C5- to C8-type carbides. These classes of carbides areharder and have more heat resistance, propertiesthat are needed when cutting stainless. The dis-cussion in this chapter has stated that the cuttingof stainless is more difficult and generates moreheat. Cutting materials that withstand these cir-cumstances are needed. The C2- to C4-type car-bides are not well suited for stainless steels be-cause they do not have the heat resistanceneeded. There are always exceptions. For exam-ple, it is possible to use the C2- to C4-type car-bides when using the older cam-operated multi-spindle and Swiss machines, for which machinespeeds are limited. Tooling manufacturers suchas Kennametal, Sandvik, Iscar, etc. have theirown proprietary grade designations, but manycorrespond to the C5 to C8 types.

Coatings. Tool coatings have contributed tomachining improvements. Coatings add a very

hard layer on the tool surface that will provideadditional lubricity between the tool and chip aswell as potentially providing heat resistance forthe tool. Like the carbide, there are certain coat-ings that are more beneficial than others. Tita-nium nitride (TiN) was a great coating when itwas first introduced, but further development inthis area has created other coatings that workeven better. Grades like titanium-aluminum-nitride (TiAlN) and aluminum-titanium-nitride(AlTiN) are great coatings for turning, form cut-ting, cutoff, grooving, drills, reamers, andmilling. These tools are usually subjected tohigher temperatures from high speeds, deepcuts, and limited coolant. The aluminum in thecoating breaks down and combines with the surrounding oxygen to form aluminum oxide(Al2O3), a material with higher hardness andbetter thermal resistance. When tooling will notbe subjected to much heat, such as with taps,high wear resistance coatings such as titaniumcarbonitride (TiCN) coatings are good.

Coolants. Stainless cutting is hot, has highfrictional forces, and has tendencies of themetal to stick to the tools. Coolants need to re-move this heat and provide lubrication to reducefriction and minimize BUE. Coolants availableare petroleum based, semisynthetic, synthetic,water soluble, and the new type: vegetablebased. All of these coolants need to be highlyfortified for use in stainless machining. For ex-ample, petroleum-based oils need high sulfur,chlorine, and fat levels for lubricity at highertemperatures. The other coolants need high-pressure additives, again to help lubricity. It isalways a good idea to discuss the options withthe coolant supplier.

REFERENCES

1. Machining Data Recommendations, in Met-als Handbook, desk ed., 2nd ed., J.R. Davis,Ed., ASM International, 1998, p 917–950

2. T. Kosa and R.P. Ney, Sr., Machining ofStainless Steels, in ASM Handbook, Vol16, Machining, ASM International, 1989, p 681–707

SELECTED REFERENCE

• www.ugitech.com

CHAPTER 16

Surface Finishing

Summary

SURFACE TREATMENTS are extremelyimportant to the end user, and they are totallywithin the end user’s control and specification.They include the proper cleaning of stainless,the various means of descaling after thermaltreatment, and the choice and application of sur-face finishes. The cost ramifications of improperor suboptimal surface treatments are immensebecause of the possibility of them compromis-ing corrosion performance, which is a charac-teristic for which the end user pays dearly.

Introduction

Surface finishing is usually very important forstainless steel. The underlying economic justifi-cation for using stainless steel is that it does notcorrode if properly specified for the environmentit faces. Thus, its surface appearance remainsnormally intact throughout its life. This appear-ance should therefore be aesthetically pleasing,even in an industrial setting, and the surface fin-ish should not detract from its performance.

Raw stainless surfaces resulting from rollingand annealing operations are not considered at-tractive and are used only for functions in whichaesthetics are a negligible consideration. Evenso, there are surface treatments required forstainless steel intended for such uses. The stain-less surface must be freed of:• Oxides resulting from annealing, joining, or

hot forming• Accumulated ambient foreign material• Applied process materials, such as forming

lubricants, fluxes, etc• Contamination from other materials, espe-

cially iron.

Beyond such considerations, one can alsocreate surfaces on the stainless that enhance itsbeauty and performance. The surface can bemade reflective or matte, ground or mechanicallypatterned, coated, painted, plated, or oxidized.It can be treated chemically or electrolytically.The surface can be altered on an atomic basis,sometimes producing profoundly different me-chanical and corrosion-resistant properties. Allof these are discussed in this chapter.

Function of Surface Treatments

Removal of Oxide Scale

Oxide scales form on stainless steel duringannealing, hot-forming, and joining operations.Removal of this scale is important to proper cor-rosion resistance. This is because the chromiumin the steel oxidizes much more readily thanother elements, so the surface of the steel underthe oxide is chromium depleted and thereforehas lost possibly a significant amount of corro-sion resistance. An oxide scale is quite differentfrom a passive film. Scales deplete chromium;passive films enrich the surface in chromium byselective loss of iron.

Oxide scales are arguably best removed bypickling. Pickling is the chemical dissolution ofthe oxide scale. The acids commonly used are ni-tric (HNO3), hydrofluoric (HF), and sulfuric(H2SO4). HCL, which is commonly used topickle carbon steel, is not recommended for stain-less because it locally attacks (i.e., pits) the sur-face. HNO3 is an oxidizing acid that by itself doesnot pickle stainless. It is used in combination withHF to modulate the attack by the strongly reduc-ing action of the HF. This combination allowsgood control of pickling rates by varying the ratioof the two acids. H2SO4 is used alone. It is often




used as a preliminary pickling because of its ag-gressive, and less-controllable, action.

Pickling involves both scale dissolution andmetal dissolution. Thus, pickling can be consid-ered a deliberately imposed corrosion process.The scale formed during hot rolling may be anembedded scale with a minimal chromium-depleted layer beneath the scale. Annealing andshotblasting are required before pickling so thatthe acid can penetrate to the chromium-depletedregion under the scale and selectively dissolvethis layer. Nearly all the action of pickling de-rives from undercutting of the oxide scale anddissolution of the chromium-depleted layer. It isalso vital that this layer be removed as it dimin-ishes the corrosion resistance of the surface.Simply removing the scale mechanically maybe insufficient if only the oxide is removed andthe compromised (depleted) metal surface isleft. Metal removal rates are proportional to thechromium level. Thus, sensitized material willundergo grain boundary attack if pickled. Allstainless has slightly greater oxidation in thegrain boundaries when it is oxidized, and thisleads to “ditching” of the grain boundaries dur-ing pickling, resulting in the matte appearanceof a pickled surface, as seen on two-dimensionalflat-rolled material.

Typical concentrations for H2SO4 pickling are8 to 15% by weight. Bath temperatures are gen-erally 150 °F (65 °C). The rate of pickling de-creases rapidly as iron builds up in the bath, andthe bath must be replaced for efficient pickling.The attack of H2SO4 on the base metal can besevere, and undissolved constituents can remainon the surface as “smut.” This smut must bephysically removed or dissolved by subsequentHF/HNO3 pickling.

The HF/HNO3 pickling is carried out between120 and 140 °F (50 and 60 °C). Higher tempera-tures cause excessive HF evaporation and canalso lead to visible emission of nitrogen oxides.The NO2 (nitrogen dioxide) produced duringthe reduction of the HNO3 (oxidation of themetal) is visibly brown-red. The rate of NO2formation increases at higher temperatures.HNO3 concentrations are normally from 10 to25%. HF concentrations can vary from 1% forlight scales to 8% for maximum aggressivenessand difficult-to-pickle alloys.

The HF provides:

• a complexing agent for iron and chromium• destabilization of the passive film• stabilization of the redox potential

The HNO3 provides:

• a source of H+ ions• stabilization of the passive film• elevation of the redox potential• an oxidizing agent of the base metal• a dissolving agent for the scale

The nitrates carried off from this picklingprocess are an environmental problem. It hasbeen shown that the HNO3 can be eliminatedand replaced by hydrogen peroxide (H2O2).This is the proprietary, patented UG3P processdeveloped by Ugine (Ref 1) or Henckel’s“CleanOx” process.

Electrolytic pickling is commonly used in theproduction of cold-rolled stainless. This processuses alternating positive and negative polariza-tion in baths using sulfates or nitrates. The H+

ions are produced by the polarization instead ofby an acid, so neutral solutions can be used.

The low solubility of trivalent chromium(Cr+3), typically found in oxide scale, makessuch scale hard to remove. Use of strong acidsand complexing agents is required to get pick-ling to occur at an economically acceptable rate.Hexavalent chromium (Cr+6) is much morereadily soluble in aqueous solutions, for exam-ple, oxidizing treatments such as Kolene,* amolten mixture of nitrate salt and hydroxide, orelectrolytic conversion treatments such as elec-trolytic sodium sulfate. Both treatments convertthe trivalent chromium to hexavalent form, al-lowing for easy scale dissolution. Their usemay, however, lead to environmental problemsas hexavalent chromium compounds are toxicand heavily regulated.

Pickling may be assisted by prior treatment ofthe scale in molten 85% sodium hydroxide(NaOH), 14% sodium nitrate, and 1% sodiumchloride bath. The strong oxidizing action ofsuch a bath chemically alters the chromiumoxide in the scale to a more soluble oxide with-out attacking the metal, making it more easilyremoved by subsequent pickling.

Cleaning

Stainless steel is very resistant to chemicalsof many kinds, permitting it to be cleaned bymany aggressive agents. Because contaminationis “on” the surface of stainless rather than “in”the surface as is the case with many materials,simply using the cleaning agent appropriate for

* Kolene is a registered trademark of Kolene Corporation.

Chapter 16: Surface Finishing / 195

the contaminant will permit very thoroughcleaning of the stainless surface. The stainlesssurface will not be harmed by cleaning as longas strong halides, iron utensils, and abrasivesthat alter the surface finish are avoided. Table 1summarizes some cleaning recommendationsbased on the contaminant to be cleaned.

The greatest controversy regarding cleaningstainless steels involves the use of sodiumhypochlorite. Many highly qualified expertsflatly disapprove of its use because it can easilyleave a residue of chloride on the stainless sur-face, leading to localized corrosion. But, in real-ity, health considerations make this position un-tenable. Sanitation concerns in the food serviceindustry take precedence over this prohibition.In fact, every stainless steel, as explained in the“Corrosion and Oxidation” section of this Vol-ume, is resistant to some level of chloride con-centration at a given temperature and pH. Thus,short-term, room temperature use of hypochlo-rite bleaches will generally not harm most typesof stainless steel if well rinsed after application.Field experience has shown that stainless steelswith less than 16% chromium can be harmed byhypochlorite bleaches. Damage occurs mainlyon abrasively polished surfaces in alloys with16% chromium, such as 430, but does not occuron roll-finished or bright-annealed surfaces.Stainless steels with slightly higher chromiumand nitrogen, such as 201, 301, and 304, are notattacked unless concentrated chloride solutionsare permitted to stay on the surface, particularlyin crevices.

The recommendation not to use abrasivesdoes not include soft abrasives such as calcium

carbonate or brush material that is sufficientlysoft not to mar the stainless surface. The user isencouraged to test whether a product meets thisrequirement by testing it on a small, preferablyunexposed, area. Blasting the surface with car-bon dioxide pellets is a rapidly growing processfor removing paint and other adherent, softcoatings and deposits without damaging the sur-face of stainless steel.

These cleaning recommendations apply to in-dustrial, architectural, and domestic uses ofstainless. For special levels of cleanliness re-quired for medical, pharmaceutical, or semicon-ductor applications, refer to the chapters dealingwith those applications.

Passivation is a very commonly used surfacetreatment to remove surface contamination, no-tably iron, and to form a passive film. The filmforms of its own accord when a clean surfaceencounters moist air, but film formation can beaccelerated by controlling the environment.

The surface to be passivated should first becleaned by one of the methods discussed. Thisallows uniform passivation and avoids contami-nation of the passivating solution, especially bychlorides, which can cause a rapid attack. Thosestainless steels with more than about 17%chromium can be pickled in 20% HNO3 at 50 to60 °C (140 °F). Precipitation-hardenable (PH),martensitic, and straight chromium grades withless than 17% chromium should have 2.2 g/Lsodium dichromate added to that solution toavoid attack. The free-machining grades arethe most easily attacked of all types. Theyshould be immersed in 5% NaOH for 30 min at75 °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 solvent

Wax and oil polishes minimize fingerprinting. Glasscleaner is appropriate for mirror finishes.

Grease, oil Hydrocarbon solvent Alkaline cleaners may also be used in severe casesbut may require cleaning the entire surface tomaintain visual uniformity.

More severe stains, discolorations, and rust stains

Nonscratching creams or polishes Do not use HCl-containing products. Hypochlorite bleaches must be well rinsed toavoid pitting.

Hard water scale, mortar 10–15% phosphoric acid, sulfamic acid-containing, or oxalic-containing cleansers

Neutralize with ammonia, rinse, and dry. Do not use HCl-containing products.

Oxides, heat tint If severe, treat by pickling; Scotchbrite©, stainless scouring pad, or nonscratchingcream or polish

If abrasives must be used, they should blendwith existing surface finish in size and direction.

Paint Alkaline, trisodium phosphate, or hydrocarbon solvent

Follow manufacturer’s directions


nitric/dichromate passivation is performed. Thisshould be followed by a rinse, another 30-minNaOH treatment, and a final rinse.

In recent years, citric acid has become a pop-ular replacement for HNO3 because it avoidsthe problem of toxic nitrates and hexavalentchromium. When citric acid is used, a 10%solution is applied for the same time and tem-perature as with HNO3. The use of NaOH be-fore and after is still recommended for low-chromium and free-machining alloys. Refer toASTM documents A 967-01, A 380-99, B 912-00, as well as Federal Specification QQ-P-35.

Brightening

Stainless steels can be brightened by chemi-cal or electrolytic action, which selectively dis-solves the surface in such a way that it becomesmicroscopically smoother. Electropolishing isthe most effective means of accomplishing this.It causes the surface roughness to decrease byapproximately one-half in Ra (the arithmeticalaverage of surface peaks and valleys as meas-ured over a straight line). Electropolishing, likepickling, also selectively removes any exposedchromium-depleted regions, leaving only thebulk alloy with the intended corrosion resist-ance on the surface. Apparent pitting resistanceis thus increased. The resulting surface is brightand cleanable and provides the optimal corro-sion resistance that a given alloy can achieve.The reflectivity of the surface is a function ofthe preelectropolished surface.

The most commonly used electropolishingsolution is 40% H2SO4 and 45% phosphoricacid (H3PO4), balance water, used at 90 °C(194 °F) with a current density of between 1.0and 3.0 amp/m2. Other baths using perchloricacid are technically good but carry the risk ofexplosion.

Coloring

Stainless steel passive films are so thin thatthey are quite invisible. Oxide films are thickerand through optical interference can cause differ-ent colors. The films are formed thermally in aprocess known as heat tint. The color depends onoxide film thickness, which is a function of thetime at a given temperature, well metal composi-tion, and oxygen partial pressure. The colors typ-ically range from light yellow formed at 300 °C(570 °F) through violet formed at 420 °C (790°F) and dark blue formed at 600 °C (1110 °F).Reproducibility is sometimes a problem.

Electrochemical methods permit more uni-form and reproducible color and tougher films.The most prominent is the International NickelCompany (INCO) process. In this two-stepprocess, the steel is first immersed in a mixtureof one part chromic acid, two parts H2SO4, andone part water at 30 °C (85 °F). This is followedby an electrochemical treatment with the samebath or with H3PO4 substituted for H2SO4.

The colors produced are correlated to thick-ness and treatment parameters (Ref 2) as shownin Table 2. Rocha-Fila et al. confirmed thatthese coloring treatments did not degrade thepitting corrosion resistance since their forma-tion mechanism more closely resembles passivefilm formation than oxidation.

Aesthetic Surface Finishes

When steel is used in other than the as-annealed and pickled state, it is often for aes-thetic rather than functional reasons. The majorexception to this is temper-rolled strip, whichhas a bright surface from cold rolling, butwhose normal use is strictly functional. The hot-rolled or cold-rolled annealed and pickled sur-face finish is nonuniformly dull and unattractiveto most observers. When stainless steel first ap-peared on the market, a highly polished surfacewas the paragon, as showcased on the ChryslerBuilding in New York City (Fig. 1).

This surface finish was very expensive to pro-duce since it had to be done by polishing andbuffing. Soon, the more economical abrasivepolishing became the standard, and it had thebenefit of removing the many cosmetic defectsthat then were common to the manufacture ofstainless. Much higher levels of surface qualitywere made possible with the development of theSendzimir mill to the bright anneal process per-mitted much higher levels of surface quality.

Table 2 Parameters for oxide film coloring ofstainless steel

Heat treatment time ColorOxide film

thickness, nm

10 Gold brown 7015 Brown/red/blue 10020 Brown/blue 12025 Green/blue 14030 Green/gold/blue 16535 Gold/green 18540 Gold/green/brown 21050 Red/brown 250


Mirror finishes could be achieved with coldrolling followed by bright annealing. This hasnot entirely replaced sheet buffing to obtainmirror finishes, but sheet buffing has been rele-gated to those few applications where a highdegree of perfection is demanded. The mainmethods of producing aesthetic surface finisheson stainless steel are abrasive finishes and rolledfinishes. The latter are superior in uniformityand corrosion resistance, but the polished fin-ishes are still more common.

Polished Finishes

Polishing is carried out with coarse- tomedium-sized abrasives that are bonded to aflexible backing. This is distinct from grinding,in which abrasives are bonded rigidly to eachother on a rigid backing. Buffing is done byvery fine abrasives, which are loose and do theirwork by being forced along the surface by a softmaterial. The standard rule is that materialsused in finishing must not permit iron or ironoxide particles to come in contact with thestainless surface or passivation layer.

Stainless steel surface finishing is governed by:

• High strength and high work-hardening raterequire more power for metal removal.

• High surface hardness influences whichabrasive will be effective.

• Low thermal conductivity can cause exces-sive temperatures during processing anddistortion.

• Oxidation (heat tinting) that occurs overabout 250 °C (480 °F)

• Residual stress due to surface working, es-pecially in austenitic alloys

Grinding is a relatively coarse procedure withusefulness that is largely confined to defect re-moval. Polishing for aesthetic purposes is gen-erally done with abrasive coated belts and isdone to both coils and sheets. There are no offi-cial roughness values for the various surfacefinishes. ASTM merely describes the finish by agrit with which it is typically made. Table 3shows the varying surface roughnesses that aretypical of polished stainless.

Neither the producer nor the customer is pro-tected by a clear specification as of this writing.Thus, the producer is encouraged to publishstandards, and the customer should specify min-imum and maximum Ra values for his purpose.The typical number 4 polish varies in roughnessfrom the beginning of the coil to the end be-cause of the wear of the abrasive particles onthe belts. The difference can be more than 10μin. (0.25 μm) in Ra, is quite visually apparent,and will make adjacent pieces of stainless fromthe same coil look different. This may be objec-tionable for certain products, especially appli-ances and architectural panels.

Polish appearance varies also with the pres-ence or absence of lubricant. Lubrication duringpolishing (i.e., wet polishing) gives sharper cutsand less heat tint from frictional heating, result-ing in more surface brightness. Polish appear-ance also depends on the length of the grit linecaused by an individual particle of abrasive. Thelength of the grit line varies with the speed of thematerial as it passes the rotating abrasive belt,

Table 3 Polished finish designations based ongrit sizes to achieve target surface roughnesses

Finish number Grit number Ra max, µm Ra max, µin.

3 60–120 1.0 40

4 120–180 0.75 307 240–320 0.30 88 500 0.15 4

Fig. 1 View of the Chrysler Building in New York. ©iStockphoto.com/stevenallen


and the rotational speed and diameter of thebackup roll for the abrasive belt. This combina-tion of variables, while seemingly controllable,usually varies enough so that no two polishedsheets look identical to the trained eye. Thisvariability also causes problems in field repair.

Repair of a polished surface damaged, forexample, by welding or scratching is quitechallenging. A perfect match and blend to thesurrounding original surface is an art, and nopractitioner of that art can accomplish it per-fectly. If the repaired area is small, then it canbe acceptable, if not undetectable, but for largeareas, repolishing of the entire surface is oftenthe best remedy.

A solution to the problem of matching pol-ished surfaces is to use very long grit lines forfinishes of the same roughness. By eliminatingthe variable of grit length, only correct pressureand grit size are required to achieve good visualmatching. This can be done with a belt sander.These finishes are called “hairline” in Japan and“grainline” and other names in the UnitedStates. Lack of uniformity, difficulty of repair,and a decrease in corrosion resistance are thechief drawbacks of polished finishes.

Bright Annealing

Aesthetic finishes that do not depend on abra-sion are derived from bright annealing. Brightannealing is annealing in a very low oxygenatmosphere, either dissociated ammonia* orhydrogen/argon. This process was originallydeveloped as a means of producing bright fer-ritic trim for automobiles. That use has largelypassed out of favor, but the process itself is in-trinsically superior to annealing in air becauseno oxide is formed only to be later removed,usually at significant expense. Because the orig-inal product was strip, many older bright-an-nealing facilities are narrower than normal sheetwidth. The wider, more modern lines are highspeed and wide enough to produce bright sheet.Bright-annealed sheet will only be as bright andflawless as the cold-rolled sheet that is an-nealed. Therefore, a quality product, generallyas mirror-like as possible, must be producedwith great care. The better producers have pro-prietary methods of prior pickling, annealing,

and rolling by which the surface is made flaw-less and very smooth. Final brightening occurswith temper passing after bright annealing.

Rolled Finishes

The bright-annealed and temper-passed sur-face may be used in its mirror-like condition,or it can be used as a basis for rolled finishes,which can take on any appearance and can beengraved onto a temper mill roll. All polishedfinishes can be duplicated as rolled finishes.But, in addition, many other designs such asreplicating fabric or leather, geometric de-signs, or matte finishes can also be made. Thepattern on the roll is impressed into the stain-less surface, and elastic flattening of the rolland the stainless cause about 50% less Ra onthe stainless than on the roll. This effect in-creases with the yield strength of the stainlessbeing processed. The benefits of rolled finishesare:

• They can be made identically from coil tocoil since roll engraving is quite precise.

• They retain the enhanced corrosion resist-ance of bright-annealed material.

• They are less expensive to produce, so thefinal product is generally priced lower thanpolished material.

In theory, such finishes can be made from air-annealed material, but finishes applied by roll totwo-dimensional surfaces that have been dulledby pickling are somewhat gray and indistinct. Ifthe pickling is kept mild enough to retain surfacebrightness, then it is possible insufficient prick-ling has occurred to remove the chromium-depleted layer caused by the air anneal. Some ofthe finishes produced by rolling are shown in thearchitectural chapter (see Chapter 18). These fin-ishes can be applied to all the normal sheetalloys, including the lean duplex alloys. This haspermitted rolled finishes to be used on UNS32003 for building exteriors in Doha, Qatar,where corrosive conditions require a PREN(percentage chromium equivalent) of 25.

The most critical applications from a surfaceperfection point of view use rolled finishesrather than abrasively finished number 4 pol-ishes. These finishes also permit type 430 stain-less steel to be used successfully in restaurantapplications without corrosion issues wherepolished type 430 material previously hadcorroded. There are also instances for whichpolished UNS S31600 was corroding in coastal

* At one time, “dissociated ammonia” referred to a 3:1(mole ratio) mixture of hydrogen and nitrogen produced bythe catalytic decomposition of anhydrous ammonia. Today,this composition is prepared by mixing gases. For manystainless steels, the 3:1 H2:N2 can be varied without difficulty.


architectural applications in the United States.When the abrasively polished surface wasreplaced with rolled finish UNS S30400(Koolline), there was no further corrosion. Be-cause alloying is such a high component of costin stainless, it makes sense to employ rolled fin-ishes whenever possible. However, it must benoted that when manual polishing is used to re-pair scratches or other damage in rolled finishmaterial, its corrosion resistance may be reduced.

Surface Alteration

In carbon steels, surface chemistry can bechanged to affect certain properties. Carburiz-ing and nitriding are examples of suchprocesses. Simply using these processes onstainless cannot be done because these ele-ments combine too strongly with chromium ascarbides or nitrides, dramatically reducing thecorrosion resistance. There have been modifi-cations to carburizing and nitriding that permitaustenitic stainless steels to have very high car-bon levels implanted to a thin surface layer. By

exposing an activated stainless surface to ahigh carbon fugacity at 470 °C (880 °F) foraround 200 h a 50 nm thick layer with 12 at.%carbon in supersaturation can be achieved with-out carbide formation. The properties of thislayer are phenomenal; the hardness is 1000Vickers 25 compared to 200 for the base alloy(Ref 3). In addition, the corrosion resistance in-creases significantly. Only by such supersatura-tion with carbon could it be determined thatcarbon, like nitrogen, is a powerful antipittingalloying element when kept in solid solution.These processes are at the very initial stages ofcommercial use.

REFERENCES

1. Stainless Steels, Les Editions de Physiques,1989

2. C. Rocha-Filo et al., J. Braz. Chem. Soc.,Vol. 15 (No. 4), 2004, p 472–480

3. Y. Cao, F. Ernst, and G. Michal, ActaMater., Vol 51, 2003, p 4171–4181

CHAPTER 17

Welding

NEARLY ALL WELDING of stainless steelis done by end users or processors. Like thermalprocessing, it is complex in theory and practice.This chapter gives a basis for understanding theinfluence of alloy composition and metallurgyon the welding process, which must be re-spected as a process that combines melting, re-fining, and thermal processing. Knowledge ofeach aspect is required for the process to be de-signed and executed properly.

The welding and joining of stainless steels re-quires knowledge of both the technology of thewelding or joining process and the response ofthe steel to the thermal and mechanical effectsof the process. The welding process must, ofcourse, produce a sound joint, but it must alsoresult in the weld and its surrounding affectedmetal having correct strength, toughness, corro-sion resistance, etc. for the intended serviceconditions. This chapter does not attempt toteach welding. The main objective is to showhow standard welding technology is correctlyapplied to stainless steels.

The foremost special consideration of weld-ing stainless steel as opposed to carbon steel isthat the chromium in stainless steel, which iswhat makes it stainless, must be protected fromoxidation, so that:

1. It stays in solution as a corrosion-resistingelement.

2. It does not form refractory oxides that woulddiminish weld soundness.

Welding Characteristics of Stainless Steels

Austenitic stainless steels are readilywelded by nearly all welding techniques. The

characteristics of austenitic stainless steels thatdistinguish them from ordinary carbon steels inwelding are:

• Austenitic stainless steels have lower thermalconductivity and higher thermal expansionthan carbon steels or ferritic stainless steels,which can localize the heating, thus increas-ing the potential for residual stress andtherefore hot cracking.

• Stainless steels contain readily oxidizedchromium, which must be protected.

• Surface oxidation during welding depleteschromium in all types of stainless steel fromthe underlying surface, resulting in reducedcorrosion resistance unless this layer is re-moved.

• The possible formation of chromium car-bides in the heat-affected zone (HAZ) cancause susceptibility to grain boundary corro-sion (sensitization).

• The possible precipitation of intermetallicphases in the HAZ can lower toughness andcorrosion resistance.

• There is increased microsegregation in thefusion zone with increasing alloy content.

• There are thermodynamically metastableconditions due to the low diffusion rates inthe face-centered cubic (fcc) matrix.

The influence of carbon has been well ad-dressed using low-carbon versions of all gradeswhenever welding involves significant time be-tween 600 and 900 °C (1110 and 1650 °F). Thisprevents rapid precipitation by reducing the su-persaturation of carbon. The older method ofpreventing sensitization is to stabilize the alloyswith titanium, as in type 321, or with niobium,as in 347. This is foolproof only if carbon levelsare low, less than 0.04%, since TiC can dissoci-ate at elevated temperatures and not be able to




recombine successfully with titanium duringcooling, permitting a thin zone of sensitizationcalled knife-line attack. Fortunately, most 321and 347 are produced with carbon levels below0.03%. The higher carbon-stabilized alloys andthe high-carbon (>0.03%) unstabilized alloysmust be annealed after welding to redissolvechromium carbides if the cooling was suffi-ciently slow for the carbides to have formed.This is avoided only in thin-gauge (>1.5 mm,0.06 in.) material or when the HAZ is drasti-cally reduced, as in laser welding.

The high thermal expansion of austeniticstainless steel can cause high residual stressaround welds, which may require annealing toeliminate. Another serious threat posed by ther-mal stresses is hot cracking. This can occur tomaterial that has just solidified when geometricconstraints to contraction imposed by the sur-rounding material imposed act on weak grainboundaries. This weakness occurs when thesteel solidifies in an austenitic mode. Whenaustenite freezes, it strongly rejects sulfur to theintergranular areas, where it forms weak films.This is solved by balancing the composition sothat alloys solidify first as ferrite, which doesnot reject the sulfur, forcing it to precipitate assulfide inclusions within the grains. This ap-proach is highly effective but cannot be used for

some highly alloyed grades with compositionsthat do not permit a ferritic solidification mode.In such alloys, sulfur and other contaminants,such as phosphorus, oxygen, zinc, and copper,must be excluded from the weld zone. Welds ofless highly alloyed austenitics, generally thosewith less than 20% chromium, which are bal-anced to freeze in a ferritic mode, retain someferrite at room temperature, normally between3 and 10%. This is not harmful since the ferriteis richer in chromium and in molybdenum, ifpresent.

The amount of ferrite expected can be meas-ured by magnetic devices and estimated fromthe Schaeffler diagram, a useful empirical map-ping of weld metal phase composition shown inFig. 1. This diagram has an arbitrary coolingrate resembling that of tungsten inert gas (TIG;described in a separate section of this chapter)welds. Faster or slower cooling will change therelative amounts of ferrite and austenite becauseof the need for diffusion to achieve the most sta-ble phase balance. Very rapid cooling, as withlaser welding, tends to make austenitic weldsless ferritic and has the opposite effect in duplexalloys.

The Schaeffler diagram has been improvedby the Welding Research Council’s adoption ofthe modification shown in Fig. 2, which super-

Fig. 1 The Schaeffler diagram. Source: Ref 1

Chapter 17: Welding / 203

imposes the solidification mode as a function ofthe composition. The crucial line on this dia-gram is dotted-dashed line AF, which delineatesthose compositions that solidify in a primaryferrite mode, precluding the problem of inter-granular solidification cracking.

Another problem particular to the morehighly alloyed grades is the formation of inter-metallic phases from long cumulative exposureto temperatures in the 600 and 900 °C (1110 and1650 °F) range (coincidentally, the same as forcarbide precipitation). The slow diffusion of al-loying elements in austenitics makes this alesser problem than in ferritics or duplex. Thisadverse precipitation is largely prevented in themodern, nitrogen-alloyed grades, so these al-loys are recommended if extensive welding isplanned. The more highly alloyed grades alsosuffer from greater microsegregation during so-lidification. This causes austenitic dendriticcores to have lower chromium and molybde-num content and consequently lower corrosionresistance. Thus, the welds have lower resist-ance to localized corrosion. This is addressed byusing more highly alloyed filler metal or by so-lution annealing the welds.

Restricting heat input to under 16 kJ/mm(400 kJ/in.) and interpass temperature to under150 °C (300 °F) helps to minimize each ofthese risk factors inherent to the more highlyalloyed austenitic grades. Note that the influ-ence of microsegregation of alloying elementsis separate from and in addition to the negativeinfluence of sulfur on the corrosion resistanceof welds. Austenitics at the alloy level of 316and above should not have sulfur above0.001% for these alloys to deliver the expectedcorrosion resistance.

The austenitic stainless steel weld metal com-position can be altered by the gases to which themolten base metal is exposed. Lack of shieldingcan lead to oxygen combining with chromiumand other elements, creating slag and depletingthe alloy of needed elements. Thus, oxygen-freegas mixtures are used to exclude the ambient at-mosphere from the molten pool during electricarc welding. Inert gases provide the barrier,while the addition of 3 to 5% by volume of ni-trogen gives the necessary partial pressure toensure that welds will not be depleted of vitalnitrogen content. Figure 3 (Ref 1) shows the in-fluence of nitrogen content of the shielding gas

Fig. 2 Welding Research Council’s (WRC’s) 1992 constitution diagram


on corrosion resistance of a highly alloyedaustenitic grade. Excess nitrogen in the shield-ing gas (e.g., more than 10%) can cause poros-ity in the weld, and greater than 5% is detrimen-tal to the life of the tungsten electrode.

The heat from welding can produce a surfaceoxide composed mainly of iron and chromium.The underlying surface can be significantly de-pleted of chromium because of the loss ofchromium to this scale and therefore signifi-cantly lower in corrosion resistance. Pits canstart in this thin layer and propagate into soundmetal beneath. For heat-tinted surfaces, thedarker the tint, the stronger will be the effect. Tofully restore corrosion resistance, the area mustbe ground to remove the oxide and any depletedbase metal. This should be followed by acidpickling, which completes the removal of theoxide and depleted zone.

Duplex stainless steels differ from austeniticstainless steels in their metallurgical response towelding mainly because their approximately50% ferrite causes greater thermal conductivityat lower temperatures, and ferrite has greaterdiffusion rates. These alloys solidify in a com-pletely ferritic mode, and since ferrite rejectslittle sulfur on solidification, hot shortness is nota problem. So, compared to austenitic stainlesssteels, duplex stainless steels have the followingdistinguishing factors:

• The ferritic solidification mode of duplexstainless steels provides very good hotcracking resistance. The rapid cooling ofwelds produces welds and HAZ with moreferrite than the parent metal by quenching inthe high-temperature ferrite.

• Duplex alloys are more sensitive to prob-lems in the HAZ because their generallyhigh chromium and molybdenum contentplus their ferritic content make the precipita-tion of embrittling intermetallic phases morerapid than in austenitics, so minimizing thetotal time at high temperature is the overrid-ing concern.

• While carbide sensitization is not an issuewith the duplex alloys, the formation of in-termetallic phases can cause loss of corro-sion resistance.

• Duplex, like all stainless types, must be pro-tected from oxidation by shielding gas, andsince nitrogen is a crucial alloying element,especially in duplex alloys, it must be acomponent of the gas mixture.

• Cleaning before and after welding is equallyimportant in duplex as in austenitics.

Modern duplex alloys derive their impressivestrength, toughness, and corrosion resistancefrom their nearly equal percentage of ferrite andaustenite. The nitrogen content of the austenitebrings its corrosion resistance up to that of theferrite phase, which is richer in chromium andmolybdenum. Nitrogen additions partition tothe austenite and thus both strengthens it and in-creases its corrosion resistance to close to thatof the ferrite. The early duplex alloys had a ten-dency to form excessive ferrite when weldedand formed embrittling intermetallic phasesrather rapidly. The additions of larger amountsof nitrogen stabilized the austenite to highertemperatures, so welds did not become so fer-ritic. The nitrogen also decreased the speed atwhich intermetallic phases form, enlarging thetime window for welding without their precipi-tation. And, by promoting greater austenite for-mation at high temperature, the addition of high(>0.12%) nitrogen actually reduces the ten-dency for chromium nitride precipitation. De-spite these advances, the key precaution inwelding duplex alloys is to prevent the forma-tion of embrittling phases while preserving asclose to a 50/50 austenite/ferrite structure aspossible. Minimizing time at red heat tempera-tures (500 to 900 °C, 930 to 1650 °F) is the ob-jective. But, sufficient time must be spent aboveabout 1000 °C (1830 °F) to promote the forma-tion of sufficient austenite. If the weld cannot beannealed, increased nickel filler metal (e.g.,2209 with 2205 base metal) should be used.Thus, joint preparation must be done correctlyand not left to the welder to correct using time-consuming remedial procedures.

Fig. 3 Effect of weld shielding gas composition on crevicecorrosion resistance of autogenous welds in AL-6XN

alloy tested per American Society for Testing and Materials(ASTM) G-48B at 35 °C (95 °F)


Duplex stainless steels, because of their mod-erate thermal expansion and higher thermalconductivity, can tolerate relatively high heatinputs since these factors determine the stressintensity that will be generated by thermal gra-dients. However, excessively low heat inputscan result in fusion zones that are predomi-nantly ferritic, with a resultant loss of toughnessand corrosion resistance. At the other extreme,heat inputs that are too high lead to the forma-tion of embrittling intermetallic phases. Thisissue concerns the HAZ, which must dwell inσ-forming temperatures for some period oftime. The key is to limit the time at those tem-peratures by not permitting interpass tempera-tures to exceed 150 °C (300 °F) because work-piece temperature has the greatest influence ontime at σ-forming temperatures. It is prudent toimpose this limitation when qualifying the weldprocedure and then monitoring the productionwelding interpass temperature electronically toensure qualifying procedures are not more le-nient than are those of production.

Postweld stress relief is not needed for duplexweldments and indeed could be harmful be-cause of the danger of embrittlement. Full an-nealing can be done and can restore the originalphase balance and composition that gives theoptimal toughness and corrosion resistancefound in wrought material.

Ferritic stainless steels can be split into twogroups for purposes of welding: the older semi-ferritic group and the more prevalent stabilizedferritic group. The first group, in whichchromium is between 16 and 18% with carbonup to 0.08%, is exemplified by the alloy 430.These alloys form appreciable amounts ofaustenite when heated above 800 °C (1470 °F).Unless they are cooled extremely slowly, moreslowly than can be done in welds, the austenitetransforms to martensite, which is very brittle.

The stabilized grades commonly use titaniumor niobium to combine with the carbon and ni-trogen, which otherwise would cause the forma-tion of the high-temperature austenite, render-ing the alloys ferritic at all temperatures.

The salient metallurgical characteristics forwelding of the two groups are:

• Both groups offer good thermal conductivityand low thermal expansion.

• Both groups require protection from oxida-tion by shielding gases. The stabilized groupshould not be exposed to nitrogen.

• The semiferritic group will form martensite,which requires annealing to eliminate.

• The stabilized group can lose toughness viaexcessive grain growth.

• The grades more highly alloyed withchromium and molybdenum can form α' andσ, leading to embrittlement.

The semiferritic alloys such as 430, 434, and436 are seldom welded and often called un-weldable. The reason is that the welds are in-variably partially martensitic and thus normallybrittle. Only very specially controlled composi-tions of 430 can be welded successfully, andthese are not generally available commercially.While the technical remedy for this is simplyannealing, it is seldom economically viable. It israre to see any welding more extensive thanspot welding of unexposed surfaces with thesealloys. If for some reason they must be used andwelded, then the techniques for weldingmartensitic stainless steels should be employed.

The stabilized ferritic stainless steels arecommonly welded. The levels of stabilizing ele-ments required to prevent austenite formationand sensitization are well known and are re-flected in the alloys’ chemistry specifications.For 409, the required titanium level is Ti > 0.08+ 8(C + N), while the requirement for the higherchromium 439 is 0.20 + 4(C + N). These areempirical relationships that take into accountthat some titanium oxidizes before it can stabi-lize carbon and nitrogen. Niobium can replacesome titanium. This is discussed in detail inChapter 8 on ferritic stainless steels. Because ofthe low toughness these alloys have in largecross sections, these alloys are only rarely seenwith minimum section size of more than 3 mm(0.11 in.) and normally have sections less than 2mm (0.08 in.). Thus, successful welding is sim-plified to making a sound, well-shielded weldwithout producing excessive grain growth in theHAZ. In practice, this can be achieved by limit-ing heat input to less than 6 kJ/cm. An empiricalrelationship between grain diameter D and heatinput E (kJ/cm) has been reported (Ref 2). Inthe fusion zone, the relationship is:

D = 206 × E – 585.6 (Eq 1)

In the HAZ, it is:

D = 29.6 × E – 50.6 for up to 6.6 kJ/cm (Eq 2)

and

D = 75 × E – 350 above 6.6 kJ/cm (Eq 3)

The light gauges ensure sufficiently shorttimes at high temperature that precipitation of


intermetallic phases should not be a concern,even though they can form, especially in super-ferritic alloys.

The impact properties of ferritic stainlesssteels are always a concern because their transi-tion temperature can become elevated to ambi-ent levels. It has been determined that there ex-ists an optimum level of titanium around 0.10%,which ensures this minimum transition temper-ature (Ref 3). Because it is difficult to have lowenough carbon plus nitrogen to stabilize at thistitanium level, dual stabilization with titaniumand niobium as well as not having excessiveheat input are the best way to ensure weldtoughness.

Especially in the superferritics, maintainingthe benefits of having the fairly precise balanceof carbon plus nitrogen to the stabilizing ele-ments titanium and niobium requires that nei-ther carbon nor nitrogen come into contact withthe weld pool. Likewise, oxygen must be rigor-ously avoided because it will quickly depletethe essential titanium, which is even more read-ily oxidized than chromium. Extraordinary sur-face cleaning at and near the weld will pay divi-dends in final quality.

Martensitic stainless steels vary little in alloycontent, ranging from 11 to 18% chromium withsmall amounts of nickel and molybdenum.Their carbon content ranges from 0.10 to over0.30%. Thus, the major challenge they presentis avoiding the potential cracking, which ismost likely to occur in the HAZ from stressescaused by the austenite-to-martensite transfor-mation on cooling. Since this transformationcannot be avoided, the desired approach is tostart with a well-tempered or annealed materialand then preheat and maintain high interpasstemperatures. For low carbon levels, below0.10%, preheat can be omitted, but between0.10 and 0.20% carbon, preheating to 250 °C(480 °F) is advised and for higher carbon levels,300 °C (570 °F). The problem becomes moresevere with increasing carbon level because thetransformation takes place at lower tempera-tures in more brittle material. Even with pre-heating, distortion may be encountered. For allnormal uses of martensitic stainless steels, afinal heat treatment is required to achieve thequenched and tempered properties for whichthese alloys are designed.

Aside from the cracking consideration,martensitic welding considerations are similarto, but less stringent than, those of low-alloystabilized ferritic stainless steels with regard to

cleanliness and shielding. If mechanical re-quirements permit, the use of austenitic (309L)weld filler metal should be considered. The softjoint may deform to accommodate thermalstrains and thus minimize weld cracking.

Precipitation-Hardening Stainless Steels.Last, precipitation-hardening (PH) stainlesssteels, while very complex metallurgically, arestraightforward from a welding perspective.Obviously, any heat treatment to achieve theproperties of which these alloys are capablemust be a final step. The considerations in weld-ing them are:

• Shielding must be sufficient to prevent lossof oxidizable alloying elements such as tita-nium, aluminum, and, of course, chromium.

• Filler metal must match the base metal iflike properties are required.

• Postweld heat treatment solution annealingmust be adequate to homogenize weld solid-ification segregation.

• Austenitic PH grades are fully austenitic andsubject to hot short cracking.

• The high aluminum or titanium contents ofmany PH alloys cause their welds to be“slaggy,” and these slaggy welds have are ir-regular with objectionable recesses,crevices, or prominences.

These alloys are easily welded and not proneto cracking or developing embrittling phases.But, because these alloys are designed for ex-treme mechanical performance, it is essential topreserve their correct chemistry by shieldingwith a fully inert gas mixture. If mechanicalproperties equal to that of the base metal are notrequired in the weld, then austenitic filler, suchas 309L, can be used.

Table 1 summarizes the major metallurgicallyimportant parameters for the various types ofstainless alloys. It is prudent to consult with themanufacturer’s data sheets for specific recom-mendations on alloys that they produce as theyare often privy to test data and user experiencethat cannot be found elsewhere in the literature.

Material Selection and Performance

Stainless alloys that are prone to precipitationof intermetallic phases require special preweld-ing consideration. Such alloys include duplex,superferritic, and superaustenitic alloys. Anyamount of time for which these alloys havebeen exposed to temperatures at which inter-


metallic phases form without full subsequenthomogenization anneal is time that the weldercannot use to complete a satisfactory weld be-fore precipitation occurs. Thus, accurate knowl-edge of material history is vital. Likewise, vari-ations within specification of nitrogen contentinfluence the time it takes intermetallic phasesto form. Once a welding procedure is qualifiedfor an alloy with given nitrogen content, use oflower nitrogen alloys would not be prudent en-gineering practice.

Austenitic stainless steels that are intendedfor autogenous welding are often specified withelevated sulfur levels, on the order of 0.005 to0.015%. This is done to improve weld penetra-tion through the so-called Marangoni effect.

This effect exploits the temperature-dependentsurface concentration of sulfur in the weld pool,which causes a decreased surface tension to-ward the hotter center of the pool, causing themolten pool to flow toward the center on thesurface and then flow downward, shooting thehottest metal to the bottom of the weld pool, asshown in Fig. 4. This speeds welding and mini-mizes weld and HAZ width, which is a goodthing. The effect on corrosion resistance is lessdesirable since the abundant MnS inclusionsthat result from the higher sulfur levels decreasepitting resistance. This decrease in corrosion re-sistance can only be eliminated by a long an-neal. Unfortunately, the pipe purchaser cannotknow if the pipe has had a sufficient anneal.

Table 1 Welding parameters for various stainless steels

Alloy group FillerHeat input

kJ/cm(max)Shielding

gas PreheatInterpass

maxPostweld heat treat

Austenitic . . . 20–40 Ar+2% O2,Ar/3% CO2/2% H2He+7.5%Ar+2.5 CO2

150 oC 150 oC None or full anneal

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

argon + 3–5% N2:no O2

50 oC 100 oC None or full anneal

PH grades Same as base alloy 20–40 Argon/helium no . . . Full solution anneal

Martensitic

410 410, 308, 309L 20–40 Ar+2% O2,He+7.5%Ar+2.5 CO2

250oC 250 oC min Slow cool

420 420, 308, 309L, 310 20–40 Ar+2% O2,He+7.5%Ar+2.5 CO2

250 oC 250 oC min Anneal

440 440, 308, 309L, 310 20–40 Ar+2% O2,He+7.5%Ar+2.5 CO2

250 oC . . . . . .

Supermartensiic Same as base metal 20–40 Argon/helium no . . . Full solution anneal

Ferritic

430 430, 309L 20–40 Ar+2% O2,He+7.5%Ar+2.5 CO2

no . . . Subcritical anneal

434 309 Mo L 20–40 Ar+2% O2,He+7.5%Ar+2.5 CO2

no . . . Subcritical anneal

409 410L, 308, 309L 6.0 Ar+2% O2,He+7.5%Ar+2.5 CO2

no n.a. none

439 439L, 309L, 316L 6.0 Ar+2% O2,He+7.5%Ar+2.5 CO2

no n.a. none

Superferritic 29-4C 6.0 Argon/helium no n.a. None or full anneal2003, 2101, 2304,19-D

2209 5–25 Argon + 3% N2 no 150 oC None or full anneal

2205 2209 5–25 Argon + 3% N2 no 150 oC None or full anneal

25 Cr duplex 25Cr-10Ni-4Mo-N 5–25 Argon + 3% N2 no 150 oC None or full anneal

2507superduplex

25Cr-10Ni-4Mo-N 5-25 Argon + 3% N2 no 150 oC None or full anneal

PH, precipitation hardenable


In-line induction annealing is insufficient forthis purpose. Furnace anneals of about an hourare required. For alloys like 304L and 316L, theuser should always require material chemistrycertifications and assume that any sulfur levelsabove 0.003% are going to result in decreasedpitting resistance of 1 to 5 PREN (pitting resist-ance equivalent number), which means up to 10°C (18 °F) decrease in critical pitting tempera-ture, roughly the difference between 304 and316 in performance. This also applies to girthwelds done by the pipe user.

Welds are essentially a casting in the midst ofwrought material. In addition to inclusions de-creasing weld corrosion resistance as men-tioned, solidification segregation can also causemicroscopic regions to be poorer in corrosion-resisting alloying elements chromium, molyb-denum, and nitrogen. This effect is minimal forlow-alloy material, but for highly alloyedaustenitic grades, it is a major effect, as shownin Fig. 5. Eliminating this effect requires a thor-ough homogenization anneal.

The use of filler metal with higher corrosionresistance does not totally offset the influence ofwelding on corrosion resistance because someof the base metal melts and is not altered incomposition by the filler metal. This is calledthe unmixed zone. It is essentially a zone withproperties equal to that which would occur in anautogenous weld, that is, the corrosion resist-ance is lower depending on total alloy level andsulfur content.

Welding Processes

All stainless steels should be very clean priorto welding. The chemistries of both base metalsand filler metals are carefully formulated to pro-duce the mechanical and corrosion propertiesthat these alloys have been designed to produce.Virtually any contaminant can either interferewith the welding procedure or detrimentally

alter the composition of the welded joint, whichin turn can alter corrosion and mechanical prop-erties and compromise the entire structure.Moisture, paint, dirt or grease, oil, and oxidesall can negate good material, good weldingtechnique, and good procedural qualification.Cutting fluids, especially sulfurized oils, are es-pecially detrimental and should be removedcompletely prior to welding. Preheating is neverstrictly forbidden since it is required to elimi-nate moisture.

Joint design does not differ in principle fromthat of other steel weldments. There is, howeveran increased need for dimensional uniformityfor the alloys susceptible to intermetallic precip-itation since minimizing time at temperature is apriority, and variations in joint geometry impedethe swift completion of the weld. This is alsotrue for alloys that are susceptible to excessivegrain growth, such as the stabilized ferritics, orto sensitization.

Figure 6 shows some joint designs appropri-ate to stainless steels, including the more sensi-tive alloys. These, like all joint designs, aim toensure full penetration without burn through.

Gas tungsten arc welding (GTAW)/tung-sten inert gas (TIG) is commonly used for theautomated production of stainless steel pipe andtube, as well as manual short runs. It is versatileand generally used when thicknesses are lessthan 6 mm (0.2 in.). It can produce very high-quality welds. A constant-current power supply ispreferred. It is best performed with the DCSP (di-rect current straight polarity) electrode negative

Fig. 4 Metal flow directions in a weld pool with (left) andwithout (right) sulfur. Source: Adapted from Ref 4

908580757065605550454035302520151050

−5

194185176167158149140131122113104958677685950413223

Crit

ical

pitt

ing

tem

pera

ture

in 6

% F

eCI 3

, °F

Crit

ical

pitt

ing

tem

pera

ture

in 6

% F

eCI 3

, °C

Molybdenum, wt%

Unwelded

Welded

1 2 3 4 5 6 7

Fig. 5 The influence of molybdenum on critical pitting tem-perature. Source: Adapted from Ref 5


technique. It is helpful to incorporate a high-frequency circuit to aid in establishing the arc.Thoriated electrodes containing 1.7 to 2.2% tho-ria are recommended because they have betteremissive properties and provide better arc stabil-ity at higher currents. If consumable electrodesare used, the shielding gas precludes the need forcoatings. The weld metal alloys are not necessar-ily the same as the parent alloys but are chosenbased on their ability as weld metals to providethe most acceptable corrosion and mechanicalproperties. This sometimes means usingaustenitic filler with a ferritic base or highernickel content in an austenitic or duplex base tocompensate for the solidification rate or inher-ently lower corrosion resistance of the weld.

The shielding gas must replicate the con-trolled gas mixtures used to refine stainless steeland establish the original composition. Theweld pool exposes a great deal of surface area tothe atmosphere in a very turbulent manner. Gasflows, usually 12 to 18 L/min, must be adequateto prevent air infiltration by aspiration or turbu-lence before arc contact, ideally until tempera-tures cool to below oxidation temperatures.

For manual GTAW using a filler wire, thewire should be fed continuously into the weldpool. Intermittent wire addition can lead to cre-ation of zones of essentially autogenous weld,negating many of the benefits of filler metal ad-dition. Moving the tip of the wire in and out ofthe protection of the gas shield is especially

t

d

t

d

d

a

d

a

d

a

k

d

a

k

d

a

r = 6-8mm

r = 6-8mm

Fig. 6 Joint designs. Courtesy Ugine S.A.

Thickness Gap d, Root K, Groove Process th, mm (in.) mm (in.) mm (in.) Bevel α(°)

GTAW 3–5 1–3 . . . . . .

GMAW 3–5 1–3 . . . . . .

SMAW 3–4 1–3 . . . . . .

SMAW 4–15 1–3 1–2 55–65

GTAW 3–8 1–3 1–2 60–70

GMAW 5–12 1–3 1–2 60–70

SAW 9–12 0 5 60

SMAW >10 1.5–3 1–3 55–65

GMAW >10 1.5–3 1–3 60–70

SAW >10 0 3–5 80

SMAW >25 1–3 1–3 10–15

GMAW >25 1–3 1–3 10–15

SAW >25 0 3–5 10–15

GTAW >3 0–2 . . . . . .

GMAW >3 0–2 . . . . . .

SMAW >3 0–2 . . . . . .

SMAW 3–15 2–3 1–2 60–70

GTAW 25–8 2–3 1–2 60–70

GMAW 3–12 2–3 1–2 60–70

SAW 4–12 2–3 1–2 70–80

SMAW 12–50 1–2 2–3 10–15

GTAW >8 1–2 1–2 10–15

GMAW >12 1–2 2–3 10–15

SAW >10 1–2 1–2 10–15

GMAW, gas metal arc welding; GTAW, gas tungsten arc welding; SAW, submerged arc welding; SMAW, shielded metal arc welding


bad. The hot tip can carry oxides and nitridesinto the weld, defeating the action of the shieldgas and impairing weld quality.

Gas metal arc welding (GMAW) is arcwelding in which a consumable electrode pro-vides larger amounts of filler weld metal thanpractical in GTAW. There are three GMAWtechniques:

• Pulsed arc transfer• Spray transfer• Short-circuiting transfer

Pulsed arc transfer employs a power sourcethat is switched rapidly to provide transfer ofweld metal droplets at regular intervals. Spraytransfer uses a high current to form a stream offine drops from the end of the electrode. This isdone with high power, resulting in a large fluidweld pool, and therefore limits the technique tohorizontal orientations and thick material.Short-circuiting transfer uses arc contact withthe workpiece at low power to melt the elec-trode, after which the short circuit is broken,and material transfer ceases. The technique cre-ates a minimal weld pool and is viable in manyorientations. It is a low-heat process suitable forthin material but may cause lack of penetrationdefects if used for thick-section welding.

For all GMAW processes, excessive protru-sion of the wire should be avoided; otherwise,the full benefit of the inert gas shielding may belost.

Submerged arc welding (SAW) employs aconsumable electrode immersed in a conductiveflux that acts as a protective shield from the at-mosphere. The arc is struck through the flux,and gravity deposits the molten metal to theworkpiece. The large weld pool has high heatinput and can deposit large amounts of metalrelatively quickly. Thus, SAW may be prefer-able to multipass techniques for alloys such asduplex for which time at temperature is limited.It is restricted to horizontal orientations and re-quires postweld slag (flux) removal.

Shielded metal arc welding (SMAW) isdone manually with short lengths (“sticks”) ofcoated electrodes. This method has great versa-tility with some trade-off in cost and quality.This last aspect is arguable, but the lack ofshielding gas may introduce oxygen to the weldmetal, which can be detrimental to toughness.

Flux cored wire (FCW) welding is a methodthat is able to accommodate a large range ofthickness and orientations while providing highdeposition rates. The equipment is the same as

for GMAW, but the consumable electrode, theFCW filler metal, has a flux core that supple-ments the shielding gas. Because of the flux, theshielding requirements are reduced; gases canbe argon/25% carbon dioxide for horizontalwelding with current and voltages from 150 to200 amp and 22 to 38 V, respectively. Verticalwelds can use 100% carbon dioxide with am-perage of 60 to 110 amp and voltage of 20 to 24V. Flow rates of gas are 20 to 25 L/min. It ispossible to get high-carbon welds, which maynot resist corrosion as well as desired, so as al-ways, weld qualification, including corrosionevaluation, is critical.

Oxyfuel gas welding (OFW), “torch” weld-ing, uses oxygen to accelerate fuel (typicallyacetylene) combustion to produce temperaturesthat can melt steels. By controlling the fuel-airmixture, the flame can be made nonoxidizingfor low-alloy steels. However, these “neutral”flames can simultaneously oxidize and carbur-ize stainless steels. Thus, the OFW process isnot suitable for use with stainless steels.

Laser welding has become a major produc-tion method when it can be automated, as forpipe and tube or high-production manufactureditems, such as air-bag canisters. Metallurgically,it resembles resistance welding in that bothhave minimal HAZ and very high quenchingrates, both of which can have a pronounced ef-fect on some types of stainless steel. The effectis to undercool the molten metal and suppressthe transformation that would normally occur.So, an austenitic alloy that normally solidifies ina ferritic mode before transforming to austenitefreezes directly as austenite. The freezing is sorapid that the normal hot shortness of austeniticsolidification is avoided, so quality is not com-promised. In fact, laser welds quench the mate-rial so rapidly that corrosion resistance is en-hanced since inclusions cannot nucleate andgrow. Duplex alloys, on the other hand, freezein their high-temperature ferrite structure be-cause the fast quench prevents the nucleationand growth of austenite. Unless this ferrite isheated to permit austenite to form, lower-tough-ness welds will result. Ferritic, martensitic, andPH alloys are not harmed by the rapid quench.

Resistance welding is readily done on mosttypes of stainless steel. Allowance must be madefor the lower thermal and electrical conductivityof stainless steels compared to other commonmaterials. Most resistance welds, including bothseam and spot welds, have deep, tight crevicesadjacent to the welds. The possibility of crevice


corrosion in these regions should be consideredwhen contemplating the use of spot welds instainless materials. The possibility of entrap-ment of foreign material and the difficulty of re-moving it from such crevices should also beconsidered, especially in equipment for foodhandling, pharmaceutical production, etc.

High-frequency induction welding ofstainless steel is more difficult than for low-alloy steel because of the refractory nature ofchromium oxide, which has a higher meltingtemperature than does the stainless base metal.This is opposite from the situation in low-alloysteels, for which the iron oxide melts at a lowertemperature than does the iron base metal. Thepresence of this refractory oxide on the surfacesto be joined makes it more difficult to obtain adefect-free weld.

Thermal cutting of stainless steels is rou-tinely practiced, but the processes and parame-ters used are determined by the refractory na-ture of the chromium oxides that form onstainless steels. The high temperatures attain-able with lasers or plasma arc torches providegood cutting action, and these processes are fre-quently used. To expand the range of thick-nesses that can be cut or to increase cuttingspeed, supplemental oxygen or nitrogen blastjets may be used. Stainless steels may also becut using oxyfuel equipment if supplementaliron powder is used. Combustion of the iron in-creases the temperature, while the iron oxidehelps flux the refractory chromium oxide. Ther-mally cut edges of stainless steel usually requiresubsequent cleaning, typically by grinding ormilling. Chemical cleaning of all surfaces of cutpieces to remove heat tint, fume deposits, andother contaminants is advisable.

Soldering and brazing are possible with allstainless steels. Soldering is done below 450 °C(840 °F), while brazing is done above 450 °C(840 °F). Solders are generally alloys of tin andbismuth, lead, silver, or antimony or combina-tions of several of these. Brazes are normally ei-ther silver based or nickel based. Thechromium-rich oxide coating must be removedby a suitable flux for bonding to occur. Fluxesare typically acid type with chlorides. Thus,after the soldering or brazing, the flux must bethoroughly removed to prevent subsequent pit-ting corrosion. Brazing temperatures must bechosen to avoid ranges at which unfavorablephases form. The best range can be determinedfrom examining temperature ranges to beavoided in the thermal processing chapter

(Chapter 13) of this book. Brazes and soldersrarely match the corrosion resistance of stain-less steels, and careful attention should be givento the potential for galvanic and other forms ofcorrosion when considering the use of solderedor brazed joints with stainless steels.

Welding Practices

Safety must always be considered whenwelding. In addition to the normal hazards(which are not discussed here) associated withwelding, welding of stainless steels presents aspecial hazard: hexavalent chromium. The fumecreated by welding stainless steel contains sig-nificant concentrations of chromium trioxideand other forms of hexavalent (Cr+6) chromium.Hexavalent chromium is a carcinogen and regu-lated by the Occupational Safety and HealthAdministration (OSHA). Exposure to and in-halation of stainless steel welding fumes mustbe avoided. The product exposure limit forhexavalent chromium is 5 μg/m3 as of Decem-ber 31, 2008. Refer to OSHA for further updateson this limit. Use of fume extraction equipmentis generally the preferred method of minimizinghexavalent chromium exposures. Positioningand operation of the fume extraction devicemust be done precisely to ensure effective fumeremoval while avoiding excess turbulence,which can cause loss of effective inert gasshielding of the weld pool. Thermal cutting ofstainless steels also generates hexavalentchromium, and similar procedures are requiredto minimize exposure during such operations.

Nondestructive Evaluation (NDE) is usedalmost universally to ensure weld quality. All ofthe standard NDE techniques used with othermaterials are applicable to stainless steel weld-ments. Allowance must be made for the differ-ing physical properties of stainless steels, andappropriate reference defect standards must beprovided. However, one technique—magneticparticle inspection—is problematic. The pres-ence of bands of persistent austenite in marten-sitic or PH stainless steels can lead to spuriousdefect indications. For this reason, magneticparticle examination of stainless steel welds isbest avoided.

Recent developments in stainless steel havebeen made with weldability as a major consid-eration. Highly alloyed, low-carbon martensiticalloys for line pipe have been developed with


the express purpose of use in the as-welded con-dition. The low carbon makes welds of this ma-terial that are tough and do not require temper-ing, so girth welds in the field are possible.

Likewise, the lean duplex alloys have verydelayed precipitation of intermetallic phases be-cause of their higher nitrogen and lowerchromium and molybdenum contents. Thismakes welding of these alloys much more fool-proof than with the early duplex alloys, such asS31803. The dual-stabilized ferritic alloys havetougher welds than those stabilized with only ti-tanium or niobium.

New developments in welding also have animpact on stainless steels. The friction stir weld-ing (FSW) process offers the promise of reliablesolid-state joining. By avoiding melting andresolidification, issues associated with solute redistribution are eliminated. The relatively lowtemperatures involved essentially eliminate

generation of weld fume (see the discussion ofsafety). Other new welding processes, such asmultiple (GTA or GMA) torch welding, laser-assisted GMA or GTA welding, etc. promisegreater productivity.

REFERENCES

1. D.J. Kotecki, Welding of Stainless Steels,Welding, Brazing, and Soldering, Vol 6,ASM Handbook, ASM International, 1993,p 677–707

2. B. Aziez and R. Feen, Sheet Metal Ind., 1,1983, p 28–34

3. S.D. Washko and J.F. Grubb, Proc. Int’lConf on Stainless Steel, 1991, Chiba, ISIJ

4. Stainless Steels, Les Editions de Physiques,1992, p 786

5. A. Garner, Corrosion, 37, 1981, p 178

CHAPTER 18

Architecture and Construction

Summary

STAINLESS STEEL IS OFTEN EMPLOYEDas an architectural material; the material can it-self be viewed as a metaphor for architecture, adiscipline that must balance aesthetics, economy,and structural integrity. Stainless steel’s uniquecombination of beauty, strength, and economymakes it a remarkably appropriate material foruses as diverse as sculpture and concrete rein-forcing bar. However, stainless steels are com-plex; they come in many different grades (chemi-cal analyses), and these grades have varyingstrengths, appearance, resistance to corrosion,availability, and costs. The success of a buildingproject involves careful planning for theappropriate use of materials. This chapter dealswith the technology of stainless steel as it per-tains to its proper use in architecture, art, andconstruction.

Corrosion Resistance

Corrosion is the life-limiting factor for archi-tectural metals. Steel, copper, aluminum, lead,bronze, and other alloys react with the environ-ment and degrade over time, as does wood,stone, plastic, paint, and even glass. With stain-less steel, it is possible to choose a material thatcan withstand attack from the environment in-definitely. One of the first major uses of stain-less steel in architecture was in New York’sChrysler building, completed in 1930. Despitethe rather crude methods of early production,limited alloy options, and lack of applicationexperience involved in its construction, thedomed top of the Chrysler building still shinesundiminished by the harsh coastal and urbanclimate (Fig. 1). With today’s technology, one

can choose from a variety of stainless alloyswith sufficient corrosion resistance to withstandany environment.

The ability of stainless steel to resist corro-sion resides in its chromium-rich superficialpassive layer. Stainless steel by definition mustcontain slightly more than 10% Cr. The passivelayer forms spontaneously in air or water, andif it is removed, say by abrasion, it re-forms byitself. This is explained in greater technical

Fig. 1 The Chrysler building with its famous bright stainless details. Copyright © iStockphoto.com/Steven Allen.

Used with permission




detail in this book in the chapters on corrosion,but the key aspects are that the chromiumatoms on the surface of the metal react withoxygen in air and water to form with neighbor-ing iron atoms into a tight, ionically noncon-ductive layer that prevents any further oxygenpenetration. This layer is mere atoms thick andcompletely invisible.

The strength of the layer in resisting corrosiveattack is proportional to several key alloying elements: chromium, molybdenum, and nitro-gen. They contribute according to the followingformula:

PREN = %Cr + 3.3(%Mo) + 16(%N) (Eq 1)

PREN stands for pitting resistance equivalentnumber. This number can be related to resist-ance to the mildest form of corrosive attack thatstainless steel undergoes, pitting corrosion.

Pitting is a “weakest link” phenomenon inwhich corrosion begins in small, micron-sizeparts of the surface and then grows by virtue ofthe more aggressive media that form withinthem because of the corrosion reaction’s prod-ucts. Pitting occurs in environments that containchlorides. Chloride ions compete with oxygenand disrupt the integrity of the protective pas-sive layer. As alloys become richer in the alloy-ing elements mentioned, their ability to maintainthe passive layer can overcome the chlorides’ability to destroy it. The key is to choose analloy rich enough in chromium, molybdenum,and nitrogen to withstand any environment thestructure will experience. Pitting corrosion, andanother similar form of corrosion called crevicecorrosion, can be prevented by proper choice ofalloy, finish, and design. Crevice corrosion oc-curs when recessed spaces are small enough toact like a corrosion pit. The acidity within thecrevice increases because of restricted diffusionin and out of the crevice, just as happens withina pit. The buildup of iron and chloride ionsmakes a very corrosive medium that disablesthe passive film formation. A design that avoidscrevices is the best defense.

The decision criterion for material selectionwith stainless steel should then be, Which gradeand finish will exclude the possibility of pittingcorrosion at the lowest cost? Then, proper de-sign should be used to exclude the possibility ofcrevice corrosion. Other factors involvingstrength and fabrication should also be consid-ered. Table 1 ranks a number of stainless alloysby pitting resistance.

Balancing Corrosion Resistance, Processing Characteristics, and Economy

The rule of thumb for grade selection is usu-ally somewhat oversimplified to recommend theuse of types 430 and 304 on interior applica-tions, type 304 on exteriors where salt is not aproblem, and 316 where road salts or seacoasteffects make a more corrosion-resistant gradenecessary. A leading architectural metals com-pany (Ref 1) makes the following recommenda-tions, which mirror these traditional views:

• Type 304 should be used for most exteriorapplications.

• Type 316 should be used within ten miles ofsaltwater bodies. However, if the building issubject to saltwater spray, a nobler grade ofstainless steel, such as 2101 or 2003, shouldbe specified.

• In close proximity to deicing salt use, evenon nearby roadways where vehicle trafficcan create airborne particles, type 316should be used. If periodic rinsing will notoccur on all exterior surfaces, these areasmust be washed each spring. If dependablemaintenance is not predicted, a nobler gradeof stainless steel, such as 2101 or 2003,should be used.

• Specify types with low carbon, less than0.030%, if welding will be employed.

• Any grade, including type 430, may be usedin interior applications.

• In the most severe environments—high heatand humidity, low rainfall, and high salinity,such as are found in Middle Eastern coun-tries—a grade with a PREN of 25 or aboveis recommended.

These guidelines are based on the admittedlyeasy availability of these alloys and a lack ofconcern for cost during times of peak raw mate-rial prices. For projects where quick availabilityis not more important than cost, 439 and 201should be considered as viable replacements for304. Stainless steel 2003 (UNS S32003) or anequivalent lean-duplex grade can replace 316 ata cost advantage during times of high alloy cost

Table 1 Ranking of common stainless steels bypitting resistance equivalent number (PREN)Alloy 430 439 201 304 316 2101 2003 2205 2507PREN 15 17 17 19 24 26 28 35 38

Higher PREN values indicate greater pitting resistance.

Chapter 18: Architecture and Construction / 215

as have been experienced on occasion, such asduring the period of 2004 to 2007. It should benoted that the leaner alloys suggested (439, 201,and lean-duplex alloys such as 2003) can besomewhat more difficult to form. If panel de-signs call for 90° bends, this is not an issue.However, for applications requiring severeforming, as in the case of a double-lock seam ona standing seam roof, these grades can provide achallenge to the fabricator/installer. Further,these leaner grades can pose challenges withcertain finishing methods, such as abrasive pol-ishing and embossing. To avoid unwanted com-plications related to grade selection, the speci-fier should consult a competent architecturalmetals supplier. This effort will ensure a viablespecification is written that will balance costwith the necessary performance attributes tomake the part as well as resist corrosion onceinstalled.

It is valuable to know, in times of high nickelprices, that both the low-nickel 201 and the no-nickel 439 can be used in place of 304, while2101 (UNS S32101) and 2003 (lean duplexes)can replace 316. To obtain these grades usuallyinvolves working with a producing mill sincethey are not typically stocked in service centerinventories. However, any competent architec-tural metals supplier will not shy away from theuse of specialty grades where appropriate.

Surface Finish and Corrosion Resistance

Surface finish is usually an aesthetic choice,but it has a significant influence on corrosion re-sistance and must be factored into grade selec-tion. Mill finishes such as 2B and 2D are incon-sistent because they are annealed and pickled toremove oxides. These unattractive surfaces,however, have correct corrosion resistance fortheir alloy content. Welding or abrading the sur-face degrades the corrosion resistance by a sig-nificant amount. An un-heat-treated weld haslower resistance to corrosion in proportion tothe alloy content of the grade. Type 316 weldshave the corrosion resistance of wrought 304.Abrasion has a similar effect. Type 316 with aNo. 4 polish behaves like 304 with a 2B millfinish. Because welds are abraded, this com-pounds the effect.

Very smooth abrasively polished finishes mit-igate this reduced corrosion resistance, asshown in Fig. 2. The effect of reduced corrosion

resistance in abraded stainless steel surfaces isnot seen on finishes that are produced by pat-terns imprinted by hard-rolling mill rolls thathave been engraved with the desired pattern inreverse; for this reason alone, this method ofsurface finishing is recommended. These archi-tecturally useful surface finishes are producedby the preferred rolled-on, or embossed,method. In addition to their advantage in corro-sion resistance, they are extremely uniformfrom batch to batch, unlike finishes produced byabrasive-coated belts, which change in gritcoarseness with use.

Balancing Service Environment, Design Requirements, and Maintenance Considerations

An expert system has been developed that en-ables designers and specifiers to analyze thetrade-offs of climate, design requirements, andmaintenance on grade selection (Ref 3). An-swering the questions in Fig. 3 for a particularapplication yields a score that can be used toidentify an appropriate alloy according to thescale shown Fig. 4. The Nickel Institute, for-merly the Nickel Development Institute, alsooffers excellent publications on topics related toalloy selection for specific service environmentsand design requirements (Ref 4).

Reviewing a map of the salinity of rainwaterin the United States is instructive of the degreeto which geography influences corrosion sever-ity. The average atmospheric chloride levelscollected in rainwater are shown in Fig. 5 (Ref5). The highest levels occur along the coastlinesof the Atlantic and Pacific Oceans and the Gulf

Fig. 2 The decrease in corrosion resistance with increasingsurface roughness by abrasion. Source: Ref 2


Fig. 3 Stainless steel selection expert system. Source: International Molybdenum Association (Ref 3)


of Mexico. The maximum corrosion rate is re-lated to the maximum chloride in the atmos-phere. This will be related to the distance in-land, the height above sea level, and theprevailing winds (Ref 6).

Aesthetic Considerations

A correctly chosen grade of stainless steelwill have no degradation over time and, if prop-

erly maintained, will stay new looking indefi-nitely.

Surface finish aesthetics are arguably moreimportant architecturally than the influence sur-face finishes have on corrosion. Numerous fin-ishes have been developed to try to meet vari-ous objectives. Finishes vary in reflectivity,directionality, and subtlety. Figure 6 showssome of the finishes that go beyond the familiarbrushed look (Ref 7), while Fig. 7 shows specialfinishes created by one manufacturer.

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


The classic abrasively produced finishes areNo. 3, 4, and 8. These are American Society forTesting and Materials (ASTM) designations forabrasively produced finishes, which are tradi-tionally produced by abrading the surface withdifferent grit size abrasives. Finish No. 3 callsfor 80 to 100 grit abrasive; No. 4 calls for 120 to150 grit abrasive. Finish No. 8 is a mirror finishobtained by final polishing with 800 grit abra-sive. Finishes No. 3 and 4 are directional, withgrit lines typically 1 cm (0.4 in.) in length. Fin-ish No. 3 has a surface roughness average (Ra)of 0.4 to 0.8 µm (15 to 30 µin.). Specular glossat 85° is typically 40 to 60 (per ASTM D 523,

“Standard Test Method for Specular Gloss”).The standards for appearance do not existwithin specifications, only the method of pro-ducing them. There is considerable differencein appearance from sheet to sheet, coil to coil,and manufacturer to manufacturer. The greatestconsistency of appearance comes from specify-ing a brand of rolled-on finish from a givenmanufacturer. Any of the traditional finishescan be replicated by a rolled-on finish withgreater uniformity, with the possible exceptionof bright annealed having a difficult time match-ing the mirror quality of a No. 8 finish. This iscrucial in architecture, where the discovery of

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)

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


unacceptable visual nonuniformity on largeareas can be disastrous, especially when this ap-pears late in the construction process, as is nor-mally the case with exterior components.

Reflectivity or gloss can be a major consider-ation in the choice of a surface finish. Mirrorfinishes are often used for high impact, butmore diffusely reflecting surfaces are morecommon. Patterned surfaces provide consistentreflectivity from a moderately reflective 40 to60 specular gloss at 85° to a dull matte of lessthan 20, the latter having been developed forairport roofing, such as at Reagan Airport inWashington, D.C., or the Pittsburgh ConventionCenter (Fig. 8).

Flatness is a special consideration for panelswhere lack of flatness, such as by “oil-canning,”can cause a very shoddy appearance. Flatness ismeasured in I units.

Flatness (Iunits) = 2 (πH / 2L) × 105 (Eq 2)

where H is the height of the deviation from flat-ness, and L is the distance between peaks of de-viations, assuming a sinusoidal wave.

Because stainless in sheet form is usually re-flective, small deviations from flatness can bevery visible. A good standard for flatness thatprecludes visible distortion is five I units. Steelproducers have various means to produce thislevel of flatness, the most extreme of which isactually stretching the steel sheet or coil untilall distortions are eliminated. Sometimes,rather than aiming for high flatness a controlleddeviation from flatness is used, such as slightlyconcave panels or panels with a die-pressed design. Another option to ensure flatness is toback light-gauge stainless steel with a stiff material.

Deviations of sheets from squareness andstraightness (camber) are also objectionable be-cause such deviations can cause gaps betweenpanels. The degree to which this is objection-able is a function of design, and tolerances canbe held tightly at a cost. Width tolerance is normally +1/16 in./–0 in 48 in., while length isheld to +1/8 in./–0 in 10 ft or less. Maximumcamber is 3/32 in. in 8 ft. Closer tolerances canbe negotiated.

Fig. 8 The Pittsburgh Convention Center with low-gloss finish stainless steel roof


Maintenance and Repair

Maintenance is a significant cost of any struc-ture. One of the great values of stainless steel isits low cost of ownership. Stainless steel can beabused, however, and it does benefit fromproper maintenance. The main objective of themaintenance of stainless is keeping it clean.There are two reasons for this. The obvious firstreason is that whatever is soiling the surface isprobably not attractive. The second reason isthat it may harm the surface by allowing corro-sion agents to concentrate. Table 2 provides rec-ommended practices for removing various sub-stances from stainless steel surfaces (Ref 9).

One of the most common complaints aboutmaintenance of stainless is the work involved inremoving fingerprints. The oil from fingerprintsmakes an easily visible interference film on thereflective stainless surface. The typical remedyis to clean stainless steel with a solution con-taining light oil and a detergent. If the oil from ahand contacts the uniformly thin film of clean-ing oil, no visible mark is left. Alternatively,

polymer coatings are applied to stainless steel atsome producing mills to permanently provide afilm to which additional fingerprint oil cannotadd a noticeable discoloration.

The greatest ally of stainless on building exte-riors is the cleansing action of rain. Rain doesnot completely clean the surface, but it does di-lute any harmful contaminants and forestall cor-rosion from accumulated chlorides. Without thebenefit of cleansing by rain, stainless exteriorsshould be washed during routine window wash-ing operations.

Given the importance of cleaning, the follow-ing design considerations are recommended:

• Designs that can collect dirt, such as hori-zontal surfaces and recesses, should beavoided.

• Designs that create uneven flow or drainagepatterns producing uncleansed areas shouldbe avoided.

• Sheltered areas and areas subject to splatter,especially roadside spatter, should be de-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 solution in warm clean water. Apply with aclean sponge, soft cloth, or soft-fiber brush,then rinse in clean water and dry.

Satisfactory on most surfaces

Fingerprints Detergent and warm water; alternatively, hydrocarbon solvent

Proprietary spray-applied polishes available to clean and minimize re-marking

Oil and grease marks Hydrocarbon solvent Alkaline formulations are also available with surfactant additions.

Stubborn spots, stains, and light discoloration; water marking;

light rust staining

Mild, nonscratching creams and polishes. Apply with soft cloth or soft sponge; rinse off residues with clean water and dry.

Avoid cleaning pastes with abrasive additions. Cream cleaners are available with soft calcium carbonate additions. Avoid chloride-containing solutions.

Localized rust stains caused by carbon steel contamination

Proprietary gels or 10% phosphoric acid solution (followed by ammonia and waterrinses) or oxalic acid solution (followed bywater rinses)

Small areas may be treated with a rubbing block comprising fine abrasive in a hard rub-ber or plastic filler. Carbon steel wool andpads that have previously been used on carbon steel should not be used. A testshould be carried out to ensure that the original surface finish is not damaged.

Adherent hard water scales and mortar/cement splashes

10–15 vol% solution of phosphoric acid. Use warm, neutralize with diluted ammonia solution, rinse with clean water and dry

Proprietary formulations available with surfactant additions. Avoid the use of hydrochloric acid-based mortar removers.

Heat tinting or heavy discoloration Nonscratching cream or polish. Apply with soft cloth or soft sponge. Rinse offresidues with clear water and dry.

Suitable for most finishes.

Nylon-type pad Use on brushed and polished finishes along thegrain.

Badly neglected surfaces with hardened accumulated grime deposits

A fine abrasive paste as used for car body refinishing. Rinse clean to remove all paste material and dry.

May brighten dull finishes. To avoid a patchy appearance, the whole surface may need tobe treated.

Paint, graffiti Proprietary solutions or solvent paint stripper depending on paint type. Use soft, nylon orbristle brush on pretreated material.

Apply as directed by manufacturer

(a) Cleaning agents should be approved for use under the relevant national environmental regulations and should be prepared and used in accordance with the com-pany’s or supplier’s health and safety instructions. Source: Adapted from Ref 9


• Contamination by rust from carbon steel iscorrosive and must be removed.

• Designs should facilitate easy access forcleaning.

While any structure would benefit from theseguidelines, they are especially valuable in maxi-mizing the benefits of stainless steel.

Repair of more severe damage done to sur-faces, such as scratching, is difficult. If a deco-rative surface pattern has been damaged, thechallenge is in trying to replicate it in the field.Very few surface finishes are wholly repairablein the sense that they can be repaired in a spotso that the repair is invisible. The reason forthat is mainly that abrasive finishes are appliedby rotating belts, and it is virtually impossibleto match the pressure, grit size, and arc of con-tact that created the original surface. It can bemore nearly done to a surface with a rolled-onfinish, which has a very consistent grit lengthand depth. The exception to this rule are abra-sively applied or rolled-on long-grain finishes.These have very long grit lines, so grit lengthis easily duplicated with a belt sander, theusual tool available for field repairs. Evenwelds can be removed and reblended to be in-distinguishable from the surrounding originalsurface. The ability to be repaired should be atop criterion in the choice of a surface finishwhose appearance is critical and that may besubject to damage.

Fabrication Considerations

Fabrication and joining of stainless steel em-ploy the same techniques as for carbon steel andother metals. The specifics of cutting, forming,joining, soldering, and welding are described inthe processing section of this book and are notrepeated here. The main distinction in the use ofstainless steel in this regard is that its higherstrength and corrosion resistance permit the useof lighter gauges. This in turn permits designs inhollow or rolled-formed sections, which havehigher stiffness and low weight and potentiallylower overall cost than using less-expensivemetals. A second aspect of higher strength andlighter gauge is greater spring back in formingoperations, such as press braking.

Certain processing principles related to archi-tectural and building applications of stainlesssteels should be emphasized:

• Separation of tools and work areas betweenthose used for stainless steel and carbonsteel is prudent. Contamination of stainlesssurfaces with carbon steel from welding,grinding, and cutting can stain the surface ofstainless steel and result in corrosion. Thiscan be remedied by passivation, but it ismuch better to avoid it in the first place.

• Welding is better done in the shop than inthe field. Correct filler metals must be used,and proper weld finishing is essential. Theability of contactors to produce sound, at-tractive welds is an indicator of their overallcompetence with stainless steel.

• Fasteners used with stainless steel shouldalso be stainless steel. Galvanized steel, car-bon steel, and aluminum will corrode morereadily than the stainless, and this corrosionis aggravated by galvanic contact. The re-sulting corrosion products are also harmfulas well as unsightly. Fasteners should not bepermitted to cause distortion of flat panels.

Additional Service Considerations

Fire resistance is an important considerationin buildings. Stainless steel is the only com-mon building material that remains strong andtough at temperatures encountered in fires. Or-dinary carbon steel undergoes a phase changeat about 760 °C (1400 °F). This change inatomic structure results in a sudden shrinkageof more than 1 linear percent. This can literallypull a building apart. When this occurs to astructure already weakened by heat, as carbonsteel is, catastrophic failure ensues. Austeniticstainless steel keeps the same atomic structureand remains much stronger than carbon steel atelevated temperatures. Thus, austenitic stain-less steel has great value as a material forstructures that must retain structural integrityin a fire.

Tests have been conducted on glass-rein-forced plastic, aluminum, galvanized steel, andaustenitic stainless steel ladders under load andexposed to flame temperatures of more than1000 °C (1830 °F). The plastic and aluminumladders failed in less than a minute. The galva-nized carbon steel lasted 5 min, while the stain-less steel remained intact (Ref 10). If the needfor fire resistance is serious, stainless steel becomes the material of choice. It is used onoffshore oil platforms for stairways, ladders,walkways, handrails, gratings, floor systems,


firewalls, blast walls, living modules, and soforth.

Ecological considerations are never trivialwhen considering a construction material.Many materials used in buildings degrade envi-ronmentally, usually by corrosion, and enter thegeneral environment. Asbestos, lead-basedpaints, lead coatings, and others are once-ac-cepted materials whose long-term effects havebeen dangerous and costly.

Stainless steel, because it does not corrodewhen properly used, does not enter the environ-ment. While this seems obvious, it has been theobject of interdisciplinary studies that havedemonstrated its innocuousness even underconditions of heavy acid rain on freshly abradedsurfaces (Fig. 9) (Ref 11).

Stainless is a material that will never comeback to haunt an architect years later. Its intrin-sic raw material content value ensures that evenwith the end of the life of a structure, the stain-less in the structure will be recycled.

Concrete Reinforcing Bar

Concrete reinforcing bar is one of the leastglamorous uses of stainless steel in structures.

In bridges, parking garages, and other concretestructures, saltwater can penetrate the cementover time. If the internal rebar corrodes, the expansion of the corrosion products spalls the concrete, leading to the failure of thestructure.

This can be delayed by treating the concreteto repel the incursion of water or by coatingthe carbon steel rebar with epoxy. These areless than 100% effective. A more certain ap-proach is to use stainless steel rebar. The stain-less steel for this duty need not resist pittingcorrosion, which affects a tiny percentage ofthe steel volume. Therefore, an inexpensive,low-nickel grade, such as 409, 430, or 201, canbe used. Most of the work to date has beenwith more expensive grades, such as 316 and2205. The lean duplexes are ideal for this ap-plication because of their high strength, resist-ance to corrosion and SCC, and moderate cost.The use of even these alloys reduces the long-term cost of these structures, so the futureadoption of less-expensive stainless steelsholds great promise. The more enlightenedtransportation departments in the United King-dom; Ontario, Canada; and Michigan, NewJersey, and Oregon in the United States haveled this development.

Fig. 9 Graphic depicting low release of metal ions from two grades of stainless steel (304 and 316) to the environment, based on a4-yr multidisciplinary research project involving both field research and laboratory studies. Source: Ref 11


REFERENCES

1. Stainless Steel Selection Criteria, Contrar-ian Metal Resources, www.metalresources.net, accessed June 2008.

2. Bulletin of the National Dairy Federation189, 1985, p 3–12

3. C. Houska, “Which Stainless Steel ShouldBe Specified for Exterior Applications?” In-ternational Molybdenum Association, www.imoa.info, accessed June 2008

4. The Nickel Institute, www.nickelinstitute.org, accessed June 2008

5. H. Guttman, Atmospheric and WeatheringFactors in Corrosion Testing, AtmosphericCorrosion, W.H. Ailor, Ed., John Wiley andSons, 1982, p 51

6. R.B. Griffin, Corrosion in Marine Atmos-pheres, Corrosion: Environments and In-

dustries, Vol 13C, ASM Handbook, ASMInternational, 2006, p 42–60

7. Guide to Stainless Steel Finishes, 3rd ed.,Euro Inox, 2005, www.euro-inox.org, ac-cessed June 2008

8. Contrarian Metal Resources, www.metalre-sources.net, accessed June 2008

9. Care and Maintenance of Stainless Steels,Leda-Vannaclip, www.l-v.com.au, accessedJune 2008

10. The Nickel Institute, www.nickelinstitute.org, accessed June 2008

11. D. Berggren et al., Release of Chromium,Nickel and Iron from Stainless Steel Ex-posed Under Atmospheric Conditions andthe Environmental Interaction of TheseMetals, European Confederation of Iron andSteel Industries, Oct 2004, www.eurofer.org, accessed June 2008

CHAPTER 19

Automotive and Transportation Applications

Summary

THE ADEQUATE DURABILITY and lifespan of cars, trucks, or any transport systemrequires freedom from corrosion. This has re-quired subsystems, such as those for exhaustand fuel, to resist more corrosive environmentsfor longer periods of time. The main result hasbeen a strong growth in the use of the leanerferritic stainless steels in many components.As more exotic propulsion systems and fuel,such as fuel cells and ethanol, emerge, stain-less steels may be required to endure the corro-sive environments.

Introduction

The use of stainless steel in automobilesused to be mainly a story of decorative appli-cations: wheel covers and trim with a minoramount used for valves and hose clamps. How-ever, as automobiles became more sophisti-cated technically and as durability and envi-ronmental demands grew, the role of stainlessbecame increasingly functional and less orna-mental. Stainless alloys in common automo-tive use now are generally highly engineeredfor their specific application and representsome of the most highly evolved applicationsengineering in any use of stainless. An exami-nation of the preferred practices in materialsselection in automotive systems is an excellentexample of the rule of using the simplest andlowest alloy content grade that can do the job.Because automotive and steel-producing engi-neers have collaborated so well, both partieshave benefited greatly, as have consumers, and

the auto manufacturers have become thelargest users of stainless steel.

Exhaust systems constitute the largest use ofstainless steel in the automotive market, butthere are other important applications that can-not be ignored: valves and gaskets, hoseclamps, seat belt and air bag components, tub-ing, hardware, and filters. And, there will benew applications that respond to new socioeco-nomic needs, such as for greater crash worthi-ness, lighter weight, or resistance to the corro-sion of new fuels. But, since exhaust systemscurrently predominate, they are covered first.

Exhaust Systems

Laws enacted in the United States in the1970s mandated automotive emission stan-dards that could be met only with catalyticconverters. The only practical materials thatcould withstand the temperatures of the hotend of an exhaust system using a catalytic con-verter were stainless steels. From Allegheny-Ludlum’s MF-1 evolved a succession of fer-ritic alloys that grew in sophistication to meetthe increasing needs of corrosion resistance,oxidation resistance, creep, thermal fatigue re-sistance, and formability. Soon, entire exhaustsystems were made of stainless; often, theywould last the life of the vehicle, rendering obsolete an entire muffler replacement indus-try. The compelling need of the automotive in-dustry for economy drove the widespreadadoption of argon oxygen decarburization(AOD), the continuous caster, and other high-volume methods of the carbon steel industry.So, even while the traditional automotive uses




of stainless steels—wheel covers and other or-namental trim—faded to nearly nothing, theuse of stainless in automobiles grew to about30 kg (65 lb) per vehicle by the turn of the cen-tury. The stainless steel industry was changedfrom a boutique industry to a mass productionindustry by its embracing the needs of the au-tomotive market.

An exhaust system normally consists of amanifold to collect exhaust gases, a catalyticconverter to reduce NOx and CO emissions, anda muffler; each of these are connected by pip-ing. Each component of the system has differentrequirements for formability, resistance to oxi-dation, resistance to external corrosion, resist-ance to internal corrosion, and mechanicalproperties. At the outset, it should be noted thatferritic stainless steels, as opposed to austenitic,are optimal for oxidation resistance, especiallycyclic oxidation. It is not a difference in theoxide scale.

The reason is that the thermal expansion offerritic stainless more closely matches that ofthe oxide scale than does that of austenite. Thisprevents the fracturing and spalling of the scale.The intact scale of ferritic stainless is thus pro-tective up to the temperature at which oxygendiffusion through the scale becomes greatenough so that “breakaway” oxidation occurs.At the breakaway temperature level, scalegrowth is no longer parabolic with time but be-comes linear and therefore no longer protective.The temperature of this breakaway increases aschromium content increases. We will see thatother alloying elements can also improve thisperformance.

Not only are the steels in exhaust systems al-most exclusively ferritic, they are also stabi-lized by titanium or niobium. This prevents sen-sitization and makes all the chromium contentuseful as alloy. Titanium stabilization greatlyimproves corrosion resistance by removing notjust the carbon and nitrogen, but also the oxy-gen and sulfur from solution. This sharply im-proves resistance to pitting corrosion. Niobiumis used to costabilize and fight creep. These al-loys are therefore essentially interstitial free andhave excellent formability, which the designs ofexhaust system components require.

Their formability is further enhanced, whennecessary, by low additions of tramp substitu-tional alloying elements such as manganese,nickel, and copper. Special thermomechanicalprocessing is also used to optimize texture andgrain size.

The metallurgy of ferritic stainless is dis-cussed in depth in Chapter 8. Rather than reex-plain these concepts here, we revisit only themain points that are relevant to alloy selectionfor exhaust systems.

• As chromium level increases, so does re-sistance to oxidation and corrosion, butyield strength also increases, and ductilitydecreases.

• Alloying with silicon, aluminum, andmolybdenum also increases oxidation resist-ance, but these elements have the samedetrimental effect on ductility while increas-ing hot strength.

• Niobium above that needed for stabilizationis a powerful solid solution hardener and iseffective at high temperatures.

• Ferritic stainless steels have very anisotropicforming properties. They resemble high-formability carbon steels in that they tendnot to thin when stretched, which greatly as-sists in formability.

• Ferritic alloys can form a hard, brittle phasecalled α’ in a process commonly called 885°F (or 475 °C) embrittlement. This is only afactor in alloys with chromium of 18% ormore, especially those containing molybde-num and aluminum. Cold work acceleratesthe formation of this phase.

• The σ phase does not readily occur in alloyscontaining less than 20% Cr, so it is not aconsideration for exhaust systems unless sil-icon or molybdenum are also elevated.

• Coating ferritic steel with aluminum is effec-tive in preventing oxidation and corrosion.

All these factors come into play in the designof exhaust systems. Because the alloys haveevolved so well to fit the individual require-ments for each component, we discuss them seg-ment by segment through the exhaust system.

The exhaust manifold collects the hot, com-busted gases from the engine and deliversthem to the front pipe. The exhaust manifoldmust possess good high-temperature strengthand resistance to thermal fatigue. It must alsobe able to resist oxidation at the exhaust tem-perature, which can reach 950 (C (1740 (F).Exhaust manifolds had previously been heavycastings but are generally now formed fromstamped sheet stainless steel or formed fromwelded tubing that may have a double-wallstructure to insulate the gases from heat loss,which could preclude successful catalytic con-version downstream.

Chapter 19: Automotive and Transportation Applications / 227

As the highest-temperature component of theexhaust system, the exhaust manifold must pos-sess the greatest resistance to high-temperatureoxidation damage. Risk of such damage is dueto the intermittent use of vehicles, which causescyclic oxidation and the ensuing spalling of theoxide scale. There are numerous alloying ap-proaches for optimizing the ferritic stainless al-

loys for spalling resistance. All approaches in-volve raising chromium content but use differ-ent techniques to enhance the effect ofchromium. Table 1 lists a number of the gradesof stainless steel commonly used in exhaust sys-tems and where they are used. The alloys arelisted in order of increasing severity of the re-quirements for each major system component.

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

Component

Service temperature

RequirementsAlloys currently used, common name

(related designation)ºC ºFExhaust manifold 750–950 1380–1740 High-temperature

strength, thermal fatigue strength, oxidation resistance,formability

• T439HP (UNS S43035, dual-stabilized 439)• 18CrCb (DIN 1.4509, 18CrCb)• 441 (DIN 1.4509)• 304/304L/304H (UNS S30400, S30403,

S30409)• 321 (UNS S32100)• 309S (UNS S30908)• 310S (UNS S31008)• 332Mo (S35125)• 600 (N06600)• 601 (N06601)• 625 (N06625)

Front pipe 600–800 1110–1470 High-temperature strength, thermal fatigue strength, oxidation resistance,formability

• 409 ALMZ (aluminized 409)• T439HP (UNS S43035, dual-stabilized 439)• 18CrCb (DIN 1.4509, 18CrCb)• 441 (DIN 1.4509)• 436S (type 436S)• 444 (UNS S44400, T441)• 433 (T443)

Flexible pipe 600–800 1110–1470 High-temperature strength, thermal fatigue strength, oxidation resistance,formability, salt attack resistance

• 304/304L (UNS S30400, S30403)• T321 (S32100)• 316/316L (S31600/S31603)• 316Ti (S31635)• 332Mo (S35125)• 625 (N06625)

Catalytic converter shell 600–800 1110–1470 High-temperature strength, salt attack resistance, formability

• 409HP (UNS S40930, dual-stabilized 409)• T439HP (UNS S43035, dual-stabilized 439)• 441 (DIN 1.4509)• 18CrCb (DIN 1.4509, 18CrCb)• 444 (UNS S44400, T441)• 433 (T443)

Catalytic converter substrate

1000–1200 1830–2190 Oxidation resistance, thermal shock resistance

• ALFA-IV (FeCrAl)

Center pipe 400–600 750–1110 Salt damage resistance

• 409HP (UNS S40930, dual-stabilized T409)• 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 100–400 210–750 Corrosion resistance, from inner and outer surface

• 409HP (UNS S40930, dual-stabilized T409)• 409 ALMZ (aluminized 409)• T439HP (UNS S43035, dual-stabilized T439)• 436S (T436S)• 441 (DIN 1.4509)• 18CrCb (DIN 1.4509, 18CrCb)• Type 304/304L (UNS S30400, S30403)

Tailpipe 100–400 210–750 Corrosion resistance, from inner and outer surface

• 409HP (UNS S40930, dual-stabilized T409)• 409 ALMZ (aluminized 409)

Source: Adapted from Allegheny Technologies Inc.


The best choice for a given design is not obvi-ous. We attempt to simplify the choices.

Thus, the basic alternatives for exhaust sys-tem alloys are:

• Straight chromium alloying at 11 to 12%with stabilization by titanium or niobium,the basic type 409 (UNS S40920)

• Straight chromium alloying at 17 to 18%with stabilization by titanium or niobium,the basic type 439 (UNS S43036)

Either of these basic alloys can enjoy en-hanced oxidation resistance by additional alloy-ing with molybdenum, aluminum, or silicon. Inaddition, they can be coated with hot-dippedaluminum-silicon alloy to increase oxidation re-sistance.

Chromium or molybdenum alloy additionsincrease corrosion resistance, whereas alu-minum or silicon additions do not improve thattrait. Aluminum coating is a powerful corrosionfighter, and it has the aesthetic benefit of notshowing red rust.

Use of molybdenum or niobium enhanceshigh-temperature strength. Alloys with theseadditions are thus useful for manifolds with adesign that constrains expansion and contrac-tion, making thermal fatigue a problem.

All alloying additions detract from formabil-ity and toughness, as well adding to basic mate-rial costs. Thus, the objective must be to useonly those alloying elements that are indispen-sable to performance.

The front pipe connects the exhaust manifoldto the flexible joint and experiences nearly thesame temperatures as the exhaust manifold, butnot the same risk of thermal fatigue. To reduceexhaust noise, a double-wall pipe is sometimesused for this component.

The flexible joint is the one segment of theexhaust system for which austenitic stainlesssteels are preferred. The function of the flexi-ble joint is to prevent vibration from the en-gine from being transmitted to the rest of theexhaust system. It consists of a double-wallpipe in a bellows configuration with an outercovering of braided stainless steel wire. It musthave very good high-temperature fatiguestrength to withstand the cyclic stress of the vi-bration it absorbs. The material used musthave exceptional formability to be formed intoa bellows. The greater hot strength and forma-bility of austenitic steels thus prevails. Theflexible joint is also exposed to road salt in

some regions, so it must resist hot salt corro-sion. This may force the use of 316L versus thenormal choice of 304L.

The catalytic converter is the next compo-nent of the exhaust system. It exposes the ex-haust gases to noble metal catalysts, whichcomplete the combustion of the gases to formless-noxious compounds. This is an exothermicreaction at temperatures equal to those in theexhaust manifold. Thus, the housing, while notrequiring great hot strength, must resist oxida-tion. The catalyst itself is supported by a ce-ramic and ferritic stainless steel carrier thatmust resist thermal shock and possess low heatcapacity for rapid heating. Exotic alloys of20% Cr with 5% Al are used for the carriers.The housing is generally made of a 17% Cr fer-ritic stainless. The converter is usually directlybeneath the passenger compartment, so a heatshield of type 409 is used to separate it fromthe floor.

The center pipe conveys the converted gasesto the muffler. The cooling exhaust gases nolonger present a major oxidation threat, but thecondensing water vapor creates an internal cor-rosion risk, and road salt presents an externalone. However, a simple grade such as 409should generally provide sufficient resistance tothis environment.

The muffler, next in line, presents only a cor-rosion issue. The muffler must withstand corro-sion from the outside, the worst of which comesfrom road salt or coastal salt sources. Internalcorrosion is also a major consideration becausecondensing exhaust gases create a hostile,acidic environment. After startup, the heating ofthe muffler to temperatures above 100 °C (212°C) evaporates these condensates, and internalcorrosion ebbs. On short runs, this may notoccur. This represents a worst case for internalcorrosion. The dual internal and external corro-sive attacks require the use of aluminized stain-less for best performance.

The tailpipe is exposed to view in most vehi-cles, and its appearance is therefore important.For this reason, an austenitic such as 304 can beused, as can chromium plating or aluminizing.The object here is to avoid visible corrosion.

Truck exhaust systems are beginning to re-quire similar technical sophistication as theiremissions come under increased regulation.However, they do not present any challengesnot already confronted and solved for passengervehicles.


Structural Components

The driving forces of durability, safety, andweight reduction have spawned other, more var-ied applications for stainless in automotive en-gineering. Across the board, the main distin-guishing trait of stainless that qualifies it as theoptimal material is its corrosion resistance, butthis characteristic would be insufficient in manycases without considering mechanical proper-ties. Indeed, even if stainless steel were not cor-rosion resistant, its superior strength and tough-ness would qualify it for many automotiveapplications.

Austenitic stainless steels are the toughest andstiffest practical materials available to the auto-motive engineers. Common 301 can be coldworked to yield strengths anywhere from its an-nealed level of about 300 MPa (44 ksi) up to2000 MPa (290 ksi). In this higher-strength con-dition, it has become the standard material forseat belt anchors and hose clamps. Type 301 inthe annealed condition is actually the originaltransformation induced plasticity (TRIP) steel asit can be tailored to have a controlled level ofaustenite stability. This allows it to transform ata known rate to martensite during deformation,giving not only a very high work-hardening rate

but extraordinary resistance to localized thin-ning, necking, and therefore fracture. Whencrash worthiness becomes a prime considera-tion, then this characteristic makes 301, or itslow-nickel counterpart 201, an ideal material forstructural, energy-absorbing components sinceaustenitic stainless can be rivaled for such appli-cations only by heat-treated alloy steel, titanium,or aircraft aluminum alloys, all of which aremore expensive, less durable, or less formable.

Tables 2 and 3 show the properties of specificvariations on basic 301 developed by Out-okumpu and how they stack up against themost competitive carbon steels, dual-phasesteels, and TRIP steels (Ref 1). The value of amaterial as an energy-absorbing structure (i.e.,one that enhances crash worthiness) is meas-ured by the energy it can absorb per unit ofmass. The kinetic energy of a collision that astructure can absorb in deformation is propor-tional to its strength multiplied by the amount itcan deform before fracturing. The superiorityof metastable stainless steels (i.e., 201 and 301,those that most easily transform to martensiteduring deformation) is shown in Fig. 1. Evenwith its lower density, aluminum falls far shortof austenitic stainless in energy absorption perunit weight.

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

Type Thickness, mm0.2% proofstrength, MPa

Ultimate tensilestrength, MPa

True stress at ultimate tensilestrength, MPa

Uniform elongation, % Total elongation, %

Carbon steels

TRIP 700 1.58 473 703 818 16.4 17DP 750 1.48 513 811 920 13.4 18.8DP 800 1.44 573 896 976 8.9 9.9

Austenitic stainless steels

HyTens X 1.16 306 937 1429 52.5 59.3HT 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 105DP 750 1.131 101DP 800 1.32 74

Austenitic stainless steels

HyTens X 0.536 364HyTens 1000 1.726 269Source: Ref 1


These exceptional strength-to-weight and en-ergy absorption-to-weight characteristics permitautomotive engineers to reduce weight and in-crease crash worthiness while designing vehi-cles with greater life span—because corrosionresistance “comes along for the ride,” as it were.

Some components in which these virtues aremost readily exploited are bumper systems (Ref2). Porsche uses austenitic stainless steel forfront and rear side members, internal push rodson front and rear axles, and lower rear wish-bones in its Carrera GT.

Another manufacturer, Audi, engineered vari-ous components of austenitic stainless steel intoits otherwise aluminum-intensive A6 series. Theuse of stainless steel in strategic componentsenables greater weight reduction than thatwhich the vehicle would have in all aluminum.Volvo and Saab have designed austenitic stain-less steel bumper systems that also serve to re-duce overall vehicle weight.

While it is probably apparent to the reader thatessentially any body component can be made instainless and be made better in stainless, thequestion of when doing so is a better engineer-ing decision involves economic considerations.Large automotive companies generally havelarge fixed investments in painting and coatingsystems to protect entire bodies from corrosion.The incremental savings of eliminating coatings

on individual body components is thus essen-tially nil. However, if the entire system is stain-less and the investment is avoided, then the ini-tial cost of a stainless body actually can be lowerthan one in coated carbon steel. This is the expe-rience of Italian bus manufacturers, who beganin the 1980s using 304 stainless steel in buses.Now, buses are 80% stainless.

Designers began the conversion to gain thenormal advantages the stronger stainless gives:over 10% lighter weight and over 10% im-provement in crash worthiness of the passengercompartment, the accompanying savings in fuelconsumption, and the virtual elimination ofbody maintenance. With essentially the entirebody now in stainless, coating and paintingcould be eliminated. A stainless bus body isshown in Fig. 2.

This swung the economic pendulum to stain-less in a major way. Now, not only was thelong-term cost of operating the bus lower, butthe initial cost of the bus was lower. The eco-nomic analysis is shown in Table 4 (Ref 2).

The design key was to use rectangular 304stainless structural tubing, which allowed strong,stiff sections to be welded into space frames. It isonly a matter of time until this is improved on bythe use of 201 (with 3 to 4% Ni instead of 8 to9% Ni) to lower cost and cold working of thetubing to achieve higher strength levels.

Fig. 1 True stress-true strain curves for 301 variants (HyTens X and HyTens 1000) versus two duplex carbon steels (DP750 andDP800) and a transformation induced plasticity (TRIP) steel (TRIP700). Source: Ref 1


Microcars are now a familiar sight in Europe.These vehicles are prized for their ability to bedriven and parked in very small or congested lo-cations. Their economy of operation is also amajor attraction. These considerations combineto make stainless the best material for many oftheir components. Figure 3 shows a stainlesssteel microcar frame.

The design by the famous design house Pinin-farina employs a stainless frame to give maxi-mum torsional stiffness and crashworthinesswhile eliminating painting entirely.

Other Automotive Components

Stringent emissions controls regulations, ledin the United States by the state of California,have made manufacturers reexamine the suit-ability of polymeric fuel tanks. These tanks con-tribute more to the required maximum 2 g/dayof hydrocarbon emissions than is tolerable, so a

few manufacturers, such as Volkswagen, haveinstalled stainless steel fuel tanks in their vehi-cles (Ref 2).

Trucks

Over-the-road trailers are an excellent exam-ple of stainless steel being used for utilitarianpurposes. Trailers used for hauling foodstuffs orcorrosive materials are now constructed almostentirely of stainless steel; lined carbon steeltanks are now largely obsolete. The engineeringbasis for this is the same as for buses: highstrength, no coating costs, and a product withlong life and low maintenance costs. Structuralmembers in trailers are typically 304, whiletanks may also be 316L for corrosion resistancewhen the transported material requires it. Tankwrappers are often made of bright annealed andbuffed 304. Manufacturers of trailers would bewell advised to consider upgrading to duplexgrades such as 2003 or 2205 or to cold-workedaustenitic stainless, which would permit majorweight reduction. This weight reduction woulddirectly translate into greater load-carrying ca-pacity because the payload of liquid-carryingtrailers is limited by total gross weight. Theability to add a few thousand more pounds ofpayload would quickly pay back a small pre-mium in material cost.

Normal cargo-carrying trailers also use somestainless where corrosion is problematic, suchas in doors and door frames. Weight reduction isless important in these trailers, which reachmaximum load at a volume limit rather than aweight maximum.Fig. 2 Stainless steel bus bodies. Source: Ref 2

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

Cost of capital 10.00%Inflation rate 5.00%Real interest rate 4.76%Desired LCC duration 20.0 yearsDowntime per maintenance/replacement event 1.0 dayMonetary unit U.S. $Value of lost production 101 U.S. $/day

Stainless steel Carbon steelMaterial costs 3.331 1.391Fabrication costs 25.322 28.582Other installation costs 2.185 4.050Total initial costs 30.838 32.023Maintenance costs 0 1.448Replacement costs 0 2.897Lost Production 0 57Material-related costs 0 0Total operating cost 0 4.402Total LCC cost 30.838 36.425Source: Ref 2


Rail Transport

Passenger trains have exploited the highstrength-to-weight and toughness qualities ofthe 301 family of stainless steels for manyyears. The corrosion resistance of these alloysmakes them corrosion free in long use, obviat-ing the need for painting and lowering mainte-nance costs. As with any other major use of type301, a 5 to 10% increase in economy could beachieved if type 201 were used instead of 301.No loss in performance would occur. The transi-tion to 201 has not occurred simply because ofinertia and resistance to change on the part ofdesigners and producers.

Hopper cars made of 12% Cr martensiticstainless steels, typically 409Ni and 3Crl2, haveexcellent abrasion and corrosion resistance aswell as high strength and therefore greater load-carrying capacity. Curiously, the use of cold-worked austenitic stainless in railcars, which has

been successful for decades, has not been carriedover into trucks and buses even though it is tech-nically feasible to economically produce struc-tural sections in very high-strength stainless.

REFERENCES

1. R. Andersson, E. Schedin, C. Magnusson, J. Ocklund, and A. Persson, The Applicabil-ity of Stainless Steels for Crash AbsorbingComponents, ACOM, No. 3–4, AvestaPo-larit AB, 2002

2. F. Capelli, V. Boneschi, and P. Viganò,“Stainless Steel: A New Structural Automo-tive Material, Vehicle Architectures: Evolu-tion Towards Improved Safety, Low-Weight,Ergonomics, and Flexibility,” paperpresented at Florence ATA 2005, 9th Interna-tional Conference (Florence), May 2005,www.centroinox.it, accessed June 2008

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

CHAPTER 20

Commercial and Residential Applications

Summary

STAINLESS STEEL HAS BECOME the es-sential material for products related to food,health care, and laundry because it combinesstrength and durability with an unexcelled abil-ity to be cleaned, disinfected, and sterilized.These qualities have long been apparent to com-mercial food, laundry, and health care profes-sionals and have increasingly carried over intoequivalent domestic areas as consumers havebecome more aware of the benefits of stainless.

Introduction

The last 20 years have seen the long-standingpervasive commercial use of stainless steel forfood preparation and serving; laundry; heating,ventilation, and air conditioning (HVAC); andother appliances penetrate the domestic marketfor the same types of goods. Whether this is afad of an increasingly affluent consumer or a re-flection of more design engineers and con-sumers being more interested in lasting valuethan they were in the “throwaway” society thatpreceded that period remains to be seen. Stain-less has been increasingly identified with high-quality, high-end products. But, the case forvalue rather than fad seems to be stronger if thelessons of the harshly pragmatic automotive in-dustry, in which decorative use of stainless hasvirtually disappeared while utilitarian uses havemushroomed, are any indication.

The case for using stainless in appliances ofall types, whether they are commercial or resi-dential, relates to stainless being able to providethe best value over the intended service life.

Stainless is without rival for ruggedness anddurability. Steel and aluminum corrode. Glass,stone, and ceramics break. Plastic is weak. Thesecond, even more important, reason is thatstainless steel is essentially benign from a hy-gienic viewpoint. Stainless steel itself is inert,both chemically and biologically, with respectto food. Further, it provides minimal harbor forunwanted biologic growth as do more porousmaterials. Stainless also competes quite well es-thetically with other materials, offering the de-signer numerous surface finishes. Last, stainlessis very amenable to nearly all manufacturingtechniques. Its lack of need of coatings oftenmakes components made from stainless less ex-pensive to produce than equivalent designs thatmust be coated with paint, porcelain, or metal.

Food Contact Qualifications

Setting aside cost, esthetics, and manufactur-ing considerations, a food contact material mustfirst meet three criteria: It must be chemicallyinert, biologically inert, and cleanable and ableto be disinfected.

Chemical neutrality is achieved by a mate-rial when the material does not enter into thefood with which it comes in contact. This hasbecome an increasing concern as the effects ofions or chemicals released from food prepara-tion materials have been viewed as potentialtoxic or disease agents. Medical knowledge isnot sufficiently advanced to convince con-sumers of the harmlessness of such contami-nants, so it is preferable to demonstrate the ab-sence of contamination if one is to win thepublic confidence in a food contact material.




Stainless steel contains many constituent ele-ments. Were they to enter the food with whichthe stainless came in contact, then stainlesswould be a poor food contact material. The dis-tinguishing characteristic of stainless, however,is the spontaneous passive film, which is so sta-ble chemically. This film acts as a barrier to cor-rosion, which would result in metal release.Stainless therefore is effectively inert. Testshave been made of the rates at which metal ionscan enter foodstuffs (Ref 1). Table 1 shows thevastly lower rates of metal ion release fromstainless than from aluminum and carbon steel,both of which are permissible, if not optimal,food contact materials. Aluminum releases alu-minum ions into solution of both cooking oiland 3% acetic acid at nearly equal rates of15 mg/cm3 in 30 days. Carbon steel releasesiron at over 100 mg/cm3 in the same period.Stainless, however, releases less than 0.010mg/cm3 of iron. This is complemented by otherstudies showing that the transfer of ions fromfood contact vessel to food is diminishingly andnegligibly small. Tests have been conducted onstainless steels, types 304, 439, and 444, thathad both industrial finishes (2B and BA) as wellas freshly abraded and air-aged finishes. Thesesample steels were subjected to boiling solu-tions of oils, alcohols, water, and 3% aceticacid. None caused the transfer of eitherchromium or nickel to exceed the statutory 0.1ppm level (Ref 2). Nickel levels of variousfoods before and after cooking have been scien-tifically measured to assess the possibility ofleaching of that ion from 304 stainless steel(Ref 2). No increase was noted from the natu-

rally occurring level of up to 0.3 μg/g in cerealsto 1.1 μg/g in meat and fish.

These negligible levels of leaching simply in-dicate that foodstuffs are a benign chemical tostainless steel. Nevertheless, it is necessary toapply the correct assessment of the corrosivityof the foodstuff in question. In food production,as opposed to preparation for serving, more ex-treme levels of acidity and salinity can be en-countered. Nippon Steel reported (Ref 3) thatmaterials used in the manufacture of soy sauce,which can have 15% salt, must withstand pro-longed contact at 45 °C (115 °F). Under suchconditions, 316 stainless pits in about 1 day,while higher alloy grades, the 6Mo alloys, ofwhich their YUS 270 is one (equivalent to UNSS31254), are projected to last 20 years beforepitting.

This is significant because only pitting corro-sion is likely to release metals ions into a foodsubstance. So, while guidelines exist for theminimum alloy content permissible for normalfood contact, such as those promulgated by theNational Science Foundation (NSF), one muststill verify the corrosion due to a particularlyaggressive food ingredient. Choosing the propergrade of stainless, based on pH, salinity, andtemperature, is the responsibility of the designengineer. Referred to chapters in this book onboth corrosion and individual alloy families forguidance in choosing an alloy based on resistingpitting in a given environment. This havingbeen said, no alloy greater in pitting resistancethan 304 is required in residential or commer-cial cooking food contact. The higher alloy requirements come from the more aggressive

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

Material Time

Metal migration during indicated time, µg/cm3

Iron Chromium Aluminum Nickel

Austenitic stainless 30 min10 days20 days30 days

2.44.22.72.3

0.120.220.220.28

≤ 0.190.220.19

≤ 0.19

≤ 0.12≤ 0.12≤ 0.12

0.31Ferritic stainless 30 min

10 days20 days30 days

3.07.38.66.6

0.430.400.710.87

≤ 0.190.19

≤ 0.19≤ 0.19

≤ 0.12≤ 0.12≤ 0.12≤ 0.12

Aluminum 30 min10 days20 days30 days

4.918.217.931.3

0.933.425.58

12.40

9305,3007,160

15,350

≤ 0.12≤ 0.12≤ 0.12

0.22Carbon steel 30 min

10 days20 days30 days

8,43057,70062,900

112,000

0.627.406.82

14.00

2.726.724.036.9

≤ 0.12≤ 0.12≤ 0.12≤ 0.12

Source: Ref 1

Chapter 20: Commercial and Residential Applications / 235

concentrations and exposure periods that can befound in food-processing plants.

Certain abuses can even damage stainlesscookware. Very high temperatures, such as canoccur when unattended pans have their liquidsboiled away, can damage stainless but are moreharmful to less-rugged alloys such as copperand aluminum.

Stainless steel is primarily composed of iron,chromium, and nickel along with small amountsof manganese, silicon, and molybdenum. It con-tains trace amounts of copper, aluminum phos-phorus, and sulfur. Each of these elements isnaturally occurring in food. Each can be foundin a typical multivitamin/multimineral supple-ment. Stainless is essentially devoid of heavymetals, such as lead and mercury, which are va-porized at the temperatures at which stainless isrefined. Even if toxic metals were somehow tobe made to contaminate stainless, the passivefilm would prevent their release. All these fac-tors combine to make stainless the most chemi-cally neutral metal found in food contact. Alter-native alloys, such as copper and aluminum,actively leach into foods. Copper and aluminumhave been linked to but not demonstrated tocause Alzheimer’s disease.

Biological Neutrality. Microorganisms ad-here to solid surfaces. When a clean surfacecomes in contact with food, a surface deposit isformed from the food. The film may also con-tain molecules left from previous cleaning anddisinfecting. The formation of this film is pre-sumably influenced by material characteristicssuch as roughness, although there are no spe-cific studies on this. However, microorganismsadhere to this film and, as colonies of themgrow, form a biofilm. This film consists of lay-ers of microorganisms that can produce an exo-cellular polymeric matrix, which protects thecolony from cleaning and disinfecting. Geomet-ric factors also can protect these colonies.Rough surfaces are intuitively more difficult toclean. The ability to maintain a microscopicallysmooth surface is an asset in stainless that poly-meric, enamel, and mineral surfaces lack. Stain-less steel is much less roughened by abrasion,keeping the surface smooth (Ref 4).This will beseen to influence its ability to be cleaned anddisinfected.

There is some technology to go beyond bio-logical neutrality in the use of coatings that ac-tively discourage or eliminate growth of mi-croorganisms. Polymeric coatings impregnatedwith silver ions have been developed and com-

mercialized (Ref 5). Silver ions, like copperions, are powerful antimicrobial agents. Thecombination of such a coating with stainless asa corrosion-proof substrate may represent themaximum in hygienic and chemical protectionand is already being used in medical applica-tions where such concerns exceed those in ordi-nary food contact situations.

Cleanliness. A necessary quality in any mate-rial considered for food contact is the ability tobe cleaned. This includes the removal of bothorganic and inorganic substances. The most im-portant objective of cleaning is to remove thevisible and invisible materials that can provide agrowth medium for microorganisms. Thisprocess is distinguished from disinfection,which is the reduction of the microbial popula-tion to a satisfactory level. What this level is de-pends on the standards of hygiene in force. And,although cleaning can and does reduce the pop-ulation of microorganisms, true bacteriologicalcleanliness is obtained only after disinfection.The combination of cleaning and disinfecting isimportant. Studies have shown that the efficacyof disinfectants is weaker on bacteria that havebeen established in a surface biofilm than onbacteria in suspension. The most complete formof disinfecting is sterilization, whose objectiveis the complete removal of all microbial life andviruses.

The purpose of cleaning stainless steel is torid it of contamination. Various stainless manu-facturers and associations have identified anumber of effective of cleaning products (Ref 2,6, 7):

• Alkalines, which dissolve fats and oils• Chelating or sequestering agents, which ag-

glomerate contaminants. These are often or-ganic acids such as citric acid or oxalic acidand amine acids such as sulfamic acid andethylene diamine tetraacetic acid (EDTA) orsalts of these compounds.

• Hydrocarbon solvents• Water with soap, detergent, trisodium phos-

phate, or other surface active agents, whichemulsify

• Dilute oxidizing acids like nitric acid• Mild acids such as phosphoric acid

The effectiveness of a cleaner relates mainlyto the contaminant to be removed. Some trialand error may be required for a given contami-nant. Some precautions are worth mentioning.Abrasive cleaners should be used with caution.The abrasive size and hardness must be chosen


so that the stainless surface finish is not affectedin an unwanted manner. If the abrasive is harderthan the stainless or coarser than the stainlesssurface roughness, the underlying finish can bedisturbed. Care should also be taken to cleanwith the polish grain if a polished surface isbeing cleaned. Also, cleaners containing chlo-rides are common. Their use is not recom-mended on stainless. Use of hydrochloric(muriatic) acid is especially detrimental. Ifchloride-containing cleaners are used, then thor-ough rinsing should be conducted to avoidchloride concentration through evaporation, es-pecially in crevices. Steel wool or steel brushesshould not be used on stainless under any cir-cumstances as iron residue interferes with theintegrity of the passive film.

The ability of stainless to be cleaned is bestmeasured by the actual removal of bacteriacolonies. This has been done to compareunabraded and abraded (to simulate new vs.used) stainless steel, enameled steel, mineralresin, and polycarbonate materials, which canbe used for sinks, counters, food prep tables,etc. (Ref 4). Figure 1 shows that the reduction inbacteria count by the same cleaning technique isten times more effective on stainless than on theother material types. Abrasion did not degradethe ability of stainless to be cleaned as it didsofter materials. The surface of stainless, evenwith the seemingly protected recesses due to

abrasive polishing, permits bacterial colonies tobe removed. The greater roughness of the othermaterials may serve to protect the bacterialcolonies from shear forces and provide greaterspecific surface area on which the colonies canbond.

Disinfection. The ability of a surface to bedisinfected is measured by the concentration ofa given disinfectant required for a specific re-duction in bacterial population. Numerousstudies have been published (e.g., Ref 8–10)showing that glass and stainless steel have equalaptitude for disinfection, and that polyesters,polyurethanes, rubber, and aluminum all re-quired about one to two orders of magnitudegreater concentrations of disinfectant for thesame result. These results indicate why stainlessis so essential to the food industry. Stainless canbe disinfected quite readily, which allows thegreat invisible liability of food-borne diseasesto be minimized.

The effectiveness of sodium hypochlorite as adisinfectant is inarguable, also. So, despite itspotential corrosivity, it will be commonly used.Taking this into account requires that commer-cial and residential food equipment be able towithstand some chloride level greater than oth-erwise projected. Industry practice in the UnitedStates has shown that corrosion problems occurat an unsatisfactory level with mechanicallypolished 430 but not with bright-annealed 430.

Fig. 1 Bacterial retention as a function of material and cleaning time. Source: Ref 4


Thus, alloys with less than 16% Cr should notbe used unless corrosion can be accepted. Al-loys containing 16% Cr can be used with opti-mal surface finish. Alloys as low in carbon as12% are used for cutlery applications whereslight corrosive attack can be accepted. This is anecessary trade-off required to achieve highhardness for good cutting edge retention.Higher chromium grades such as 304 can beused even with mechanically polished surfacefinishes. From a cost-effectiveness point ofview, there is no reason to use more expensivealloys than 430 or 201 in the vast majority ofcommercial and residential kitchen and laundryapplications from a corrosion standpoint as longas surface finishes that have not been producedby abrasive polishing are specified. Many suchfinishes are widely used. In North America, therolled-on replicas of No. 4 finish, Koolline,Lustrite, etc., are quite common, while in Eu-rope the bright-annealed finish has been pre-ferred. Both of these are preferable to mechani-cally polished finishes.

The food industry, an immense consumer ofstainless steel, could do more than any other in-dustry to help conserve nickel by specifying al-loys such as 430, 439, and 201 as their standardalloys as well as by specifying nonabrasive fin-ishes. This can be done with no loss of function-ality or change of appearance and could save23% to 50% in material cost.

Applications

Cookware. Any interaction between a foodcontact material and the food is most likely tooccur during the cooking process when temper-atures are greatest. Only glass and stainless areexcellent food contact materials. And, sincecookware must be flexible enough to handle anypotential food, the choice of material for cook-ware must be the most conservative. For thisreason and because of the brittleness of glass,stainless is the material of choice. The qualitiesdiscussed make aluminum and copper less de-sirable. Both leach into food. Copper can betinned to combat this. The tin also corrodes overtime but has very low toxicity. A larger draw-back is the expense of retinning copper utensils.Aluminum is known as a toxic metal, with itstoxicity causing symptoms similar to those ofAlzheimer’s and osteoporosis (Ref 11). Thesetwo metals do have one advantage over stain-less, however: their thermal conductivity. High

thermal conductivity in a cooking utensil mini-mizes differences in temperature across the sur-face in contact with the food, permitting bettercontrol of the cooking process. The solution tothe problem of thermal conductivity is to makecomposite materials. Stainless can be bonded tocopper and aluminum, which allows the stain-less to be on both the food contact surface aswell as the exterior, with an inner layer of cop-per or aluminum effectively spreading the heat.Aluminum and copper are nearly equally effec-tive as inner conductive layers. Premium cook-ware features them both. The “sandwich” is theoptimal design because it optimizes heating uni-formity even more than using aluminum or cop-per alone would since the high conductivityinner core functions as an isotherm. The unifor-mity is the more important consideration thanthe absolute thermal conductivity or even thethermal diffusivity.

In a triple layer, the choice of the non-food-contacting stainless is less stringent. Some-times, the exterior is made of a ferritic stainlesssteel. The ferromagnetism of ferritic stainlesssteel makes it ideal for induction heating. Al-loys such as 436 have been used for this appli-cation, while 304 is the pervasive choice for thefood contact surface. This is despite the fact that201 or 301 are quite adequate for this applica-tion. It is also possible to produce a magneticcarbon steel core with stainless bonded to bothsheet surfaces. The exposed edges are rolled toshield them from corrosion.

Nonstick coatings, such as polytetrafluoreth-ylene (PTFE), are very popular because of theirnonstick qualities. Above 350 °C (660 °F) thesecoatings give off toxic fumes. This is a dangerfor certain types of cooking, such as wok cook-ing or blackening, but more likely to be encoun-tered by accidentally high temperatures abovethose intended. Since they can be scratched andare not impermeable, their use does not alter thechoice of the material to which they are applied.

Kitchen Appliances. Every type of commer-cial kitchen appliance can be, and usually is,made of stainless steel, as are premium domes-tic kitchen appliances. This choice is based ondurability and ease of cleaning and disinfecting.And, because many commercial appliances arevisible to the customer, aesthetics are also adriving force. Choice of alloy for a given appli-ance is a crucial cost factor. As was noted that430 is marginal for kitchen use, because of theprevalence of chloride-containing cleaners, un-less it has been bright annealed. All austenitics


are satisfactory under normal use. Designersseem to generally neglect the possibility thattheir equipment may be used in coastal cli-mates. In the high ambient salinity of coastalclimates, corrosion will occur unless 304 with abright-annealed finish or a brushed finish rolledonto a bright-annealed 304 is used. Mechani-cally polished 304 stainless is inadequate forcoastal environments. These are the sameguidelines used for architectural applications.Figure 2 shows how different alloys withstandcoastal conditions. The corrosion on 430 wouldbe considered excessive, while that on the 201and 304 is acceptable given that some routinecleaning would have prevented the corrosionthat is present on these samples, which were ex-posed to coastal salt and humidity for 10 yearsin North Carolina (Ref 12).

In the vast majority of ambient conditions,coastal salinity is not a problem. This applies toinland conditions or coastal conditions whereinterior environments are protected by air con-ditioning or adequate cleaning of the stainless ispracticed. This is normally the case for com-mercial equipment. Under these conditions, 201is quite adequate, and the use of 304 representswasteful overengineering. This choice is sup-ported by decades of use by the major manufac-turers of commercial appliances. Many who arelarge enough to specify their desired grade onbills of materials rather than simply buyingfrom service center inventories have routinelyused 201 and realized an approximately 8%lower cost before surcharges. Use of 201 versus304 reduces surcharges by almost 50%, whichcan be a much larger savings than the base pricesavings. Smaller manufacturers are often pre-cluded from these savings because of the gen-eral, if inexplicable, practice of service centersnot stocking 201 despite its being the most cost-effective general-purpose stainless grade. Theextended nickel price elevation from 2004 on-ward has a good chance of changing that situa-tion as end users rebel against surcharges,which cannot be passed on to their customers. Ithas been pointed out that there is an array of201-type grades, and that this is a drawback totheir wider adoption.

I recommend following American Society forTesting and Materials (ASTM) A240 and speci-fying UNS S20100 when substituting for 304 asthis has very similar performance in forming,welding, and appearance to 304 and can bemost easily interchanged without complicationsin manufacturing and field performance. For

Fig. 2 Stainless steel samples exposed on a North Carolinabeach for 10 yr. Source: Ref 12


parts made by deep drawing, substitution is stillvery possible, but deep-drawn grades are morefinely tuned to specific process paths and mustbe more tightly specified than general-purposegrades. The more commonly used alloys for ap-pliances are listed in Table 2.

The greatest savings comes, of course, fromusing ferritic grades, and they should be usedwhenever forming requirements permit, whichis the majority of the time, since most appliancecomponents experience little more than cutting,bending, and welding. There are important pre-cautions, however. Mechanical polishing resultsin unacceptable corrosion resistance, and thelow work hardening rate of ferritics causes themechanical polish to take on a different colorshade. This subtle difference can be magnifiedto objectionable levels when a mechanicallypolished ferritic stainless, such as 430, is putside by side with an austenitic such as 201 or304. This can be solved by specifying rolled-onfinishes, which look the same on ferritics andaustenitics. These finishes also supply the addedcorrosion resistance that makes alloys such as430 acceptable. It is still preferable to use adual-stabilized grade such as 468, which can bewelded without adverse corrosion effects andhas high formability and corrosion resistance atas little as half the cost of 304 when alloy sur-charges are factored in. Use of dual stabilizationpermits keeping titanium levels to a minimum,making it possible to avoid TiN-caused surfacedefects, which occur if significant TiN precipi-tation occurs before solidification in the originalsteel production. This occurrence is strictly athermodynamic phenomenon related primarilyto the titanium and nitrogen levels, whichshould be minimized so that the product of tita-nium times nitrogen is less than 0.0025 when

concentrations are in weight percent. This is dif-ficult to achieve for 17% Cr alloys if stabiliza-tion is by titanium alone.

Interior or working parts of appliances, to thedegree they require high cleanability or contactfood, are also often made of stainless. This isespecially true of dispensing machines, such asfor beverages, ice cream, and ice. Stainless inte-riors are often found in refrigerators and dish-washers. In the case of dishwashers, forming re-quirements are often severe enough to requirethe use of austenitic stainless. Rolled-on fin-ishes are generally preferred. Not a small reasonfor this is that this finish requires only a singletemper pass to both flatten and provide the fin-ish. This yields very consistent forming charac-teristics, meaning much lower breakage duringpress-forming operations. Rolled-on finishesalso have very high visual consistency, which isusually a very important quality criterion for ap-pliance manufacturers.

Canisters, chafing dishes, serving pans, etc.are generally made from austenitic stainlesssteel, which lends itself to the typical deep-forming operations used in their manufacture.Coatings are rarely used. If antimicrobial coat-ing were to be used in food contact, this wouldbe an ideal application since already cookedfood is most often in the intermediate tempera-ture danger zone at which bacteria can multi-ply. Food preparation tables also fit into thiscategory.

Appliance facades are increasingly usingstainless. These include refrigerators, stoves,microwaves, drawers, etc. Shelves and exhausthoods also benefit from being made of stainless.The drivers here are cleanability, durability,and esthetics. There are important visual con-siderations in these applications. Consistent,

Table 2 Stainless steels commonly used for appliances

Alloy UNS No.

Composition, %

C N Cr Ni Mn Si Mo Ti/Nb

201 S20100 0.15 0.25 16.0–18.0 3.5–5.5 5.5–7.5 1.00 . . . . . .301 S30100 0.15 . . . 16.0–18.0 6.0–8.0 2.00 1.00 . . . . . .304 S30400 0.08 0.10 18.0–20.0 8.0–10.5 2.00 1.00 . . . . . .316 S31600 0.08 0.10 16.0–18.0 10.0–14.0 2.00 1.00 2.0-3.0 . . .430 S43000 0.12 . . . 16.0–18.0 0.75 1.00 1.00 . . . . . .439 S43035 0.07 0.04 17.0–19.9 0.50 1.0 1.0 . . . 0.20 + 4 × (C + N),

to 1.10468 S46800 0.030 0.030 18.0–20.0 0.50 1.00 1.00 . . . Ti + Nb: 0.20 + 4 ×

(C + N), to 0.80436 S43600 0.12 . . . 16.0–18.0 . . . 1.00 1.00 . . . Nb + Ta: 5 × C, to 0.70444,YUS190

S44400 0.025 0.035 17.5–19.5 1.0 1.0 1.0 1.75–2.50 Ti + Nb: 0.20 + 4 ×(C + N), to 0.80

29-4C S44735 0.025 0.025 28.0–30.0 0.5 1.00 0.75 3.5–4.5 Ti + Nb: 0.20 + 4 ×(C + N), to 0.80


defect-free surface finishes are paramount. Thisagain can really only be achieved by rolled-onfinishes since abrasively polished finishes varyexcessively in roughness, reflectivity, and color.Panel flatness is often very important and an-other benefit from rolled finishes. If visiblewelds are required, as is often the case for prod-ucts such as hoods and counters, then specialfinishes with very long polish grains have amajor advantage in that the weld can be groundand polished with a belt sander of the appropri-ate grit size so that the weld blends impercepti-bly with the adjoining original surface. This is apractical impossibility with abrasively polishedfinishes and very difficult with rolled finishes.Freedom from fingerprinting can be anothervaluable attribute for faÁade applications. Thiscan be obtained on stainless by the mill applica-tion of a thin, bonded polymer film. All barestainless finishes show fingerprints. With un-coated stainless, it is best avoided by using min-eral oil-based cleaners.

Although very high alloy stainless steels areused for high-temperature kitchen applications,such as heating element sheathing (AmericanIron and Steel Institute [AISI] type 334), it isseldom used for oven interiors because it doestake on a heat tint when exposed to tempera-tures above 300 °C (570 °F). Range tops, whichsee lower temperatures, are normally stainless.Outdoor cooking grills, because they must en-dure exterior environments without corrosion,are almost always 304 or a similar grade. Heattint does not occur with these to a problematicdegree. Gas burner manifolds are also stainless.In this case, ferritics are required because of theneed for extreme high-temperature oxidation re-sistance and the desirability of a low coefficientof thermal expansion. The preferred alloys arethose developed for automotive exhaust sys-tems, variations on 409 and 439.

Flatware and cutlery were among the origi-nal uses of stainless. Stainless filled the gap be-tween carbon steel, which was hard but whoserusting was an obvious problem, and silver,which was soft and whose cost prohibited itsuse to all but the wealthy few. Cutlery is the do-main of martensitic stainless steel. The corro-sion resistance of martensitic grades cannot beimproved above modest levels, never reachingthat of 304, but this criterion is secondary tohardness because of the need to keep a sharpcutting edge. Maximum corrosion resistance isachieved in the as-quenched condition. But,some toughness is a valuable but not crucial

characteristic, so most cutlery is tempered atlow temperatures. The vast majority of require-ments for high-quality cutlery are satisfied by420 stainless. If greater cutting edge retention isdesired, then more or harder carbides are engi-neered into the martensitic matrix. This is doneby adding more carbon and chromium, as isfound in 440A and to a greater extent in 440C.The wear resistance added by carbides is pro-portional to their hardness and amount. Thechromium carbides of these straight-chromiummartensitic stainless steels are very hard, 1800HV, versus the 1100 HV hardness of iron car-bides. The addition of higher levels of carbonties up chromium so that it cannot add to corro-sion resistance, however, so that it can becomebarely rust resistant. Furthermore, at high car-bon levels, carbides precipitate in the liquid andare much coarser. These large carbides can pullout during edge honing, making a ragged ratherthan a fine, smooth cutting edge.

However, vanadium and tungsten have evenharder carbides, 2800 and 2100 HV, respec-tively. Through conventional casting and hotworking, only a small amount of these carbidescan be introduced into the matrix. The problemis that if primary carbides form during solidifica-tion, they tend to be coarse and to embrittle thealloy. Hard particles are much more useful forwear resistance if they are small and widely dis-persed. To a degree, this refinement of the pri-mary carbides can be achieved by raising nitro-gen levels. These problems can be circumventedby the use of powder metallurgy, which permitsthe solidification step on a macroscale to beskipped. Larger volume fractions of hard car-bides such as vanadium carbide and tungstencarbide can be added and dispersed. Table 3 liststhe martensitic alloys used for cutlery. It is rea-sonable to say that most of these alloys far ex-ceed the requirements of food preparation.

Improved corrosion resistance of these alloysis achieved by adding molybdenum at the ex-pense of chromium, which would cause exces-sive δ-ferrite retention if it were raised. This canbe seen in alloys above the basic 420.

Flatware has no hardness requirement, sograde selection is based on the need for per-ceived quality. At the high end is 304, which hasall the corrosion resistance that could be neededfor flatware. However, type 301 is commonlyused also, as are ferritic steels, such as 430, forlow-cost flatware. Depending on the shape ofthe final utensil, material is stamped or forgedand then finished.


Many kitchen utensils are also made entirelyor in part with stainless. Type 304 is the alloymost commonly used, but again any of thestainless steels with at least 16% Cr are ade-quate, and grade selection depends on formingand joining requirements.

Laundry appliances have converted signifi-cantly to stainless. This trend began in Europewith the development of high spin speed, hori-zontal axis washing machines. These washersuse far less water and energy to achieve higherlevels of cleaning with less damage to clothing.These features have eroded the share enjoyed byvertical axis, agitator-type washers, whose lowspeed allows them to be constructed of low-strength materials such as plastic or porcelain-coated steel. The stresses induced by the highspin speeds, which are necessary in horizontalaxis machines to take water removal from 80%to 95%, require the strength of stainless steel.Porcelain-coated carbon steel obviously can bestrong, but the coating is cracked by strains thatthe steel itself easily tolerates. An additional ben-efit to stainless over porcelain is that stainlessstarts smooth and becomes even smoother withuse, while porcelain becomes quite abrasive overtime as wear opens voids with edges that can bequite sharp and cause significant damage toclothing. Washer tubs and drums are made ofboth ferritic and austenitic stainless. The selec-tion is based on forming requirements rather thancorrosion or strength. If components can be madeby bending rather than stretching, then the lower-cost ferritics can be used. Ferritics should be a17% stabilized grade, such as 439 or 468, andaustenitics can be 201, 301, or 304. Unstabilizedferritic alloys, such as 430, should never be usedin welded applications. Dryers are less challeng-ing, and it is difficult to make a strong case forthe functional value of stainless. Those designsthat use stainless will last longer and be gentler to

clothing. Those, along with the implied quality ofstainless, are the main drivers for its use.

Heating and Water Heating. With the de-velopment of high-efficiency, natural-gas-fired,forced-air furnaces, stainless has come into do-mestic use as a heat exchanger material. Thesefurnaces gain their extra efficiency by condens-ing water from combustion gas exhaust. Thiscondensate can, depending on incoming air,contain corrosion elements, which has led to theuse of very highly alloyed ferritic stainless steelin their construction. Alloy 29-4C (UNSS44735) was the original alloy used nearly uni-versally in the United States. The worst conse-quence of perforation by pitting could be the re-lease into the home of toxic gas, so pittingcorrosion must not be allowed.

The intermediate efficiency furnaces (80 to90%) require the use of corrosion-resistant ventpipe to prevent corrosion from condensation inthe flue. High-temperature plastics were tried,but failed joints in them caused their recall afterseveral fatalities were reported. High-efficiency(90% or higher) furnaces can use low-tempera-ture plastic pipe, but these units require the useof a corrosion-resistant secondary heat ex-changer to recover the latent heat of vaporiza-tion of the water from combustion. Alloy 29-4Cwas the original choice for most secondary heatexchangers, but at least one used alloy 6XN(UNS N08367) alloy for formability. Somemanufacturers have always used lower-alloyedstainless steels.

The proper handling of combustion productsis an interesting problem in materials selection.The variability of the use environments leads toa huge spread in corrosion conditions and mate-rials performance. In the end, one has to balancematerials selection between cost (fortunately,29-4C alloy is nickel free) and probability offailure. Given the number of units produced and

Table 3 Stainless steels commonly used for cutlery

Alloy Designation Form

Composition, %

C Mn S Si Cr Mo Ni Other

420 UNS S42000 Wrought 0.15 min 1.00 0.030 1.00 12.0–14.0 . . . . . . . . .4116 DIN 1.4116, nominal Wrought 0.50 . . . . . . . . . 14.5 0.65 . . . 0.15 V440A UNS S44002 Wrought 0.60–0.75 1.00 0.030 1.00 16.0–18.0 0.75 . . . . . .440C UNS S44004 Wrought 0.95–1.20 1.00 0.030 1.00 16.0–18.0 0.75 . . . . . .BG-42 Nominal composition Wrought 1.15 . . . . . . 0.3 14.5 4.0 . . . 1.2 VATS-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 VCPM S60V Nominal composition PM 2.15 0.40 . . . . . . 17.0 0.40 . . . 5.5 VCPM S90V Nominal composition PM 2.20 . . . . . . . . . 13.0 1.0 . . . 9.0 V


the potentially serious consequences of failure,failure rates must be less than 10– 4, while fail-ure rates much less than 10– 6 are impossible toverify and hard to justify. In any case, the com-petition is always between various stainlesssteels. Galvanized steel will not work. The issueis difficult enough for natural-gas-fired fur-naces. Oil fired is a developing situation forwhich there is no good consensus. Wood burn-ers and other unconventional furnaces (such ascorn burners) present additional challenges, andanswers are even less obvious.

Water heaters are sometimes made of stain-less steel. It is not uncommon for water to havea sufficient level of chlorides to lead to stresscorrosion cracking if an austenitic stainless isused. Therefore, the recommended alloy for thisapplication is UNS S44400. More recently, leanduplex alloys have been developed, such as2101 and 2003, which can perform quite wellwithout corrosion or stress corrosion cracking.More highly alloyed duplex alloys such as 2205are more expensive but would work well.

REFERENCES

1. M.J. Julio, M.L. Martin, and J.M. Baena,Cation Migration Tests in Metal Containers,Innovation Stainless Steel (Florence), Oct1993, p 1.221–1.226

2. “Stainless Steel in Contact with Food,”Document Ugine, June 1996

3. “The Application of High CorrosionResistance Stainless Steel YUS270 inFood Processing Facilities and Equipment,”Nippon Steel Technical Report 87, Jan2003

4. J.T. Holah and R.H. Thorpe, Bacteria Re-tention on Cleaned Surfaces, J. Appl. Bacte-riol., Vol 69, 1990, p 599–608

5. Agion Technologies, www.agion-tech.com,accessed June 2008

6. Removal of Stains and Discolourations, Out-okumpu, www.outokumpu.com, accessedJune 2008

7. “The Care and Cleaning of Stainless Steel,”Specialty Steel Industry of North America,www.ssina.com, accessed June 2008

8. E.P. Kysinski et al., J. Food Processing, Vol55, 1992, p 246–251

9. A.A. Mafu et al., J. Dairy Sci., Vol 73,1990, p 3428–4332

10. P. Gelinas and J. Goulet, Can. J. Microbiol.,Vol 29, 1983, p 1715–1730

11. R.A. Goyer, Toxicity of Metals, Propertiesand Selection: Nonferrous Alloys and Spe-cial-Purpose Materials, Vol 2, ASM Hand-book, ASM International, 1990, p 1233–1269

12. Allegheny Ludlum research, as presented inD.S. Bergstrom and C.A. Botti, AL201HPTM (UNS S20100) Alloy: A High-Performance, Lower-Nickel Alternative to300 Series Alloys, Stainless Steel World,KCI Publishing, 2005

CHAPTER 21

Marine Systems Applications

Summary

AS RECENTLY AS THE 1960s AND 1970s,handbooks on stainless steel were stating that“stainless steels are not stainless in seawater,”and “successful prolonged corrosion-free serv-ice of stainless steel in seawater requires sophis-ticated corrosion engineering, or enormous goodfortune” (Ref 1). The advances in stainless steelmade since then have thankfully made thesestatements obsolete. Not only have basic corro-sion problems been solved, stress corrosioncracking also can be avoided. More impres-sively, this can be done with alloys withstrengths much higher than those of the alloys,such as 316, that they replace and that have beenonly marginally successful in marine environ-ments. The inertia in changing from the weaker,less-corrosion-resistant, more expensive aus-tenitic stainless steels is large because of lessavailability of the newer, better alloys, and lackof familiarity with their benefits. Those whounderstand and use these newer duplex alloyswill be rewarded. This chapter reviews the majormarine applications of stainless steels, includingdesalination equipment, shipping containers,and heat exchangers that handle seawater.

Desalination

At one time not long ago stainless steel wasthought to be an inadequate to marginal materialfor use in seawater. Its use in heated seawaterwas therefore all the more suspect. This was firstchanged with the development of superferriticand superaustenitic alloys. The superferritic al-loys such as Seacure (UNS S44660) and 29-4C(UNS S44735) are quite resistant to seawater,even at high temperature. Their low toughness

restricts their use to items of rather thin gauge,less than about 1.0 to 2.0 mm, depending onalloy. Thus, their use is limited to tubing. Super-austenitic alloys can be used at any thickness, although they are a costly material. The successstory for stainless steel and seawater and there-fore desalination is that of duplex stainless steel.With the same corrosion resistance as any superaustenitic or superferritic alloy, it has nearlydouble the strength plus resistance to stress cor-rosion cracking. And while duplex stainless steelis not a cheap material, it does contain much lessnickel than an equivalently corrosion resistantaustenitic stainless steels, which is a major costsaving factor.

Desalination technology is relatively new ifone ignores the fact that distillation has beenaround for a very long time. Desalination in com-mercially viable quantities began with multi-stage flash technology in the 1950s. The underly-ing principle of this process is the evaporation ofwater vapor from salt water with the subsequentcondensation of the salt-free water vapor. In themulti-stage flash (MSF) approach feedwater isheated and the pressure is lowered so that thewater “flashes” into steam. A variation on thistechnology is multiple-effect distillation (MED),another low-temperature distillation process. Thedifferences in all distillation-based systems reduce to the efficiency of the design in minimiz-ing energy consumed per unit of pure water out-put. All distillation processes require heating ofthe input water and some process power.

The other basic engineering approach to de-salination is reverse osmosis (RO). The inven-tion of polymer membranes that could separatethe salt ions from the water made this technol-ogy possible. No thermal energy is required. Thewater is pumped at high pressure through thesepermeable membranes physically separating the




salt from the water. The change in salt concen-tration across the membrane is a function of thepressure and the membrane itself. A secondtreatment may be used to improve water quality.

The distillation methods require about 5kWh/metric ton of water output, while the ROmethods require twice that. The distillationmethods require another 20 kWh of thermal en-ergy from some source for feedwater heating,while the RO method requires none. Thus theability to find energy from cogeneration or asource such as solar, etc may determine whichprocess is preferred.

Materials Selection for Desalination

Materials used for distillation processes haveevolved from use of type 316 (UNS S31600)stainless steel, first as lining and then ascladding. The superaustenitic alloys, the 6Movariations, came next because they truly solvedthe corrosion problem, but at a price. Then, sep-arately the duplex alloys were developed, withthe first market the petroleum industry, whosedemands and research made these alloys possi-ble. It was not a stretch to see that high-strengthalloys that could withstand seawater in offshoreapplications could do well on land as well. Togive full credit, the pulp-and-paper industry wasalso beginning to employ duplex stainless steels

for their processes. Type 316 stainless steel haspassed from consideration as a material for han-dling brackish water or seawater.

In distillation systems, the rule of thumb isthat 2205 alloy (UNS S32205), with its pittingresistance equivalent number (PREN) of 35, issufficient for seawater up to 20 °C (70 °F); al-loys 2507 (UNS S32750) or Zeron 100 (UNSS32760) should be used for seawater at elevatedtemperatures or high salinity. For the output offresh water, lesser alloying is required. Stainlesssteel types 304 (UNS S30400), 316 (UNSS31600), 2101 (UNS S32101), 2003 (UNSS32003), or even 439 (UNS S43035) may beused depending on the combination of salinityand temperature of the output water.

Besides their high corrosion resistance forlower total alloy cost, the duplex stainless steelshave higher strength, which is a significant fac-tor since distillation plants are large. The use ofduplex allows wall thickness reductions thatbring about larger savings than those basedsolely on their cost per unit weight. Figure 1shows the difference among the candidate stain-less steels in corrosion resistance (Ref 2). Theviable materials for seawater are those that canwithstand roughly 20,000 ppm Cl– level at theappropriate temperature.

The strengths of the various candidate materi-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

Refer to the appropriate design code for yourparticular application to find minimum proper-ties. The reader is cautioned that duplex longitu-dinal properties are slightly lower than thetransverse properties that testing requires.

Pumps for seawater follow the same guide-lines as piping, tanks, and all other components.Cast or wrought duplex are the alloys of choice.

Shipping

The major uses of stainless steel in shippingare in bulk storage containment. Cargos rangefrom food and beverages to chemicals and liquidnatural gas (LNG). Practice in the past has beento use austenitic grades of stainless with cathodicprotection when necessary to address inadequatecorrosion resistance. However, since 2000 ma-rine chemical tankers have become the largestconsumer of duplex stainless steel. The reasonfor this is that cargo tanks ideally have the widestpotential range of cargos possible. This range isdefined by corrosion resistance. This factor aloneis reason to choose duplex over austenitic alloyssuch as 316L (UNS S31603) or 317L (UNSS31703). An equally decisive factor is strength.With codes permitting the tank’s design to bebased on yield strength, the use of duplex al-loys—with strengths about double those ofaustenitic steels—permits significant weight re-duction. This is a major economic factor for shipowners in that dead weight can be replaced byfee-paying cargo at the same operating cost.These incremental revenues, over the life of thevessel, are many times the original cost of thematerial. Based on the high value for strength inship economics, it would seem that the highest-strength alloys, such as 2507 (UNS S32750),

may be justified based on strength alone; theirexceptional corrosion resistance would be simplyan excellent side benefit. Corrugated stainlessbulkheads are positioned within the carbon steelhull. The stiff corrugated bulkheads are them-selves structural strengtheners for the entire ship.The vertical corrugations also facilitate tankcleaning as internal stiffeners are eliminated.

Cryogenic containers are still the bastion ofaustenitic stainless steels. As leaner austeniticalloys have become less expensive than 9% Nialloy steel, a martensitic grade, they have be-come the material of choice. In this case, the201 types are preferred to 304 because 201 hasgreater strength at the cryogenic operation tem-perature and is, of course, less expensive. Theexpanding market for LNG has made oceantransport increasingly important because largedisparities in prices often are due to the diffi-culty in transporting it. The two best materialsare UNS S20153 and S20400, which performequally well. If higher strength is valuable to adesign for cryogenic uses, then UNS S21904(21-6-9 or Nitronic 40) could be used. Thisalloy has yield strengths of 460 MPa (65 ksi) atroom temperature and 1200 MPa (175 ksi) at–196 °C (–320 °F). It is completely resistant tomartensite formation.

Other shipboard systems benefit equally fromthe use of duplex stainless steel. This extends topiping, hardware, propellers, shafts, etc.

Heat Exchangers

Coolers for captive water systems such as forpower plants often need to resist corrosion bybrackish water or seawater. To the extent thatthese are thin-wall tubing, ferritic alloys such as

Table 1 Typical analyses and properties of major marine alloys

Alloy UNS

Composition, %

PREN(a)

Yield strength Tensile strengthElongation,

%Cr Mo N Ni MPa ksi MPa ksi

2101 S32101 21.5 0.3 0.22 1.5 26 515 75 650 94 402003 S32003 20.5 1.5 0.18 3 29 515 75 725 105 402205 S32205 22 3 0.17 5 35 515 75 760 110 352507 S32750 25 4 0.27 7 42 550 80 800 116 35304L S30403 18 0 0.05 8 18 220 32 520 75 50316L S31603 16 2 0.05 10 24 220 32 520 75 50317L S31703 18 3 0.05 14 29 230 33 540 78 456XN N08367 21 6 0.22 24 45 380 55 760 110 45254SMO S31254 20 6 0.20 18 43 380 55 750 109 45Zeron 100 S32760 25 3.5

(+0.75 W)0.27 7 42 550 80 750 109 35

(a) PREN, pitting resistance equivalent number.


Seacure (UNS S44660) or 29-4C (UNS S44735)have been used quite successfully. If thickertubes are required, then the equivalent duplex oraustenitic alloys can be used. This would includetypes 2003 (UNS S32003), 2205 (UNS S32205),or 2507 (UNS S32750) duplex stainless steels,depending on salinity and temperature; foraustenitics, the 6Mo alloys such as 254SMO(UNS S31254) and AL6XN (UNS N08367) maybe used. The duplex alloys have the advantageof lower cost. Both are resistant to stress corro-

sion cracking to very high temperature andsalinity.

REFERENCES

1. Peckner and I. Bernstein, Stainless SteelHandbook, McGraw-Hill, 1966, p 37-1

2. Stainless Steel for Desalination Processes,Feb 2006, Outokumpu, www.outokumpu.com, accessed June 2008

CHAPTER 22

Petroleum Industry Applications

Summary

THE PETROLEUM INDUSTRY has had todeal with increasingly hostile environments inits search for new supplies of oil. And that pe-troleum, when found, often contains harmful in-gredients. The result is increasing demand forsteels with greater strength and corrosion resist-ance. Martensitic and duplex stainless steelshave provided the corrosion resistance andstrength to deal with higher levels of hydrogensulfide, carbon dioxide, chlorides, and acidity.This chapter reviews the selection of stainlesssteels for petroleum applications, including oilcountry tubular goods (OCTGs), line pipe, off-shore platforms, and refinery equipment.

Introduction

The petroleum industry has driven large seg-ments of the steel industry since both their be-

ginnings. Demand for steel for drill pipe, cas-ing, and tubing has led to many developments,such as the technology for producing high-quality seamless and welded pipe and tubing.Pipeline needs have fueled the market for high-strength, low-alloy plate. Offshore productionin often-hostile environments has presented se-vere material challenges. And, as the light sweetcrude that was easily found and produced onland is exhausted, future supplies of hydrocar-bons are increasingly likely to contain sulfides,carbon dioxide, and saltwater in sufficientamounts to make corrosion a top priority in se-lecting materials. For reference in this chapter,Tables 1 through 5 list the relevant alloys forpetroleum industry applications. Many, but notall, of these alloys are listed in the National As-sociation of Corrosion Engineers (NACE)MR0175, Sulfide Stress Corrosion Cracking Re-sistant Metallic Materials for Oil Field Equip-ment; the tables in this chapter also includesome newer alloys not in the NACE document.

Table 1 Ferritic stainless steels for petroleumindustry applicationsUNS Common name

S40500 405S40900 409S43000 430S43035 439S43400 434S43600 436S44200 442S44400 444 (18-2)S44500 …S44600 446S44626 26-1 Ti, E-BriteS44627 26-1 CbS44635 26-4-4, MonitS44660 Seacure, SC-1S44700 29-4S44735 29-4CS44800 29-4-2S46800 468Note: See Appendix 1 for alloy compositions. Source: Adapted from NACEMR0175, “Sulfide Stress Corrosion Cracking Resistant Metallic Materials forOil Field Equipment”

Table 2 Martensitic stainless steels for petroleum industry applications

UNS Common nameHardness,

HRC, max(a)

J91150 CA15 . . .J91151 CA15M . . .J91540 CA6 NM . . .K90941 9Cr 1Mo . . .S14125 S/W 13Cr 28S41000 410 22S41426 13CRS . . .S41427 … 29S42000 420 22S42400 F6NM 23S42500 15Cr 22

JFE KL-12G . . .JFE KNHP12Cr . . .Nippon NT-CRS . . .Nippon NT-CRSS . . .420M . . .L80 13 Cr . . .

Note: See Appendix 1 for alloy compositions. (a) As specified in NACEMR0175. Source: Adapted from NACE MR0175, “Sulfide Stress CorrosionCracking Resistant Metallic Materials for Oil Field Equipment”




The modern dilemma that makes stainless nec-essary is the addition presence of wet carbondioxide, which is extremely corrosive to carbonand alloy steel. As if this is not a sufficient ma-terial problem, sometimes the wetness is fromsaltwater, which further aggravates corrosivity.This corrosion problem is compounded by theaccelerating influence of high temperature indeeper formations. What is the answer to thecorrosion problem? Inhibitors, coatings, ca-thodic protection, or more corrosion-resistantmaterials are the main responses. The first threeresponses are not always practical. They alsorepresent an ongoing cost rather than a one-timecost. Each situation must be evaluated regard-ing which is the optimal solution.

Combating Corrosion in Alloys for Petroleum Applications

Alloying steel with chromium, copper,molybdenum, and nickel can lower the corro-sion rate of steel by a factor of 10,000. Figure 1shows the influence of chromium alone, whichproduces a 100-fold reduction in corrosion ofsteel in seawater and carbon dioxide (Ref 1).

Molybdenum is the most powerful alloyingaddition to magnify the benefit of chromium.The effects of copper and nickel are also verysignificant, as Fig. 2 (Ref 2) indicates. These ad-ditions must be made in a very balanced way ifa tough, fully martensitic structure is to bemaintained. Carbon must be kept low to avoidthe formation of chromium carbides duringtempering, which would counteract the benefitof the chromium. Nickel is necessary to preventσ-ferrite formation, which reduces toughness.

The chapters on martensitic and precipitationhardening stainless steels discuss this in detail.The martensitic stainless steels used for theseapplications are resistant to carbon dioxide-enhanced corrosion up to partial pressures of100 atm, after which further alloying is neces-sary. This cannot be achieved with a martensiticstructure, but the duplex alloys have the corro-sion resistance and strength to work in thisregime. They have high annealed strength andcan also be cold worked to higher strength levels.

If hydrogen sulfide is present, the selectionprocess can become more difficult. High-strength martensitic steels are susceptible tobrittle delayed failure in the presence of hydro-gen sulfide. Being stainless does not by itselfprovide immunity. If localized corrosion occurs,hydrogen uptake ensues, and delayed failurefollows. Only keeping hardness below well-established levels can render a martensitic alloyimmune. If the localized corrosion can be pre-vented, however, then the stress corrosioncracking (SCC) cannot be initiated. Molybde-num alloying expands the pH and chloriderange from which an alloy can be free of the pit-ting corrosion that initiates SCC, as shown inFig. 3 (Ref 2). Martensitic steels of all typeshave a maximum in susceptibility to SCC viahydrogen embrittlement near room temperature.Duplex alloys and austenitic alloys becomesusceptible at higher temperatures and do not

Table 3 Precipitation hardening stainless steelsfor petroleum industry applications

UNS Common nameHardness,

HRC, max(a)

S13800 13-8 PH . . . S15500 15-5 PH 33S15700 15-7 PH 32S17400 17-4 PH 33S17700 17-7 PH . . .S35000 AM-350 . . .S35500 AM-355 . . .S45000 Custom 450 31S45500 Custom 455 . . .S46500 Custom 465 . . .S66286 A-286 35

Custom 465 (275) . . .Custom 475 . . .

Note: See Appendix 1 for alloy compositions. (a) As specified in NACEMR0175. Source: Adapted from NACE MR0175, “Sulfide Stress CorrosionCracking Resistant Metallic Materials for Oil Field Equipment”

Table 4 Duplex stainless steels for petroleumindustry applications

UNS Common name PREN(a)

J93345 Escoloy 31-47J93370 CD4MCu 30-34J93380 Z100 38-46J93404 958 39-47S31200 44LN 30-36S31260 DP3 34-43S31500 3RE60 27-31S31803 2205 (old) 30-36S32001 19D 20-24S32003 2003 27-31S32101 2101 25-29S32205 2205 (new) 34-38S32304 2304 23-27S32404 U50 27-32S32520 52N+ 37-48S32550 255 32-44S32750 2507 38-44S32760 Zeron 100 40-46S32803 2803Mo 33-41S32900 329 26-35S32906 2906 36-45S32950 7-Mo Plus 32-43S32977 AF 918 39-46S39274 DP3W 39-47

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

Chapter 22: Petroleum Industry Applications / 249

Fig. 1 Influence of chromium on the corrosion rate of steel inenvironments experienced by oil country tubular

goods. Test conditions: synthetic sea water; CO2 partial pres-sure, 0.1 MPa; test temperature, 60 (C °140 °F); test duration,150 h; flow velocity, 2.5 m/s; specific volume, 800 mL/cm2.SSC, stress corrosion cracking. Source: Ref 1

Fig. 2 Influence of copper and nickel on the corrosion rate ofmartensitic stainless alloys used for oil country tubular

goods. Source: Ref 2

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 46N08028 Sanicro 28 39 S31266 B66 59N08031 Nicrofer 3127 hMo 54 S31600 316 . . .N08320 20Mod 38 S31603 316L 25N08366 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 64S20400 Nitronic 30 (204L) . . . S33000 330 . . .S20430 204 . . . S33400 334 . . .S20500 205 . . . S34565 4565 54S20910 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”


exhibit the same increasing susceptibility withstrength. So, when hydrogen sulfide, which en-hances hydrogen uptake, levels exceed about10–2 atm, the martensitic alloys should no longerbe used, and the duplex alloys are preferred. Astemperatures and hydrogen sulfide partial pres-sures increase, alloying must also, until at 1 atmof hydrogen sulfide nickel base alloys are re-quired. Figure 4 shows this progression with thealloy recommendations of Sumitomo. The re-quirements behind this diagram are generic.Any producer’s alloys must comply with thisdiagram’s regions, which have been defined byNACE. Stainless steels are required above acertain carbon dioxide level for all levels of hy-drogen sulfide. Martensitic alloys, commonlycalled “13Cr,” are the first step up from alloysteels. At higher levels of carbon dioxide andhydrogen sulfide, duplex alloys are required,with the 22CR alloys such as UNS S32205 usedat temperatures up to 200 °C (390 °F) and the25CR alloys such as UNS S32507 at tempera-tures up to 250 °C (480 °F). Nickel-base alloysare required at hydrogen sulfide levels above 1atm partial pressure.

The NACE recommendations of suitable ma-terials are defined by MR0175. Table 6 summa-rizes these recommendations. The reader is en-couraged to refer to the latest version of thisdocument for further details.

Oil Country Tubular Goods

Oil country tubular goods (OCTG) includethe drill pipe, casing, and tubing and associ-ated hardware used to construct oil and gaswells. Drill pipe is used to twist the drill bitand convey drilling fluids to the point of con-tact and flush away debris. Casing is put inplace to stabilize the well walls, while tubing isplaced within the casing to carry oil and gas tothe surface. Each of these components sees sig-nificant stresses, and high strength-to-weightmaterials are needed. Drill pipe is in tension,torsion, and compression alternately throughoutits life. Casing hangs from the wellhead underits own weight for distances from hundreds ofmeters to 7000 or 8000 m and must withstandvery high collapse as well as burst pressures.Well strings, the exact sequence of size andstrength pipe for each level of the well, are opti-mized for the conditions of each well. The vari-ety of strengths and sizes are standardized bythe American Petroleum Institute. The use ofthe highest strengths has always been limited byhydrogen embrittlement accelerated by hydro-gen sulfide, so that the maximum hardness for agiven material must be strictly adhered to whenhydrogen sulfide is present.

The terms 13Cr, 22Cr, and 25Cr are com-monly used in the industry even though thisgreatly oversimplifies the alloying, and there-fore performance, options that exist. The 13Cralloys are a family of martensitic stainlesssteels. The 22Cr and 25Cr alloys are duplexgrades. The former are used in the quenchedand tempered condition, while the duplex alloysare used as annealed or cold worked.

The 13Cr grades began as simply variationson 420, which is a straight-chromium marten-sitic often used for cutlery. This alloy, while farbetter (about 100 times) than alloy steel in cor-rosion resistance, has nearly the least corrosionresistance of all stainless steels. To achievehigher corrosion resistance molybdenum isadded. Molybdenum at 1% increases resistanceto general corrosion in a sodium chloride/hydrogen sulfide/carbon dioxide environment byabout tenfold. Another 1% increases it another

Fig. 3 Influence of molybdenum on susceptibility to stresscorrosion cracking in solutions containing (a) 3.5%

NaCl and (b) 0% NaCl. Source: Ref 2


tenfold. The 2% level of molybdenum alsogreatly reduces pitting, which in turn eliminatesthe initiation point of SCC. Simply addingmolybdenum would cause the alloy to have ex-cessive δ-ferrite, which cannot transform tomartensite and would therefore reduce mechan-ical properties. Thus, nickel must be added tocounter the ferrite stabilizing effect, unfortu-nately, but necessarily increasing the cost. Thenickel does help lower the general corrosionrate. Carbon and nitrogen in these alloys arekept at low concentrations. These alloys are oth-erwise almost identical to precipitation harden-ing martensitic stainless steels without the pre-cipitating phase.

Martensitic alloys are susceptible to SCC by ahydrogen embrittlement mechanism. This sus-ceptibility is strongly temperature dependent. Itdecreases with temperature from a maximum atambient to none at around 100 °C (210 °F). If thehydrogen sulfide level exceeds 0.03 atm, then22Cr alloys should be used rather than 13Crbecause of this risk. Hydrogen sulfide may becontained in the petroleum, or it may come fromsulfate-reducing bacteria, introduced by flooding,for example. This can cause a sulfide-free systemto become sulfide rich after the fact and make ini-tial materials choice wrong after the fact.

The 22Cr and 25Cr alloys have significantlyhigher resistance to chlorides and wet hydrogen

Fig. 4 Alloy suitability as a function of H2S and CO2 partial pressure. Source: Ref 1


sulfide and can resist SCC at ten times higherconcentrations than the 13Cr alloys. Besides theinherently greater corrosion resistance that de-rives from the chromium, molybdenum, and ni-trogen levels of the 22Cr and 25 Cr alloys (seethe chapters on corrosion in this book), the du-plex alloys have very fine grain size and aroughly 50/50 mixture of ferrite and austenite.This acts as a crack arrestor should one phase besusceptible to cracking while the other is not.

There have been no reported downhole fail-ures of annealed or cold-worked duplex alloys.There was one instance of very high-strengthtubing cracking after cathodic contact with car-bon steel casing. This was after removal fromthe well and after handling damage had oc-curred. The affected microstructure was foundto be high (70%) in ferrite (Ref 3).

For corrosion resistance above that furnishedby superduplex materials such as the 25Cr al-loys, super austenitic alloys fill a gap beforenickel base alloys are needed. These alloysachieve a tenfold increase in hydrogen sulfideresistance and very elevated SCC resistance.These are the so-called 6Mo grades. The moreadvanced of them contain high levels of nitro-

gen. The more common alloys are UNS S32654and N08367.

The recent development of lean duplex alloyshas not yet made its way into OCTGs. These al-loys offer an inherent alloy savings over the13Cr grades in nickel and molybdenum contentwhile offering better corrosion and SCC resist-ance. Their strength levels in the annealed con-dition, 450 MPa (65 ksi), are lower than thoseof the martensitic alloys, 600 MPa (87 ksi), sofor most downhole applications they will re-quire cold working. It is likely, however, thatthese alloys will see their first service as linepipe, where they will not need to be cold workedto higher strength levels to be widely used.

Line Pipe and Flow Lines

With the awesome cost of corrosion, thecase for stainless line pipe is easily made.Whether to use stainless depends on vulnera-bility of carbon steel. This evaluation is madebased on the carbon dioxide, hydrogen sulfide,water, salinity, temperatures, pressures, flowconditions, and so forth. The normal basis for

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

UNSCommon

name PREN(a)Temperature,

ºC (ºF) pH, min H2S, kPaHardness,

HRCCl–,

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 90,000N08926 . . . . . . 121 (250) 3.5 700 . . . 60,700S15500 15-5 . . . . . . . . . 3.4 33 . . .S15700 15-7 . . . . . . . . . . . . 32 . . .S17400 17-4 . . . . . . . . . 3.4 33 . . .S20910 . . . . . . 66 (150) . . . 100 . . . 35S41000 410 . . . . . . 3.5 10 22 . . .S41425 . . . . . . . . . . . . 10 28 . . .S41426 . . . . . . . . . 3.5 10 27 . . .S41427 . . . . . . . . . 3.5 10 29 6,000S41429 . . . . . . . . . 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,

type 3a(Ni+2Mo) >30 . . . . . . . . . . . . . . .

Superaustenitic,type 3b

>40 121 (250) . . . 700 . . . 5,000149 (300) . . . 310 . . . 5,000171 (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”


these calculations follows that published by C.de Waard of Shell (Ref 3). The competing tech-nology when corrosion dangers arise is the useof corrosion inhibitors, cathodic protection, orto line carbon steel with a protective coating.The use of inhibitors is subject to the risk of ve-locity limitations, temperature limitations, andsimply of the inhibitor working appropriately,not to mention cost. Cathodic protection iscostly and complex. Coatings can be damagedby numerous occurrences in acidity, mechani-cal damage, or fluctuations in temperature orpressure. Use of a stainless corrosion-resistantalloy can have well-defined and controlledcosts and performance over the life of an instal-lation. Line pipe differs from downhole in hav-ing strength requirements more in line with thatof annealed duplex alloys. These requirementsgave birth to modern duplex alloys, startingwith UNS S31803 and evolving to UNSS32205 as the value of higher nitrogen becameunderstood.

Nitrogen not only enhances corrosion resist-ance, but also suppresses the formation of unde-sirable and embrittling intermetallic phases thatmight otherwise form at welding temperatures.It also keeps the desirable austenite/ferrite ratiosin weld metal.

Since the development of the first widely ac-cepted duplex alloys, more alloys haveemerged. Superduplexes, such as UNS S32750,have become accepted alloys. Then, the need toimprove costs led in the 1990s to the use ofmartensitic alloys with high levels of nickel andmolybdenum, which at the time were lowercost. The emergence in the early 2000s of leanduplex alloys provided strength and more corro-sion resistance with lower nickel levels, givingthem a cost advantage during periods of highnickel cost.

The main attribute required by line pipe thatis not as important in OCTGs is weldability.This is not an overwhelming challenge for du-plex alloys, but for martensitic alloys, it re-quires a very low interstitial level so that themartensite is self-tempering and ductile in theas-welded condition. This can be achieved bystabilizing the alloy with small amounts of tita-nium. It would appear that under current condi-tions that alloy 2101 (UNS 32101) has acost/performance edge over the martensiticcompetition and should for the long term. Themain ingredients required in a duplex forstrength and corrosion resistance are chromium

and nitrogen, both relatively inexpensive alloy-ing elements. Alloys S32001 and S32101 arewell formulated for medium and high levels ofcorrosion resistance required for wet carbondioxide, hydrogen sulfide, and trace chlorides.The main precaution for duplex alloys is main-taining a nominally 50/50 mixture of ferrite andaustenite with no embrittling intermetallicphases. The modern alloys have high (greaterthan 0.14%) nitrogen, which helps to preserveaustenite levels after welding and suppress in-termetallic formation. Nevertheless, minimiza-tion of time above 350 °C (660 °F) is important.This tendency increases with chromium andmolybdenum content, which is another reasonwhy the lean duplex alloys are so attractive.

For subsea use, 22Cr duplex generally re-quires cathodic protection because of the risk ofcrevice corrosion. 25Cr duplexes are used with-out cathodic protection. Duplex pipelines havebeen in service in the North Sea since the 1970s.

Umbilical Tubing and Risers

Increasingly, wells are located undersea. It isstandard practice to control and monitor thesewells via bundled umbilical tubing. The tubingcan provide hydraulic and electrical power, con-trol and adjust pressure, carry communications,and even introduce chemical to the well. Thedepth of wellheads can increase collapse pres-sures to levels beyond the capability of thermo-plastics, which has led to the use of duplexstainless steel because of its strength and resist-ance to corrosion and SCC. When resistance toseawater is the main concern, the rule of thumbis that a pitting resistance equivalent number(PREN) of 35 or greater is required, whereas re-sistance to crevice corrosion requires a PRENof at least 40. This has made the superduplexUNS S32750 the standard. Such a critical itemas an umbilical may seem like a poor applica-tion on which to economize, but again the leanduplexes offer possibilities to do so. By zinccoating lean duplexes such as alloy 19D (UNSS32001) and 2101 (UNS S32101), very longservice lives can be safely extrapolated. Thesealloys are being promoted on their lower sus-ceptibility to σ formation during welding, and ifwelding thermal cycles cannot be controlledthat may be an issue, but superduplex seems tohave become a pervasive choice because it issuperbly reliable.


Risers are now produced in coiled tubing ofover 100 mm (4 in.) diameter, so that very eco-nomical long lengths are feasible.

Platforms

Platforms present a special case in which thecosts of maintenance are high, the corrosion en-vironment is severe, and the penalty for excessweight is also high. A savings of 1 ton in weighttopside can save over $100,000 in steel in thesubsea jacket. This leads to a rapid payback forthe use of materials that are sufficiently resistantto corrosion such that corrosion loss allowancecan be eliminated. Both titanium and stainlessalloy UNS S32750 are equal candidates for thisservice, depending on availability and currentalloy prices. Except in rare cases, stainless steelwins the cost battle between these alloy systems.

Almost any structure is a candidate for stain-less topside processing: piping, pumps, flanges,fittings, etc. Hardware of any type and construc-tion materials benefit from being stainless. Sea-water systems often employ 22Cr duplex withcathodic protection or unprotected 25Cr duplex.

A wise preventive action is to paint stainlessthat is covered by insulation or similar material,which otherwise can result in concentrationcells and consequent pitting.

Liquefied Natural Gas Vessels

Liquefied natural gas (LNG) is becoming anincreasingly important commodity as the valueof stranded gas makes it economically desirableto convert it to a transportable state. Convertingnatural gas to a cryogenic liquid presents a ma-terial problem. Vessels to contain it must havestrength and toughness at temperatures below–150 °C (–240 °F). The traditional material, 9%Ni martensitic steel, has become expensivecompared to the lower-nickel austenitic stain-less steels, such as 201LN (UNS S20153),which have no transition temperature andstrengthen with decreasing temperature. Alloy201LN is cheaper, easier to weld and fabricate,and of course is stainless, which 9% Ni steel isnot. The extreme ductility of 201LN comparedto martensitic steel gives it a decided advantagein terms of rupture resistance, which is a majordesign and political concern with this poten-tially explosive commodity.

Refinery Equipment

Corrosion resistance is a major factor in thechoice of materials in refinery operations. As wediscussed, crude oil itself is sometimes a verycorrosive 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(SCAs)

Preheaters, distillation tower

405, 409, 410 …

Vacuum fractionalization SCA, chlorides Towers 405, 410, 316 Depending on crude corrosivity

Condensers S44735, 2205 Depending on chloride levelCoker SCA, H2S Coke drums 409 Depending on crude corro-

sivityGas plants H2S, water, Cl–, ammonia Compressor coolers,

reboiler tubesAL-6XN, 2205, 2507 …

Trays 410S, 316L …Amine plant Ammonia, MEA, DEA Reboilers, trays, filters,

condenser tubing304L, 316L …

Sulfuric acid alkylation Sulfuric acid Contactor, mixer 20Cb3 Low pH excursions possibleDilute sulfuric acid Effluent piping 316L …

Hydrotreating H2S, ammonia, PTA(a) Hot sections 321, 347 Long exposure at high tem-perature

General 410S, 304 …Catalytic re-forming High-temperature strength

neededReactor internals 304 HCl catalyst regeneration

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,

6MoDepending on temperature,

risk of chloridesSour water stripping Sulfuric acid, ammonium

bisulfide, chloridesStripper 304, 20Cb3, 2205 Severity depends on pres-

ence of sulfuric acid

(a) PTA, polythionic acid. Source: Ref 4


chemicals used in refining and the temperaturesused may further aggravate that corrosivity. Theaggressive chemical agents that refinery materi-als must withstand include wet hydrogen sulfideand carbon dioxide, napthenic acids, polythionicacids, chlorides, sulfuric acid, and alkalines aswell as simple oxidation. Sometimes, tempera-tures of use are such that embrittling or sensitiz-ing phase transformations may occur. Table 7lists some major refinery processes and the ma-terials used in them (Ref 4).

Most of these situations are discussed else-where in this book in detail. One that is quitespecific to refinery applications is polythionicacid (PTA) attack. These acids usually form ac-cidentally when sulfide corrosion products reactwith moisture and air. The attack is intergranu-lar, and materials respond to it much as they doto the Strauss test. The remedies are to preventthe inadvertent formation of PTA and to avoid

using austenitics, which are prone to grainboundary chromium depletion by sensitization,and instead use low-carbon grades and stabi-lized grades.

REFERENCES

1. Sumitomo Products for the Oil and Gas In-dustries, www.sumitomometals.co.jp, ac-cessed June 2008

2. H. Asahi et al., “Development of HighChromium Stainless Line Pipe,” NipponSteel Technical Report 72, January 1997

3. C. de Waard and U. Lotz, “Prediction ofCO2 Corrosion of Carbon Steel,” Paper 69,presented at Corrosion/93, National Associ-ation of Corrosion Engineers, 1993

4. C.P. Dillon, Corrosion Resistance of Stain-less Steels, Marcel Dekker, 1995

CHAPTER 23

Chemical and Process Industry Applications

Summary

ENGINEERS IN THE PROCESS industriesmust have materials that can contain a huge vari-ety of chemical species at many temperatures,pressures, and flow rates. This is applied corro-sion engineering combined with physics andstructural design. It is obvious that this task de-pends on the availability of corrosion data, morethan can be presented here. This chapter coverswhat data are necessary and how they can befound.

Introduction

The need to work with hostile chemicals be-gins with the manufacture of those chemicals. Itwas in the production of nitric acid that stainlesshad its first industrial application. These are in-dustries with purely need-driven material chal-lenges. New processes are constantly in devel-opment, and they present new environments inwhich materials must perform. The choices arehighly pragmatic. In an industrial environment,the costs of a poorly performing material can bewell known by its effect on downtime, mainte-nance, liability, etc. The essential knowledge iswhich materials will work.

The selection of materials for the chemicaland power industries is first a study of corrosionresistance, including resistance to stress corro-sion cracking (SCC). Strength plays a secondaryrole but can be an important cost factor. Theseconsiderations may occur at very high or verylow temperatures, in which case corrosion resist-ance may become oxidation resistance andstrength may mean creep strength.

Single- and Dual-Environment Systems

Under ideal conditions, a material may needto resist one single major corrosion threat. If themost potentially damaging species can beclearly identified, then candidate materials canbe found by searching published data. Thesedata are available freely online from Web sites(such as Ref 1 and 2) or for a charge fromsources such as the National Association ofCorrosion Engineers (NACE; Ref 3) and ASMInternational (Ref 4). It is difficult for any pub-lished data to keep up with the latest develop-ments. The testing alone of new materials cantake a long time, and then it must wait for publi-cation. All materials are not covered, especiallywhen a manufacturer publishes data on propri-etary alloys and excludes competitive materials.That having been said, any improvements overstandard alloys will first be reported by the de-velopers of the alloy, and they will logicallytout its strongest points. For this reason, dia-logue with the primary steel producers is en-couraged. No one has more exposure to the lat-est trends in applications.

A single-environment system is typically onein which the aggressive chemical species is theonly consideration. This is normally the case forpiping, tanks, or reaction vessels holding thespecies or materials immersed in the aggressivespecies.

A dual-environment system is typically en-countered in heat exchangers, but it must alsobe extended to single-environment systems inwhich the exposure of the nonreactant side ofthe material to the ambient environment cannotbe neglected, as in the case of marine ambientenvironments.




The challenges that must be met are primarilyensuring adequate corrosion resistance and sec-ondarily having acceptable mechanical proper-ties. The corrosion issues run the full gamut ofpotential forms of corrosion:

• General corrosion• Pitting corrosion and crevice corrosion• Intergranular corrosion• Stress corrosion cracking• Erosion corrosion

In addition to these forms of corrosion associ-ated with liquids, there are considerations of gasphase attack, which may be oxidation, sulfida-tion, or attack by other gases.

Mechanical design considerations are nor-mally limited to static stress allowances. Previ-ously, handbooks dealt very lightly with thistopic because all the normally recommendedsteels had similar strength. The proliferation ofduplex stainless steels has changed that. Now,high-strength alloys of high corrosion resistanceand SCC resistance are available and are mak-ing traditionally chosen stainless steels less thanoptimal.

Corrosion Types

A designer wants to deal with general corro-sion. Its rate can be predicted, and thicknesscan be chosen to allow for its occurrence. Cor-rosion data for general corrosion are normallypresented in isocorrosion charts. These presentthe temperatures and concentrations for agiven environment at which various materialswill exhibit the same corrosion rate. This rateis most often 0.1 mm/yr, an amount that can bethought of as a tolerable level for many uses.Figure 1 shows an isocorrosion chart for stain-less steels in sulfuric acid (Ref 1). The data areclear when presented in this fashion. It can fur-ther be appreciated that in general reducingalloy performance to a mathematical formula,such as the pitting resistance equivalent num-ber (PREN) equation, would not be reasonablesince the relative performance of alloyschanges considerably with concentration.Thus, the design engineer must rely on experi-mentally developed data. Since these data areavailable both online and in print, no attemptwill be made to reproduce them fully here. Ex-amples are given in Tables 1 and 2 (Ref 1).Such tables are very useful, although the pre-sentation is not visually compact. The inclusion

of carbon steel and titanium gives a valuableframe of reference for the engineer.

If the forms of localized corrosion discussednext can be avoided, the corrosion tables aresufficient to guide the designer to a reasonableselection of candidate materials for any processin which the chemical species involved havebeen identified. If the data have not been devel-oped for a certain environment, then the tablesgive a first approximation of which materialsmay be resistant from examination of similarenvironments, and a final decision can only bereasonably made though direct corrosion testingof candidate materials. Refer to the chapters oncorrosion for a more thorough discussion ofuniform corrosion.

Pitting and Crevice Corrosion

Stainless steel is unique among metals and al-loys in that it derives its corrosion resistancefrom constituent alloying elements working to-gether to form a thin passive layer that, whenintact, is highly resistant to corrosion. Thestrength of the passive layer in resisting attackby halide ions, which are the most disruptiveions to the layer, is proportional principally tothe chromium, nitrogen, and molybdenum con-tents of the alloy. This relationship follows theformula:

PREN = %Cr + 3.3%Mo + 30%N (Eq 1)

This formula is one of the commonly usedversions, none of which is universally correct.Both tungsten and carbon can increase pittingresistance, while sulfur diminishes it. This isdiscussed in the corrosion section of this book.

Fig. 1 Isocorrosion chart for sulfuric acid. Source: Ref 1

Chapter 23: Chemical and Process Industry Applications / 259

The important consideration is that this formulaassumes that the key alloying elements are ho-mogeneously distributed in solution. This willonly be true if correct thermomechanical pro-cessing occurs because, thermodynamically,these alloys are not used in an equilibrium con-dition. Were they to attain equilibrium, say byoverheating, alloy segregation by precipitationcould occur, causing localized loss of corrosion

resistance, which is what causes pitting.Chromium is very reactive: Its affinity for oxy-gen makes the passive film strong. Pitting hasnearly always been associated with manganesesulfide inclusions, and although there is still de-bate over the precise mechanism, it appears thatchromium depletion at the metal-inclusion in-terface is to blame. Eliminating inclusions byeliminating either manganese or sulfur improves

Table 1 Corrosion table for sulfuric acid (H2SO4)

Concentration, %Temperature, °C

0.1100 = BP

0.520

0.550

0.5 100 = BP

120

150

170

185

1100 = BP

220

250

260

320

335

350

Carbon steel 2 2 2 2 2 2 2 2 2 2 2 2 2 2 213% Cr steel 2 2 2 2 2 2 2 2 2 2 2 2 2 2 218-2 (UNS S44400) 2 0 2 2 0 2 2 2 2 0 2 2 0 2 23R12 (UNS S30400) 2 0 1 2 0 1 1 2 2 0 1 1 0 1 13R60 (UNS S31600) 1 0 0 1 0 0 0 1 1 0 0 0 0 0 018-13-3 1 0 0 1 0 0 0 1 1 0 0 0 0 0 017-14-4 1 0 0 1 0 0 0 0 1 0 0 0 0 0 02RK65

(UNS N08904)0 0 0 1 0 0 0 0 1 0 0 0 0 0 0

Sanicro 28 (UNS N08028)

. . . 0 0 0 0 0 0 0 0 0 0 0 0 0 0

254SMO (UNS S31254)

. . . 0 0 . . . 0 0 0 0 1 0 0 0 0 0 0

654 SMO (UNS S32654)

. . . 0 0 . . . 0 0 0 0 0 0 0 0 0 0 0

SAF 2304 (UNS S32304)

1 0 0 . . . 0 0 0 0 1 0 0 0 0 0 0

SAF 2205 (UNS S31803)

. . . 0 0 1 0 0 0 0 . . . 0 0 0 0 0 0

SAF 2507 (UNS S32750)

. . . 0 0 . . . 0 0 0 0 0 0 0 0 0 0 0

Titanium 1 0 0 1 0 0 1 1 1 0 0 1 0 0 1


385

3100 = BP

520

535

560

575

585

5101 = BP

1020

1050

1060

1080

10102 = BP

2020

2040

Carbon steel 2 2 2 2 2 2 2 2 2 2 2 2 2 2 213% Cr steel 2 2 2 2 2 2 2 2 2 2 2 2 2 2 218-2 (UNS S44400) 2 2 2 2 2 2 2 2 2 2 2 2 2 2 23R12 (UNS S30400) 2 2 1 1 2 2 2 2 2 2 2 2 2 2 23R60 (UNS S31600) 1 2 0 0 1 1 2 2 0 1 1 2 2 0 118-13-3 1 2 0 0 0 1 2 2 0 1 1 2 2 0 117-14-4 1 2 0 0 0 1 2 2 0 0 1 2 2 0 12RK65 (UNS

N08904)0 1 0 0 0 0 1 2 0 0 0 1 2 0 0


. . . 1 0 0 0 0 0 2 0 0 0 0 2 0 0

254SMO (UNS S31254)

. . . 1 0 0 0 0 1 2 0 0 0 0 2 0 0

654 SMO (UNS S32654)

0 0 0 0 0 0 0 2 . . . . . . 0 0 . . . 0 0

SAF 2304 (UNS S32304)

. . . 1 0 0 0 0 0 2 0 0 0 2 2 1 2

SAF 2205 (UNS S31803)

. . . 1 0 0 0 0 0 2 0 0 0 1 2 0 0

SAF 2507 (UNS S32750)

. . . 1 0 0 0 0 . . . . . . 0 0 0 0 2 0 0

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 0.1–1.0 mm/yr. The material is not corrosion proof but useful in cer-tain 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


Table 1 (continued)


2050

2060

2080

20100

3020

3040

3060

3080

4020

4040

4060

4090

5020

5040

5070

Carbon steel 2 2 . . . 2 2 2 2 . . . 2 2 2 2 2 2 213% Cr steel 2 2 . . . 2 2 2 2 . . . 2 2 2 2 2 2 218-2 (UNS S44400) 2 2 . . . 2 2 2 2 . . . 2 2 2 2 2 2 23R12 (UNS S30400) 2 2 . . . 2 2 2 2 . . . 2 2 2 2 2 2 23R60 (UNS S31600) 1 2 . . . 2 1 2 2 . . . 2 2 2 2 2 2 218-13-3 1 1 . . . 2 1 1 2 . . . 2 2 2 2 2 2 217-14-4 1 1 . . . 2 1 1 2 . . . 2 2 2 2 2 2 22RK65

(UNS N08904)0 0 1 2 0 0 1 . . . 0 0 1 2 0 0 2


0 0 . . . 2 0 0 1 . . . 0 0 1 2 0 0 1

254SMO (UNS S31254)

0 0 . . . 2 0 0 1 2 . . . 1 . . . . . . 0 1 . . .

654 SMO (UNS S32654)

0 0 0 2 . . . . . . . . . . . . 0 0 0 . . . 0 0 . . .

SAF 2304 (UNS S32304)

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

SAF 2205 (UNS S31803)

0 1 2 2 0 1 2 2 2 2 2 2 2 2 2

SAF 2507 (UNS S32750)

0 0 1 2 . . . 0 1 2 0 1 2 2 1 1 2

Titanium 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2


6020

6040

6070

7020

7040

7070

8020

8040

8060

8520

8530

8540

8550

9020

9030

Carbon steel 2 2 2 2 2 2 2 2 2 0 1 2 2 0 113% Cr steel 2 2 2 2 2 2 2 2 2 1 1 2 2 0 118-2 (UNS S44400) 2 2 2 2 2 2 2 2 2 1 1 1 2 0 13R12 (UNS S30400) 2 2 2 2 2 2 2 2 2 1 1 1 2 0 03R60 (UNS S31600) 2 2 2 2 2 2 1 2 2 1 1 1 2 0 018-13-3 2 2 2 2 2 2 1 2 2 1 1 1 2 0 117-14-4 2 2 2 2 2 2 1 2 2 1 1 1 2 0 12RK65

(UNS N08904)0 1 1 0 1 1 0 1 2 0 0 1 1 0 0


0 0 1 0 0 1 . . . 1 1 0 0 0 0 0 0

254SMO (UNS S31254)

0 1 . . . 0 1 . . . 0 1 2 0 . . . . . . . . . 1 . . .

654 SMO (UNS S32654)

0 1 . . . 0 1 . . . . . . . . . . . . . . . . . . . . . . . . 1 . . .

SAF 2304 (UNS S32304)

2 . . . . . . . . . . . . . . . . . . . . . . . . 1 1 . . . . . . . . . 1

SAF 2205 (UNS S31803)

2 2 2 1 . . . . . . 2 2 2 1 . . . . . . . . . 1 1

SAF 2507 (UNS S32750)

. . . . . . . . . . . . 2 2 . . . 2 2 1 1 . . . . . . 0 0

Titanium 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2


9040

9070

9420

9430

9440

9450

9620

9630

9640

9650

9830

9840

9850

9880

Carbon steel 2 2 0 2 2 2 0 1 2 2 1 1 2 213% Cr steel 2 2 0 1 2 2 0 1 2 2 1 1 2 218-2 (UNS S44400) 2 2 0 0 2 2 0 0 1 2 0 1 2 23R12 (UNS S30400) 2 2 0 0 1 1 0 0 0 1 0 0 2 23R60 (UNS S31600) 1 2 0 0 0 1 0 0 0 1 0 0 0 218-13-3 1 2 0 0 1 1 0 0 1 1 0 0 1 217-14-4 1 2 0 0 1 1 0 0 1 1 0 0 1 22RK65

(UNS N08904)1 2 0 0 1 1 0 0 1 1 0 1 1 2


0 1 0 0 0 0 0 0 0 1 0 0 0 1

(continued)



the potential at which passive film breakdownoccurs. This is especially important for welds,which, if not annealed, can have maximumdeleterious segregation by both inclusions andsolidification segregation. All austenitic andduplex stainless alloys have best corrosion re-sistance when quenched from the solution an-nealing temperature. The precipitation harden-ing, martensitic, and ferritic alloys are morecomplicated but are less relevant to this topic. Ifinformation on them is needed, they are dis-cussed in detail in their respective chapters.

When a crevice is permitted to exist, it mim-ics the pH-altering action found within pits inwhich transport restriction leads to a buildup ofmetal and hydrogen ions and oxygen depletion.All alloys undergo crevice corrosion underless-aggressive conditions than those requiredto induce pitting, so care must be taken to avoidcrevices.

Intergranular Corrosion

Intergranular corrosion is a problem that canbe avoided entirely by correct alloy selectionand proper thermal processing. The principlecause of grain boundary attack is alloy deple-tion at the grain boundaries. The most familiarform of this problem occurs when austeniticalloys having carbon levels above 0.03% arewelded. The region near the weld where tem-peratures reach 600 to 900 °C (1100 to 1650 °F)may have carbon migrate to and along grainboundaries, the fast diffusion paths, where itcombines with less-mobile chromium atomsand precipitates as chromium carbide. Thislowers the chromium level in solution, result-ing in poor corrosion resistance only at thegrain boundaries. This is easily prevented by se-lecting alloys with low carbon levels. Duplexalloys, curiously, undergo chromium carbideprecipitation under the same conditions but donot undergo significant chromium depletionbecause the neighboring ferrite grains, in whichchromium diffuses more rapidly, contributechromium, mitigating the depletion. Precipita-tion segregation of all types, not just by carbides,must be guarded against. Sigma phase, nitrides,secondary austenite, and others can cause localbreakdown of corrosion resistance if alloys areheated to a dangerous temperature for sufficienttime. It is important to learn these potential

Table 1 (continued)


9040

9070

9420

9430

9440

9450

9620

9630

9640

9650

9830

9840

9850

9880

254SMO (UNS S31254)

1 . . . . . . . . . . . . . . . 1 . . . . . . . . . . . . . . . 0 2

654 SMO (UNS S32654)

2 2 . . . . . . . . . 2 0 1 . . . 2 . . . . . . 1 1

SAF 2304 (UNS S32304)

1 . . . . . . . . . . . . . . . 1 . . . . . . 0 . . . . . . 0 1

SAF 2205 (UNS S31803)

1 . . . 0 . . . . . . . . . 0 0 1 . . . 0 0 1 1

SAF 2507 (UNS S32750)

0 . . . 0 0 0 1 0 0 0 1 0 0 0 1

Titanium 2 2 2 2 2 2 2 2 2 2 2 2 2 2


Table 2 Corrosion table for fuming sulfuricacid (oleum), H2SO4 + SO3

Conc. H2SO4, %Conc. SO3, %Temperature, °C

100760

1001160

10011100

1006020

1006070

1006080

Carbon steel 0 0 2 . . . . . . . . .13% Cr steel 0 0 2 . . . . . . 218-2

(UNS S44400)0 0 . . . . . . . . . . . .

3R12 (UNS S30400)

0 0 1 0 0 0

3R60 (UNS S31600)

0 0 0 0 0 0

18-13-3 0 0 . . . . . . 0 . . .17-14-4 0 0 . . . 0 0 . . .2RK65

(UNS N08904)0 0 . . . 0 0 . . .

Sanicro 28 (UNSN08028)

. . . . . . . . . . . . . . . . . .

254SMO (UNS S31254)

. . . . . . . . . . . . . . . . . .

654 SMO (UNS S32654)

. . . . . . . . . . . . . . . . . .

SAF 2304 (UNS S32304)

. . . . . . . . . . . . . . . . . .

SAF 2205 (UNS S31803)

. . . . . . . . . . . . . . . . . .

SAF 2507 (UNS S32750)

. . . . . . . . . . . . . . . . . .

Titanium 2 2 2 2 2 2

Notes: 0, corrosion rate of less than 0.1 mm/yr. The material is corrosion proof. 1,corrosion rate of 0.1–1.0 mm/yr. The material is not corrosion proof but useful incertain cases. 2, corrosion rate of more than 1.0 mm/yr. Serious corrosion. Thematerial is not usable. Conc., concentration. Source: Adapted from Ref 1


vulnerabilities by reviewing the metallurgy ofany alloy selected for service.


The theory of SCC is still under debate. Thereader will find the arguments confusing as thedebate generates more heat than light. We willskip the theory; it can be found in the corrosionchapters. SCC, like excessive general corrosionor pitting, is avoided by referring to publishedtest data from the corrosion tables. If a materialmust be used where a risk of SCC occurs, thenstress levels must be managed to stay below thethreshold stress for SCC. Figure 2 shows howvarious alloys resist SCC as a function of chlo-ride concentration and temperature, the twomost important aggravating factors. Materialcomparisons are made difficult because tests arenormally run at a given fraction of a material’syield strength. Thus, the data in Fig. 2 (Ref 1)must be interpreted. Higher-strength duplex al-loys, while having better SCC performance thanaustenitics of equal corrosion resistance (e.g.,316 vs. 2304), have much better SCC resis-tance. Furthermore, the stress at which failurewill occur is much higher since the yieldstrength at which the testing takes place is abouttwice as high for duplex alloys. SCC also ex-hibits a threshold stress below which failuredoes not occur. This is about 60% of tensilestrength for duplex and about 30% for

austenitics. Designing within this limit is sensi-ble practice. And, if alloy selection uses a ruleof avoiding situations in which pitting canoccur, SCC will also be avoided even if stressexcursions occur since in general pitting is anecessary precondition for SCC.

Erosion

Flow velocities can reach levels at which ero-sion becomes problematic, especially if hardparticles are suspended in a fluid. Assuming thatthe material has sufficient corrosion resistanceto survive well in the static environment, thebest performance under erosive conditions isobtained by materials with higher surface hard-ness. Accordingly, the duplex perform betterthan austenitic alloys of the same corrosion re-sistance level.

Specific Environments

The list of specific environments againstwhich stainless steels are sufficiently resistantto select for use in the chemical process indus-tries is too long to provide here. Some of themost important specific corrosives, such asnitric, sulfuric, phosphoric, hydrochloric, andorganic acids and others, are covered in thechapter on corrosion. The main caution to thedesigner is to make sure that the source mate-rial from which design guidance is sought iscurrent. Many otherwise excellent handbooksare somewhat obsolete in that they do notinclude the very importance duplex stainlesssteel family or only include the oldest alloys inthe group, such as 2205 (UNS S32205). Manynew alloys now exist that range in corrosionperformance from that of 316 to that of the6Mo-plus-N austenitics. These alloys are us-able in all gauges, have high strength andtoughness, resist SCC, and can achieve thecorrosion resistance levels of any ferritic oraustenitic alloy. They can also provide signifi-cant savings in alloy cost at the same corrosionlevel because they have lower nickel levels.

REFERENCES

1. Sandvik Materials Technology, www.smt.sandvik.com, accessed June 2008

Fig. 2 Stress corrosion cracking (SCC) resistance in neutralchloride solutions containing 8 ppm oxygen. Testing

time, 1000 h. Applied stress equal to proof strength at testingtemperature


2. Outokumpu Corrosion Handbook for Stain-less Steels, www.outokumpu.com, accessedJune 2008

3. P.A. Schweitzer, Corrosion Resistance Tables, 5th ed., National Association of Cor-rosion Engineers, NACE 37755, 2004 \aq2\

4. D.B. Anderson and B.D. Craig, Handbookof Corrosion Data, 2nd ed., ASM Interna-tional, 1995

CHAPTER 24

Pulp-and-Paper Industry Applications

Summary

THE PULP-AND-PAPER INDUSTRY hasseen more benefits from developments in stain-less steel than any other industry. The harshchemicals used in this industry called for bettermaterials than the normal austenitic stainlesssteels without the expense of the 6Mo grades.This need has been met through the use of theduplex alloys, which have become the newstandard.

Introduction

The proximity of the Scandinavian paper in-dustry to that region’s specialty steel industryhas been symbiotic. As a result of the strong interaction between engineers having well-specified needs for improved materials and met-allurgists capable of providing them, the ad-vances in materials in the pulp-and-paperindustry have been a model of rapid technologytransfer and innovation. Beginning in 1988, du-plex stainless steels first went into production inkraft digesters, and there has been no turningback in the replacement of austenitic stainlesssteels by duplex. So, a discussion of the materi-als selection for the pulp-and-paper industry haschanged from a fairly complicated analysis ofwhich austenitic steel to use while guardingagainst stress corrosion cracking and pittingcorrosion and when to use clad materials forcost savings, to a fairly simple discussion ofwhich duplex stainless steel is most economicalfor a given piece of equipment. Since this revo-lution occurred in the 1990s before the latestsurge in nickel prices, it is safe to say that futurepulp-and-paper projects will be essentially the

exclusive domain of duplex stainless steels be-cause of their lower cost per unit of corrosionresistance, high strength, and near immunity tostress corrosion cracking. Pricing changesmainly with alloying element costs, principallythose of nickel and molybdenum. At prices be-tween the highs and lows of the first decade ofthe 2000s, duplex costs have been roughly one-third less than that of an equivalent corrosion-resisting austenitic stainless. This factored inwith strength nearly double that of the equiva-lent austenitic make them an overwhelminglysuperior choice for pulp-and-paper equivalentexcept if very special corrosion requirementsdiffer from the norm, such as in bleaching.

Paper-Making Processes

The kraft (German for “strong”) process wasintroduced in 1937, replacing the sulfite process.In the kraft process, the lignin-connecting woodfibers are dissolved under conditions of elevatedtemperature and pressure in acidic conditions ofpH 2.0 to 4.0. This leaves a long fiber, which en-ables paper of high strength, hence the namekraft. Over the years the materials used for thevessels, called digesters, in which this process iscarried out have been sequentially carbon steel,stainless steel, and stainless steel clad onto car-bon steel. In the previous sulfite process, acid-resistant brick vessels were used. Now, the di-gesters, essentially large vertical tanks, areconstructed of 2205 (UNS S32205) as a rule (seeFig. 1).

The digestion is typically carried out at 150and 180 °C (300 and 360 °C) and 10 to 12 bar.The pH of the sulfate is around 2.0 to 4.0. Inthis environment, 316L can survive, but it




requires maintenance and has a finite life. The2205 is twice as resistant to corrosion, 0.005mm/yr versus 0.011 mm/yr (Ref 1). In thenonchloride environment, molybdenum is notan essential alloying element, so the introduc-tion of the use of 2101 (UNS S32101) or 2304(UNS S32304) is a logical cost-saving movewithout strength or corrosion compromises. Thereduction in wall thickness allowed by thehigher-strength duplex depends on the engi-neering code required. The American Society ofMechanical Engineers (ASME) code require-ment is based on tensile strength and permitsonly a 24% reduction in wall thickness, whilethe total kjeldahl nitrogen (TKN) code, basedon yield strength, would allow a 46% reduction.This large a difference in strength levels re-quired by codes is unfortunate and reflects anorientation to materials in which the yield/tensile ratio is closer to unity, unlike either du-plex or austenitic stainless steel. In the more un-usual case of digesters using the sulfite process,the materials selected would be the same.

As one proceeds downstream in the process,environments change greatly, but the optimalmaterials remain duplex for various reasons.The subsequent stage is blow tanks in which the

pulp suspension is injected at high velocity. Theenvironment is a mixture of alkaline liquid,while the vapor phase can contain organic acids.The hardness of the duplex helps mitigate ero-sion, while the alloy level is beneficial againstcorrosion. 2205 is the alloy of choice here, but2003 (UNS S32003) would suffice.

The next step, washing and screening, hasseen increasingly severe environments as closedsystems required for pollution control have be-come more common. This has rendered the pre-vious choice of carbon steel untenable. Thisstage also sees erosion potential from hard par-ticles, such as sand, in the pulp. The optimumsolution is a lean duplex such as 2101, 2304, or2003.

The delignification of the pulp comes next.This oxygen process dates from the 1970s. Atfirst, highly alloyed austenitic alloys wereused. Subsequently, it was found again that du-plex performed better in that they were suffi-ciently corrosion resistant, but also offeredfreedom from stress corrosion cracking as wellas materials savings because of their higherstrength.

The bleaching of the pulp is important formany types of paper, and this can be done bychlorine bleaches or ozone/peroxide bleaches.The chlorine bleaching now must generally bedone in closed systems, which results in abuildup of chloride levels to a point at whichcorrosion levels are unacceptable unless veryhighly alloyed materials are used. The 6Mogrades have been successful, but now they canbe replaced by duplex alloys such as 2507(UNS S32750), which again save cost by virtueof their higher strength.

Bleaching can be accomplished without chlo-rine in the so-called TCF, totally chlorine free,process. This reduces the corrosivity of the en-vironment as the ozone and hydrogen peroxideused in the process are relatively harmless tostainless steel. Alloys such as 316 are adequatefor this environment, but lean duplex, 2101 or2304, offer cost reductions through their greaterstrength.

In plants that use recycled paper and mechan-ical wood chip processing, the materials selec-tion criteria remain the same. Duplex stainlesshas become the clear choice.

Further downstream, containers and processequipment benefit equally from duplex down tothe handrails and walkways. This wholesale useof duplex can make plants nearly maintenance

Fig. 1 The first kraft digester fabricated from alloy 2205. Courtesy of Outokumpu

Chapter 24: Pulp-and-Paper Industry Applications / 267

free from a corrosion point of view, a dramaticchange in an industry in which the thousand-foldgreater corrosion rates of carbon steel presentedoperators with endless equipment downtimeproblems.

Additional detail about corrosion challengesand the use of stainless steels in the pulp-and-paper industry can be found in Ref 2.

REFERENCES

1. A. Tuomi et al., Duplex America 2000 Con-ference, Houston, KCI Publishing, 2000

2. H. Dykstra et al, Corrosion in the Pulp andPaper Industry, Corrosion: Environmentsand Industries, Vol 13C, ASM Handbook,ASM International, 2006, p 762–802

APPENDIX 1

Compositions



270 / Stainless Steels for Design EngineersTa

ble

A1.

1C

ompo

siti

on o

f aus

teni

tic

stai

nles

s st

eels

Nam

eD

esig

nati

on(a

)

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

SO

ther

Oth

er

201

S201

000.

150.

2516

.0–1

8.0

3.5–

5.5

5.5–

7.5

. . .

1.00

0.06

00.

030

. . .

. . .

201L

S201

030.

030.

2516

.0–1

8.0

3.5–

5.5

5.5–

7.5

. . .

0.75

0.04

50.

015

. . .

. . .

201L

NS2

0153

0.03

0.25

16.0

–18.

03.

5–5.

55.

5–7.

5. .

.0.

750.

045

0.01

5. .

.. .

.G

all-

Toug

hS2

0161

0.15

0.08

–0.2

015

.0–1

8.0

4.0–

6.0

4.0–

6.0

. . .

3.00

–4.0

00.

040

0.04

0. .

.. .

.20

2S2

0200

0.15

0.25

17.0

–19.

04.

0–6.

07.

5–10

.0. .

.1.

000.

060

0.03

0. .

.. .

.23

0 E

ZS2

0300

0.08

. . .

16.0

–18.

05.

0–6.

55.

0–6.

50.

51.

000.

040

0.18

–0.3

5C

u 1.

75–2

.25

. . .

205

S205

000.

12–0

.25

0.32

–0.4

016

.5–1

8.0

1.0–

1.75

14.0

–15.

5. .

.1.

000.

060

0.03

0. .

.. .

.N

itron

ic 3

0S2

0400

0.0–

30.

15–0

.30

15.0

–17.

01.

5–3.

07.

0–9.

0. .

.1.

00.

040

0.03

0. .

.. .

.N

itron

ic 3

2S2

4100

0.08

0.20

–0.4

016

.5–1

9.0

0.5–

2.5

11.0

–14.

0. .

.1.

000.

060

0.03

0. .

.. .

.N

itron

ic 3

3S2

4300

0.08

0.20

–0.4

017

.0–1

9.0

2.25

–3.7

511

.5–1

4.5

. . .

1.00

0.06

00.

030

. . .

. . .

Nitr

onic

40

(219

)S2

1900

0.08

0.15

–0.4

019

.0–2

1.5

5.5–

7.5

8.0–

10.0

. . .

1.00

0.06

00.

030

. . .

. . .

21-6

-9 L

CS2

1904

0.04

0.15

–0.4

019

.0–2

1.5

5.5–

7.5

8.0–

10.0

. . .

1.00

0.06

00.

030

. . .

. . .

Nitr

onic

50

S209

100.

060.

20–0

.40

20.5

–23.

511

.5–1

3.5

4.0–

6.0

1.5–

3.0

1.00

0.04

00.

030

Nb

0.1–

0.3

V0.

1–0.

3N

itron

ic 6

0S2

1800

0.10

0.08

–0.1

816

.0–1

8.0

8.0–

9.0

7.0–

9.0

. . .

3.50

–4.5

00.

040

0.03

0. .

.. .

.Te

nelo

nS2

1400

0.12

0.35

17.0

–18.

50.

7514

.5–1

6.0

. . .

0.30

–1.0

00.

060

0.03

0. .

.. .

.C

ryog

enic

Tene

lon

S214

600.

120.

25–0

.50

17.0

–19.

05.

0–6.

014

.0–1

6.0

. . .

1.00

0.06

00.

030

. . .

. . .

Ess

hete

1250

S215

000.

15. .

.14

.0–1

6.0

9.0–

11.0

5.5–

7.0

. . .

1.20

0.04

00.

030

Nb

0.75

–1.2

5V

0.15

–0.4

0 21

6S2

1600

0.08

0.25

–0.5

017

.5–2

2.0

5.0–

7.0

7.5–

9.0

2.0–

3.0

1.00

0.04

50.

030

. . .

. . .

216L

S216

030.

030.

25–0

.50

17.5

–22.

07.

5–9.

07.

5–9.

02.

0–3.

01.

000.

045

0.03

0. .

.. .

.30

1S3

0100

0.15

16.0

–18.

06.

0–8.

02.

0. .

.1.

000.

045

0.03

0. .

.. .

.30

1LN

S301

530.

030.

10–0

.20

16.5

–18.

06.

0–8.

0. .

.. .

.0.

045

0.03

0. .

.. .

.30

2S3

0200

0.15

. . .

17.0

–19.

08.

0–10

.02.

0. .

.1.

000.

045

0.03

0. .

.. .

.30

2Cu

S304

300.

08. .

.17

.0–1

9.0

8.0–

10.0

2.0

. . .

1.00

0.04

50.

030

Cu

3.0–

4.0

. . .

302B

S302

150.

15. .

.17

.0–1

9.0

8.0–

10.0

2.0

. . .

2.00

–3.0

00.

045

0.03

0. .

.. .

.30

3S3

0300

0.15

. . .

17.0

–19.

08.

0–10

.02.

00.

60 o

ptio

nal

1.00

0.04

50.

15 m

in. .

.. .

.30

3Se

S302

230.

15. .

.17

.0–1

9.0

8.0–

10.0

2.0

0.60

opt

iona

l1.

000.

045

0.06

min

Se 0

.15

min

. . .

303

Plus

XS3

0310

0.15

. . .

17.0

–19.

07.

0–10

.02.

5–4.

50.

60 o

ptio

nal

1.00

0.04

50.

25 m

in. .

.. .

.30

4S3

0400

0.08

0.10

18.0

–20.

08.

0–10

.52.

0. .

.1.

000.

045

0.03

0. .

.. .

.30

4LS3

0403

0.03

0.10

18.0

–20.

08.

0–10

.52.

0. .

.1.

000.

045

0.03

0. .

.. .

.30

4HS3

0409

0.04

–0.1

0. .

.18

.0–2

0.0

8.0–

10.5

2.0

. . .

1.00

0.04

50.

030

. . .

. . .

304N

S304

510.

080.

10–0

.16

18.0

–20.

08.

0–10

.52.

0. .

.1.

000.

045

0.03

0. .

.. .

.30

4HN

S304

520.

080.

16–0

.30

18.0

–20.

08.

0–10

.52.

0. .

.1.

000.

045

0.03

0. .

.. .

.30

4LN

S304

530.

030.

10–0

.16

18.0

–20.

08.

0–10

.52.

0. .

.1.

000.

045

0.03

0. .

.. .

.30

4BI

S304

240.

080.

1018

.0–2

0.0

12.0

–15.

0. .

.2.

00.

75. .

.. .

.B

1.0

0–1.

20

. . .

153M

AS3

0415

0.04

–0.0

60.

12–0

.18

18.0

–19.

09.

0–10

.00.

8. .

.1.

00–2

.00

. . .

. . .

Ce

0.04

. .

.30

5S3

0500

0.12

. . .

17.0

–19.

010

.5–1

3.0

2.0

. . .

1.00

. . .

. . .

. . .

. . .

Cro

nife

r 18

15S3

0600

0.01

8. .

.17

.0–1

8.5

14.0

–15.

52.

00.

23.

75–4

.25

. . .

. . .

Cu

0.50

. .

.R

A85

HS3

0615

0.16

–0.2

4. .

.17

.0–1

9.5

13.5

–16.

02.

0. .

.3.

2–4.

00.

045

0.03

0A

l 0.8

–1.5

. .

.30

8S3

0800

0.08

. . .

19.9

–21.

010

.0–1

2.0

2.0

. . .

1.00

0.04

50.

030

. . .

. . .

253M

AS3

0815

0.05

–0.1

00.

14–0

.20

20.0

–22.

010

.0–1

2.0

0.8

. . .

1.4–

2.0

. . .

. . .

1.0

Al

0.03

–0.0

8 C

e30

9S3

0900

0.20

. . .

22.0

–24.

012

.0–1

5.0

2.0

. . .

0.75

0.04

50.

030

. . .

. . .

309S

S309

080.

08. .

.22

.0–2

4.0

12.0

–15.

02.

0. .

.0.

750.

045

0.03

0. .

.. .

.(c

ontin

ued)

Not

es: A

ll co

mpo

sitio

ns in

clud

e Fe

as

bala

nce.

Sin

gle

valu

es a

re m

axim

um v

alue

s un

less

oth

erw

ise

note

d. (

a) U

nifie

d N

umbe

r Sy

stem

, UN

S nu

mbe

rs a

re S

or

N f

ollo

wed

by

5 di

gits

.

Appendix 1: Compositions / 271

Tabl

e A

1.1

(con

tinu

ed)

Nam

eD

esig

nati

on(a

)

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

SO

ther

Oth

er

309H

S309

090.

04–0

.10

. . .

22.0

–24.

012

.0–1

5.0

2.0

. . .

0.75

0.04

50.

030

. . .

. . .

309C

bS3

0940

0.08

. . .

22.0

–24.

012

.0–1

6.0

2.0

. . .

1.00

0.04

00.

030

Nb

10xC

to 1

.10

. . .

309H

Cb

S309

410.

04–0

.10

. . .

22.0

–24.

012

.0–1

6.0

2.0

. . .

1.00

0.04

00.

030

Nb

10xC

to 1

.10

. . .

309S

iD

IN 1

.482

80.

04–0

.10

0.11

19.0

–21.

011

.0–1

3.0

2.0

. . .

1.50

–2.5

00.

040

0.01

5. .

.. .

.31

0S3

1000

0.25

. . .

24.0

–26.

019

.0–2

2.0

2.0

. . .

1.00

0.04

50.

030

. . .

. . .

310S

S310

080.

08. .

.24

.0–2

6.0

19.0

–22.

02.

0. .

.1.

000.

045

0.03

0. .

.. .

.31

0HS3

1009

0.04

–0.1

0. .

.24

.0–2

6.0

19.0

–22.

02.

0. .

.1.

000.

045

0.03

0. .

.. .

.31

0Cb

S310

400.

08. .

.24

.0–2

6.0

19.0

–22.

02.

0. .

.1.

000.

045

0.03

0N

b 10

xC to

1.1

0 . .

.31

0HC

bS3

1041

0.04

–0.1

0. .

. 24

.0–2

6.0

19.0

–22.

02.

0. .

.1.

000.

045

0.03

0N

b 10

xC to

1.1

0. .

.31

0HC

bNS3

1042

0.04

–0.1

00.

15–0

.35

24.0

–26.

019

.0–2

2.0

2.0

. . .

1.00

0.04

50.

030

Nb

10xC

to 1

.10

. . .

310M

oLN

S310

500.

020.

09–0

.15

24.0

–26.

620

.5–2

3.5

2.0

1.6–

2.6

0.5

0.02

00.

010

. . .

. . .

310S

iD

IN 1

.484

10.

200.

1124

.0–2

6.0

19.0

–22.

02.

0. .

.1.

50–2

.50

0.04

50.

015

. . .

. . .

314

S314

000.

25. .

.23

.0–2

6.0

19.0

–22.

02.

0. .

.1.

50–3

.00

0.04

50.

030

. . .

. . .

316

S316

000.

080.

1016

.0–1

8.0

10.0

–14.

02.

02.

0–3.

01.

000.

045

0.03

0. .

.. .

.31

6HS3

1609

0.04

–0.1

0. .

.16

.0–1

8.0

10.0

–14.

02.

02.

0–3.

01.

000.

045

0.03

0. .

.. .

.31

6FS3

1620

0.08

. . .

16.0

–18.

010

.0–1

4.0

2.0

1.75

–2.5

1.00

0.20

0.10

min

. . .

. . .

316L

S316

030.

030.

1016

.0–1

8.0

10.0

–14.

02.

02.

0–3.

01.

000.

045

0.03

0. .

.. .

.31

6LN

S316

530.

030.

10–0

.16

16.0

–18.

010

.0–1

4.0

2.0

2.0–

3.0

1.00

0.04

50.

030

. . .

. . .

316N

S316

510.

080.

10–0

.16

16.0

–18.

010

.0–1

4.0

2.0

2.0–

3.0

1.00

0.04

50.

030

. . .

. . .

316T

iS3

1635

0.08

0.10

16.0

–18.

010

.0–1

4.0

2.0

2.0–

3.0

1.00

0.04

50.

030

5xC

to 0

.70

Ti

. . .

317

S317

000.

080.

1018

.0–2

0.0

11.0

–15.

03.

0–4.

02.

01.

000.

045

0.03

0. .

.. .

.31

7LS3

1703

0.03

0.10

18.0

–20.

011

.0–1

5.0

3.0–

4.0

2.0

1.00

0.04

50.

030

. . .

. . .

317L

NS3

1753

0.03

0.10

–0.2

218

.0–2

1.0

11.0

–15.

03.

0–4.

02.

01.

000.

045

0.03

00.

030

max

P. .

.31

7LM

S317

250.

030.

1018

.0–2

0.0

13.5

–17.

54.

0–5.

02.

01.

000.

045

0.03

0. .

.. .

.31

7LM

NS3

1726

0.03

0.10

–0.2

017

.0–2

0.0

13.5

–17.

54.

0–5.

02.

00.

750.

045

0.03

0. .

.. .

.32

1S3

2100

0.08

0.10

17.0

–19.

09.

0–12

.02.

0. .

.1.

000.

045

0.03

0T

i 5x(

C+

N)

to 0

.70

. . .

321H

S321

090.

40–0

.10

0.10

17.0

–19.

09.

0–12

.02.

0. .

.1.

000.

045

0.03

0T

i 4x(

C+

N)

to 0

.70

. . .

330

S330

000.

08. .

.17

.0–2

0.0

34.0

–37.

02.

0. .

.0.

75–1

.50

0.03

00.

030

. . .

. . .

332

N08

800

0.03

. . .

19.0

–23.

030

.0–3

5.0

1.5

. . .

1.00

. . .

. . .

Ti 0

.15–

0.60

A

l 0.1

5–0.

60

332M

o*S3

5125

0.10

0.10

20.0

–23.

031

.0–3

5.0

1.0–

1.5

2.0–

3.0

0.75

. . .

. . .

Nb

0.25

–0.6

0. .

.33

4S3

3400

0.08

. . .

18.0

–20.

019

.0–2

1.0

1.0

. . .

1.00

. . .

. . .

Ti 0

.15–

0.60

A

l 0.1

5–0.

60

347

S347

000.

08. .

.17

.0–1

9.0

9.0–

13.0

2.0

. . .

1.00

0.04

50.

030

Nb+

Ta 1

0xC

to

1.1

0 . .

.

347H

S347

090.

08. .

.17

.0–1

9.0

9.0–

13.0

2.0

. . .

1.00

0.04

50.

030

Nb+

Ta 1

0xC

to

1.1

0 . .

.

348

S348

000.

08. .

.17

.0–1

9.0

9.0–

13.0

2.0

. . .

1.00

0.04

50.

030

Nb+

Ta 0

xCto

1.1

0 C

o 0.

2

348H

S348

090.

04–0

.10

. . .

17.0

–19.

09.

0–13

.02.

0. .

.1.

000.

045

0.03

0N

b+Ta

8xC

to 1

.0. .

.

370

S370

000.

03–0

.05

0.00

512

.5–1

4.5

14.5

–16.

51.

65–2

.35

1.5–

2.5

0.5–

1.0

. . .

. . .

Ti 0

.10–

0.40

C

o 0.

0538

4S3

8400

0.08

. . .

15.0

–17.

017

.0–1

9.0

2.0

. . .

1.00

0.04

50.

030

Nb+

Ta 1

0xC

to 1

.10

Co

0.2

353M

AS3

5315

0.08

0.12

–0.1

824

.0–2

6.9

34.0

–36.

01.

0. .

.0.

6–1.

0. .

.. .

.C

e 0.

03–0

.10

. . .

(con

tinue

d)N

otes

: All

com

posi

tions

incl

ude

Fe a

s ba

lanc

e. S

ingl

e va

lues

are

max

imum

val

ues

unle

ss o

ther

wis

e no

ted.

(a)

Uni

fied

Num

ber

Syst

em, U

NS

num

bers

are

S o

r N

fol

low

ed b

y 5

digi

ts.


Tabl

e A

1.1

(con

tinu

ed)

Nam

eD

esig

nati

on(a

)

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

SO

ther

Oth

er

AC

66S3

3228

0.04

–0.0

8. .

.26

.0–2

8.0

31.0

–33.

01.

0. .

.0.

03. .

.. .

.0.

05–0

.10

Ce

Al 0

.025

In

colo

y 80

3S2

5045

0.06

–0.1

0. .

.25

.0–2

9.0

32.0

–37.

01.

5. .

.1.

0. .

.0.

015

Ti 0

.15–

0.60

0.15

–0.6

0 A

lIn

colo

y 86

4S3

5135

0.08

. . .

20.0

–25.

030

.0–3

8.0

1.0

4.0–

4.8

0.6–

1.0

. . .

0.01

5T

i 0.4

–1.0

. . .

21-4

NS6

3008

0.48

–0.5

80.

28–0

.50

20.0

–22.

03.

25–4

.50

8.0–

10.0

. . .

0.25

0.04

50.

015

. . .

. . .

21-2

NS6

3012

0.50

–0.6

00.

20–0

.40

19.2

5–21

.50

1.50

–2.7

57.

0–9.

50. .

.0.

250.

050

0.03

0. .

.. .

.21

-12N

S630

170.

15–0

.25

0.15

–0.2

520

.0–2

2.0

10.5

–12.

51.

0–1.

5. .

.0.

70–1

.25

0.04

50.

030

. . .

. . .

23-8

NS6

3018

0.28

–0.3

80.

28–0

.38

22.0

–24.

07.

9–9.

01.

5–3.

5. .

.0.

69–0

.90

0.04

50.

030

. . .

. . .

19-9

DL

S631

980.

28–0

.36

. . .

18.0

–21.

08.

0–11

.00.

75–1

.50

1.0–

1.75

0.03

–0.8

0. .

.. .

.0.

1–0.

35 T

i0.

25–0

.60

Nb

20C

b-3

N08

020

0.07

. . .

19.0

–21.

032

.0–3

8.0

0.75

–1.5

02.

0–3.

01.

000.

045

0.03

58x

C to

1.0

0 N

bC

u 3.

0–4.

020

Mo-

4N

0802

40.

03. .

.22

.5–2

5.0

35.0

–40.

01.

03.

5–5.

00.

500.

0035

0.03

5N

b 0.

15-0

.35

Cu

0.5–

1.5

20M

o-6

N08

026

0.03

. . .

22.0

–26.

033

.0–3

7.0

1.0

5.0–

6.7

0.50

0.03

00.

030

8xC

Nb

2.0–

4.0

Cu

Sani

cro

28N

0802

80.

02. .

.26

.0–2

8.0

29.9

–32.

52.

53.

0–4.

01.

000.

030

0.03

0. .

.0.

6–1.

4 C

uA

L-6

XN

0836

60.

035

. . .

20.0

–22.

023

.5–2

5.5

2.0

6.0–

7.0

1.00

0.03

00.

030

0.03

0 P

. . .

AL

-6X

NN

0836

70.

030

0.18

–0.2

520

.0–2

2.0

23.5

–25.

52.

06.

0–7.

01.

000.

030

0.03

00.

040

P. .

.JS

-700

N08

700

0.04

. . .

19.0

–23.

024

.0–2

6.0

2.0

4.3–

5.0

1.00

0.04

00.

030

8xC

to 0

.5 N

b0.

5 C

u90

4LN

8090

40.

02. .

.19

.0–2

3.0

23.0

–28.

02.

04.

0–5.

01.

000.

045

0.03

51.

0–2.

0 C

u. .

.25

4SM

OS3

1254

0.02

0.18

–0.2

219

.50–

20.5

017

.50–

18.5

01.

06.

0–6.

50.

800.

030

0.01

00.

5–1.

0 C

u. .

.45

65S3

4565

0.03

0.40

–0.6

023

.0–2

5.0

16.0

–18.

05.

0–7.

04.

0–5.

01.

0. .

.. .

.0.

10 N

b. .

.65

4SM

OS3

2654

0.02

0.45

–0.5

524

.0–2

5.0

21.0

–23.

02.

0–3.

07.

0–8.

00.

50.

030

0.00

50.

3–0.

6 C

u. .

.A

L20

N08

020

0.07

. . .

19.0

–21.

032

.0–3

8.0

2.0

2.0–

3.0

1.0

. . .

. . .

3.0–

4.0

Cu

8xC

to 1

.00

Nb+

TaA

L22

N06

022

0.01

520

.0–2

2.5

bala

nce

. . .

12.5

–14.

50.

08. .

.. .

.2.

5 C

o,0.

35 V

W2.

5-3.

5 C

roni

fer

1925

hM

oN

0892

60.

020.

15–0

.25

19.0

–21.

024

.0–2

6.0

2.0

6.0–

7.0

0.50

. . .

. . .

0.5–

1.5

Cu

. . .

Cro

nife

r 23

28. .

.0.

04. .

.22

.0–2

4.0

26.0

–28.

00.

752.

5–3.

00.

750.

045

0.03

52.

5–3.

5 C

uT

i 0.4

-0.7

N

icro

fer

3127

HM

oN

0803

10.

015

0.15

–0.2

526

.0–2

8.0

24.0

–26.

02.

06.

0–7.

00.

30. .

.. .

.1.

0–1.

4 C

u. .

.

UR

SB-8

N08

932

0.02

0.15

–0.2

524

.0–2

6.0

24.0

–26.

02.

04.

5–6.

50.

40. .

.. .

.1.

0–2.

0 C

u. .

.B

66S3

1266

0.03

0.35

–0.6

023

.0–2

6.0

21.0

–24.

02.

05.

0–7.

01.

0. .

.. .

.0.

5–3.

0 C

uW

1.0-

3.0

NIC

25

S322

000.

03. .

.20

.0–2

3.0

23.0

–27.

01.

02.

5–3.

50.

5. .

.. .

.. .

.C

N-7

MN

0800

70.

07. .

.19

.0–2

2.0

27.5

–30.

51.

52.

0–3.

01.

5. .

.. .

.3.

0–4.

0 C

u. .

.N

0832

020

Mod

0.05

. . .

21.0

–23.

025

.0–2

7.0

2.5

4.0–

6.0

1.0

. . .

. . .

. . .

. . .

Not

es: A

ll co

mpo

sitio

ns in

clud

e Fe

as

bala

nce.

Sin

gle

valu

es a

re m

axim

um v

alue

s un

less

oth

erw

ise

note

d. (

a) U

nifie

d N

umbe

r Sy

stem

, UN

S nu

mbe

rs a

re S

or

N f

ollo

wed

by

5 di

gits

.


Tabl

e A

1.2

Com

posi

tion

of f

erri

te s

tain

less

ste

els

Nam

eU

NS

desi

gnat

ion

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

ST

iN

bO

ther

405

S405

000.

08. .

.11

.5–1

4.5

0.60

1.00

. . .

1.00

0.04

00.

030

. . .

. . .

Al 0

.10–

0.30

40

0A

K a

lloy

0.05

. . .

12.0

–13.

0. .

.1.

00. .

.1.

00. .

.. .

.. .

.. .

.A

l 0.2

5 A

l40

9S4

0900

0.08

. . .

10.5

–11.

750.

501.

00. .

.1.

000.

045

0.04

56x

(C+

N)

to 0

.75

. . .

. . .

409

S409

100.

030.

030

10.5

–11.

70.

501.

00. .

.1.

00.

040

0.03

06x

(C+

N)t

o 0.

50.

17

. . .

409

S409

200.

030.

030

10.5

–11.

750.

501.

00. .

.1.

000.

040

0.03

08x

(C+

N)

to 0

.15–

0.50

. . .

. . .

409

ultr

afor

mA

K a

lloy

0.02

0.02

010

.5–1

1.7

0.50

0.75

. . .

1.00

. . .

. . .

8x(C

+N

). .

.. .

.46

6S4

0930

0.02

0.02

010

.5–1

1.75

0.05

1.00

. . .

1.00

0.04

00.

030

0.8+

8x(

C+

N)

Ti+

Nb

. . .

. . .

409C

bS4

0940

0.06

. . .

10.5

–11.

70.

501.

00. .

.1.

0. .

.. .

.10

xC

to 0

.75

Nb

. . .

. . .

409N

iS4

0975

0.03

0.03

10.5

–11.

70.

5–1.

01.

00. .

.1.

00.

040

0.03

0. .

.. .

.. .

.11

Cr–

Cb

AK

allo

yty

pica

l0.

010

0.01

511

.35

0.20

0.25

. . .

1.30

. . .

. . .

. . .

0.35

. . .

12 S

RA

K a

lloy

typi

cal

0.02

00.

015

12.0

. . .

. . .

. . .

. . .

. . .

. . .

0.30

0.60

AL

1.2

Alf

a I

AT

I al

loy

typi

cal

0.02

5. .

.13

.0. .

.0.

035

. . .

0.03

. . .

. . .

0.40

. . .

AL

3.0

Alf

a II

AT

I al

loy

typi

cal

0.02

5. .

.13

.0. .

.0.

035

. . .

0.03

. . .

. . .

0.40

. . .

AL

4.0

4724

Out

ukum

puty

pica

l0.

08. .

.13

.5. .

.0.

70. .

.1.

0. .

.. .

.. .

.. .

.A

l 1.0

429

S429

000.

12. .

.14

.0–1

6.0

0.75

1.00

. . .

1.00

0.04

00.

030

. . .

. . .

. . .

430

S430

000.

12. .

.16

.0–1

8.0

0.75

1.00

. . .

1.00

0.04

00.

030

. . .

. . .

. . .

430F

S430

200.

12. .

.16

.0–1

8.0

. . .

1.25

. . .

1.00

0.06

00.

150

min

. . .

. . .

. . .

430S

eS4

3023

0.12

. . .

16.0

–18.

0. .

.1.

25. .

.1.

000.

060

0.06

0. .

.. .

.SE

0.1

543

0Ti

S430

360.

100.

0416

.0–1

9.5

1.00

1.00

. . .

1.00

0.04

00.

030

0.20

+4x

(C+

N)

to 1

.10

. . .

AL

0.15

439

S430

350.

070.

0417

.0–1

9.0

0.50

1.00

. . .

1.00

0.04

00.

030

0.20

+4x

(C+

N)

to 1

.10

. . .

. . .

439L

TS4

3932

0.03

0.03

17.0

–19.

00.

501.

00. .

.1.

00.

040

0.03

00.

20+

4x(C

+N

)to

0.7

5 T

i+N

b

. . .

Al 0

.15

439

HP

439

ultr

afor

mA

TI,

AK

allo

ys0.

010.

0117

.50.

20.

35. .

.0.

45. .

.. .

.0.

35. .

.. .

.

468

S468

000.

030

. . .

18.0

–20.

00.

501.

00. .

.1.

000.

040

0.03

0T

i+N

b:0.

20+

4x(C

+N

)to

1.1

0

. . .

. . .

18 C

r–C

bA

K a

lloy

typi

cal

0.02

0. .

.18

.0. .

.0.

30. .

.0.

45. .

.. .

.0.

250.

55. .

.

(con

tinue

d)

Not

es: A

ll co

mpo

sitio

ns in

clud

e Fe

as

bala

nce.

Sin

gle

valu

es a

re m

axim

um v

alue

s un

less

oth

erw

ise

note

d


Tabl

e A

1.2

Com

posi

tion

of f

erri

te s

tain

less

ste

els

Nam

eU

NS

desi

gnat

ion

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

ST

iN

bO

ther

18SR

AK

allo

yty

pica

l0.

015

. . ..

17.3

00.

250.

30. .

... .

... .

... .

.0.

25. .

.A

l 1.7

4742

Out

ukum

puty

pica

l0.

08. .

.18

.0. .

.0.

7. .

.1.

3. .

.. .

.. .

.. .

.A

l 1.0

434

S434

000.

12. .

.16

.0–1

8.0

. . .

1.0

0.75

–1.2

51.

00.

040

0.03

0. .

.. .

.. .

.43

6S4

3600

0.12

. . .

16.0

–18.

8. .

.1.

00.

75–1

.25

1.0

0.04

00.

030

. . .

Nb+

Ta5x

C:0

.70

. . .

441, 45

09,

430J

1L

S441

000.

030

. . .

17.5

–18.

5. .

.1.

00. .

.1.

00.

040

0.03

00.

1–0.

69x

C0.

3–1.

0. .

.

442

S442

000.

20. .

.18

.0–2

3.0

0.6

1.0

. . .

1.0

0.04

00.

030

. . .

. . .

. . .

436S

AT

I al

loy

typi

cal

0.01

0.01

517

.30.

30.

201.

20.

4. .

.. .

.8x

(C+

N)

min

. . .

. . .

444,

YU

S 19

0-E

MS4

4400

0.02

50.

035

17.5

–19.

51.

01.

00.

75–1

.25

1.0

0.04

00.

030

Ti+

Nb:

0.20

+4x

(C+

N)

to 0

.80

. . .

. . .

433

AT

I al

loy

typi

cal

0.01

. . .

20.0

0.25

0.30

. . .

0.4

. . .

. . .

. . .

10x(

C+

N)

. . .

4762

Out

ukum

puty

pica

l0.

08. .

.24

.0. .

.0.

7. .

.1.

4. .

.. .

.. .

.. .

.A

l 1.5

453

AT

I al

loy

typi

cal

0.03

. . .

22.0

0.3

0.3

. . .

0.3

. . .

. . .

0.02

. . .

0.60

Al

0.10

RE

ME

-Bri

te, 2

6-1

S446

270.

010.

015

25.0

–27.

50.

500.

400.

75–1

.25

0.40

0.02

00.

020

. . .

0.5–

0.20

0.2

Cu

0.5

Cu+

Ni

Mon

itS4

4635

0.02

50.

035

24.5

–26.

03.

5–4.

51.

003.

5–4.

50.

750.

040

0.03

0T

i+N

b:0.

20+

4x(C

+N

)to

0.8

0

. . .

. . .

Sea–

cure

S446

600.

025

0.03

525

.0–2

7.0

1.5–

3.5

1.00

2.5–

3.5

1.00

0.04

00.

030

Ti+

Nb:

0.20

+4x

(C+

N)

to 0

.80

. . .

. . .

29-4

CS4

4735

0.02

5. .

.28

.0–3

0.0

0.5

1.00

3.5–

4.5

0.75

0.04

00.

030

Ti+

Nb:

0.20

+4x

(C+

N)

to 0

.80

. . .

. . .

446

S446

000.

200.

2523

.0–2

7.0

0.6

1.50

. . .

1.00

0.04

00.

030

. . .

. . .

. . .

Not

es: A

ll co

mpo

sitio

ns in

clud

e Fe

as

bala

nce.

Sin

gle

valu

es a

re m

axim

um v

alue

s un

less

oth

erw

ise

note

d.

Tabl

e A

1.2

(con

tinu

ed)


Tabl

e A

1.3

Com

posi

tion

of m

arte

nsit

ic s

tain

less

ste

els

Nam

eU

NS

desi

gnat

ion

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

SO

ther

403

S403

000.

15 m

ax. .

.11

.5–1

3.5

. . .

1.00

0.50

0.50

0.04

00.

030

. . .

410

S410

000.

15 m

ax. .

.11

.5–1

3.5

. . .

1.00

1.00

1.00

0.04

00.

030

. . .

410S

S410

030.

03. .

.10

.5–1

2.5

1.5

1.00

. . .

. . .

0.04

00.

030

. . .

410

S410

080.

08. .

.11

.5–1

3.5

. . .

1.50

1.00

1.00

0.04

00.

030

. . .

410C

bS4

1040

0.18

max

. . .

11.5

–13.

5. .

.1.

001.

001.

000.

040

0.03

0N

b 0.

05–0

.30

412

S410

030.

030

max

. . .

10.5

–12.

51.

51.

501.

001.

000.

040

0.03

0. .

.41

4S4

1400

0.15

max

. . .

11.5

–13.

51.

25–2

.50

1.00

1.00

1.00

0.04

00.

030

. . .

414

mod

S414

250.

050.

06 0

.12

12.0

–15.

04.

0–7.

00.

5–1.

00.

600.

600.

040

0.00

5C

u 0.

3041

5S4

1500

0.05

max

. . .

11.5

–14.

03.

50–5

.50

0.50

–1.0

0.60

0.60

0.04

00.

030

. . .

416

S416

000.

15 m

ax. .

.12

.0–1

4.0

. . .

1.25

1.00

1.00

0.06

00.

15 0

.30

. . .

416S

eS4

1623

0.15

max

. . .

12.0

–14.

0. .

.1.

251.

001.

000.

060

0.06

0Se

0.1

5 m

in41

8S4

1800

0.15

–0.2

0. .

.12

.0–1

4.0

1.80

–2.2

00.

500.

500.

500.

040

0.03

0W

2.50

–3.5

0 42

0S4

2000

0.15

min

. . .

12.0

–14.

0. .

.1.

001.

001.

000.

040

0.03

0. .

.41

16D

IN 1

.411

6no

min

al0.

50. .

.14

.5. .

.. .

.. .

.. .

.0.

040

. . .

. . .

420F

S420

200.

15 m

in. .

.12

.0–1

4.0

. . .

1.25

1.00

1.00

0.04

00.

15. .

.42

0FSe

S420

230.

15 m

in. .

.12

.0–1

4.0

. . .

1.25

1.00

1.00

0.04

00.

06Se

0.1

5 m

in42

2S4

2200

0.20

–0.2

5. .

.11

.0–1

3.5

0.50

–1.0

01.

000.

750.

750.

040

0.03

00.

75–1

.25

W42

4S4

2400

0.06

max

. . .

12.0

–14.

03.

50–4

.50

0.50

–1.0

0.30

–0.6

00.

30–0

.60.

040

0.03

0. .

.42

5S4

2500

0.08

–0.2

0. .

.14

.0–1

6.0

1.00

–2.0

01.

001.

001.

000.

040

0.01

0. .

.42

5 m

od. .

.0.

50–0

.55

. . .

13.0

–14.

00.

501.

001.

001.

000.

040

0.03

0. .

.T

rina

met

. . .

0.30

max

. . .

12.0

–14.

0. .

.1.

001.

001.

000.

040

0.03

0C

U 2

.0–3

.0H

P13C

r-1

JFE

nom

inal

0.02

5. .

.13

.04.

00.

45. .

.. .

.. .

.. .

.. .

.H

P13C

r-2

JFE

nom

inal

0.02

5. .

.13

.05.

00.

45. .

.. .

.. .

.. .

.. .

.N

T-C

RS

Nip

pon

nom

inal

0.03

0.04

0 12

.74.

51.

45. .

.. .

.. .

.. .

.1.

5 C

uN

T-C

RSS

Nip

pon

nom

inal

0.02

0.01

5 12

.35.

82.

0. .

.. .

.. .

.. .

.1.

5 C

uK

L-1

2Cr

JFE

nom

inal

0.01

0.01

0 11

.02.

4. .

.. .

.. .

.. .

.. .

.0.

5 C

uK

L-H

P12

Cr

JFE

nom

inal

0.01

0.01

0 12

.05.

5. .

.. .

.. .

.. .

.. .

.. .

.43

1S4

3100

0.20

max

. . .

15.0

–17.

01.

25–2

.50

1.00

1.00

1.00

0.04

00.

030

. . .

440A

S440

020.

60–0

.75

. . .

16.0

–18.

0. .

.1.

001.

001.

000.

040

0.03

0. .

.44

0BS4

4003

0.75

–0.9

5. .

.16

.0–1

8.0

. . .

1.00

1.00

1.00

0.04

00.

030

. . .

440C

S440

040.

95–1

.20

. . .

16.0

–18.

0. .

.1.

001.

001.

000.

040

0.03

0. .

.44

0FS4

4020

0.95

–1.2

0. .

.16

.0–1

8.0

0.75

1.25

1.00

1.00

0.04

00.

10–0

.35

. . .

440F

SeS4

4023

0.95

–1.2

0. .

.16

.0–1

8.0

0.75

1.25

1.00

1.00

0.06

00.

060

Se 0

.15

min

BG

-42

Nom

inal

PM

1.15

. . .

14.5

. . .

. . .

0.3

0.3

. . .

. . .

1.2

VA

TS-

34N

omin

al P

M1.

05. .

.14

.0. .

.0.

40.

350.

35. .

.. .

.. .

.14

-4 C

rMo

Nom

inal

PM

1.05

. . .

14.0

. . .

0.5

0.3

0.3

. . .

. . .

. . .

154

CM

Nom

inal

PM

1.05

. . .

14.0

. . .

0.45

0.3

0.3

. . .

. . .

. . .

CPM

S30

VN

omin

al P

M1.

45. .

.14

.0. .

.. .

.. .

.. .

.. .

.. .

.4.

0 V

CPM

S60

VN

omin

al P

M2.

15. .

.17

.0. .

.0.

40. .

.. .

.. .

.. .

.5.

5 V

CPM

S90

VN

omin

al P

M2.

20. .

.13

.0. .

.. .

.. .

.. .

.. .

.. .

.9.

0 V

Not

es: A

ll co

mpo

sitio

ns in

clud

e Fe

as

bala

nce.

Sin

gle

valu

es a

re m

axim

um v

alue

s un

less

oth

erw

ise

note

d

276 / Stainless Steels for Design EngineersTa

ble

A1.

4C

ompo

siti

on o

f sel

ecte

d pr

ecip

itat

ion–

hard

enab

le s

tain

less

ste

els

Nam

eD

esig

nati

on

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

SO

ther

Oth

er

Stai

nles

s W

S176

000.

08. .

.15

.0–1

7.0

6.0–

7.5

1.0

. . .

1.00

0.04

00.

030

Ti 0

.4–1

.2A

l 0.4

17-4

PH

S174

000.

07. .

.11

5.5–

17.5

3.0–

5..0

1.0

. . .

1.00

0.04

00.

030

Cu

3.0–

5.0

Nb

.015

–0.4

515

-5 P

HS1

5500

0.07

. . .

14.0

–15.

53.

5–5.

51.

0. .

.1.

000.

040

0.03

0C

u 2.

5–4.

5N

b 0.

15–0

.45

13-8

PH

S138

000.

050.

010

12.2

5–13

.25

7.5–

8.5

0.2

2.0–

2.5

0.10

0.01

00.

008

. . .

Al 0

.90–

1.35

Cus

tom

450

S450

000.

05. .

.14

.0–1

6.0

5.0–

7.0

1.0

0.5–

1.0

1.00

0.03

00.

030

Cu

1.25

–1.7

5 N

b 8X

CC

usto

m 4

55S4

5500

0.05

. . .

11.0

–12.

507.

5–9.

50.

500.

50.

500.

040

0.03

0C

u 1.

5–2.

5,T

i 0.0

8–1.

4N

b +

Ta 0

.1–0

.5

Cus

tom

465

S465

000.

02. .

.11

.0–1

2.50

10.7

5–11

.25

0.25

0.75

–1.2

50.

250.

015

0.01

0T

i 1.5

0–1.

80. .

.C

usto

m 4

75S1

7600

0.01

. . .

10.5

–11.

507.

5–8.

50.

504.

5–5.

50.

500.

015

0.01

0C

o 8.

0–9.

0A

l 1.0

–1.5

17-7

PH

S177

000.

09. .

.16

.0–1

8.0

6.5–

7.75

1.00

. . .

1.00

0.04

00.

030

. . .

Al 0

.75–

1.5

15-7

PH

S157

000.

09. .

.14

.0–1

6.0

6.5–

7.75

1.00

2.0–

3.0

1.00

0.04

00.

030

. . .

Al 0

.75–

1.5

AM

-350

S350

000.

07–0

.11

0.07

–0.1

316

.0–1

7.0

4.0–

5.0

0.50

–1.2

52.

5–3.

250.

500.

040

0.03

0. .

.. .

.A

M-3

55S3

5500

0.10

–0.1

50.

07–0

.13

15.0

–16.

04.

0–5.

00.

50–1

.25

2.5–

3.25

0.50

0.04

00.

030

. . .

. . .

A-2

86S6

6286

0.08

. . .

13.5

–16.

014

.0–2

7.0

2.00

1.0–

1.5

1.00

0.04

00.

030

Ti 1

.9–2

.35

V0.

10–.

050

Al 0

.35

B 0

.001

–0.0

10Fe

rriu

m S

53(a

). .

.0.

21. .

.10

.05.

50.

102.

00.

10. .

.. .

.C

o 14

W1,

V0.

3

Not

es: A

ll co

mpo

sitio

ns in

clud

e Fe

as

bala

nce.

Sin

gle

valu

es a

re m

axim

um v

alue

s un

less

oth

erw

ise

note

d. (

a) N

omin

al v

alue

Tabl

e A

1.5

Com

posi

tion

of s

elec

ted

dupl

ex s

tain

less

ste

els

Nam

eD

esig

nati

on

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

SC

uW

329

S329

000.

08. .

.23

.0–2

8.0

2.5–

5.0

1.0

1.0–

2.0

0.75

0.04

00.

030

0.75

. . .

44L

NS3

1200

0.03

0.14

–0.2

024

.0–2

6.0

5.5–

6.0

2.0

1.2–

2.0

1.0

0.04

50.

030

1.0

. . .

DP3

S312

600.

030.

10–0

.30

24.0

–26.

05.

5–7.

51.

02.

5–3.

50.

750.

030

0.02

00.

750.

1–0.

53R

E60

S315

000.

300.

05–0

.10

18.0

–19.

04.

25–5

.25

1.2–

2.0

2.5–

3.0

1.4–

2.0

0.03

00.

030

1.4–

2.0

. . .

2205

(ol

d)S3

1830

0.03

0.08

–0.2

021

.0–2

3.0

2.5–

3.5

2.0

2.5–

3.5

1.0

0.03

00.

020

1.0

. . .

19 D

S320

010.

030.

05–0

.17

19.5

–21.

51.

0–3.

04.

0–6.

0. .

.1.

00.

040

0.03

01.

0. .

.20

03S3

2003

0.03

0.14

–0.2

019

.5–2

1.–

3.0–

4.0

2.0

1.5–

2.0

1.0

0.04

00.

030

1.0

. . .

2101

S321

010.

040.

20–0

.25

21.0

–22.

01.

35–1

.70

4.0–

6.0

0.1–

0.8

1.0

0.04

00.

030

1.0

. . .

2205

S322

050.

030.

14–0

.20

22.0

–23.

04.

5–6.

51.

03.

0–3.

52.

00.

030

0.02

02.

0. .

.23

04S3

2304

0.03

0.05

–0.2

021

.5–2

3.5

3.0–

5.0

2.5

. . .

1.0

0.04

00.

040

1.0

. . .

Ura

nus

52N

+S3

2520

0.03

0.20

–0.3

524

.0–2

6.0

5.5–

8.0

1.5

3.0–

5.0

0.8

0.03

50.

020

0.8

. . .

255

S325

500.

040.

10–0

.25

24.0

–27.

06.

0–8.

01.

52.

9–3.

91.

00.

040

0.03

01.

0. .

.25

07S3

2750

0.03

0.20

–0.3

024

.0–2

6.0

6.0–

8.0

1.2

3.0–

5.0

0.8

0.03

50.

020

0.8

. . .

Zer

on 1

00S3

2760

0.03

0.20

–0.3

024

.0–2

6.0

6.0–

8.0

1.0

3.0–

5.0

1.0

0.03

00.

010

1.0

0.5–

1.0

2906

S329

060.

030.

30–0

.40

28.0

–30.

05.

8–7.

50.

8–1.

51.

5–2.

60.

50.

030

0.03

00.

5. .

.7-

Mo

Plus

S329

500.

030.

15–0

.35

26.0

–29.

03.

5–5.

22.

01.

0–2.

50.

60.

035

0.01

00.

6. .

.D

P3W

S392

740.

030.

24–0

.32

24.0

–26.

06.

0–8.

01.

02.

5–3.

50.

80.

030

0.02

00.

81.

5–2.

5A

F 91

8S3

9277

0.02

50.

23–0

.33

24.0

–26.

06.

5–8.

00.

83.

0–4.

00.

80.

030

0.02

00.

80.

8–1.

2

Not

es: A

ll co

mpo

sitio

ns in

clud

e Fe

as

bala

nce.

Sin

gle

valu

es a

re m

axim

um v

alue

s un

less

oth

erw

ise

note

d


Tabl

e A

1.6

(con

tinu

ed)

Nam

eW

roug

hteq

uiva

lent

(a)

UN

Sde

sign

atio

n

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

SO

ther

Cor

rosi

on–r

esis

ting

allo

ys

CA

-15

410

J911

500.

15. .

.11

.5–1

4.0

1.0

1.00

0.50

(b)

1.50

0.04

0.04

. . .

CA

-15M

. . .

J911

510.

15. .

.11

.5–1

4.0

1.0

1.00

0.15

–1.0

00.

650.

040.

04. .

.C

A-4

042

0J9

1153

0.40

. . .

11.5

–14.

01.

01.

000.

50(b

)1.

500.

040.

04. .

.C

A-4

0F42

0FJ9

1154

0.2–

0.4

. . .

11.5

–14.

01.

01.

00. .

.1.

500.

040.

040.

20–0

.40

SC

B-3

043

1,44

2J9

1803

0.30

. . .

18.0

–22.

02.

01.

00. .

.1.

500.

040.

04. .

.C

C-5

044

6J9

2613

0.30

. . .

26.0

–30.

04.

01.

00. .

.1.

500.

040.

04. .

.C

A-6

N. .

.J9

1650

0.06

. . .

10.5

–12.

56.

0–8.

00.

50. .

.1.

000.

040.

04. .

.C

A-6

NM

S415

00J9

1540

0.06

. . .

11.5

–14.

03.

5–4.

51.

000.

4–1.

01.

000.

040.

04. .

.C

A-2

8MW

V42

2J9

1422

0.20

–0.2

8. .

.11

.0–1

2.5

0.5–

1.0

0.5–

1.0

0.9–

1.25

1.00

0.04

0.04

0.9–

1.25

W, 0

.2–0

.3V

CB

-7C

u-1

17–4

PHJ9

2180

0.07

0.05

15.5

–17.

73.

6–4.

60.

70. .

.1.

000.

040.

042.

5–3.

2 C

u, 0

.2–0

.35N

bC

B-7

Cu-

215

–5PH

J9

2110

0.07

0.05

14.0

–15.

54.

5–5.

50.

70. .

.1.

000.

040.

042.

5–3.

2 C

u, 0

.2–0

.35N

bC

D-3

MN

2205

(S3

2205

)J9

2205

0.03

0.10

–0.3

021

.0–2

3.5

4.5–

6.5

1.50

2.5–

3.5

1.00

0.04

0.04

Cu

1.0

CD

-3M

CuN

255

(S32

550)

J933

730.

030.

22–0

.33

24.0

–26.

75.

6–6.

71.

202.

9–3.

81.

100.

040.

04C

u 1.

4–1.

9C

D-3

MW

CuN

(S32

760)

J933

800.

030.

20–0

.30

24.0

–26.

06.

5–8.

51.

003.

0–4.

01.

000.

040.

04C

u 0.

5–1.

0, W

0.5–

1.0

CD

-4M

Cu

. . .

J933

700.

04. .

.24

.5–2

6.5

4.75

–6.0

1.00

1.75

–2.2

51.

000.

040.

04C

u 2.

75–3

.25

CD

-4M

CuN

. . .

J933

720.

040.

10–0

.25

24.5

–26.

54.

7–6.

01.

001.

75–2

.25

1.00

0.04

0.04

Cu

2.75

–3.2

5C

D-6

MN

. . .

J933

710.

060.

15–0

.25

24.0

–27.

04.

0–6.

01.

001.

75–2

.25

1.00

0.04

0.04

Cu

1.75

–2.5

CE

-3M

N25

07 (

S327

50)

J934

040.

030.

10–0

.30

24.0

–26.

06.

0–8.

01.

504.

0–5.

01.

000.

040.

04. .

.C

E-8

MN

. . .

J933

450.

080.

10–0

.30

22.5

–25.

58.

0–11

.01.

003.

0–4.

51.

500.

040.

04. .

.C

E-3

031

2J9

3423

0.30

. . .

26.0

–30.

08.

0–11

.01.

50. .

.2.

000.

040.

04. .

.C

F-3

304L

J925

000.

03. .

.17

.0–2

1.0

8.0–

12.0

1.50

. . .

2.00

0.04

0.04

. . .

CF-

3M31

6LJ9

2800

0.03

. . .

17.0

–21.

08.

0–12

.01.

502.

0–3.

02.

000.

040.

04. .

.C

F-3M

N31

6LN

J927

000.

030.

10–0

.20

17.0

–21.

09.

0–13

.01.

502.

0–3.

01.

500.

040.

04. .

.C

F-8

304

J926

000.

08. .

.18

.0–2

1.0

8.0–

11.0

1.50

. . .

2.00

0.04

0.04

. . .

CF-

8C34

7J9

2710

0.08

. . .

18.0

–21.

09.

0–12

.01.

50. .

.2.

000.

040.

04N

b 8X

C m

inC

F-8M

316

J929

000.

08. .

.18

.0–2

1.0

9.0–

12.0

1.50

2.0–

3.0

2.00

0.04

0.04

. . .

CF-

1030

4HJ9

2590

0.04

–0.1

0. .

.18

.0–2

1.0

8.0–

11.0

1.50

. . .

2.00

0.04

0.04

. . .

CF-

10M

316H

J929

010.

04–0

.10

. . .

18.0

–21.

09.

0–12

.01.

502.

0–3.

01.

500.

040.

04. .

.C

F-10

MC

316H

J929

710.

10. .

.15

.0–1

8.0

13.0

–16.

01.

501.

75–2

.25

1.50

0.04

0.04

(10x

C)–

1.2

Nb

CF-

10SM

nNN

itron

ic™

60J9

2972

0.10

0.08

–0.1

816

.0–1

8.0

8.0–

9.0

7.0–

9.0

. . .

3.5–

4.5

0.04

0.04

. . .

CF-

12M

316

. . .

0.12

. . .

18.0

–21.

09.

0–12

.01.

502.

0–3.

02.

000.

040.

04. .

.C

F-16

F30

3J9

2701

0.16

. . .

18.0

–21.

09.

0–12

.01.

501.

52.

000.

170.

04Se

0.2

–0.3

5C

F-20

302

J926

020.

20. .

.18

.0–2

1.0

8.0–

11.0

1.50

. . .

2.00

0.04

0.04

. . .

CG

-6M

MN

Nitr

onic

™50

J937

900.

060.

20–0

.40

20.5

–23.

511

.5–1

3.5

4.0–

6.0

1.5–

3.0

1.00

0.04

0.04

0.1–

0.3

Nb,

0.1

–0.3

VC

G-8

M31

7J9

3000

0.08

. . .

18.0

–21.

09.

0–13

.01.

50. .

.1.

500.

040.

04. .

.C

G-1

230

8J9

3001

0.12

. . .

20.0

–23.

010

.0–1

3.0

1.50

. . .

2.00

0.04

0.04

. . .

CH

-830

9SJ9

3400

0.08

. . .

22.0

–26.

012

.0–1

5.0

1.50

. . .

1.50

0.04

0.04

. . .

CH

-10

309H

J934

010.

04–0

.10

. . .

22.0

–26.

012

.0–1

5.0

1.50

. . .

2.00

0.04

0.04

. . .

CH

-20

309

J934

020.

20. .

.22

.0–2

6.0

12.0

–15.

01.

50. .

.2.

000.

040.

04. .

.C

K-3

MC

uN25

4SM

O™

J946

530.

025

0.18

–0.2

419

.5–2

0.5

17.5

–19.

51.

206.

0–7.

01.

000.

040.

04C

u 0.

5–1.

0C

K-2

031

0J9

4202

0.20

. . .

23.0

–27.

019

.0–2

2.0

2.00

. . .

2.00

0.04

0.04

. . .

CN

-3M

904L

J946

520.

03. .

.20

.0–2

2.0

23.0

–27.

02.

004.

5–5.

51.

000.

040.

04. .

.C

N-3

MN

AL

–6X

N®

J946

510.

030.

18–0

.24

20.0

–22.

023

.0–2

7.0

2.00

6.0–

7.0

1.00

0.04

0.04

. . .

(con

tinue

d)

Not

es: A

ll co

mpo

sitio

ns in

clud

e Fe

as

bala

nce.

Sin

gle

valu

es a

re m

axim

um v

alue

s un

less

oth

erw

ise

note

d. (

a) T

he w

roug

ht e

quiv

alen

t com

posi

tion

is n

ot th

e sa

me

as th

e ca

st. (

b) M

o is

not

an

inte

ntio

nal a

dditi

on.

Tabl

e A

1.6

Com

posi

tion

of A

lloy

Cas

ting

Ins

titu

te (

AC

I) h

eat–

and

cor

rosi

on–r

esis

ting

cas

ting

allo

ys


Tabl

e A

1.6

(con

tinu

ed)

Nam

eW

roug

hteq

uiva

lent

(a)

UN

Sde

sign

atio

n

Com

posi

tion

, %

CN

Cr

Ni

Mn

Mo

SiP

SO

ther

CN

-7M

320

N08

007

0.07

. . .

19.0

–22.

027

.5–3

0.0

1.50

2.0–

3.0

1.50

0.04

0.04

Cu

3.0–

4.0

CN

-7M

S. .

.J9

4650

0.07

. . .

18.0

–20.

022

.0–2

5.0

1.50

2.5–

3.0

3.50

0.04

0.04

Cu

1.5–

2.0

CT-

15C

. . .

N08

151

0.05

–0.1

5. .

.19

.0–2

1.0

31.0

–34.

00.

15–1

.5. .

.0.

50–1

.50.

040.

04N

b 0.

5–1.

5H

eat

resi

stin

g al

loys

HA

504

J820

900.

20. .

.8–

10. .

.1.

00.

9–1.

20.

35–0

.65

0.04

0.04

. . .

HC

446

J926

050.

50

. . .

26–3

04

max

2.0

0.5(

b)1.

00.

040.

04. .

.H

D32

7J9

3005

0.50

. .

.26

–30

4–7

2.0

0.5(

b)1.

50.

040.

04. .

.H

E31

2J9

3403

0.20

–0.5

0. .

.26

–30

8–11

2.0

0.5(

b)2.

00.

040.

04. .

.H

F30

2BJ9

2603

0.20

–0.4

0. .

.19

–23

9–12

2.0

0.5(

b)2.

00.

040.

04. .

.H

H30

9J9

3505

0.20

–0.5

00.

224

–28

11–1

42.

00.

5(b)

2.0

0.04

0.04

. . .

HI

. . .

J940

030.

20–0

.50

. . .

26–3

014

–18

2.0

0.5(

b)2.

00.

040.

04. .

.H

K31

0J9

4224

0.20

–0.6

0. .

.24

–38

18–2

22.

00.

5(b)

2.0

0.04

0.04

. . .

HK

-30

. . .

J942

030.

25–0

.35

. . .

23.0

–27.

019

.0–2

2.0

2.0

0.5(

b)2.

00.

040.

04. .

.H

K-4

0. .

.J9

4204

0.35

–0.4

5. .

.23

.0–2

7.0

19.0

–22.

02.

00.

5(b)

2.0

0.04

0.04

. . .

HL

. . .

N08

604

0.20

–0.6

0. .

.28

–32

18–2

22.

00.

5(b)

2.0

0.04

0.04

. . .

HN

. . .

J942

130.

20–0

.50

. . .

19–2

323

–27

2.0

0.5(

b)2.

00.

040.

04. .

.H

P. .

.N

0870

50.

35–0

.75

. . .

24–2

833

–37

2.0

0.5(

b)2.

00.

040.

04. .

.H

P-50

WZ

. . .

. . .

0.45

–0.5

5. .

.24

–28

33–3

72.

50.

5(b)

2.0

0.03

50.

035

W4.

0–6.

0, Z

r 0.

1–1.

0H

T33

0N

0860

50.

35–0

.75

. . .

13–1

733

–37

2.5

0.5(

b)2.

00.

040.

04. .

.H

T-30

. . .

N08

603

0.25

–0.3

5. .

.13

.0–1

7.0

33.0

–37.

02.

50.

5(b)

2.0

0.04

0.04

. . .

HU

. . .

N08

005

0.35

–0.7

5. .

.17

–21

37–4

12.

50.

5(b)

2.0

0.04

0.04

. . .

HW

. . .

N08

006

0.35

–0.7

5. .

.10

–14

58–6

22.

50.

5(b)

2.0

0.04

0.04

. . .

HX

. . .

N06

050

0.35

–0.7

5. .

.15

–19

64–6

82.

50.

5(b)

2.0

0.04

0.04

. . .

Not

es: A

ll co

mpo

sitio

ns in

clud

e Fe

as

bala

nce.

Sin

gle

valu

es a

re m

axim

um v

alue

s un

less

oth

erw

ise

note

d. (

a) T

he w

roug

ht e

quiv

alen

t com

posi

tion

is n

ot th

e sa

me

as th

e ca

st. (

b) M

o is

not

an

inte

ntio

nal a

dditi

on.

APPENDIX 2

Physical and Mechanical Properties ofSelect Alloys

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

UNS Density, kg/dm3Modulus of

elasticity, GPa

Coefficient ofthermal exp.,

10–6 × K–1

Thermal conductivity,

W/M·°KSpecific heat,

J/kg·°K

Electrical resistivity,Ω·mm2/m

201 S20100 7.86 207 16.6 16.3 502 0.67301 S30100 8.03 193 16.6 16.3 500 0.73304 S30400 7.90 200 16.6 16.3 500 0.72304L S30403 7.90 200 16.6 16.3 500 0.72305 S30400 7.90 200 16.6 16.3 500 0.72316L S31603 8.00 200 16.5 14.6 480 0.74321 S32100 7.92 193 16.6 16.3 500 0.72904L N08904 7.95 190 15.3 13.2 460 0.95AL6-XN© N08367 8.06 200 15.3 11.8 474 0.89409 S40920 7.76 200 10.5 25.0 477 0.60430 S43000 7.70 200 10.3 23.9 460 0.60439 S43035 7.70 200 10.2 24.2 460 0.63468 S46800 7.76 200 10.5 25.0 477 0.60410 S41000 7.65 200 10.5 24.9 460 0.562101 S32101 7.8 200 13.5 17.0 500 0.802003 S32003 7.72 210 13.5 17.0 510 0.802205 S32205 7.8 200 14.6 16.5 500 0.802507 S32750 7.8 200 12.5 13.5 500 0.80




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

Name Condition UNS Yield strength, MPaTensile strength,

MPa Elongation, % Hardness

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

201F 2B S20100 330 700 51 89 Rb


301 tensile 2D S30100 320 850 49 88 Rb

301 1/4 hard S30100 580 900 32 25 Rb

301 1/2 hard S30100 815 1150 23 35 Rc

301 3/4 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 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

304 1/4 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 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 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 2D S40920 260 440 31 60 Rb


430 2B S43000 345 515 27 67 Rb


439 2D S43035 315 455 32 76 Rb


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 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 an-nealed, and temper passed.

APPENDIX 3

Introduction to Thermo-Calc and Instructions for Accessing FreeDemonstration Version

WITHIN THE MAIN BODY of this text-book, a number of diagrams have been plottedand attributed to a software package calledThermo-Calc. The purpose of this appendix isto give a brief introduction to Thermo-Calc, ex-plain what it is, and what are its uses. Also pro-vided are instructions for accessing a demon-stration version of the software.

What Is Thermo-Calc?

Thermo-Calc (Ref 1) is a powerful, flexiblesoftware package available from Thermo-CalcSoftware AB for performing various kinds ofthermodynamic and phase diagram calculationsfor multicomponent systems.

The software is based on the so-called CAL-PHAD (CALculation of PHAse Diagrams)method (Ref 2), which describes mathemati-cally the thermodynamics of a system through arepresentation of the Gibbs energies of the dif-ferent crystalline phases relevant to that systemand defined by the chemical composition of thesystem. Thermo-Calc minimizes the total Gibbsenergy of the system with respect to variousconstraints such as temperature, pressure, andchemical composition and thus predicts themost stable energy state (or equilibrium state)that can form. By suspending certain phases(i.e., manually removing certain selected phasesfrom the system and thus restricting the forma-tion of such phases), Thermo-Calc can also beused to investigate meta-stable equilibria-typeproblems.

Thermo-Calc is used in conjunction withthermodynamic databases containing polyno-mial functions that describe the Gibbs energiesof the different phases according to certainmodels that take into consideration nonidealchemical interactions in solution phases. Thesedatabases are based on the critical evaluation ofthermodynamic and phase equilibria data for bi-nary, ternary, and some higher-order systems,which are then assembled into self-consistentdatabases. Different databases are available fordifferent broad classifications of materials, sys-tems, or applications. For example, there aredatabases for steels and iron-based alloys; iron-based slags; nickel superalloys; aluminum,magnesium, titanium, and zirconium alloys; ce-mented carbides; nuclear materials; and more.Further information on the different databasesavailable can be found at the Thermo-Calc Website: www.thermocalc.com

The thermodynamic database for steels (Ref3), as developed by Thermo-Calc Software AB,was used in conjunction with Thermo-Calc forall the calculations made during the preparationof this book. The version of the database usedfor these calculations contains data for 20 ele-ments and 85 phases.

Although the databases are based primarilyon the critical assessments of binary, ternary,and some quaternary systems, the CALPHADmethodology provides a theoretical frameworkon which extrapolations can be made to predictthe phase equilibria for higher-order, multi-component systems (the higher the order of thesystem, the weaker the nonideal interaction




parameters become). Thermo-Calc can there-fore be used in conjunction with such databasesto make predictions for multicomponent sys-tems and alloys of industrial relevance as illus-trated by some of the examples given in themain body of this book. These calculations canbe validated against real alloy data if this infor-mation is available but is not based on (or ad-justed to fit) such higher-order alloy data.Higher order in this sense means more than fourelements (i.e., larger than a quaternary system).

Four specific types of calculation can be per-formed using Thermo-Calc, although the rangeof problems to which these can be applied isbroader:

1. Single-Point Equilibria: The temperature,pressure, composition/activity of a compo-nent (or the amount of a phase) are fixed andthe stable or meta-stable equilibrium for thespecified conditions is calculated.

2. Step: The amount of one state variable pa-rameter (or condition) can be changed, whilethe other conditions remain fixed. For exam-ple, to see how the different phases and theiramounts and compositions would vary withtemperature for a given alloy, one would“step in” temperature. Alternatively, one canvary the composition of one of the compo-nents/elements and calculate how the phaseamounts change for a fixed temperature orpredict how the solidus or liquidus wouldchange with varying alloy composition.

3. Map: Two axis variables (such as tempera-ture, pressure, composition, or activity ofthe components) are changed at the sametime. Isoplethal sections are generated byvarying temperature and composition of oneof the components. Isothermal sections arethe result of varying the amounts of two ofthe components for a fixed temperature. Ex-amples of each of these kinds of diagramsare given in the main body of the text.

4. Scheil: Thermo-Calc includes a Scheil-Gulliver model for nonequilibrium solidifi-cation and a modified Scheil model that con-siders partial equilibrium for componentsthat are selected by the user.

Applications of Thermo-Calc

Thermo-Calc is used around the world withinacademia, in government research laboratories,and by commercial industry. The software can

be used to perform calculations for most appli-cations involving phase equilibria, meta-stableequilibria, phase transformations, phase dia-grams, and various thermodynamic properties,as well as critical assessments and data evalua-tions for multicomponent systems.

While many types of calculations can bemade using Thermo-Calc, the software typicallyis used to predict:

• Stable and meta-stable phase equilibria forbinary, ternary, and higher-order systems(calculations for alloy compositions with 6,10, 15 elements are not uncommon, as illus-trated by some of the examples in the mainbody of this book).

• Amounts of phases (mass, volume and molefractions) formed (phase balance) as a func-tion of temperature, pressure, and composi-tion and also the chemical compositions ofthe phases formed

• Phase transformation temperatures such asliquidus, solidus, and solvus temperatures.Phase transformation temperatures can bepredicted based on the actual chemistry (notnominal chemistry).

• Thermochemical data such as enthalpies,heat capacity, and activities

• Driving forces for precipitation• Phase diagrams (isothermal and isoplethal

sections for multicomponent, multiphasesystems as illustrated in this book)

• Molar volume, density, and thermal expan-sion

• Scheil-Gulliver (nonequilibrium) solidifica-tion simulations

Thermo-Calc is not restricted just to model-ing the alloy. Complex systems representingprocessing, for example, can also be consid-ered. For example, another application is to cal-culate the carbon potential of multicomponentgas phase systems as a function of composition,temperature, and pressure and then predict whatphases an alloy might form at a given tempera-ture when exposed to such a carbon potential.

Thermo-Calc can thus be applied to a numberof practical problems related to metallurgy, pro-cessing, in-service performance, etc. as summa-rized by:

• Alloy Design: Modification of alloy chemis-tries to improve properties or reduce costsusing calculations to guide which composi-tions may be most suitable before preparingthem for testing

Appendix 3: Introduction to Thermo-Calc / 283

• Heat Treatment: Prediction of formationof problematic phases prior to thermal pro-cessing

• Casting: Calculation of liquidus and solidustemperatures; calculation of thermodynamicproperties of the alloy for input into castingmodeling codes

• Welding and Joining: Prediction of thephases formed at the joining of two dissimilarmaterials or the interaction with filler material

• Quality Control: Investigation of propertiesand phase balance within designated compo-sitional tolerances

More examples are available in the literature(search on key terms Thermo-Calc or CAL-PHAD). A list of published articles citingThermo-Calc is available at www.thermocalc.com.

How to Obtain a Free DemonstrationVersion of Thermo-Calc

Thermo-Calc is available in two formats:Thermo-Calc Classic, which has a commandline interface and can be run under a number ofdifferent operating systems (including Mi-crosoft Windows and Linux/Unix), andThermo-Calc for Windows, which has an easy-to-learn graphical user interface but only oper-ates in the Microsoft Windows environment.Demo versions are available for both of theseversions of the software.

The demo versions are free to use, subject tothe terms outlined in the Thermo-Calc SoftwareEnd User License Agreement. It should benoted that the demo versions are limited tousing just three elements (whereas in the fullproduct the current upper limit is 40 elements)

and are supplied with only certain small data-bases that are for demonstration purposes andthe evaluation of the software only.

A link to register and download the demon-stration version of the software can be ac-cessed via a link on the Thermo-Calc web siteat www.thermocalc.com. All fields in the regis-tration form should be completed before contin-uing to the download page, where further in-structions regarding installation of the softwarewill be provided.

On installation of the software, additionaldocumentation, including a Users Guide/Exam-ples manual in the form of PDF files will also beinstalled. Technical support for the demo ver-sions of the software is limited, but problemsrelated to installation or general inquiries can beaddressed by visiting www.thermocalc.com andlinking to their support.

The demo version will run for approximately1 month on a single computer, and installationon a network system is not supported. If youwish to run the software after the demo licensehas expired, it can be downloaded again (i.e.,obtaining a new demo license).

REFERENCES

1. J.O. Andersson, T. Helander, L. Höglund,P.F. Shi, and B. Sundman, Thermo-Calc andDICTRA, Computational Tools for MaterialsScience, Calphad, Vol 26, 2002, p 273–312,2002

2. N. Saunders and A.P. Miodownik, PergamonMaterials Series, CALPHAD (Calculation ofPhase Diagrams): A Comprehensive Guide,1, Elsevier, 1998

3. TCFE5—TCS Steel/Fe-Alloys Database,Version 5.0, 2007, Thermo-Calc SoftwareAB, www.thermocalc.com

Index

Aacetic acid

corrosion rates for various alloys of, plus formic acid,103(F)

duplex alloys, 102isocorrosion curves in, 36(F)isocorrosion performances of various alloys, 102(F)

acids, corrosion in, and bases, 31–36adsorption-induced brittleness, 51, 53adsorption-induced plasticity, 51, 52–53aeration cells, differential, 38–39aerobic bacteria, influencing corrosion, 55aesthetic finishes

Chrysler Building in New York City, 196, 197(F), 213(F)considerations, 217–219flatness, 219surface, 217–219

aging treatments, precipitation hardening stainlesssteels, 168, 170

aggressive chemical agents, refinery applications, 255Alloy Casting Institute (ACI)

composition of, heat- and corrosion-resisting castingalloys, 277(T), 278(T)

naming system, 147alloy design

avoiding unwanted phases, 94duplex alloys, 92–94Thermo-Calc software, 93, 282

alloying, sensitization, 47–48alloying elements

alloys, 1influence on alloy families, 147, 149influence on corrosion rate in contaminated sulfuric acid,

33 (F)influence on thermodynamic activity of C, N, S and O,

157(T)influence on uniform corrosion, 29–30martensitic precipitation-hardening stainless steels,

141(T)martensitic stainless steels, 130, 131(T)

alloy oxidation, behaviors, 66alloys

influence of, content on corrosion rate in hydrochloricacid, 35(F)

isocorrosion curves for, in sulfuric acid, 33(F)isocorrosion curves for, sulfuric acid with chlorides, 34(F)

alloy systemsalloying elements, 1austenitic alloys, 69, 72austenitic stainless family, 71(F)body-centered cubic (bcc) phase, 1families in perspective, 69–72most widely used, 1Schaeffler–Delong stainless steels constitution diagram,

5(F), 70(F)alpha prime

formation kinetics for duplex alloys, 94, 96(F)influence of, formation on hardness, 116(F)iron-chromium phase, 8martensite, 7–8, 73transition temperature change with, formation with

aging, 98(F)alteration, surface, 199aluminum

inclusions and pitting, 40, 41influence on thermodynamic activity of C, N, S and O,

157(T)metal migration into acetic solution from, 234(T)migration into acetic solution from stainless, aluminum

or carbon steel, 234(T)oxidation resistance, 79, 226, 228protective layer formation, 64–65

aluminum alloys, 1aluminum oxide, 191aluminum-titanium-nitride (AlTiN), 191aluminum/titanium precipitates, possible, 138(F)American Iron and Steel Institute (AISI), 240American Petroleum Institute (API), 135American Society for Testing and Materials (ASTM),

238–239American Society of Mechanical Engineers (ASME), 266anisotropy

deep drawing, 174ferritic stainless steels, 120–121Lankford ratio, 120, 175stainless long products, 179

annealingaustenitic stainless steels, 162–164bright, 198

Copyright © 2008 ASM International®. All rights reserved. Stainless Steels for Design Engineers (#05231G) www.asminternational.org

286 / Index

annealing (continued)deep drawing, 177duplex stainless steels, 170–171ferritic stainless steels, 165–166long-term, of welds, 43martensitic stainless steels, 166precipitation hardening steels, 139

anodeelectrochemical reactions, 12polarization, 20–21, 23

appliancesfacades, 239–240kitchen, 237–240laundry, 241stainless steels commonly used for, 239(T)

architecture and constructionaesthetic considerations, 217–219average chloride concentration in rainwater in United

States, 217(F)balancing corrosion resistance, processing and economy,

214–215balancing service environment, design and maintenance,

215, 217cleaning methods for uncoated stainless steel, 220(T)concrete reinforcing bar, 222corrosion resistance, 213–214design considerations, 216(F)ecological considerations, 222environment, 216(F)fabrication and joining, 221fabrication considerations, 221fire resistance, 221–222flatness, 219grades recommended by expert system, 217(F)graphic depicting low release of metal ions from 304

and 316 stainless steels, 222(F)local weather pattern, 216(F)maintenance, 220–221maintenance schedule, 216(F)roof, 219(F)ranking common stainless steels by pitting resistance

equivalent number (PREN), 214(T)repair, 221rolled-on stainless steel finishes, 218(F)salt exposure, 216(F)special finishes, 218(F)stainless steel selection expert system, 216(F)surface finish aesthetics, 217–219surface finish and corrosion resistance, 215

argon oxygen decarburization (AOD)adoption, 70alloy adjustment, 157automotive industry, 225cleanliness, 184control of nitrogen in refining by, 92ferrite, 4first commercial use, 109foundry practice, 154inclusions in steel, 40production process, 155

atmospheres, oxidation, 66–67atmospheric corrosion

uniform corrosion, 36–37atomic rearrangements, 2attraction, interatomic, 2austenite

alloying elements, 5carbide precipitation, 9carbon and nitrogen, 6, 9diffusion rates, 6face-centered cubic (fcc), 5γ-austenite in precipitation hardening alloys, 138–139interstitial elements, 6lattice expansions, 6(F)lean alloy of martensite and, 73–74mechanical properties, 7metastable state, 6phase in duplex alloy at room temperature, 91Schaeffler–Delong constitution diagram, 5(F), 70(F)Schaeffler diagram, 202(F)secondary type, 7semiaustenitic precipitation-hardenable stainless steel,

143sulfur and oxygen, 6

austenite conditioning, 169austenitic-ferritic “C” alloys, 151–152austenitic “H” alloys

high temperature HE–HP, 152–154precipitation hardening stainless steel, 170

austenitic precipitation-hardenable stainless steelscold work and aging, 146(F)composition, 145(T)corrosion resistance, 145–146mechanical properties, 145

austenitic stainless steelsalloy families, 69–72annealing, 162–164automotive structural components, 229carburization, 82composition of, 270(T), 271(T)composition of high-temperature, 82(T)compositions of commonly used lean, 72(T)compositions of corrosion-resistant, 86(T)corrosion resistance ratings, 87(T)corrosion-resistant alloys, 84–89corrosive environments, 88–89critical pitting temperature (CPT), 43, 44(F)drawability, 176ductility, 180face-centered cubic (fcc), 174family, 71(F)forming limit diagram of carbon steel and, 176(F)halogens, 82heat exchangers, 246high-temperature alloys, 79–83high-temperature mechanical properties, 82–83impact strength variation with temperature, 75(F)intermetallic phases, 82, 203isocorrosion curves for, in hydrochloric acid, 34(F)kitchen appliances, 239

Index / 287

lean alloys, 72–78machinability, 185machining setup recommendations, 183(T)martensite and austenite, 73–74mechanical properties, 74–76, 82–83mechanical properties after cold work and annealing, 163metal migration into acetic solution from, 234(T)nitriding, 82oxidation resistance, 79–81petroleum industry applications, 249(T)pitting resistance equivalent number (PREN), 43, 78, 85precipitation of carbides and nitrides, 76–78recommended thermal processing temperatures, 162(T)resilience and toughness of carbon steel vs. for automo-

tive components, 229(T)SCC (stress corrosion cracking), 49–50secondary phases in, 82(T)sensitization, 46–47soaking, 161–162stabilization, 78stainless steel in shipping, 245SCC performance, 262stress-strain curve for single crystals of stable, 53(F)surface finish, 89tensile properties of carbon steel vs. for automotive

components, 229(T)thermal processing, 161–164water vapor, 81–82weaknesses, 69welding characteristics, 201–204welding parameters, 207(T)

austenitizing, martensitic stainless steels, 131, 132(F),166–167

automotive and transportationalloy selection for exhaust systems, 226alloys for major elements of automotive exhaust sys-

tems, 227(T)automotive emission standards, 225bus bodies, 231(F)car manufacturers, 230catalytic converter, 227(T), 228center pipe, 227(T), 228decorative to highly engineered applications in automo-

biles, 225exhaust manifold and high-temperature, 227exhaust systems, 225–228ferritic stainless, 226flexible pipe, 227(T), 228front pipe, 227(T), 228fuel tanks, 231life-cycle cost calculation for stainless vs. carbon steel

for bus, 231(T)microcar frame, 231, 232(F)muffler, 227(T), 228rail transport, 232resilience and toughness of carbon and stainless steels

for automobiles, 229(T)stress-strain curves for 301 variants vs. duplex steels and

transformation steel, 230(F)structural components, 229–231

tailpipe, 227(T), 228tensile properties of carbon and stainless steels for auto-

mobiles, 229(T)trucks, 231

Bbacteria influencing corrosion, 55bacterial retention, food contact materials, 236(F)bases, corrosion in acids and, 31–36basic oxygen furnace (BOF), 156Bauschinger effect, 164biocorrosion, 55–56biological neutrality, food contact, 235bleaching, pulp, 266body-centered cubic (bcc) phase

carbon and alloy steels, 1change to face-centered cubic (fcc), 127(F)ferrite, 4, 110ferritic material, 174metals, 2

boronadditions to ferritic stainless steels, 121ferrite, 4

brazing, 211bright annealing, 198brightening stainless steels, 196buffing, 197built-up edge (BUE)

austenitic stainless grades, 185carbon and, 182coolants minimizing, 191copper and, 183ferritic stainless steels, 185grain sizes, 184iron and tool, 182nickel and, 183nitrogen and, 184precipitation hardening stainless steels, 185

buslife-cycle cost calculation for stainless vs. carbon steel,

231(T)stainless steel body, 230, 231(F)

Butler–Volmer equation, 20–21

Ccalcium

effect on machinability of 303, 188–189, 189(F)inclusions and pitting, 41

calcium-fluoride based slagelectroslag remelting (ESR), 158

“C” alloysaustenitic–ferritic alloys, 151–152corrosion resisting, 147duplex alloys, 151mechanical properties of corrosion resisting cast, 150(T)metallurgy of, 149, 151–152precipitation hardening, 151

carbide, precipitation kinetics, 7(F)

288 / Index

carbidesduplex alloys, 94flatware, 240precipitation, 76–78stainless steel, 9tooling, 191

carbonalloying element, 1austenite, 5ferrite, 3–4influence of alloying elements on thermodynamic activ-

ity of, 157(T)influence on thermodynamic activity of C, N, S and O,

157(T)influence on uniform corrosion, 29interstitial atoms of, in austenite, 6machinability of stainless steels, 182–183precipitation of carbides, 76–78precipitation rates by, content, 76(F)steel content, 156variation of martensite hardness with, 128(F)welding of austenitic stainless steels, 201

carbon diffusion, 115carbon dioxide, wet, 248carbon solubility, austenitic stainless, 76(F)carbon steel

activities and activity coefficients of elements in, 40(T)body-centered cubic (bcc) phase, 1corrosion rates of stainless vs., 135(F)metal migration into acetic solution from, 234(T)resilience and toughness of, vs. stainless steel for auto-

motive components, 229(T)tensile properties of, vs. stainless steel for automotive

components, 229(T)carburization, 82carburizing, 199casting

stainless steel processing, 158–159Thermo-Calc software, 283

casting alloysAlloy Casting Institute (ACI), 147austenitic-ferritic alloys, 151–152austenitic HE–HP alloys, 152–154chromium alloys, 148(T)chromium-nickel alloys, 148(T)composition of cast heat-resistant stainless and nickel

base alloys, 149(T)compositions of cast stainless corrosion resisting alloys,

148(T)duplex alloys, 151ferritic HA, HC, HD alloys, 152foundry practice, 154high-temperature mechanical properties of “H” alloys,

153(T)influence of alloying elements, 147, 149mechanical properties of heat-resistant stainless, at room

temperature, 152(T)metallurgy of “C” alloys, 149, 151–152metallurgy of “H” alloys, 152–154molten metal transfer, 153

naming system, 147precipitation hardening, 151room temperature mechanical properties of corrosion

resisting stainless, 150(T)welding, 154

catalytic converter, 227(T), 228catastrophic oxidation, 65cathode

effect of, polarization, 24(F)electrochemical reactions, 12mass transfer limitations, 24polarization, 20–21, 23

caustic solutions, 50center pipe of exhaust systems, 227(T), 228cerium

inclusions and pitting, 41oxidation resistance, 79–80protective layer formation, 65

Charpy V toughnesshigh-temperature austenitic alloys, 83(F)niobium-stabilized alloy, 119(F)titanium-stabilized alloy, 119(F)

chemical agents, 255chemical and process industry. See also corrosion types

corrosion table for fuming sulfuric acid, 261(T)corrosion table for sulfuric acid, 259(T), 260(T), 261(T)corrosion types, 258–262erosion, 262forms of corrosion, 258intergranular corrosion, 261–262isocorrosion chart for sulfuric acid, 258(F)pitting and crevice corrosion, 258–259, 261single- and dual-environment systems, 257–258specific corrosives, 262stress corrosion cracking (SCC), 257, 262

chemical neutrality, food contact, 233–235chemistry, machinability of stainless steels, 182–184chi

intermetallic phase, 9precipitation kinetics, 7(F)

chip breaking, 187(F)chloride concentration in rainwater, 217(F)chloride-containing solutions, 50–51chloride ion, aggressive against stainless steel, 85chlorinated oils or waxes, 179(T)chromium

alloying element, 1austenitic alloys, 72austenitic stainless steels, 85–86chromium-oxygen system volatility vs. temperature and

oxygen pressure, 64(F)corrosion resistance, 228depletion from austenite near grain boundaries, 77(F)ferrite, 3ferritic stainless alloys with low, medium and high, 110inclusions and pitting, 42, 43influence on thermodynamic activity of C, N, S and O,

157(T)influence on uniform corrosion, 29, 30(F)ion release from stainless steel grades, 222(F)

Index / 289

machinability of stainless steels, 182migration into acetic solution from stainless, aluminum

or carbon steel, 234(T)oxidation resistance, 71, 79, 80(F), 228paralinear oxidation from evaporation of chromium

superoxide, 64(F)Pourbaix diagram, 17(F)sulfide formation, 186thermodynamics of oxidation, 57–59volatile nature of Cr2O3, 63, 64(F)

chromium alloys, cast stainless, 148(T)chromium-nickel alloys, 148(T)Chrysler Building

architecture using stainless steel, 213highly polished surface, 196, 197(F), 213(F)

cleaningpassivation, 195–196recommended methods, 195(T)stainless steel, 194–196

cleaning methodsstainless steels, 235–236uncoated stainless steel, 220(T)

cleanlinessfood contact materials, 235–236machinability of stainless steels, 184

coastal climates, 237–238coatings

cookware, 237tooling, 191

cold heading, 179–180cold work, 75coloring of stainless steels, 196commercial use

applications, 237–242cookware, 237flatware and cutlery, 240–241food contact, 233–237heating and water heating, 241–242kitchen appliances, 237–240laundry appliances, 241stainless steel, 233

compositionAlloy Casting Institute (ACI) heat- and corrosion-

resisting casting alloys, 277(T), 278(T)austenitic precipitation-hardenable (PH) stainless steel,

145(T)austenitic stainless steels, 270(T), 271(T)duplex alloys commercially available, 97 (T)duplex stainless steels, 276(T)ferrite stainless steels, 111(T), 112(T), 273(T), 274(T)martensitic PH stainless steels, 140(T)martensitic stainless steels, 124(T), 125(T), 275(T)PH stainless steels, 276(T)semiaustenitic PH stainless steel, 143(T)tool and cutlery martensitic stainless steels, 134(T)

concrete reinforcing bar, 222constitution diagram

Schaeffler–Delong stainless steels, 5(F), 70(F)Schaeffler diagram, 202(F)Welding Research Council’s 1992, 203(F)

construction. See architecture and constructioncontamination, 235continuous slab casting, 158cookware, 237coolants, 191copper

acid resistance, 71machinability of stainless steels, 183

copper sulfate. See also sulfuric acid plus copper sulfatecorrosion of stainless steel and titanium in, plus sulfuric

acid, 31(F), 32(F)corrosion

combating, in alloys for petroleum industry, 248, 250definition, 11erosion, 262intergranular, 261–262isocorrosion chart for sulfuric acid, 258(F)pitting and crevice, 258–259, 261pitting resistance equivalent number (PREN), 258single- and dual-environment systems, 257–258stress corrosion cracking (SCC), 257, 262table for sulfuric acid, 259(T), 260(T), 261(T)tendency, 15–16types, 258–262

corrosion cost, 252–253corrosion kinetics

Butler–Volmer equation, 20–21introduction, 19–20mass transfer control, 21migration and ionic diffusion, 21–22mixed potential theory and polarization diagrams,

22–23passivation, 23–25Tafel regime: electrode-kinetics control, 21

corrosion ratealloys in simulated evaporator liquid, 37(F)influence of alloying element on, in contaminated

sulfuric acid, 33(F)stainless oil country tubular goods, 135(F)stainless vs. carbon steel, 135(F)Tafel slope, 23(F)vs. surface roughness, 215(F)

corrosion resistancearchitecture, 213–214austenitic precipitation-hardenable (PH) stainless steel,

145–146balancing, processing and economy, 214–215duplex alloys, 99–106, 204ferritic stainless steels, 109, 110, 121–122function of salinity and temperature, 244(F)martensitic PH stainless steels, 141–142material selection for desalination, 244–245pulp-and-paper industry, 265–267rail transport applications, 232ratings of austenitic stainless steels, 87(T)semiaustenitic PH stainless steel, 144stainless steel for refinery equipment, 254–255stainless steel in shipping, 245sulfur hurting, 188surface finish and, 215

290 / Index

corrosion resisting alloysaustenitic stainless steels, 84–89“C” alloys, 149, 151–152composition of, austenitic stainless steels, 86(T)composition of Alloy Casting Institute (ACI), 277(T),

278(T)compositions of cast stainless, 148(T)duplex alloys, 91mechanical properties of stainless, 150(T)

corrosion theorycorrosion tendency, 15–16electrochemical reactions, 11–12Faraday’s law, 12galvanic vs. electrochemical cells, 14Nernst equation, 12–14Pourbaix diagrams, 16–17standard half-cell reduction potentials vs. normal hydro-

gen electrode, 14(T)corrosion types

atmospheric, 36–37biocorrosion and microbiologically induced, 55–56chromium influence, 29corrosion fatigue, 55corrosion in acids and bases, 31–36corrosion with fatigue or fraction, 48–55crack initiation, 48–49crack propagation, 49, 52(F)crevice, 38–39, 45–46critical current density, 29, 30(F)dissimilar metals and differential aeration cells, 38–39environmental variables, 50–51environmental variables influencing uniform corrosion,

28–29grain boundary, 46–48hydrochloric acid, 33–34, 35(F)hydrogen embrittlement, 54–55influence of alloying elements, 29, 30(F)localized, 37–38material variables, 29–31, 49–50molybdenum role, 29–30nickel, 30nitric acid, 34, 35(F)nitrogen, 30organic acids, 35, 36(F)phosphoric acid, 34–35, 36(F)pitting, 39–45pitting resistance, 43–45preventing crevice, 45–46SCC (stress corrosion cracking), 48–54SCC mechanisms, 51–54sensitization, 46–48sodium chloride/carbon dioxide environment, 30strong bases, 35sulfuric acid, 31–33sulfuric acid plus copper sulfate, 31(F), 32(F)uniform, 27–37

corrosive environmentsaustenitic stainless steels, 88–89platforms, 254refinery equipment, 254–255

creep rupture strength, 83, 84(F)creep strength, 83(F)crevice corrosion

austenitic stainless steels, 85, 87corrosion type, 214, 258–259, 261critical, temperature with alloy content, 45(F)critical crevice temperature (CCT) and critical pitting

temperature (CPT), 105(F)dissimilar metals and differential aeration cells,

38–39duplex alloys, 103–104geometry, 45preventing, 45–46

critical crevice temperature (CCT), 105(F)critical current density, 29, 30(F)critical pitting temperature (CPT)

austenitic steels, 43, 44(F)critical crevice temperature (CCT) and CPT, 105(F)duplex alloys, 103, 104(F), 105(F)stainless steels for unwelded and welded material,

44(F)vs. pitting resistance equivalent number (PREN), 85(F),

104(F)cryogenic containers, 245current density, 23(F)cutlery

flatware and, 240–241martensitic stainless steels, 133–134, 240stainless steels commonly used for, 241(T)

cutting tools, 133–134

Ddeep drawing

anisotropy, 174forming stainless steel, 173, 174–179geometry, 174–175hydroforming, 177–178intermediate annealing, 177materials composition, 175(T)schematic, 174(F)strain rate, 177texture, 174tooling, 176, 178–179

defects, 160delignification of pulp, 266demand for steel, 247desalination

materials selection for, 244–245multi-stage flash (MSF), 243reverse osmosis (RO), 243–244technology, 243–244

design, balance, 215, 217designers

car manufacturers, 230pitting corrosion, 39

developmentprecipitation–hardening stainless steels,

137–138welding, 211–212

Index / 291

differential aeration cellsactive alloys, 38–39microfouling, 56schematic, 20(F)

differential aeration corrosion cell, 12diffusion

atomic rearrangements, 2ionic transport, 21–22

diffusion rates, austenite vs. ferrite, 6digesters

first kraft from alloy 2205, 266(F)pulp-and-paper industry, 265

disinfection, 236–237dissimilar metals, 38–39dissociated ammonia, 198–199dissolution

equation, 37term, 27

dryers, laundry appliances, 241dry film, forming stainless, 179(T)dual-environment system, 257duplex alloys

acetic acid, 102annealing, 170–171composition of selected, stainless steels, 276(T)compositions, 97(T)concept, 91–92corrosion resistance, 99–106corrosion-resistant “C” alloys, 151crevice corrosion, 103–104deep drawing, 178fastest-growing stainless steel family, 91–92fatigue, 98Fe-Cr-Ni phase diagrams, 92(F)formation kinetics, 96(F)formic acid, 103(F)forming and machining, 99heat exchangers, 246hot forming, 180hydrochloric acid, 100–101impact strength, 97–98impact strength variation with temperature, 75(F)iron-nickel diagrams, 93(F)machinability, 186machining setup recommendations, 183(T)mechanical properties, 94–98nitric acid, 101organic acids, 102, 103(F)partitioning of elements, 93–94petroleum industry applications, 248(T)phosphoric acid, 101–102photomicrographs, 95(F)pitting corrosion, 102–103pitting resistance equivalent number (PREN), 43PREN influencing fatigue, 98, 99(F)pulp-and-paper industry, 265, 266–267recommended annealing and stress–relieving tempera-

tures, 170(T)SCC (stress corrosion cracking), 49, 104–106sensitization, 47

soaking, 170sodium hydroxide, 101stainless steel for line pipe, 253stainless steel in shipping, 245strength, 96stress corrosion cracking (SCC) performance, 262stress-strain curves for 301 variants vs., 230(F)structure and alloy design, 92–94sulfuric acid, 100thermal processing, 170–171umbilical tubing and risers, 253–254variations of ferrite, austenite, and duplex with tempera-

ture, 98(F)welding characteristics, 204–205welding parameters, 207(T)wrought 2205 duplex microstructure, 91(F)

Eearing

deep drawing, 178measuring tendency, 178

ecological considerations, 222electrochemical cell

closed circuit, 11–12potential, 38

electrochemical corrosion, 19electrochemical reactions, 11–12electrode-kinetics control Tafel regime, 21electrodes, polarization, 20–21electrolysis cell, 14(F)electrolyte resistance, 22electrolytic cells, galvanic vs., 14electrolytic pickling, cold-rolled stainless, 194electromotive force, 13electropolishing, 196electroslag remelting (ESR), 157–158embrittlement. See also hydrogen embrittlement (HE)

alpha prime, 8high-temperature, 114σ phase at higher temperatures, 151

engineering alloysminimum mechanical properties of stainless steel,

280(T)physical properties of major stainless steel, 279(T)

environment, stainless steel selection expert system,216(F)

environmental variablesstress corrosion cracking, 50–51uniform corrosion, 28–29

epsilon martensite, 7–8, 73–74equilibrium, argon oxygen decarburization (AOD),

155–156equivalent weight (EW), 19erosion, corrosion, 258, 262expert system

recommended stainless steel grades, 217(F)stainless steel selection, 216(F)

exposure to salt, stainless steel selection expert system,216(F)

292 / Index

Ffabrication, 221facades of appliances, 239–240face-centered cubic (fcc) phase

aluminum alloys, 1austenite, 5austenitic materials, 174change to body-centered cubic (bcc), 127(F)metals, 2

Faraday’s law, 12, 21fatigue

corrosion, 55duplex alloys, 98, 99(F)

fatty oils and blends, suitability in forming stainlesssteel, 179(T)

ferritecarbon and nitrogen, 3–4carbon diffusion rate, 115chromium, 3diffusion rates in austenite vs., 6δ-ferrite in precipitation hardening alloys, 138–139hydrogen and boron, 4mechanical properties, 4molybdenum, 4phase diagram of iron chromium, 3(F)phase in duplex alloy at room temperature, 91Schaeffler–Delong constitution diagram, 5(F), 70(F)Schaeffler diagram, 202(F)stabilization with titanium, 4thermal conductivity and thermal expansion, 4–5

ferritic “H” alloys, 152ferritic stainless steels

alpha prime formation, 116, 117(F)annealing, 165–166automotive exhaust systems, 226body-centered cubic (bcc), 174carbon diffusion rate in, 115composition, 111(T), 112(T), 273 (T), 274 (T)corrosion and oxidation resistance, 109, 121–122deep drawing, 178embrittling phenomenon, 116forming limit diagrams, 176 (F)groups of low, medium and high chromium, 110, 113heat exchangers, 245–246high-temperature properties, 121hot rolling, 159impact strength variation with temperature, 75(F)intermetallic phases, 116iron-chromium phase diagrams, 113(F), 114(F)kitchen appliances, 239lowest cost and simplest stainless, 109–110machinability, 185machining setup recommendations, 183 (T)mechanical behavior, 116–117metallurgy, 113–116metal migration into acetic solution from, 234(T)petroleum industry applications, 247(T)pitting resistance equivalent number (PREN), 43recommended annealing temperatures, 165(T)

sensitization, 47soaking, 165stabilization, 109, 115, 118–120stress corrosion cracking (SCC), 49stress relieving, 166superferritics, 113texture and anisotropy, 120–121time-temperature-transformation (TTT) curve for 430,

115(F)titanium and niobium, 118titanium for carbide and nitride formation, 115toughness, 116(F), 117(F), 118–119welding characteristics, 205–206welding parameters, 207(T)

ferromagnetism, 5fingerprints, cleaning methods for uncoated stainless,

220(T)fire resistance, stainless steel, 221–222flatness, surface aesthetic, 219flatware, 240–241flexible pipe, 227(T), 228flow lines, 252–253flux cored wire (FCW) welding, 210food contact

bacterial retention by material and cleaning time, 236(F)biological neutrality, 235chemical neutrality, 233–235cookware, 237flatware and cutlery, 240–241heating and water heating, 241–242kitchen appliances, 237–240material cleanliness, 235–236metal migration into acetic solution, 234(T)qualifications, 233–237stainless steels commonly used for appliances, 239(T)stainless steels commonly used for cutlery, 241(T)surface disinfection, 236–237

formability, ferritic stainless steel, 120formic acid

austenitic stainless steels, 89corrosion in, 35corrosion rates for various alloys of acetic plus formic

acid, 103(F)duplex alloys, 102isocorrosion curves in, 36(F)

forming limit diagram (FLD), 175–176forming technology

deep drawing, 173, 174–179deep drawing materials composition, 175(T)deep drawing schematic, 174(F)duplex alloys, 99duplex stainless steel, 178ferritics, 120, 178flat, rolled stainless steel, 173–179forces for hot working, 180(F)forming limit diagram of carbon steel vs. austenitic

stainless steel, 176(F)forming limit diagrams for stainless steel categories,

176(F)hot, of stainless steel, 180

Index / 293

hydroforming, 177–178hydrogen embrittlement, 177limiting drawing ratio (LDR) vs. Lankford ratio, 175(F)optimized 409 for forming vs. normal 409, 177(F)orange peel, 178stainless long products, 179–180stainless steel, 173stretch forming, 174suitability of lubricants for use in, 179(T)surface finish, 178tooling, 176, 178–179

foundary practice, casting alloys, 154free energy, phases, 2friction stir welding, 212front pipe, alloys in automotive exhaust systems,

227(T), 228fuel tanks, 231fuming sulfuric acid, 261(T)furnace, stainless steels in, 241–242

GGallionella, 55galvanic cell

electrochemical reaction, 38schematic, 14(F)vs. electrolytic cells, 14

gas metal arc welding (GMAW)joint design, 209(F)process, 210

gas tungsten arc welding (GTAW)joint design, 209(F)process, 208–210

geometrycrevice corrosion, 45deep drawing, 174–175pitting corrosion, 39

Gibbs free energyelectrochemical reactions, 12–13oxidation, 57, 58(F)

grade selection, corrosion resistance, processing andeconomy, 214–215

graffiti, cleaning methods for uncoated stainless, 220(T)

grain boundariesaustenite, 6austenite, of martensite, 8boron additions to ferritics, 121carbide precipitation, 9, 76–77, 77(F)corrosion, 46–48defects in stainless steel, 160depletion of chromium from austenite near, 77(F)ferrite-austenite, 8

grain sizeaustenitic stainless steel annealing, 163martensitic stainless steels and toughness, 131,

132(F)material structure, 184

graphite, suitability in forming stainless steel, 179(T)grinding, coarse polishing, 197

grit sizes, 197(T)Guinier–Preston (GP) zones, 138

Hhalf-cell reactions

reduction potentials, 13–14vs. normal hydrogen electrode, 14(T)

“H” alloysaustenitic HE–HP alloys, 152–154corrosion resisting, 147ferritic HA, HC, HD, 152high-temperature mechanical properties of, 153(T)mechanical properties of heat-resistant, 152(T)metallurgy, 152–154

halogens, 82hardening

austenitic stainless steels, 75ferritic stainless steels, 116–117

heat-affected zone (HAZ)austenitic stainless steel, 207chromium carbide formation in, 201duplex stainless steels, 204ferritic stainless steels, 109, 205laser welding, 210martensitic stainless steels, 206secondary austenite, 7

heat exchangers, 245–246heat-resistant alloys

composition of Alloy Casting Institute (ACI), 278(T)compositions, 149(T)“H” alloys, 152–154mechanical properties of cast stainless, 152(T)

heat tint, coloring stainless steels, 196heat tinting, cleaning method, 220(T)heat treatment, Thermo-Calc software, 283heat treatment and conditioning, 168–170heavy-duty emulsions, 179(T)heavy metals, elimination, 157high-frequency induction welding, 211high-speed tool steels, 190–191high-temperature alloys

austenitic stainless steel, 82(T)intermetallic phases of austenitic stainless steel, 82martensitic stainless steels, 133, 134(F)mechanical properties of austenitic, 82–83, 84(F)oxidation resistance of austenitic, 79–81water vapor, 81–82

high-temperature embrittlement, 114high-temperature properties, ferritic stainless steels, 121hopper cars, 232hot ductility defects, 160hot forming, 180hot mill defects, 160hot rolling, 159–160hot Steckel mills, 159hot strip tandem mills, 159hydrochloric acid

austenitic stainless steels, 88, 89(F)corrosion in, 33–34

294 / Index

hydrochloric acid (continued)duplex alloys, 100–101influence of alloy content on corrosion rate in, 35(F)isocorrosion curves for austenitic stainless steels in, 34(F)isocorrosion curves for stainless steels in, 34(F)isocorrosion performance of duplex, 101(F)

hydrofluoric acid, 193–194hydroforming, 177–178hydrogen/argon atmosphere, bright annealing, 198–199hydrogen embrittlement (HE)

corrosion fatigue, 55crack growth, 49ferritic stainless steels, 121–122mechanisms, 54–55stress corrosion cracking, 51, 52–54

hydrogen ion reduction, 15(F), 28(F)hypochlorite bleaches, 195

Iimpact strength

duplex alloys, 97–98, 204variation with temperature for stainless steels, 75(F)

inclusion-related defects, 160inclusions

chip breaking at sulfides, 187(F)lead, selenium, tellurium, 186oxides, 188–190pitting corrosion, 40–43role in machining stainless steels, 186–190stainless steel, 10sulfur, 186–188

induction welding, high-frequency, 211ingot method, 158, 159initiation

pitting, 39–40, 43stress corrosion cracking, 48–49

interatomic attraction, thermodynamics, 2intermetallic phases

austenitic stainless steel, 82, 203ferritic stainless steels, 116stainless steel, 8–9

International Nickel Company (INCO) process, 196interstitial elements, 6ionic current, 11–12ionic diffusion, 21–22iron

body- and face-centered cubic transformations, 2electrochemical corrosion, 19ferrite, 3ion release from stainless steel grades, 222(F)machinability of stainless steels, 182migration into acetic solution from stainless, aluminum

or carbon steel, 234(T)penetration rates, 19(T)Pourbaix diagram, 16(F)pseudo-binary-phase diagram for, and sulfur, 41(F)

iron-chromiumphase diagram, 3(F)phase diagram from Thermocalc, 113(F)

phase diagrams with varying carbon, 114(F), 130(F)

phase diagrams with varying chromium, 130(F)iron dissolution, 20(F)iron reduction, 15(F), 28(F)

JJFI Steel, oil country tubular goods and line pipe

alloys, 135(T)joining

stainless steel, 221Thermo-Calc software, 283

joint design, 208, 209(F)

Kkinetics, alpha prime formation, 94, 96(F)kitchen appliances

austenitic stainless steel, 239coastal conditions, 238exposure of stainless samples to North Carolina beach,

238(F)facades, 239–240ferritic stainless grades, 239food contact, 237–240interior or working parts, 239

knife-line attack, 48, 202kraft process

paper-making, 265–267pulp-and-paper industry, 36, 265

LLankford r, earing tendency, 178Lankford ratio

anisotropy measure, 120, 175limiting drawing ratio as function of, 175

lanthanum, protective layer, 65laser welding, 210lattice expansions, 6(F)laundry appliances, 241laves, precipitation kinetics, 7(F)laves phase, 9leaching, elements from stainless to foods, 234lead, stainless steel machinability, 186lean alloys

austenitic, 72–78compositions of austenitic, 72(T)martensite and austenite, 73–74

lime content, electroslag remelting, 158limiting current

electrode reaction kinetics, 21increasing mass transfer, 23, 24(F)

limiting drawing ratio, 175line pipe

martensitic stainless, 134–135stainless steel application, 252–253

liquefied natural gas (LNG) vessels, 254localized corrosion, 37–38

Index / 295

lubricantsoxides, 189–190suitability in forming stainless steel, 179(T)

lubrication, 197–198

Mmachinability, material’s, 181(F)machining stainless steels

austenitic, 185carbides, 191carbon, 182–183chromium, 182cleanliness, 184coatings, 191coolants, 191copper, 183cross-section size, 185duplex, 186duplex alloys, 99ferritic, 185high-speed tool steels, 190–191introduction, 181–182iron, 182lead inclusions, 186machinability of stainless steel families,

185–186manganese, 183martensitic, 185material’s machinability, 181(F)molybdenum, 183nickel, 183niobium, 184nitrogen, 184oxide inclusions, 188–190physical and mechanical properties, 182–185precipitation hardening, 185–186process, 184role of chemistry, 182–184role of inclusions, 186–190selenium inclusions, 186setup recommendations for turning wrought stainless

steels, 183(T)structure, 184sulfur, 183sulfur inclusions, 186–188super stainless steels, 186tellurium inclusions, 186titanium, 184tooling and coolants, 190–191

maintenancebalancing service, design and, 215, 217stainless steel, 220–221stainless steel selection expert system, 216(F)

manganesealloying element, 1austenite, 5inclusions and pitting, 41influence on thermodynamic activity of C, N, S and O,

157(T)

machinability of stainless steels, 183sulfide formation, 186

manganese sulfidesinclusions, 41stress risers, 188x-ray examination, 189(F)

Marangoni effect, 207marine systems

corrosion resistance vs. salinity and temperature, 244(F)desalination, 243–245heat exchangers, 245–246materials for desalination, 244–245shipping, 245typical analyses and properties of marine alloys, 245(T)

martensitecarbon and nitrogen, 9composition range, 7formation, 126–127forms, 7–8, 73–74lattice expansions, 6(F)lean alloy of, and austenite, 73–74platelets from surface, 126(F)reversion of, formed by cold work, 75(F)Schaeffler–Delong constitution diagram, 5(F), 70(F)Schaeffler diagram, 202(F)tempering, 7varying hardness with carbon content, 128(F)

martensitic alloys“C” alloys, 149, 151composition of, precipitation hardening (PH) alloys,

140(T)corrosion resistance of, PH alloys, 141–142mechanical properties of, PH alloys, 139(T)microstructures of, PH alloys, 140(F)precipitation hardening stainless steels, 139–142

martensitic stainless steelsannealing, 166applications, 133–135austenitizing, 166–167composition, 124(T), 125(T)composition of, 275(T)compositions of tool and cutlery, 134(T)corrosion rates of stainless oil country tubular goods

(OCTG) alloys, 135(F)corrosion resistance, 123distinction from other alloys, 123, 126expanding austenite stability range with nickel, 131(F)flatware and cutlery, 240–241hardness variation with carbon content, 127, 128(F)high-temperature use, 133, 134(F)hot rolling, 159influence of alloying elements, 130, 131(T)iron-chromium phase diagrams, 130(F)machinability, 185machining setup recommendations, 183(T)martensite formation, 126–127OCTG and line pipe, 134–135passivation, 195petroleum industry applications, 247(T)phase structure, 127–128, 130–131

Index / 303

temperaturechromium-oxygen system volatility, 64(F)corrosion resistance vs., 244(F)critical crevice corrosion, with alloy content, 45(F)critical pitting, (CPT), 43, 44(F)impact strength variation with, for stainless steels,

75(F)partitioning ratio varying with, 93(F)standard Gibbs free energy of metal oxide formation vs.,

58(F)variation of pitting potential with, for duplex alloys,

104(F)tempering

influencing martensitic stainless hardness, 132, 133(F)

martensite, 7martensitic stainless steels, 167

tensile propertiesaustenitic precipitation-hardenable stainless steel,

145(F)austenitic stainless steels, 75

tensile strength equation, 74texture

deep drawing, 174ferritic stainless steels, 120–121

thermal conductivityduplex alloys, 205ferrite, 4–5

thermal cutting, 211thermal expansion

austenitic stainless steels, 202duplex alloys, 205ferrite, 4–5

thermal processingannealing, 162–164, 165–166, 170–171austenitic stainless steels, 161–164austenitizing, 166–167duplex stainless steels, 170–171ferritic stainless steels, 165–166martensitic stainless steels, 131–133, 166–168precipitation-hardening stainless steels, 168–170soaking, 161–162, 165, 166, 170stress relieving, 164, 166, 167–168tempering, 167

thermal stresses, predicting, 64, 65(F)Thermo-Calc

alloy design, 93, 282applications, 282–283casting, 283free demonstration version, 283heat treatment, 283iron-chromium phase diagram, 113(F)map, 282phase determination program, 2quality control, 283Scheil–Gulliver model, 282single-point equilibria, 282software package, 281–282step, 282welding and joining, 283

thermodynamicsargon oxygen decarburization (AOD), 156influence of alloying elements on, activity of C, N, S,

and O, 157(T)oxidation, 57–60stainless steel, 2

thiosulfate, 50time-temperature-transformation (TTT) diagram

high-alloy stainless steel, 94, 96(F)unstabilized 430-type alloy, 115(F)

titaniumcarbide and nitride formation, 115carbide former, 78corrosion table for, in sulfuric acid plus copper sulfate,

31(F)deoxidizer in chromium–iron alloys, 156ferritic alloy stabilization, 4, 205inclusions and pitting, 41influence on thermodynamic activity of C, N, S and O,

157(T)isocorrosion curves for, in sulfuric acid plus copper sul-

fate, 32(F)isocorrosion curves in phosphoric acid, 36(F)machinability of stainless steels, 184possible aluminum/titanium precipitates, 138(F)role in sensitization, 47stabilization of ferritic stainless steels,

118–119stabilization of ferritic steels, 226

titanium-aluminum-nitride (TiAlN), 191titanium carbonitride (TiCN), 191titanium nitride (TiN), 191tooling

carbides, 191coatings, 191coolants, 191costs in deep drawing, 176high-speed tool steels, 190–191lubricants, 189, 190materials in deep drawing, 178–179

tools, martensitic stainless steels, 133–134toughness

austenitic stainless steels, 75–76duplex alloys, 97–98, 204ferritic stainless steels, 117, 118(F), 118–119high-temperature austenitic alloys, 83(F)martensitic stainless steels, 131, 132(F)

trains, 232transient oxidation, 60transpassive dissolution, 27transpassive regime, 24transportation. See automotive and transportationtrucks, 231tubular goods, 134–135tungsten

carbides for flatware, 240influence on thermodynamic activity of C, N, S and O,

157(T)tungsten inert gas (TIG), 208–210tuyeres, oxygen injection, 156

296 / Index

martensitic stainless steels (continued)photomicrographs, 129(F)recommended annealing, austenitizing, and tempering

temperatures, 166(T)sensitization, 47smallest stainless steel category, 123soaking, 166strain energy, 126(F), 127(F)stress relieving, 167–168tempering, 167tempering and toughness, 132, 133(F)thermal processing, 131–133, 166–168tool and cutlery alloys, 133–134toughness by austenite grain size and phosphorus,

132(F)welding characteristics, 206welding parameters, 207(T)

mass transfer, 23, 24(F)mass transfer control, 21mass transport, 24–25material selection, welding, 206–208material structure, 184material variables

stress corrosion cracking (SCC), 49–50uniform corrosion, 29–31

mechanical behavior, ferritic stainless steels, 116–117mechanical properties

austenite, 7austenitic precipitation-hardenable (PH) stainless steel,

145corrosion resisting cast stainless alloys, 150(T)deep-drawing stainless steels, 175(T)duplex alloys, 94–98ferrite, 4heat-resistant cast stainless alloys, 152(T)high-temperature, of austenitic stainless steels, 82–83,

84(F)high-temperature, of “H” alloys, 153(T)lean austenitic alloys, 74–76machinability of stainless steels, 182–185marine alloys, 245(T)martensite, 8martensitic PH stainless steels, 139(T)minimum, of stainless steel engineering alloys, 280(T)semiaustenitic PH stainless steel, 144(T)stainless steels, 10

mechanismspitting corrosion, 39–40precipitation-hardening, 138stress corrosion cracking (SCC), 51–54

melting production process, 155–157metal dusting, oxidation, 67metal flow directions, 208(F)metal ions release, 222(F)metallurgy

“C” alloys, 149, 151–152ferrite stainless steels, 113–116ferritic stainless, 226“H” alloys, 152–154introduction, 1–2

metal oxidesparabolic rate constants for growth, 59(F)standard Gibbs free energy of, formation vs. tempera-

ture, 57, 58(F)metals

oxidation, 57with oxide scale, 61(F)

metastable condition, 2metastable pitting, 40microbiologically induced corrosion, 55–56microcar frame, 231, 232(F)microorganisms, food contact, 235microstructures, martensitic precipitation-hardening

stainless steels, 140(F)migration, ionic transport, 21–22mineral resin, bacterial retention, 236(F)mischmetal, 41mixed potential theory, 22–23mold powder, continuous casting, 158molybdenum

alloying element, 1austenitic alloys, 72carbide precipitation, 9corrosion resistance, 228disulfide, 179(T)ferrite, 4influence on resistance to stress corrosion cracking,

50(F)influence on thermodynamic activity of C, N, S and O,

157(T)influence on uniform corrosion, 29–30influencing critical pitting temperature in welded vs.

unwelded austenitic grade, 208(F)machinability of stainless steels, 183oxidation resistance, 226, 228protective layer formation, 64–65

muffler, automotive exhaust systems, 227(T), 228multistage flash (MSF), 243

Nnaming system, Alloy Casting Institute (ACI), 147National Association of Corrosion Engineers (NACE)

alloy listing, 247diagram showing alloy suitability, 250, 251(F)regulating high-strength alloys, 142restrictions in use recommendations for stainless steels,

252(T)natural gas, vessels for liquefied, 254Nernst equation

open circuit potential, 20thermodynamics of electrochemical reactions, 12–14

neutralitybiological, in food contact, 235chemical, in food contact, 233–235

New York City’s Chrysler Building, 196, 197(F), 213(F)

nickelalloying element, 1austenite, 5

Index / 297

austenitic alloys, 72carbide precipitation, 9corrosion rates for stainless steels and, base alloys,

80(F)influence on thermodynamic activity of C, N, S and O,

157(T)influence on uniform corrosion, 30influencing oxidation of iron-chromium alloys, 79,

80(F)ion release from stainless steel grades, 222(F)machinability of stainless steels, 183migration into acetic solution from stainless, aluminum

or carbon steel, 234(T)resistance to stress corrosion cracking, 50(F)

nickel base alloys, 149(T)niobium

carbide former, 78creep resistance, 71–72high-temperature martensitic stainless, 133, 134(F)machinability of stainless steels, 184replacing titanium, 205role in sensitization, 47stabilization, 226stabilization of ferritic stainless steels, 118–119

Nippon Steel, oil country tubular goods and line pipealloys, 135(T)

nitric acidaustenitic stainless steels, 88corrosion behavior of high-silicon alloys in concen-

trated, 35(F)corrosion in, 34duplex alloys, 101isocorrosion curve for, 35(F)pickling oxide scale, 193–194

nitridesduplex alloys, 94stainless steel, 9–10

nitridingaustenitic stainless steel and, 82surface alteration, 199

nitrogenaustenite, 5austenitic alloys, 72austenitic stainless steels, 71, 86delay in carbide precipitation by, 78(F)ferrite, 3–4influence of alloying elements on thermodynamic activ-


157(T)influence on uniform corrosion, 29, 30interstitial atoms of, in austenite, 6machinability of stainless steels, 184solubility in austenite, 77–78stainless steel for line pipe, 253

nondestructive evaluation (NDE), 211normal hydrogen electrode (NHE)

half-cell reduction potential vs., 14(T)standard hydrogen electrode, 14

North Carolina, exposure of stainless steel, 238(F)

OOccupational Safety and Health Administration

(OSHA), 211Ohm’s law, 22oil and grease marks, cleaning methods, 220(T)“oil-canning,” 219oil country tubular goods (OCTG)

influence of chromium on corrosion rate of steel, 249(F)

influence of copper and nickel on corrosion rate ofmartensitic alloys, 249(F)

martensitic stainless, 134–135stainless steels in petroleum industry, 250–252

oleum, 33open circuit potential, 13orange peel, 178organic acids

austenitic stainless steels, 89corrosion in, 35duplex alloys, 102isocorrosion curves in, 36(F)

overpotentials, 20oxidation

effect of chromium, 57–59, 60(T)effect of rare earth additions, 65effect of silicon, aluminum, and molybdenum,

64–65electrochemical nature of, 60–61influence of nickel on, of iron-chromium alloys, 79,

80(F)iron-chromium-oxygen phase diagram, 59(F)kinetics and, rates, 61–63metal dusting, 67metal with oxide scale, 61(F)oxidation-resisting grades of stainless steel, 60(T)parabolic rate constants for growth of oxides, 59(F)paralinear, from evaporation of chromium superoxide,

64(F)quasi-steady-state approximation of moving boundary

problem of internal, 66(F)reaction at anode, 12schematic predicting thermal stresses, 65(F)spalling and cracking of scale, 63–65standard Gibbs free energy of formation of metal oxides

vs. temperature, 58(F)temperature dependence of metal dusting of iron,

67(F)thermodynamics of, 57–60transient, 60under less-oxidizing atmospheres, 66–67volatile nature of Cr2O3, 63, 64(F)Wagner’s theory, 61–63

oxidation resistanceaustenitic stainless steels, 79–81ferritic stainless steels, 109, 121–122isooxidation curves, 80(F)

oxide filmcoloring stainless steels, 196parameters for, coloring stainless steel, 196(T)

298 / Index

oxideseffect calcium on machinability of 303, 188–189, 189(F)elongated, 190(F)inclusions, 10inclusions and pitting, 41machinability of stainless steel, 188–190metal with, scale, 61(F)pickling to remove, layer, 25, 193–194removal of oxide scale, 193–194stabilization of austenitic alloy, 78Ugima, in 303 matrix, 189(F)un-deformed, 190(F)x-ray examination showing Ugima, and manganese sul-

fides, 189(F)oxyfuel gas welding (OFW), 210oxygen

austenite impurity, 6ferrite impurity, 3impurity, 1inclusions and pitting, 40, 41–42influence of alloying elements on thermodynamic activ-


157(T)influence on uniform corrosion, 29steel content, 156

oxygen gas reduction, 15(F), 28(F)oxygen pressure, 64(F)

Ppaint, cleaning method for uncoated stainless, 220(T)paper-making processes

bleaching pulp, 266digestion, 265–266kraft process, 265process equipment, 266–267washing and screening, 266

partitioning elements, 93–94passenger trains, 232passivation

effect on polarization diagrams, 23–25removing surface contamination, 195–196stainless steel, 25theory, 23transpassive regime, 24

passive behavior, 27passivity, 27penetration equation, 19penetration rates, 19(T)petroleum industry

alloy suitability vs. H2S and CO2 partial pressure, 251(F)austenitic stainless steels for, 249(T)chromium influence on corrosion rate in environments

by oil country tubular goods (OCTG), 249(F)combating corrosion in applications, 248, 250copper and nickel influencing corrosion rate of marten-

sitic stainless alloys for OCTG, 249(F)demand for steel, 247duplex stainless steels for, 248(T)

ferritic stainless steels for, 247(T)line pipe and flow lines, 252–253liquefied natural gas (LNG) vessels, 254martensitic stainless steels for, 247(T)molybdenum influence on SCC susceptibility, 248,

250(F)National Association of Corrosion Engineers (NACE),

247, 248, 250, 252(T)OCTG, 250–252platforms, 254precipitation-hardening stainless steels for, 248(T)presence of wet carbon dioxide, 248refinery equipment, 254–255restrictions in stainless steel use recommended by

NACE, 252(T)stainless steels for refinery processes, 254(T)stress corrosion cracking (SCC), 248umbilical tubing and risers, 253–254

pH, corrosion tendency, 15–16phase diagrams

computer models, 2Fe-Cr-Ni, 92(F)iron-chromium, 3(F), 73(F), 113(F), 114(F), 130(F)iron-chromium-oxygen, 58, 59 (F)iron-nickel, 93(F)pseudo-binary-, for iron and sulfur, 41 (F)

phasesalloy systems, 1ferrite, 3–5free energy, 2intermetallic, of stainless steel, 8–9structure of martensitic stainless steels, 127–128,

130–131phosphoric acid

austenitic stainless steels, 88corrosion in, 34–35duplex alloys, 101–102electropolishing solution, 196isocorrosion curves in, 36(F)minimum temperatures for wet, with duplex alloys,

102(F)phosphorus

ferrite impurity, 3impurity, 1, 156–157martensitic stainless steels and toughness, 131, 132(F)

photomicrographsduplex alloys, 94, 95(F)martensitic stainless steels, 127, 129(F)

physical propertiesmajor stainless steel engineering alloys, 279(T)stainless steels, 10

picklingoxide layer removal, 25, 193–194uniform corrosion, 28

pigmented pastes, 179(T)Pilling–Bedworth ratio (PBR), 63pitting

activities and activity coefficients in liquid steels, 40(T)austenitic stainless steels, 85, 88corrosion type, 39–40, 258–259, 261

Index / 299

critical, temperatures, 44(F)geometry, 39inclusions, 40–43influence of sulfur level on, resistance, 42(F)metastable, 40mischmetal, 41passive anode polarization curve, 40(F)pit initiation, 39–40pseudo-binary-phase diagram for iron and sulfur, 41(F)resistance, 43–45“weakest link” phenomenon, 214

pitting corrosionCPT (critical pitting temperature) vs. NaCl concentra-

tion, 103, 105(F)CPT vs. pH, 103, 105(F)CPT vs. pitting resistance equivalent number (PREN),

103, 104(F)duplex alloys, 102–103varying pitting potential with temperature, 103, 104(F)

pitting resistance equivalent number (PREN)austenitic alloys, 43, 78, 85corrosion, 37, 258critical pitting temperature vs., 85(F)duplex alloys, 43, 92, 102ferritic alloys, 43influence on duplex alloy fatigue strength, 98, 99(F)pitting corrosion, 43–45, 214ranking stainless steels by PREN, 214(T)umbilical tubing and risers, 253

Pittsburgh Convention Center, 219(F)platforms, stainless steel, 254polarization

anode, 23cathode, 23influence on uniform corrosion, 29, 30(F)overpotentials, 20passivating alloys, 39(F)passive anode, curve, 40(F)stainless steel in chloride-containing solution, 40

polarization diagramseffect of cathode polarization, 24(F)effect of mass transport, 24–25, 25(F)mixed potential theory and, 22–23passivation, 23–25schematic, 22(F)schematic of passive anode polarization curve, 23, 24(F)

polishinggrit sizes for target surface roughness, 197(T)surface finishing, 197–198

polycarbonate, bacterial retention, 236(F)polythionate, 50polythionic acid, 255porosity, 52Porsche, auto components, 230postweld stress relief, 205potassium hydroxide, 88–89Pourbaix diagrams

chromium, 17(F)construction of, 16–17iron, 16(F)

powder metallurgy, 159precipitated phases

Guinier–Preston zones, 138stainless steels, 8(T)

precipitationcarbides, 9carbides and nitrides, 76–78possible aluminum/titanium, 138(F)

precipitation-hardening stainless steelsadvantage over martensitic, 137annealed condition, 139austenitic, 144–146austenitic alloys, 170cast PH alloys, 151cold work influence on aging of A–286, 145, 146(F)composition, 276(T)composition of austenitic, 145(T)composition of martensitic, 140(T)compositions of semiaustenitic, 143(T)corrosion resistance of martensitic, 141–142corrosion resistance of semiaustenitic, 144development, 137–138influence of alloying elements, 141(T)machinability, 185–186machining setup recommendations, 183(T)martensitic, 139–142martensitic grades, 168–170mechanical properties of martensitic PH alloys, 139(T)mechanical properties of semiaustenitic, 144(T)mechanism of PH, 138microstructures, 140(F)passivation, 195petroleum industry applications, 248(T)phases in stainless steel, 10possible aluminum/titanium precipitates, 138(F)presence of δ-ferrite and γ-austenite, 138–139processing routes for S15700, 142, 143(F)properties of A-286 vs. test temperature, 145(F)recommended annealing and stress-relieving tempera-

tures for martensitic grades of, 169(T)semiaustenitic, 142–144specialized family, 137stress corrosion cracking (SCC), 141, 142thermal processing, 168–170welding characteristics, 206welding parameters, 207(T)

precipitation kinetics, 7(F)prevention, crevice corrosion, 45–46production processes

basic oxygen furnace (BOF), 156casting, 158–159defects, 160electroslag remelting (ESR), 157–158hot rolling, 159–160hot Steckel mills, 159hot strip tandem mills, 159impurities, 156–157influence of alloying elements on thermodynamics,

157(T)melting and refining, 155–157

300 / Index

production processes (continued)remelting, 157–158semiaustenitic precipitation-hardenable stainless steel,

142, 143(F)stainless steel, 155thermodynamics, 156vacuum arc remelting (VAR), 157–158vacuum induction melting (VIM), 157vacuum oxygen decarburization (VOC), 156

propagationcrack, rates of metals vs. current density, 52(F)stress corrosion cracking (SCC), 49

pulp-and-paper industryduplex stainless steels, 265, 266–267kraft process, 36, 265paper-making processes, 265–267

pulsed arc transfer, 210

Qquality control, Thermo-Calc software, 283

Rrain, cleansing action of, 220rainwater, average chloride concentration, 217(F)rare earth metals

inclusions and pitting, 41protective layer formation, 65

reduction, 12reduction potential

iron and hydrogen ion reductions vs. pH, 15(F), 28(F)

iron and oxygen gas reductions vs. pH, 15(F), 28(F)

reference electrode, 13–14refinery equipment, 254–255refining production process, 155–157repair, 221residential applications

cookware, 237domestic goods, 233flatware and cutlery, 240–241heating, 241–242kitchen appliances, 237–240laundry appliances, 241water heaters, 241, 242

resistanceOhm’s law, 22pitting, 43–45

resistance welding, 210–211resistivity, 22(T)reverse osmosis, 243–244ridging, ferritics, 178rolled finishes

applications, 198–199benefits, 198

roping, ferritics, 178rust staining, cleaning method for uncoated stainless,

220(T)

SSaab, auto components, 230safety, welding, 211salinity, corrosion resistance vs., 244(F)salt exposure, stainless steel selection expert system,

216(F)sanitation, cleaning stainless steels, 195scale

metal with oxide, 61(F)spalling and cracking of, 63–65

Schaeffler–Delong constitution diagram, 5(F), 70(F)Schaeffler diagram, 202–203seawater. See also marine systems

desalination, 243–245secondary austenite, 7secondary phases, 82(T)selenium, 186semiaustenitic precipitation-hardenable stainless steels

austenite-stabilizing elements, 142compositions, 143(T)corrosion resistance, 144mechanical properties, 144(T)processing by T route, 142, 143(F)

sensitizationaustenitic, 46–47duplex steels, 47effect of alloying, 47–48ferritic, 47ferritic stainless steels, 115heat treatment vs. time, 46(F)intergranular corrosion, 9knife-line attack, 48, 202martensitic steels, 47schematic of, due to chromium-rich precipitates, 46(F)welding, 48

service, design, and maintenance, 215, 217shielded metal arc welding (SMAW)

joint design, 209(F)process, 210

shielding gaswelding austenitic stainless steel, 203–204welding parameters for various stainless steels, 207(T)

shipping, 245short-circuiting transfer, 210sigma

intermetallic phase, 8–9precipitation kinetics, 7(F)

siliconalloying element, 1content in cast alloys, 147corrosion of high- austenitic steels in nitric acid, 34, 35(F)inclusions and pitting, 41influence on thermodynamic activity of C, N, S and O,

157(T)oxidation resistance, 71, 79, 226, 228protective layer formation, 64–65

single-environment system. See also chemical andprocess industry

aggressive chemical species, 258

Index / 301

slabs, 158slip dissolution, 51soaking

austenitic stainless steels, 161–162duplex stainless steels, 170ferritic stainless steels, 165martensitic stainless steels, 166

soap-fat pastes, 179(T)sodium chloride/carbon dioxide environment, 30sodium hydroxide

austenitic stainless steels, 88–89corrosion in, 35–36corrosion rates of duplex alloys, 101(F)corrosion rates of duplex alloys with contaminated envi-

ronment, 101(F)duplex alloys, 101

sodium hypochloritecleaning stainless steels, 195disinfecting stainless, 236–237

software package, Thermo-Calc, 2, 281–283soldering, 211solution treatment, precipitation-hardening stainless

steels, 168–170specialization, precipitation-hardening stainless steels,

137Sphaerotilus, 55spots, cleaning method for uncoated stainless, 220(T)spray transfer, 210stability

expanding austenite, with nickel, 131(F)lean alloy of martensite and austenite, 73–74

stabilizationferrite, 4ferritic stainless steel, 109, 115, 118–120ferritic steels for exhaust systems, 226lean austenitic alloys, 78

stacking fault, 73–74stainless long products

cold heading, 179–180hot forming, 180

stainless steel alloysaustenite, 5–7ferrite, 3–5

stainless steels. See also casting alloysactivities and activity coefficients of elements in, 40(T)bacterial retention by material and cleaning time,

236(F)carbides, 9casting, 158–159casting alloys, 147, 149categories for oxidation resistance, 59–60classifications by sulfur content, 187–188cleaning methods for uncoated, 220(T)composition of austenitic, 270 (T), 271 (T)composition of duplex, 276(T)composition of ferrite, 273(T), 274(T)composition of martensitic, 275(T)composition of precipitation-hardenable (PH), 276(T)concrete reinforcing bar, 222corrosion rates of, vs. carbon steel, 135(F)

corrosion table for, in sulfuric acid plus copper sulfate,31(F)

deep drawing, 173, 174–179defects in, hot–rolled bands, 160flat, rolled, 173–179hot rolling, 159–160inclusions, 10, 186–190isocorrosion curves for, in sulfuric acid plus copper sul-

fate, 32(F)lubricants for forming, 179 (T)machinability, 185–186machining setup recommendations, 183 (T)melting and refining, 155–157minimum mechanical properties of, engineering alloys,

280 (T)nitrides, 9–10oxidation-resisting grades, 60(T)passivation, 25penetration rates, 19(T)physical properties of major, engineering alloys, 279 (T)precipitated phases, 8 (T)precipitation-hardening process, 10precipitation kinetics in 316, 7 (F)properties, 10ranking by pitting resistance equivalent number (PREN),

214(T)ranking common, by PREN, 214(T)refinery processes, 254(T)remelting, 157–158resilience and toughness of carbon steel vs. for

automotive components, 229(T)Schaeffler–Delong constitution diagram, 5 (F)selection expert system, 216(F)tensile properties of carbon steel vs. for automotive

components, 229(T)thermodynamics, 2welding parameters, 207(T)

stains, cleaning method for uncoated stainless, 220(T)standard Gibbs free energy, 57, 58(F)standard hydrogen electrode, 14Steckel mills, hot, 159Steel Founder’s Society of America, 147strain energy, martensitic stainless steels, 126(F), 127(F)strain rate, deep drawing, 177stress corrosion cracking (SCC)

advantages of duplex alloys, 105–106austenitic stainless steels, 87–88corrosion form, 258crack initiation, 48–49crack propagation, 49crack propagation rates of metals vs. current density,

52(F)debating mechanisms, 105dilation of austenite due to hydrogen in solution, 53, 54(F)duplex alloys, 91, 104–106environmental variables, 50–51ferritic stainless steels, 121influence of molybdenum on resistance, 50(F)martensitic precipitation-hardening (PH) stainless steels,

141, 142

304 / Index

UUgima oxide

coating cutting tool and lubricant, 189–190comparing 304L chips with and without, 190(F)machinability by sulfur levels with and without, 190(F)x-ray showing, 189(F)

umbilical tubing and risers, 253–254uniform corrosion. See also corrosion types

environmental variables influencing, 28–29material variables, 29–31stainless steel, 27–28

United States, chloride concentration in rainwater,217(F)

unmixed zone, 208

Vvacuum arc remelting (VAR), 157–158vacuum induction melting (VIM), 157vacuum oxygen decarburization (VOD)

cleanliness, 184refining process, 156

vanadiumcarbides for flatware, 240high-temperature martensitic stainless, 133, 134(F)

Volvo, auto components, 230

WWagner’s theory, 61–63washer tubs and drums, 241water heaters, 241, 242waterline corrosion, 38water marking, cleaning method for uncoated stainless,

220(T)water vapor, 81–82wax-base pastes, 179(T)wax or soap plus borax, 179(T)weather pattern, stainless steel selection expert system,

216(F)weldability, 253welding

austenitic stainless steels, 201–204cast stainless alloys, 154characteristics of stainless steels, 201–206duplex stainless steels, 204–205

ferritic stainless steels, 205–206flux cored wire (FCW), 210gas metal arc welding (GMAW), 210gas tungsten arc welding (GTAW), 208–210high-frequency induction, 211joint design, 208, 209(F)laser, 210martenistic stainless steels, 206material selection and performance, 206–208metal flow directions in weld pool, 208(F)new developments, 212nondestructive evaluation (NDE), 211oxyfuel gas welding (OFW), 210parameters for various stainless steels, 207(T)practices, 211–212precipitation-hardening (PH) stainless steels, 206processes, 208–211recent developments, 211–212resistance, 210–211safety, 211Schaeffler diagram, 202(F)sensitization, 48shielded metal arc welding (SMAW), 210soldering and brazing, 211submerged arc welding (SAW), 210thermal cutting, 211Thermo-Calc software, 283tungsten inert gas (TIG), 208–210Welding Research Council’s 1992 constitution diagram,

203(F)weld shielding gas composition and crevice corrosion

resistance, 204(F)Welding Research Council, constitution diagram,

203(F)welds

influence of sulfur on pitting resistance of unannealed,42(F)

long-term annealing, 43wet carbon dioxide, 248

Yyield strength

austenitic precipitation-hardenable stainless steel, 145(F)equation, 74high-temperature austenitic alloys, 83, 84(F)

yttrium, 65

302 / Index

stress corrosion cracking (continued)material variables, 49–50mechanisms, 51–54petroleum industry, 248resistance, 257, 262(F)stress-strain curve for single crystals of austenitic steel

with and without hydrogen, 53(F)susceptibility of martensitic stainless steels, 123susceptibility to, with oxygen and chloride content for

304 stainless, 51(F)theory, 262threshold stress for, for various alloys, 88(F)varying resistance to, with nickel content, 50(F)water heaters, 242zones of susceptibility, 48(F)

stress relief annealing (SRA), 48stress relieving

austenitic stainless steels, 164ferritic stainless steel, 166martensitic stainless steels, 167–168

stress risers, 188stress sorption, stress corrosion cracking, 51, 53stretch forming, 174strip casters, 158–159strip tandem mills, 159strong bases

austenitic stainless steels, 88–89corrosion in, 35–36

structureduplex alloys, 92–94machinability of stainless steels, 184phase, of martensitic stainless steels, 127–128, 130–131

submerged arc welding (SAW)joint design, 209(F)process, 210

submerged entry nozzle, 158sulfate process, pulp-and-paper, 36sulfides

inclusions, 10inclusions in stainless steel, 186–187size and shape and machinability, 187stabilization of austenitic alloy, 78

sulfite process, 265sulfur

austenite impurity, 6comparing machinability, 190(F)effect on stainless machinability, 187, 188(F)ferrite impurity, 3impurity, 1, 156inclusions and pitting, 40, 41influence of alloying elements on thermodynamic activ-


157(T)machinability of stainless steels, 183metal flow directions in weld pool with and without,

208(F)pitting resistance of unannealed welds, 42(F)pseudo-binary-phase diagram for iron and, 41(F)stainless steel machinability, 186–188

sulfuric acidaustenitic stainless steels, 88corrosion in, 31–33corrosion table, 259(T), 260(T), 261(T)corrosion table for fuming, 261(T)duplex alloys, 100electropolishing solution, 196influence of alloying element on corrosion rate in con-

taminated, 33(F)isocorrosion, 88(F)isocorrosion chart for, 258(F)isocorrosion curves for various alloys in, 33(F)isocorrosion curves for various alloys in, with chlorides,

34(F)isocorrosion curves of duplex grades, 100(F)isocorrosion rates for various stainless steels, 32(F)oleum, 33pickling oxide scale, 193–194

sulfuric acid plus copper sulfatecorrosion table for stainless steels and titanium in, 31(F)isocorrosion curves for stainless steel and titanium in,

32(F)sulfurized or sulfochlorinated oils, 179(T)superaustenitic stainless steels, 164superferritics, 113super stainless steels, 186surface finishing

aesthetics, 217–219aesthetic surface finishes, 196–199austenitic stainless steels, 89bright annealing, 198brightening, 196cleaning, 194–196coloring, 196and corrosion resistance, 215deep drawing, 178–179function of surface treatments, 193–196introduction, 193parameters for oxide film coloring of stainless, 196(T)passivation, 195–196pickling, 193–194polished finishes, 197–198recommended cleaning methods, 195(T)removal of oxide scale, 193–194rolled finishes, 198–199rolled-on finishes, 218 (F)surface alteration, 199

surface roughness, 215(F)surface treatments

brightening, 196cleaning, 194–196coloring, 196passivation, 195–196removal of oxide scale, 193–194

TTafel slope, 23(F)tailpipe, automotive exhaust systems, 227(T), 228tellurium, 186

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

This publication is copyright © ASM International®. All rights reserved.

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

Mail Customer Service, ASM International 9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA

Email [email protected]

In Europe

American Technical Publishers Ltd. 27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX, United Kingdom Telephone: 01462 437933 (account holders), 01462 431525 (credit card) www.ameritech.co.uk

In Japan Neutrino Inc. 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.