应力腐蚀开裂

www.elsevier.com/locate/corsci

Theeffectofsensitizingtemperatureonstresscorrosioncrackingoftype316austeniticstainlesssteelinhydrochloricacidsolution

RokuroNishimuraa,*,IingMusalamb,YasuakiMaedaaabDepartmentofAppliedMaterialsScience,CollegeofEngineering,OsakaPrefectureUniversity,

1-1Gakuen-cho,Sakai,Osaka599-8531,Japan

ResearchandDevelopmentCentreforMetallurgy,KompleksPUSPIPTEK,Serpong,Tangerang,

Indonesia

Received17May2001;accepted5June2001

Abstract

Thestresscorrosioncracking(SCC)ofacommercialausteniticstainlesssteeltype316wasinvestigatedasafunctionofsensitizingtemperature(750–1300K)andtesttemperature(333–373K)in0.82kmol/m3hydrochloricacid(HCl)solutionbyusingaconstantloadmethod.Fromtheappliedstressdependenceofthreeparameters(iss:steadystateelongationrate,tss:timeintervalofSCC-dominatedfailure,tf:timetofailure),therelationshipsbetweenappliedstressandthethreeparametersweredividedintothreeregionsthataredominatedbyeitherstress,SCCorcorrosion.IntheSCC-dominatedregion,thelogarithmofisswasalinearfunctionofthelogarithmoftfirrespectiveofappliedstressandtesttemperature,althoughitsslopedependeduponsensitizingtemperature.Thisresultshowedthattheissbecameausefulparameterforpredictionoftfaswellasthecaseofthesolutionannealedspecimens.Fur-thermore,atthemostseveresensitizationwithasensitizingtemperatureofaround1000K,theslopeofthelinearfunctionoflogissvs.logtfshowedaminimum,thevalueoftss=tfwasamaximumandthefractureappearancewasanintergranularmode.Onthebasisoftheresultsobtained,theeffectofsensitizationonSCCwasdiscussedincomparisontotheresultsforthesolutionannealedtype316andaqualitativeintergranularSCC(IGSCC)mechanismwasinferred.Ó2002ElsevierScienceLtd.Allrightsreserved.

Keywords:Austeniticstainlesssteeltype316;Sensitization;Stresscorrosioncracking;Prediction;Steadystateelongationrate;Corrosionelongationcurve;IGSCCmechanism

*Correspondingauthor.Tel./fax:+81-722-54-9323.

E-mailaddress:nishimu@ams.osakafu-u.ac.jp(R.Nishimura).

0010-938X/02/$-seefrontmatterÓ2002ElsevierScienceLtd.Allrightsreserved.PII:S0010-938X(01)00117-2

1344R.Nishimuraetal./CorrosionScience44(2002)1343–1360

1.Introduction

Inapreviouspaper[1],thestresscorrosioncracking(SCC)oftheausteniticstainlesssteeltype304sensitizedintherangeof800–1300Kfor86.4kswasin-vestigatedin0.82kmol/m3hydrochloricacid(HCl)solutionasfunctionofappliedstress(r)andtesttemperaturebyusingaconstantloadmethod.Fromtheresultsobtained,itwasfoundthatthesteadystateelongationrate(issm/s)obtainedfromthecorrosionelongationcurvebecametherelevantparameterforthepredictionoftimetofailure(tf).However,theslopeofthelinearequationbetweenlogissandlogtfchangedwithsensitizingtemperatureandbecamealwayssmallerthanthatforthesolutionannealedspecimens,theslopesofwhichwereÀ1underaconstantappliedstresscondition(butenvironmentvariable)andÀ2underaconstantenvironmentcondition(butrvariable)[2,3].Itwasalsofoundthatthefractureappearance(transgranularorintergranular),theratioofthetimeintervalofSCC-dominatedfailure(tss)totfandSCCsusceptibilitydependeduponsensitizingtemperature.Toexplaintheseresults,wequalitativelyinferredtheintergranularSCC(IGSCC)mechanism,inwhichthegrainboundarysliding(GBS)wasthekeyfactor.

ItwasalreadyconfirmedthatalthoughtheSCCsusceptibilityofthesolutionannealedtype316wasmuchsmallerthanthatofthesolutionannealedtype304in0.82kmol/m3HCland0.82kmol/m3sulfuricacid(H2SO4)solutions[2–4],theirSCCbehaviorcouldbeexplainedwiththeproposedTGSCCmechanism[5]irrespectiveofthedifferencesinanionspeciesandthematerials.However,itwasnotclearwhethertheeffectofsensitizationontheSCCbehavioroftype316becomesthesameasthatofthesensitizedtype304.ThefirstpurposeofthepresentpaperwastoelucidatewhethertheuseofissastheparameterofSCCpredictiondescribedabovecanbeappliedtotheausteniticstainlesssteeltype316sensitizedatvarioustemperaturefor86.4ksaswellasthesensitizedtype304[1].Inaddition,fromthecomparisonbetweentheresultsofthesensitizedandthesolutionannealedtype316specimens,thesecondpurposewastoclarifywhetherornottheIGSCCmechanismproposedforthesensitizedtype304couldbeappliedtotheSCCbehaviorofthesensitizedtype316specimens.

2.Experimental

Thespecimensusedwereacommercialtype316austeniticstainlesssteel(yieldstrength:333MPaandultimatetensilestrength:7MPa),whosegeometryforSCCexperimentswasshowninFig.1.Thechemicalcomposition(mass%)wasasfollows:C0.054,Si0.67,Mn1.38,P0.030,S0.005,Ni11.16,Cr17.21,Mo2.21.Thespecimensweresolutionannealedat1373Kfor3.6ksunderanargonatmosphereandthenwaterquenched,whichwerecalledthesolutionannealedspecimens.Fur-therthesolutionannealedspecimensweresensitizedattemperaturesintherangeof750–1300Kfor86.4ks(24h)undertheargonatmosphereandthenwaterquenched.Priortotheexperiments,thesensitizedandsolutionannealedspecimenswerepol-ishedto1000gritemerypaper,degreasedwithacetoneinanultrasoniccleanerandwashedwithdistilledwater.Afterthepretreatment,thespecimenswereimmediately

setintoaSCCcell.Thetestsolutionusedwas0.82kmol/m3HClsolution,andwaspreparedfromdistilledwaterandguaranteedgradereagent.Thetesttemperaturewasusedintherangeof333–373K,butmainly353Æ0:5K.Allexperimentswerecarriedoutunderanopencircuitcondition.

Alever-typeconstantloadapparatus(leverratio1:10)towhichthreespecimenscanbeseparatelyandsimultaneouslyattachedwasusedwithacoolingsystemonthetoptoavoidevaporationofthesolutionduringexperiments.Thespecimenswereinsultedfromrodandgripbysurfaceoxidizedzirconiumtube.AchangeinelongationofthespecimensundertheconstantstressconditionwasmeasuredbyaninductivelineartransducerwithanaccuracyofÆ0.01mm.3.Results

3.1.Corrosionelongationcurve

Fig.2showsarepresentativeexampleofthecorrosionelongationcurvesatvariousconstantappliedstressesforthetype316specimenssensitizedat923Kfor86.4ksin0.82kmol/m3HClsolutionat353K.Thecorrosionelongationcurveswerefoundtoconsistofthreeregionsuptofailurewithaninitialsuddenriseofelongation;primary,secondaryandtertiaryregionsaswellasthoseofthesolutionannealedspecimens(type304andtype316)[2–4]andthesensitizedtype304spec-imens[1]obtainedunderthesameexperimentalconditions.Fromthecurves,wegotthreeparametersthatarethesteadystateelongationrate(iss¼I=tm/s)inthesec-ondaryregion,thetimeintervalofSCC-dominatedfailureorthetransitiontime(tss)betweenthesecondaryandtertiaryregions,andthetimetofailure(tf).Thetssisthetimewhenanelongationstartstodeviatefromalinearincreaseinthesecondaryregion.Asalreadyreported,whenthefailureofthespecimenstakesplacebySCC,thethreeregionsinthecorrosionelongationcurve,whichreflectsonSCCprocessuptofailure,correspondtocracknucleation,steadycrackpropagationandterminalcrackpropagationperiods,respectively[1–4].

3.2.Appliedstressdependenceofthreeparameters(iss,tf,tss)

Fig.3showsarepresentativeexampleofappliedstress(r)vs.logtfcurveforthespecimenssensitizedat923Kfor86.4ksin0.82kmol/m3HClsolutionat353K,

wherethedashedlineistheresultforthesolutionannealedspecimensunderthesameexperimentalcondition.Therelationshipforthesensitizedspecimenswasdividedintothreeregionsaswellasthoseforthesolutionannealedspecimens(type304andtype316)[2–4]andforthesensitizedtype304specimens[1],whichweredesignatedbyarabicnumerals1–3inFig.3.Themaximumappliedstress(rmax)inregion2becamelargerforthesensitizedspecimensthanforthesolutionannealedspecimens,whiletheminimumappliedstress(rmin)inregion2becamethesameasthatforthesolutionannealedspecimens.Thedifferenceintfbetweenthesensitizedandthesolutionannealedspecimensinregion2tendedtobecomelargewithincreasingr.Fig.4showstherelationshipbetweenrandlogissforthespecimenssensitizedat923K,wherethedashedlineistheresultforthesolutionannealedspecimens.ItwasfoundthattherelationshipwasalsodividedintothreeregionscorrespondingtothoseinFig.3,showingthatthedistinctionbetweenregions2and3becameclearercomparedtothatinFig.3.Inaddition,thevalueofissinregion2becamesmallerthanthatforthesolutionannealedspecimensatanappliedstressoflessthanabout400MPa,whereasthevalueoftfbecameshortincomparisontothatforthesolutionannealedspecimensasshowninFig.3.

Fig.5showstherelationshipbetweenrandtheratiotsstotfforthespecimenssensitizedat923K,wherethedashedlineistheresultforthesolutionannealedspecimens.Thevalueoftss=tfforthesensitizedspecimensinregion2heldconstantindependentofr,butbecamelargerthanthatforthesolutionannealedspecimens.Ontheotherhand,thevaluesinregions1and3becamelargerthanthatinregion2.Withregardtotheothersensitizedspecimens,furthermore,theirappliedstressde-pendencesofthethreeparametersweresimilartothatforthesensitizedspecimens

(923K,86.4ks),althoughthermaxandthevalueoftss=tfweredependentuponsensitizingtemperatureasdescribedbelow.

Fromtheappliedstressdependenceofthethreeparametersforthesolutionan-nealedandsensitizedspecimens,thethreeregions(1to3)forthesensitizedspeci-menswereconsideredtocorrespondtothestress-dominated,SCC-dominatedandcorrosion-dominatedregions,respectively,asalreadyreported[1–4].Fig.6showsthemaximumstress(rmax)andtheminimumstress(rmin)intheSCC-dominatedregion(region2)asafunctionofsensitizingtemperature,wherethetrianglesinFig.6standforthoseofthesolutionannealedspecimens.Thermaxbecamelargerthanthatforthesolutionannealedspecimensinthesensitizingtemperaturerangeof750–1250Kandshowedamaximumatasensitizingtemperatureofaround1000K.Ontheotherhand,therminkeptconstantindependentofsensitizingtemperature,whichwasthesameasthatforthesolutionannealedspecimens.Itwasconfirmedbyusingatensilemachinethatthebehavioroftheultimatetensilestrengthatroomtem-peratureforthespecimenssensitizedatvarioussensitizingtemperatureshadthesimilartendencyasthatofthermaxinFig.6,sothattheincreaseinthermaxwasrelatedtothatintheultimatetensilestrengthaswellasthecaseofthesensitizedtype304specimens[1].

Fig.7showsthevalueoftss=tfintheSCC-dominatedregionasafunctionofsensitizingtemperature,wherethetriangleinFig.7standsforthatofthesolution

annealedspecimens.Thevalueoftss=tfbecamelargerthanthatforthesolutionannealedspecimensinthesensitizingtemperaturerangeof750–1250Kandshowedamaximumatasensitizingtemperatureofaround1000K,correspondingtothatofrmaxinFig.6.Basically,theSCCbehaviorobtainedforthesensitizedtype316specimenswasconcludedtobecomethesameasthatofthesensitizedtype304,withregardtothebehaviorofrmax,rminandthevalueoftss=tf.3.3.Testtemperaturedependenceofthreeparameters

Toinvestigateatesttemperaturedependenceofthethreeparametersforthesensitizedspecimens,aconstantrintheSCC-dominatedregionwasused.Fig.8

showsarepresentativeexampleoftherelationshipsbetweenthreeparametersandareciprocaloftesttemperatureforthesensitizedspecimen(923K,86.4ks)underaconstantrof490MPa.Thelogtfandlogissbecameagoodstraightlineagainstareciprocaloftesttemperature,whilethevalueoftss=tfheldconstantindependentoftesttemperatureandwasthesameasthatobtainedasafunctionofrinFig.5.Thesimilarresultswereobtainedfortheothersensitizedspecimens.Althoughthevalueoftss=tfdependeduponsensitizingtemperature,itssensitizingtemperaturedepen-dencebecamethesameasthatinFig.7.3.4.Fractureappearances

Fig.9showstherepresentativeSCCfractureappearancesofthesensitizedspecimens.Thefractureappearance(a)ofthesensitizedspecimens(923K,86.4ks,r¼392MPa,Â1000)wasentirelyintergranular,whichwasthesameasthatofthespecimenssensitizedat1023Kfor86.4ks.However,that(b)ofthesensitizedspecimens(1123K,86.4ks,r¼392MPa,Â1000)waspredominantlytransgranular.Thesamefractureappearancewasobtainedforthespecimenssensitizedatasen-sitizingtemperatureofabove1123Kandbelow823K.

R.Nishimuraetal./CorrosionScience44(2002)1343–13601351

Fig.9.Therepresentativefractureappearancesofthesensitizedtype316,where(a)923K,86.4ksand(b)1123K,86.4ks(Â1000).

4.Discussion

4.1.Predictionoftimetofailure

Fig.10showsarepresentativeexampleoftherelationshipbetweenlogissandlogtf(r:variable)forthespecimenssensitizedat923K,whichisreplottedfromFigs.3and4.ThedashedlineinFig.10istheresultintheSCC-dominatedregionforthesolutionannealedspecimens.ItwasfoundthattherelationshipintheSCC-domi-natedregionbecameagoodstraightlinewithaslopesmallerthanthat(À2)forthesolutionannealedspecimens.Suchastraightlinecouldbeapplicabletotherelationshipsfortheothersensitizedspecimens,althoughtheirslopesdependeduponsensitizingtemperature.Hencewegotthefollowingequationforthesensitizedspecimens,

logiss¼ÀalogtfþC1

ðr:variableÞ

ð1Þ

wheretheslope,Àa,dependsuponasensitizingtemperatureandC1isaconstantdependingontheenvironmentalfactors(pH,testtemperature,concentrationandsoon)andsensitizingtemperature.Fig.11showsthevalueofainEq.(1)asafunctionofsensitizingtemperature,wherethetriangleisthatforthesolutionannealedspecimens.Thevalueofaforthesensitizedspecimensbecamesmallerthanthatforthesolutionannealedspecimens(a¼2)inthesensitizingtemperaturerangeof750–1200K,showingtheminimumatasensitizingtemperatureofaround1000K.Fig.12showsarepresentativeexampleoftherelationshipbetweenlogissandlogtf(r:constant)forthespecimenssensitizedat923K,whichisreplottedfromFig.9.ThedashedlineinFig.12istheresultforthesolutionannealedspecimens.Itwasfoundthattherelationshipalsobecameagoodstraightline.Theothersensitized

specimenshadasimilarlinearfunctionasthatinFig.12.Therefore,weobtainedthefollowingequation,

logiss¼ÀblogtfþC2

ðr:constantÞ

ð2Þ

wheretheslope,Àb,dependsuponasensitizingtemperatureandC2isaconstantdependingontheenvironmentalfactorsandsensitizingtemperature.Fig.13showsthevalueofbinEq.(2)asafunctionofsensitizingtemperature,wherethetriangleinFig.13isthatforthesolutionannealedspecimens.Thevalueofbwasfoundtobecomesmallerthanthat(b¼1)forthesolutionannealedspecimensinthesensi-tizingtemperaturerangeof750–1200K,showingtheminimumatasensitizingtemperatureofaround1000K.

Theabovelinearequationsshowthattheissofthesensitizedtype316specimensbecomesausefulparameterforpredictionoftfaswellasthesensitizedtype304specimens,andthesolutionannealedtype304andtype316specimens,becausetheisscanbeobtainedatatimewithin10–20%oftffromthecorrosionelongationcurve.

Hereitshouldbeemphasizedthattheslopesofthelinearequationsdependeduponthedegreeofsensitizationaswellasthoseofthesensitizedtype304,althoughtheslopesofthesolutionannealedspecimensheldconstantirrespectiveoftheanion

À

speciessuchasClÀandSO24andthematerials(type304andtype316)[2–4].4.2.Effectofsensitization

ThemaindifferencesintheSCCbehavioroftype316betweenthesolutionan-nealedandthesensitizedspecimensaresummarizedbelow:

(a)thermaxandthevalueoftss=tfintheSCC-dominatedregionarelargerforthesensitizedspecimensthanforthesolutionannealedspecimens,

(b)thevaluesofaandbinEqs.(1)and(2)aresmallerforthesensitizedspecimensthanforthesolutionannealedspecimens,

(c)thefractureappearanceisintergranularinthevicinityofthesensitizingtem-peraturearound1000Kandistransgranularforthesolutionannealedspecimens.Theseresultsarethesameasthoseobtainedforthesensitizedtype304underthesameexperimentalconditions[1].

Itiswellknownfortheausteniticstainlesssteelsoftype304andtype316thattheCrcarbidesandCrdepletionzoneareformedat/alongthegrainboundariesby

R.Nishimuraetal./CorrosionScience44(2002)1343–13601355

sensitization[6],thedegreeofwhichdependsuponsensitizingtemperature.ItisevidentthatthedifferencesdescribedabovearecausedbyeithertheformationofCrcarbidesandCrdepletionzoneorlittleformation.ItisconsideredthatCrcar-bidescouldactasanobstacleofdislocationmovement[7],bywhichthesensitizedmaterialswouldbecomeharderthanthesolutionannealedmaterial,similarlytoprecipitationhardeningmaterials.Thiswouldleadtotheincreaseintheultimatetensilestrengthofthesensitizedspecimens,whichispresumedtocontributetotheincreaseinrmax.Inthecaseofthesolutionannealedspecimens,thegrainboundarieswithlittleCrcarbidesdonotserveastheobstacleofthedislocationmovementandhencethedislocationscanmoveoverseveralgrainsbyappliedstress.TheCrdepletionzoneservesasapreferentialsiteforcorrosionattack,whilethepreferentialsiteofthesolutionannealedspecimensisslipstepsemergingonthesurface.

Onthebasisoftheconsiderationdescribedabove,thesensitizingtemperaturedependenceofthevalueoftss=tfisqualitativelyrelatedtootherfactorsasfollows.Theappliedstress,r,ispresentedasWload=Ainitial,whereWloadisaconstantloadandAinitialistheinitialcrosssectionalareaofthespecimens,whileatruestress(rtrue)attss,ispresentedasWload=Ass,whereAssisthecrosssectionalareaattss.ThertruewouldhaveaconstantvalueindependentofWloadforeachsensitizedspecimensbecauseoftheconstantvalueoftss=tfintheSCC-dominatedregion,butitwouldincreasewiththeincreaseinrmax,dependinguponthesensitizingtemperature.UnderaconstantloadconditionðWload¼constantÞ,itwasconsideredthattheratiooftsstotfforthesolutionannealedspecimenscorrespondstotheratioofLsstoLfðLss=LfÞ[8],whereLssisthecrackpropagatinglengthuntiltssandLfisthetotalcrackpropagatinglengthuptotf.Therefore,aslongastheidenticalgeometryofthespecimensisused,theLssneedstoincreasefortheincreaseinrtrue,thatis,thereductioninAss.ThisimpliesthatthevalueofLss=Lfwouldincreaseproportionallywiththeincreaseinrmax(orrtrueattss)andconsequentlythevalueoftss=tfshouldincreasecorre-spondingly.

Thepreferentialsitesforcorrosionattackaretheslipstepsforthesolutionan-nealedspecimensandtheCrdepletionzoneforthesensitizedspecimens,atwhichsitesacrackinitiationwouldtakeplacefollowedbyacrackpropagation.Thedif-ferenceinthecrackinitiationsitewouldbetheprimarycauseforthechangeinfracturemodebetweenthesensitizedandthesolutionannealedspecimens.Inthenextsection,thedifferenceintheslopesandthechangeinfracturemodewillbedescribedinmoredetailthroughaqualitativeSCCmechanism.

4.3.QualitativeIGSCCmechanismandexplanationoftheresultsobtained

4.3.1.AproposalofIGSCCmechanism

TheIGSCCmechanismwillbedescribedinmoredetailstoexplainthepresentresultsbelow,whereinadditiontoalocalstresscausedbyappliedstress,alocalstresscausedbyaninteractionbetweenfilm,dissolutionandGBSatcracktipsispresumedtoleadtoaruptureofthefilm[1].Fromthepreviouspaper[1],withregardtoafilmrupture-formationeventduringatimeintervalofDtintheSCC

1356R.Nishimuraetal./CorrosionScience44(2002)1343–1360

process,anetlocalstressatcracktips(rtip),anetlocalelongation(In)andacor-rosioncurrentdensityatcracktips(js)areexpressedbelowrespectively.

rtip¼rappþrgbs

ð3Þ

whererappiscausedbyappliedstressandrgbsiscausedbytheinteractionbetweenthefilmandGBS.Thenarateofrtipisprovidedas,

drtip=dt¼drapp=dtþdrgbs=dtAsforIn,itispresentedas,

In¼IrþIgbs

ð5Þ

whereIristheelongationrelatedtofilmruptureandIgbsistheelongationcausedbyGBSduringfilmformationuptothenextfilmrupture.ArateofInisexpressedas,

dIn=dt¼dIr=dtþdIgbs=dtThejsisalsoprovidedas,

js¼jfþjdjs/j0ðt=t0Þ;

Àn

ð4Þ

ð6Þ

ð7Þ

jforjd/j0ðt=t0Þ

Àm

ð8Þ

wherejfisthefilmformationcurrentdensity,jdisthedissolutioncurrentdensity,tisatimeduringafilmrupture-formationevent,andj0,t0,nandmaretheconstants,butdependuponsensitizingtemperature.

Furthermore,itisconsideredthatthereareacriticalfracturestress(rF)foreachsensitizedspecimenstocausefilmruptureandathresholdstress(rth)belowwhichnoSCCtakesplace.HerethertipcouldreachrFatacertainbalancebetweenarateoffilmformation(aninhibitingeffectofthefilmonelongation)andadissolutionrate(anenhancingeffectofdissolutiononelongation).Weconsiderqualitativelytheinterrelationshipbetweenrtip,Inandjsbelow.

AslidingdisplacementbyGBS(S)isprovidedas[9],

S¼nb

ð9Þ

wherenisthenumberofdislocationonthegrainboundariesandbistheBurgersvector.FromEq.(9),arateofGBS(slidingrate)isgivenas,

dS=dt¼bdn=dt

ð10Þ

ThedS=dtinEq.(10)shouldbeassociatedwithjdwhichacceleratesdn=dtbyin-jectionofdefectsintograinboundaries[10].Therefore,thefollowingrelationwouldbespeculated,

dS=dt/dn=dt/jd=zF

ð11Þ

Theincreaseinthedegreeofsensitizationleadstotheincreaseinjd(orjs),sothatdS=dtalsoincreaseswithincreasingthedegreeofsensitization.

Ontheotherhand,itisreportedthattherateofrtip,drtip=dt,isrelatedtosanddS=dt[11],

R.Nishimuraetal./CorrosionScience44(2002)1343–13601357

drtip=dt/sdS=dtð12Þ

wheresistheappliedshearstressasr¼2sunderauni-axialcondition.FromEqs.(11)and(12),thefollowingequationisderived,

drtip=dt/sjd=zF

ð13Þ

Eq.(13)impliesthattherateofrtipisproportionaltos(orr)andjd.

Underaconstantappliedstresscondition,rappisinstantaneouslyattainedassoonasaloadisappliedandtheniskeptconstantduringthesteadycrackpropagationperiod(correspondingtothesecondaryregioninthecorrosionelongationcurve).Therefore,drapp=dtwouldbecomezerointhesteadycrackpropagationperiodandEq.(4)issimplifiedas,

drtip=dt¼drgbs=dt

ðinsteadycrackpropagationperiodÞ

ð14Þ

Consequently,therateofrtipissubjectedtoonlydrgbs=dtinthecorrosiveen-vironments.Ifatestsolutionisanon-corrosiveenvironmentsuchasdistilledwater,drtip=dtisexpressedas,

drtip=dt¼drgbs=dt¼0

ðinnon-corrosiveenvironmentÞ

ð15Þ

Becausejsiszeroornegligiblysmallandasaresultdrgbs=dtbecomeszero.

ItisevidentthatthebehaviorofrtipisassociatedwiththatofIn,sothatthefollowingrelationispresumed,

drtip=dt/dIn=dt¼dIr=dtþdIgbs=dt

ð16Þ

wheretheisscouldbeassociatedwithanaveragerateofDIn=Dt,thatis,iss/DInð¼DIrþDIgbsÞ=Dt.TheIrwouldbeinstantaneouslyattainedjustafterfilmruptureaswellasrappandthenkeepconstantduringthesteadypropagationperiod,dIr=dt¼0.Therefore,fromEqs.(13),(14)and(16),thefollowingrelationshipwouldbeestablished,

jd=zF/dIgbs=dt/drgps=dt

ðinsteadycrackpropagationÞ

ð17Þ

Hereitshouldbenotedagainthatthefilmformationatcracktipsisprerequisitetotheformationofrgbs,whilethejdisprerequisitetothedevelopmentofrgbs.4.3.2.AppliedstressandtesttemperaturedependencesofSCCsusceptibility

Inthissection,weconsidertheeffectsofrandtesttemperatureontheSCCbehaviorofeachsensitizedspecimensonthebasisoftheIGSCCmechanismde-scribedintheprevioussection.Undertheconstantenvironmentconditions,exceptr,ifthebehaviorofjs(orjd)foreachsensitizedspecimenswouldbethesamein-dependentofr,drgbs=dtcouldincreasewithincreasings(orr)fromEqs.(13)and(14).ThismeansthatthetimeforrtiptoreachrFbecomesshorterbecauseofrF¼constant.Inaddition,theincreaseinrleadstotheincreaseinrapp,whichmeansthatthecontributionofrgbstortiplessens.Therefore,itisconcludedthattheincreaseinr,leadingto(a)theincreaseindrgbs=dtand(b)theincreaseinrapp,contributestothedecreaseinDt.Inotherwords,thefrequencyofthefilm

1358R.Nishimuraetal./CorrosionScience44(2002)1343–1360

rupture-formationeventincreaseswithincreasingrtoleadtothedecreaseintfandtheincreaseiniss.

Ifrbecomeslargerthanrmax,rappwouldbecomelargerthanrFandhencethergbswouldhavelittleneededforthefractureofthespecimens,whichcorrespondstothecaseinthestress-dominatedregion.Ontheotherhand,ifrbecomeslessthanrmin,thecontributionofrapptortipwouldbecomeverysmallandhencetheincreaseinrtiptoreachrFissubjectedtoonlythatofrgbs.Inthiscase,atimeforrtiptoreachrFwouldbecomelongandjswouldbecomecomparabletoageneralcorrosioncurrentdensityonthesurface.Asaresultthefractureofthespecimenswouldtakeplacebythereductionincrosssectionalareathroughgeneralcorrosion.Thisisthecaseinthecorrosion-dominatedregion.

Undertheconstantappliedstresscondition(risconstant),butundervariableconditionsofenvironment,thatis,rapp¼constantandhencedrapp=dt¼0,sothatdrtip=dt(ordrgbs=dt)dependsupononlyjdfromEqs.(13)and(14).Asthetesttemperatureincreases,jdforeachsensitizedspecimensincreasescorrespondingly,wherejddoesnotshowthesamebehaviorasthetesttemperaturechanges,showingthatnorminEq.(8)woulddecreasewithincreasingtesttemperatureandviceversa.Consequently,Dtwoulddecreasebytheincreaseindrgbs=dtwithincreasingthetesttemperatureandviceversa,toleadtothedecreaseintfandtheincreaseiniss.Thusthetesttemperaturedependenceofissandtfcanbequalitativelyexplainedforeachsensitizedspecimens.

Furthermore,intheterminalcrackpropagationperiodaftertss,theincreaseinrtruebycrackpropagationbeginstoaffecttheelongationbehavior(therapidincreaseinelongation),whichmeansthattheincreaseinrappaffectstheelongationbehavior.Therefore,thecontributionofrapptortipincreaseswithtimeandthatofrgbsde-creases.Finally,rappitselfreachesrFtoleadtoamechanicalfractureofthespeci-mens.

4.3.3.SensitizingtemperaturedependenceofSCCsusceptibility

ThedifferenceofSCCsusceptibilitybetweenthespecimenssensitizedatdifferentsensitizingtemperaturesisdiscussedundertheconstantappliedstress(rapp¼constant)andenvironmentconditionsinthissection.Itisrecognizedthatjdincreaseswithincreasingthedegreeofsensitization,sincethedegreeofCrdepletionincreasescorrespondingly.Thisimpliesthatasthedegreeofsensitizationincreases,drgbs=dtfromEq.(17)increasewiththeincreaseinjdandthenDtdecreaseswithincreasingthedegreeofsensitization.Ontheotherhands,themagnitudeofrFisconsideredtodependuponthedegreeofsensitization,correspondingtothatofrmax.TheincreaseinrFwouldcontributetotheincreaseinDtunderthesameexperi-mentalconditions,sincerappisconstantandhencethecontributionofrgbsforrtiptoreachrFincreases.Thus,toexplaintheeffectofthedegreeofsensitizationonthebehaviorofrgbs,thebehaviorofjdandrFwiththereverseeffecteachothermustbetakenintoconsideration.However,itispresumedthattheincreaseinjdwouldbesuperiortotheincreaseinrF,toexplainthefactthatthetfbecomestheshortestattheseverestsensitizingtemperatureofaround1000K.

R.Nishimuraetal./CorrosionScience44(2002)1343–13601359

Furthermore,inthecaseofthesolutionannealedspecimenssuchastype304,type316,type310[12]andtype430(ferriticstainlesssteel)[13],wheneverthevalueofisshasalargerone,thevalueoftfbecomesalwayssmaller.ThismeansthatthevalueofissisproportionaltotheDIn=Dt.However,whenthevaluesofissandtfforthesensitizedspecimensarecomparedwiththoseforthesolutionannealedspecimensatacertainconstantappliedstress,asshowninFigs.3and4,thevalueofissbecomessmallerthanthatforthesolutionannealedspecimensatanappliedstressoflessthan400MPa,whereasthevalueoftfissmallerforthesensitizedspecimensthanforthesolutionannealedspecimens.ThismaysuggestthattheproportionalrelationshipbetweenInandI(thenetelongationinFig.2)forthesensitizedspecimensisdifferentfromthatforthesolutionannealedspecimens.Inotherwords,theelongationbe-haviorbyGBSisdifferentfromthatbydislocationmovementinthegrainsforTGSCC[5]withrespecttoitsappliedstressdependenceanditsmagnitude.Theincreaseoftheformerelongationwouldbecomesmallerthanthatofthelatterbe-causeoftheexistenceofCrcarbidesunderanidenticalappliedstresscondition.Therefore,thefollowingrelationshipwouldbepresumed,

iss/aDIn=Dt

ð18Þ

whereaisamodifiedfactoranddecreaseswithincreasingthedegreeofsensitizationintherangeof0–1.EvenifthevalueofDIn=Dtbecomesthesameforthesensitizedandthesolutionannealedspecimens,thevalueofisswoulddecreasewithincreasingthedegreeofsensitizationandviceversa.

4.3.4.Changeinfracturemode

Thefracturemodeofthesensitizedtype316specimenschangesfromintergran-ulartotransgranular,dependinguponthedegreeofsensitization.Thiswouldbeexplainedasfollows.Asthedegreeofsensitizationdecreases,thedegreeofCrde-pletionandCrcarbidesdecreasescorrespondingly.Thedislocationmovementre-strictedwithinonegrainbyCrcarbidescouldbegintotakeplaceoverseveralgrainspartlywithdecreasingthedegreeofsensitizationtocausethecompetitionbetweenGBSanddislocationmovement.WhendislocationmovementbecomessuperiortoGBS,thefractureappearancewouldbecomepredominantlytransgranular.

Inaddition,atasensitizingtemperatureofabove1100Kandbelow850K,thesensitizationaffectsslightlytheslopesinEqs.(1)and(2),andthevalueoftss=tfasshowninFigs.7,11and13,butlittleintergranularfracturemodeappears.Mulford,etal.[14]pointedoutthatsolutesegregationsuchasphosphorousandsulfuratgrainboundariestakesplaceevenforthesolutionannealedspecimensaswellasforthesensitizedspecimens.Inthesesensitizingtemperatureranges,therefore,wewillneedtoconductfurtherexperimentstoclarifytheroleinsolutesegregationonSCCbehavior.

5.Conclusions

TheSCCcharacteristicsofthesensitizedtype316specimensaresummarizedasfollows:

1360R.Nishimuraetal./CorrosionScience44(2002)1343–1360

(1)Curvesofappliedstress––threeparameters(steadystateelongationrate,iss,timetofailure,tf,andtheratioofthetimeintervalofSCC-dominatedfailure,tss,totf)curvesaredividedintothreeregions,thatis,thestress-dominated,SCC-domi-natedandcorrosion-dominatedregions,respectively.

(2)Thevalueoftss=tfissmallerthanthatofthesolutionannealedspecimenoverthewholesensitizingtemperature,butitholdsconstantindependentofappliedstressandtesttemperature.

(3)TherelationbetweenissandtfobtainedasafunctionofappliedstressbecomesagoodstraightlinewithaslopeoflessthanÀ2ofthesolutionannealedspecimens.Underaconstantappliedstresscondition,italsobecomesagoodstraightlinewithaslopeoflessthanÀ1ofthesolutionannealedspecimen.However,theslopeforthesensitizedspecimensdependsuponsensitizingtemperature.

(4)Fromtheitem(3),thesteadystateelongationrate,iss,becomestheparameterforpredictionoftimetofailure,tf.

(5)Thefracturemodebecomespredominantlyintergranularatasensitizingtemperatureofaround1000K.

(6)ThepresentresultsobtainedareexplainedintermsofGBS,Crcarbides,Crdepletion,filmformationandcorrosioncurrentdensityatcracktipsalongwiththequalitativeIGSCCmechanismalreadyproposedforthesensitizedtype304speci-mens.

References

[1][2][3][4][5]

R.Nishimura,I.Katim,Y.Maeda,Corrosion,inpress.R.Nishimura,K.Kudo,Corrosion45(19)308.R.Nishimura,Corrosion46(1990)311.R.Nishimura,Corros.Sci.34(1993)1859.

R.Nishimura,K.Yamakawa,ProceedingsofPassivityandLocalizedCorrosion,in:M.Seo,B.MacDougall,H.Takahashi,R.G.Kelly(Eds.),ElectrochemicalSocietyPV99–27,ElectrochemicalSocietyInc.,Pennington,NJ,USA,2000,pp.624–633.

V.Cihal,IntergranularCorrosionofSteelsandAlloys,MaterialsScienceMonographs,vol.18,Elsevier,Amsterdam,1984.

G.E.Dieter,MechanicalMetallurgy,seconded.,McGraw-Hill,NewYork,1976,p.195.R.Nishimura,Corrosion49(1993)796.

H.E.Evans,MechanismsofCreepFracture,Elsevier,Amsterdam,1984,p.169.B.Gu,W.-Y.Chu,L.-J.Qiao,C.-M.Hsiao,Corros.Sci.36(1994)1437.H.L.Logan,J.Res.Natl.Bur.Stand.48(1952)99.R.Nishimura,Corros.Sci.34(1993)1463.R.Nishimura,Corrosion48(1992)882.

R.A.Mulford,E.L.Hall,C.L.Briant,Corrosion39(1983)132.

[6][7][8][9][10][11][12][13][14]