Nanomechanical characterization of nanostructured bainitic steel: Peak Force Microscopy and Nanoindentation with AFM

3/3/2016 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForceMicroscopyandNanoindentationwithAFM http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4658537/ 1/14 Goto: SciRep.2015;5:17164. Publishedonline2015Nov25.doi:10.1038/srep17164 PMCID:PMC4658537 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForce MicroscopyandNanoindentationwithAFM LuciaMoralesRivas, AlejandroGonzálezOrive, CarlosGarciaMateo, AlbertoHernándezCreus, FranciscaG.Caballero, andLuisVázquez DepartmentofPhysicalMetallurgy,NationalCenterforMetallurgicalResearch(CENIMCSIC),Avda.GregoriodelAmo8,28040Madrid,Spain DepartamentodeQuímicaFísica,FacultaddeQuímica,UniversidaddeLaLaguna,38206Tenerife,Spain InstitutodeCienciadeMaterialesdeMadrid,(ICMMCSIC),Cantoblanco,28049Madrid,Spain Email:[email protected] Received2015Jun2;Accepted2015Oct1. Copyright©2015,MacmillanPublishersLimited ThisworkislicensedunderaCreativeCommonsAttribution4.0InternationalLicense.Theimagesorotherthirdpartymaterialinthisarticleareincluded inthearticle’sCreativeCommonslicense,unlessindicatedotherwiseinthecreditline;ifthematerialisnotincludedundertheCreativeCommonslicense, userswillneedtoobtainpermissionfromthelicenseholdertoreproducethematerial.Toviewacopyofthislicense,visit http://creativecommons.org/licenses/by/4.0/ Abstract Thefullunderstandingofthedeformationmechanismsinnanostructuredbainiterequiresthelocalcharacterization ofitsmechanicalproperties,whichareexpectedtochangefromonephase,bainiticferrite,toanother,austenite. ThisstudybecomesachallengingprocessduetothebainiticnanostructurednatureandhighYoung’smodulus.In thiswork,wehavecarriedoutsuchstudybymeansofthecombinationofAFMbasedtechniques,suchas nanoindentationandPeakForceQuantitativeNanomechanicalMapping(PFQNM)measurements.Wehave addressedcriticallythelimitsandadvantagesofthesetechniquesandbeenabletomeasuresomeelastoplastic parametersofbothphases.Specifically,wehaveanalyzedbyPFQNMtwonanostructuredbainiticsteels,witha finerandacoarserstructure,andfoundthatbothphaseshaveasimilarYoung’smodulus. Nanostructuredbainitebelongstoanewgenerationofadvancedsteelswithimprovedmechanicalproperties, presentingthehigheststrength/toughnesscombinationseverrecordedinbainiticsteels(2.5 GPa/30 MPa m ) .Thedifficultytoanalyzetheflowbehaviorofnanostructuredbainitearisesfromthecomplexityofits structureandfromthecombinationofdifferentdeformationstrengtheningmechanisms,includingthemechanically inducedtransformationofausteniteintomartensite.Understress,nanostructuredbainiterespondsasacomposite likestructurewiththestressandstrainpartitioningbetweenthephasesbainiticferriteandretainedaustenite,as reportedforothermultiphasesteels . Thedeterminationoflocalelasticpropertiesisapreviousandcrucialsteptobeabletounderstandhowmechanical partitioningoccursandthus,howthiscanaffectthemacroscopicflowbehaviorofthematerial.Infact,whena multiphasematerialissubjectedtoamacroscopicmechanicalload,attheelasticregime,thephasewithahigher Young’smoduluswouldundertakemostofthestress,exertingthusashieldingeffectoverthephasewithalower E.Thiscompositetypebehaviorinfluencesthepointatwhichyieldingstartsandthewayittakesplace,i.e.,which phasereachesfirstitsplasticregime.Moreover,elasticpropertiesareanimportantfactorinfluencingthe developmentofresidualstressesformedasaconsequenceofinteractionsamongotherssuchastime,temperature, deformationandmicrostructure . Elasticpropertiesareexpectedtodifferfromonephasetoanotherandbothfromthecorrespondingmacroscopic valueobtainedfromamechanicaltest.Sincetheirphasesarenotpossibletobeprocessedasmonophasematerials, theelasticpropertiesofeachphasemustbeindividuallymeasured.Therefore,traditionalbulkmeasurementsof 1 2 a,1 2 1 3 1 2 3 a 1/2 1,2,3,4 5,6,7 8 3/3/2016 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForceMicroscopyandNanoindentationwithAFM http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4658537/ 2/14 Goto: elasticpropertiesbasedonultrasonicresonance,whichisnotphasesensitive,arenotsuitableformultiphase structures .Instead,andspecificallyforthedeterminationofYoung’smodulus,twodifferentexperimental techniquesarecommonlyemployed: 1Insitudiffractiontests,wheretheYoung’smoduluscanbeestimatedbyconsidering,intheelasticregime,the evolutionofthelongitudinallatticestrainwiththestress.ThistechniquegivesinformationoftheaverageYoung’s modulusforeachphaseandforaspecificfamilyofplanesconsidered .So,basedondiffractiontechniques, differentYoung’smodulusvaluesarecalculatedfromtheaveragelatticespacingofagroupofgrainswithplane normals[hkl]directedalongthescatteringvector,whichincludesallcrystalorientationsrotatedaroundthat particular[hkl] .Therefore,Young’smodulusthusobtainedreflectstheanisotropicnatureofthesinglecrystal. Ontheotherhand,thereisalsoanaveragingwhichoriginatesfromthelargevarietyofmechanicalboundary conditionsofthisparticulargroupofgrains,eachsurroundedbyothergrainsofunknownorientation .Thefactthat thestressundertakenbyagivengroupofgrainsisnotdirectlymeasuredisadisadvantageofthismethod.This mustbeestimatedassumingaparticularstresspartitioningmodel,suchasReussmodelthatimplicitlyassumesthe macroscopicstress . 2DeterminationoflocalYoung’smodulushasbeenintensivelyperformedbymeansofnanoindentation.The analysistypicallyconsistsintheexaminationoftheunloadingcurve,forwhichanalyticalsolutionsfordifferent indentergeometriesareused .Thismethod,however,doesnotaccountforthepileup,sinkin, andtipbluntingeffects,whichchangethevalueofcontactarea,akeyparameteronwhichresultsarestrongly dependent .Moreover,whendealingwithananocrystallinestructure,asitisourcase,theYoung’smodulus measuredthroughthisprocedureislikelytobeanaverageofthebulkmaterial,sincethefootprintgeneratedcan haveasizesimilarandevenhigherthanthemicrostructuralfeaturestobemeasured.Inaddition,inthecaseof nanostructuredbainite,asasteelpresentingTransformationInducedPlasticity(TRIP)effect,theindentationmight inducethemartensitictransformationofaustenite,sotheinitialYoung’smodulusvalueisexpectedtodifferfrom themeasuredoneduetothisphasetransition. Inprinciple,AFMbasedtechniquescouldbeappositetoaddressthestudyofthenanobainiticlocalelastic propertiesatthenanoscale.AFMworkingatdifferentmodessuchasforcecurveanalysisorPeakForce QuantitativeNanomechanical(PFQNM)couldbeapropertechniquetoobtaintheYoung’smodulusvaluesof thesenanocrystallinestructuresduetoitsimprovedlateralresolutionandcapabilitiestoperformveryshallow indentations(i.e.atthenanoscale).Moreover,inthePFQNMmode,itwouldbepossibletoobtainsimultaneously morphologicalandstiffness(i.e.Young’smodulus)imagesofthesamearea.Therefore,localspatialvariationsof Young’smodulusduetotheheterogeneityofthemicrostructurecouldbeobserved. Theaimofthisworkistodeterminethelocalelasticpropertiesofnanoscalecompositelikebainiticstructures, usingthecombinationofAFMbasedtechniques.Inspecial,thesuitabilityofPFQNMtoobtaintheYoung’s modulusdistributioninthismaterialhasbeenexplored.Aparticularemphasishasbeenpaidtodiscernwhether bainiticferriteandretainedaustenitehavesimilarordissimilarYoung’smodulus.Forthispurpose,twodifferent samples,transformedintobainiteatdifferenttemperatures,havebeenanalyzedbyPFQNM.Inaddition,results havebeencomparedtothoseobtainedfromtheanalysisoftheloadingforcecurveofsingleindentationsperformed byAFM.Resultshavebeendiscussedintermsofthenatureandscaleofthemicrostructuresandhaveprovided newexperimentalevidenceonthemechanicalbehaviorofnanoscalecompositelikebainiticstructures. MaterialandExperimentalProcedure Material ThechemicalcompositionofthesteelisFe1.0C–2.5Si–0.75Mn–0.12Ni1.0Cr0.03Mo0.2Cu(wt.%).Theheat wasindustriallymanufacturedviaelectricarcfurnaces.Oncesolidified,theingotwasreheatedat1200 °Candhot rolledtoa35 mmbar,whichwasafterwardslowlycooleddowninafurnacetoavoidcracking.After austenitizationat950 °Cduring60 minand15 minforHT250andHT350,respectively,twoisothermal transformationtreatmentswereselected:250 °Cduring16 h(sampleHT250),toachieveananoscalebainitic structure,and350 °Cduring480 min(sampleHT350),toachieveasubmicronbainiticstructure.Both temperaturesareinthebainiticrangeandtimesarelongenoughtoensuretheendofthetransformation. 9 10 11 9 10 12,13,14,15,16,17,18,19 20 3/3/2016 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForceMicroscopyandNanoindentationwithAFM http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4658537/ 3/14 FortheAFMexperimentsthesamplepreparationofundeformedmaterialconsistedofafirststageofgrounding, followedbypolishingusing3 μmand1 μmdiamondpasteandseveralcyclesofetchingusing2%Nital(2%nitric acidinethanol)andpolishing,inordertoremovethedeformedlayer.Sampleswerefinallypolishedusingcolloidal silicasuspension.Afterthat,twodifferentsurfacefinishingprocesseswereappliedperheattreatment:etching (HT250andHT350withetchedsurface)orcleaning(HT250andHT350withpolishedsurface).Etchingwas carriedoutusing2%Nital.UponthecorrosivechemicaletchingwithNital,theetchantremovesthetoplayerofthe metalinaselectiveway,dependingonthephaseandthecrystalorientation,asopposedtothesocalledcolor etching,inwhichathinfilmisbuiltonthetopofthemetal .Inturn,cleaning,intendedtoremoverestsofsilica colloidal,wasmadeusingasolutionof96 mlwater,2 mlammoniawater(30%)and2 mlhydrogenperoxide(3%), basedon . TEMspecimenswereslicedfrom3mmdiameterrodsoftheheattreatedmaterial,mechanicallythinnedto0.06 mm,andthentwinjetelectropolishedtoperforationusingamixtureof5%perchloricacid,25%glyceroland70% ethanolat10 °Cat45 V.ThesampleswereexaminedonaTEMJEOL2010transmissionelectronmicroscope operatedat200 keV. QuantitativeXRDanalysiswasusedtodeterminethelatticeparameterofferriteandausteniteinHT250and HT350samples.TheywerethenstepscannedinaBruckerAxsD8XraydiffractometerusingunfilteredCoKα radiation.XraydiffractionmeasurementswerecarriedoutwithaBrukerAXSD8diffractometerequippedwitha CoXraytube,GoebelmirroropticsandaLynxEye6.LinearPositionSensitiveDetectorforultrafastXRD measurements.Thistypeofradiationisespeciallysuitedforironrichsamplestoavoidthestrongfluorescence arisingfromcopperradiation,andtoproducehighresolutiondata.Acurrentof30 mAandavoltageof40 kVwere employedastubesettings.OperationalconditionswereselectedtoobtainXRDprofilesofsufficientquality: namely,optimalcountingstatistics,narrowpeaksanddetectionofthesmalldiffractionpeaksofminorphases.The XRDdatawerecollectedovera2θrangeof35–135°withastepsizeof0.01°.Thelatticeparameterofthephases wasdeterminedfromtheRietveldmethodtoattainahighdegreeofprecision . AdvancedAFMtechniques Asnotedabove,themainaimofthisworkistodeterminethemechanicalpropertiesofnanostructuredsteelsatthe nanoscaleinordertobeabletoassesswhetherbothphaseshavedifferentYoung’smodulusornot.Accordingly, themostsuitabletechnique,inprinciple,toaddresssuchinvestigationistherecentlyintroducedatomicforce microscopymode,i.e.,peakforcequantitativenanomechanical,PFQNM,imaging,whichallowstomapthelocal elasticpropertieswithlateralnanometerresolution .Thismethodisbasedontheacquisitionofforcecurves recordedateachpixelofthetopographicimage.Theforcecurvesareanalyzedinstantaneously.Then,provided thatparameterssuchascantileverspringconstantandtipgeometryarecalibrated,thequantificationofthenano mechanicalpropertiesispossible.InthepresentconfigurationthisanalysisisdonethroughtheDerjaguin–Müller– Toporov(DMT)fitmodeloftheretracingcurve forasphericalindenter,whichaccountsfortheadhesion effectsbetweenthesamplesurfaceandthetip,accordingtotheequation: Thus,Fistheappliedforce,Risthetipradius,disthedeformationvalueatagivenforce,F themaximum adhesionforce,andE isknownastheeffectiveelasticmodulus,whichisdefinedas: WithE(ν)andE (ν )beingtheYoung’smodulus(Poisson’sratio)ofthesampleandindenter,respectively. Asthesecalculationsaredoneinrealtimeforeveryforcedistancecurveobtainedateveryimagingpixel,the topographyandnanomechanicalpropertiescanbesampledsimultaneously.Moreover,thelatterdataarealso displayedastopviewimagejustasthetopographicalone(thezaxisbeingtheYoung’smodulus,E).Thismethod thenallowstomeasurethelocalnanomechanicalpropertieswiththesamelateralresolutionthaninthe 21,22 23,24 25 26,27 28,29 adh * i i 3/3/2016 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForceMicroscopyandNanoindentationwithAFM http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4658537/ 4/14 topographicalimage.Thisfactturnsitasverysuitabletechniquetoassessthesepropertiesinthenanostructured bainiticsteels.However,itisworthnotingthatthistechniquehasbeenmostlyemployedtostudythenano mechanicalpropertiesofsoftsurfacessuchaspolymers ,biologicalentitiesasfibrils ,livingcells and evennanobubbles .Incontrast,scarcestudieshavebeendevotedtosamplethenanomechanicalpropertiesof hardsurfaces.Thus,PFQNMhasbeenappliedtostudytheYoung’smodulusofWC/aCcoatings,inthe100– 200 GParange ,nanostructuredGesurfaces,inthe50–100 GParange ,andhardenedcementpaste,upto100 GPa . Specifically,quantitativemappingwasperformedatroomtemperaturewithaMultimode8(Veeco)andusinga NanoscopeVcontroller(Bruker).SampleswereimagedbyusingAFMoperatinginPFQNMinairatascanrate of1.0–1.2 Hz.TheAFMprobeusedinthesestudieswasadiamondtip,modelMDNISPHS,(Veeco),witha resonantfrequencyof63 kHz,springconstantof402 N/m,andnominalradiusof45 nm.Theloadingforcesduring themeasurementswerekeptcloseto2 μNrange.Thediamondtipwascalibratedinitiallyonsapphiresurfacesin ordertosetthedeflectionsensitivity.Oncecalibrated,thetipwasusedforPFQNMimaging.Itshouldbenoted thatthelaserspotwaskeptatthesamelocationonthecantileverduringallthemeasurements.Finally,itshouldbe commentedthatthediamondtipcaneventuallybecomecontaminatedordirtyduringtheQNMmeasurements. Whenthisoccurs,thetipiscleanedbyperformingindentationsonagoldsurface. Forthesecondsetofexperiments,i.e.,thenanoindentationanalysis,aNanoscopeIIIa(Veeco)equipment operatingintappingmodewasemployed.Inthiscase,adiamondprobe(DNISP,Veeco)witharesonant frequencyof64 kHz,nominalradiusofcurvatureof45 nm,andaspringconstantof247 N/mwasused.Thesame calibrationandcleaningproceduresasthosedescribedintheQNMexperienceswerefollowedalsointhiscase. Now,themeasurementprocedurewasasfollows:First,atappingimagewasobtainedandthenanarrayof indentationswasperformedbychangingthelateralpositionofthetipoverthesurface.Ateachspotofthearraya singleforcecurve,atarateof3.5 Hzandcontrollingthemaximumloadforce(thehighestvalueappliedwasof 200 μN),wasdone.Inafinalstep,thetipwaschangedforasiliconone,withanominalradiusofcurvatureof8 nm,inordertolocatetheareainwhichtheindentationarraywasmadeandthenmeasureitincontactmodein ordertohavemorereliableinformationontheindentationgeometry(i.e.,areaanddepth)sincethediamondtipis toowide.Theanalysisoftheforcecurves(seebelow)intermsofthequantificationoftheYoung’smodulusis doneattheinitialstagesoftheloadingcurve,i.e.attheelasticregime,byapplyingtheHertzmodel,inwhich adhesioneffectsarenotaccountedfor,fordifferentindentergeometries .Specifically,wehavefittedtheloading forcecurvesforspherical,conicalandberkovitchindentergeometries,andwehavefoundthattheoneleadingto thebestfittingswasthesphericalone.Thisfactallowsustobettercomparetheanalysisofthesingleindentations withthedataobtainedbyPFQNMsincebothanalysesaredoneforthesametipgeometry. Forthisgeometry,therelationshipbetweentheappliedforce,F,andtheindentation,δ,is: whereRistheindenterradiusandE isknownastheeffectiveelasticmodulus,whichwasdefinedineq.2. Providedthatthecantileverhasbeencalibratedonahardsurface(sapphire),itsforceconstantandRvaluesare known,thecurveF = F(δ)canbeobtainedandthenfromalogarithmicplotthevalueofEcanbederived.Atthis point,itisworthtocommentontheerrorsinvolvedinthesemeasurements.Themainsourceoferrorsisthevalue ofR,whichisquitedifficulttoknowexactly(aswellastheexactindentershapeattheveryendofthetip). However,inthisworkwearemostlyinterestedindetectingdifferencesbetweentheEvaluesonthedifferent phasespresentinthenanostructuredsteelthanindeterminingtheirexactvalues. Thegeometryofthetipwasmeasuredfollowingthisprocedure:Asampleofgold,consideredasaperfectlyplastic metal,wasnanoindentedwiththediamondtip,andthefootprintswerethenscannedusingafinertappingtip, Fig.1a.Arelationbetweendepthandareaofthetipwasobtained,plottedinFig.1b.Thedepth,h,ismeasured withrespecttothelowestheightvalueofthetopographicalmapscannedovertheindentation.Theprojectedareais theareaofthefootprintsectionparalleltotheview’splane,ateverydepthvalue,whereastherealareaaccountsfor thewholesurfacebelow,i.e.,thegoldsurfaceincontactwiththetip.Besidesthetargetsamples,HT250and 30,31 32 29,33 34 35 36 37 38 * 3/3/2016 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForceMicroscopyandNanoindentationwithAFM http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4658537/ 5/14 Goto: HT350,fusedsilicawasusedasareferencesample,nanoindentedwithAFMinordertoruleoutpossibleartifacts inforcecurvesnotrelatedtothematerialmechanicalbehavior. Figure1 Topographicalmapofthesurfaceofagoldsample,scannedwithafinetip, andshowingthenanoindentationfootprintmadebyadiamondtip(a),with itscorrespondingareadepthrelationship(b). Atypicalarray,performedonHT350,ofsevenrowseachcontainingelevennanoindentations,asmeasuredin contactmode,isdisplayedinFig.2.Itisobservedthattheindentationrowspresentinsomecasessomesortof distortionregardingtheirlateralposition.Thisisduetothelateraldriftofthepiezoelectric.Theholesaremostly∼1 μmapartfromeachother.Ateachpointthecorrespondingforcecurvehasbeenrecordedandanalyzedinorderto obtaintheEvalue(seebelow). Figure2 TopographicalmapoftheetchedsurfaceofHT350sample. Results Initialmicrostructure Nanostructuredbainiteisformedbyabodycenteredcubic(bcc)nanoscalematrix(ferrite)andafacecentered cubic(fcc)secondphase(austenite).BothphasescanbeobservedinthesecondaryelectronSEMimage,Fig.3,of etchedsurfacesamples.Austenite(γ),whichisthedispersedsecondphase,hastwodifferentmorphologiesnamely, thinfilmsbetweentheplatesofferrite(α),bothinthenanoscalerange,andcoarseblocks,uptothesubmicron scale.FromFig.3,itisevidenttheeffectoftheheattreatmenttemperatureonthescaleofthemicrostructure:The sampletreatedathighertemperature,HT350,exhibitsmuchcoarsermicrostructuralfeaturesthanHT250. MicrostructuraldifferencesbetweenbothsamplesgobeyondthoseobservedbySEM,andhighermagnification techniquesarerequiredtodetectthecomplexityandnanoscalenatureofthesestructures.Bainitereactionoccurs viaadisplaciveanddiffusionless,solidsolidphasetransformation,wherethereisnochangeinthechemical compositionbetweentheparentandproductphase.Transformationisaccompaniedbyplasticrelaxationofthe shapechangeoccurringasaconsequenceofthementioneddisplacivegrowthofbainite,whichtakesplacesvia generationofboth,dislocationsintheaustenite/bainiticferriteinterfaceaswellasviamicro/nanotwinsinthe austeniteincontactwithbainiticferrite.Thisplasticrelaxationproducestheappearanceofinhomogeneusresidual microstrains.TheTEMimageofHT250inFig.4showsbainiticferriteplates(brighterregions)betweenfilmsof retainedaustenite(darkerregions),exhibitingevidencesofdislocationdebrisandotherdefectsas nanotwins .Oncethediffusionlessgrowthofabainiticferritesubunithasended,theexcessofcarbon partitionsintothesurroundedaustenite.Theprocesscontinuesbysuccessivenucleationofsubunitsuntilthecarbon concentrationoftheresidualaustenitereachesthevalueatwhichthefreeenergyofferritebecomeslessthanthatof austeniteofthesamecomposition;thetransformationstopsatthatpointbecauseitisthermodynamicallyimpossible forthetransformationtoproceedbyadiffusionlessreaction .Carbonenrichmentmakestheaustenitethermally stable,avoidingmartensitictransformationuponcoolingatroomtemperature.Carbonliesnotonlyatdefectfree solidsolution,butalsoatthehugeamountofdislocations,boundaries,andclusterspresent.However,during bainiticreaction,massivecarbideprecipitationisavoidedthankstotheuseofSiasalloyingaddition.Therefore,the twofinalphasesarethedepictedbainiticferriteandretainedaustenite,whoselatticeparametersarelistedinTable1 .Asithasbeenrecentlyproven,bainiticferriteistetragonalratherthancubic,increasingitscarbon solubility . Figure3 SecondaryElectronSEMmicrographofHT250sample(a)andHT350(b). Nonetchedregionscorrespondtoaustenite(γ)whichdisplaystwodifferent morphologies,blockandthinfilms,whereasetchedregionscorrespondto 39,40,41 42 43,44,45 3/3/2016 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForceMicroscopyandNanoindentationwithAFM http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4658537/ 6/14 bainiticferrite(α).... Figure4 TEMmicrographofsampleHT250,detailingthepresenceofaustenite (γ)andbainiticferrite(α),anddefects. Table1 Latticeparametersofaustenite(a )andbainiticferrite(a andc ). Elasticproperties Loadingforcecurvesrecordedduringthenanoindentations,areanalyzedattheirverybeginning,i.e.,withinthe elasticregime.ThemethodconsistsinfittingthefirststageofthecurvetotheHertziansolutionofEq.(1)sothatE canbeobtained,assumingthatR = 45 nm;K = 247 N/m;ν = 0.3. Thenanoindentationprocesswasfirstperformedonahomogeneousandisotropicsample,fusedsilica,usedasa reference,withaknownEof75 GPaandaPoisson’sratio,ν,of0.17.Theforcecurvesanalysisresultedinan averageEvalueof75 ± 15 GPa.Theobtainederrorrangeshouldbe,inprinciple,mostlyattributedtotheintrinsic limitationsofthetechnique. ThesameprocedurewasfollowedforsampleHT250.Theaimofthisstudywastobeabletoidentifythephase natureofeachspotinwhichanindentationhadbeenperformed.Inprinciple,thiswouldbepossiblebecausethe etchingprocesswithNitalwouldhaverevealedbothphases,i.e.,theaustenitelyingatmoreelevatedlocationsthan theferrite,thelatterbeingpreferablyetched.However,thistaskprovedtoberatherdifficultduetothe nanostructurednatureofthesteel.Thus,mostnanoindentationfootprintsinHT250werefoundovermorethanone phaseoratgrainboundariesturningthecorrespondingphaseidentificationquiteambiguous.Anexampleisshown inFig.5awhereitisobservedthateventhoseindentationsperformedontheaustenitephase,i.e.thesmooth plateaulikeregions,arecloseto,orevenaffectto,etchedzones.Thisproblemalsoisevidentinthecentral indentationthatwasmadeonapresumedferritenanodomain.Inthiscase,itisalsoclearthattheindentationhas beenmadeonalocallyroughorsteppedsurface.Thus,itshouldbestressedthatitwasstatisticallyquite improbabletoindentunambiguouslyinasinglephase. Figure5 DetailsoftopographicalmapsoftheetchedsurfaceofHT250withthree numberednanoindentationsproducedbyAFM(a);andtheircorresponding forcecurves(b). InFig.5bthecorrespondingforcecurves,plottedasforceversusindentation,aredisplayedinwhichtheHerztian regimeisshown.Also,itisworthnotingthatattheendofthisregimethereisaclearcrossovertoalinearregime, whichisrelatedtotheyieldingprocess(seebelow).However,whenweanalyzedtheHertzianregime,weobtained Evaluesthatwereappreciablylowerthantheexpectedones.Thisfactcouldbeduetotherelativelyjagged morphologyonwhichmostoftheindentationswereperformedasaconsequenceoftheetchingofthe nanostructuredbainite.InordertoovercomethisproblemandtotrytomeasurereliablytheEvalueofbothphases, wefollowedthestrategyofstudyingthesamesteeltreatedatahighertemperature,inordertoobtainacoarser bainitestructure(sampleHT350).Thegoalofthisapproachwastoperformunambiguouslytheindentationson welldefinedphases.TypicalexamplesoftheseexperimentsareshowninFig.6a,b.Now,nanoindentation footprintscanbeseenoverclearlyidentifiedphases.However,onlyinthecaseofaustenite,ithasbeenpossibleto identifyunambiguouslythephaseinahighnumberofnanoindentationfootprints.Thecorrespondingforcecurves areshowninFig.6c.ThereisagoodfittingbetweenthetheoreticalHertziancurveandtheexperimentaldataupto thecriticalpointwhereplasticregimestarts,atwhichbothcurvesdiverge.Theonsetoftheplasticregimeandits evolutionwillbediscussedlater,afterelasticpropertieshadbeenevaluated. γ α α 3/3/2016 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForceMicroscopyandNanoindentationwithAFM http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4658537/ 7/14 Figure6 DetailsoftopographicalmapsoftheetchedsurfaceofHT350inwhichfootprintsof nanoindentationsarenumbered(a)and(b);andtheircorrespondingforcecurves(c).The nanoindentationfootprint6(b),inferrite,ispresentedtogetherwithitssurface... FromFig.5,EvaluesofbainiticferriteinHT250resulttobeextremelylow,comparedtothosereportedin literaturewhereformoststeelsEhasavalueof200 ± 15 GPa .Theseresultscanbeduetothefinerscaleofthe HT250microstructure,implyingmoreartifactscausedbytheetchedinducedtopography.Ifwerestrictourstudyto thecaseoftheaustenitephasewhichiscoarserandflat,theanalysisofthewholenanoindentationarraysrevealeda widespreadofvaluesforHT250austenite:50 GPa < E < 200 GPa.Incontrast,forausteniteofHT350,therewere moreanalyzablecurves,enoughtobuildareliablenormalizedEdistribution,Fig.7,withEvaluesofaustenite, between120and160 GPa,i.e.,slightlylowerthanthemacroscopicEvalue,about180 GPa.Thespreadinthe measuredEvaluesmaycomefromdifferentsources.Asexplained,anapprox.errorof20%isinherenttothe technique.ButdifferencesinEvaluesmayalsoarisefromthecomplexityofthebainiticmicrostructures, comprisingtwodifferentandheterogeneousphases,eachconsistingoffeatureswithdifferentcrystallographic orientationsandalsoaheterogeneouscarbondistributioninbothsolidsolutionandlocatedatdefects.Infact,when consideringonlytheelasticanisotropy,Ehasbeenreportedtovaryinarangeofapprox.50 GPadependingonthe planefamilyconsidered,bydiffractiontechniques,insimilarmicrostructures .However,thefactthattheEvalues obtainedfortheaustenitephaseonHT350arehigher,andclosertotheexpectedones,thanthosemeasuredon HT250,suggeststhatotherissuescancomeintoplaysuchasthesurfacetopography,notonlytheroughnessbut alsothelocalmorphology.Furthermore,theAFMindentationarrayexperimentscanalsoimplyadditional problemsorlimitationssuchastheeventualcontaminationofthetipasrelativelyhighloads(intheAFMrange)are applied.Besides,thediscreetsampling,thecomplexityofthebainiticmicrostructuresaswellastheblindcharacter oftheexperimenthamperobtainingreliablestatisticaldataonbothphases,particularlyontheferriteone. Figure7 NormalizeddistributionofEvaluesobtainedbyAFM nanoindentationsinausteniteofHT350. Thus,atthisstageofourwork,wedecidedtoaddressthisstudybymeansofPFQNM.Thistechniqueallowsto measurethetopographyandEvalueofthesurfacesimultaneouslyandcontinuously,i.e.onthewholeimagedarea. Furthermore,wehaveoperatedthemicroscopeusingextremelylowloads,at2 μN,whichisconsiderablylower thanthoseusedintheindentationexperiments.Thisfactreducestheprobabilityoftipdegradationand contamination.Inanycase,oncethetipbecomescontaminatedthereisasharpreductionintheEvaluesthatallows ustostopthemeasurementandtoproceedtocleanthetipbyindentingthegoldsurfaceascommentedinthe experimentalsection.Thisisasortofinsitucheckingofthetipstatus. First,weanalyzedtheHT250sample.InFig.8a,baredisplayedthetopographicalandEmapsmeasured simultaneouslybyPFQNM.Thetopographyshowsthejaggedmorphologycommentedabovewithseveral crevicesanddeepgroovesproducedasaconsequenceofthepreferentialetchingoftheferritephase.This morphologyresultsinanaverageroughnessof12 nm.InFig.8bthecorrespondingEmapisdisplayedinwhich theausteniteplateaulikephaseappearstohaveahighE,inthe130–190 GParange,whereasdarkpatcheswith extremelylowEvalues,closeto20 GPainsomecases,arealsomeasured.TheEimageleadstoanormalizedE distributionplotasthatdepictedinFig.8c,withasharppeakcenteredat30 GPaandabroaderonewithinthe60– 220 GParange.AcarefulinspectionofFig.8a,bshowsthatthelowEpatchesareclearlyrelatedtotheetched zones.Therefore,thestraightforwardconclusionwouldbetoassociatethemtotheferritephase.However,the extremelylowEvaluesreachedatsomespotssuggestthatamoredetailedanalysisisrequired.Thus,wedecidedto obtaintheslopeimagecorrespondingtoFig.8a.ThisisshowninFig.8d(seecaption).Fromthecomparisonof imagesFig.8b,d,itbecomesevidentthatthereisadirectcorrelationbetweentheslopeofthemorphologyandthe 46 47 3/3/2016 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForceMicroscopyandNanoindentationwithAFM http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4658537/ 8/14 obtained(low)Evalues.Thiswasconfirmedbyperformingcrosssectionsofbothimagesalongthesamepaths (notshown).Therefore,lowEvalueswereobtainedwhensharpdiscontinuitiesorslopeswerepresentinthe surfacemorphology.ThisfactpreventsustoassociatethelowEvaluestotheferritephase.Incontrast,asthe austenitephasehasasmoothandflatmorphologytheEvaluemeasuredontheseregions,130–190 GPa,ismore reliable.However,eveninthiscasesomeinfluenceofthemorphologycannotbediscardedastheflataustenite plateausstilldisplayacertainroughnessatthenanoscalelevel. Figure8 Thetopographicalmap(a)andEmap(b)ofsampleHT250withetched surface,measuredbyPFQNM;togetherwithitsnormalizedEdistribution (c)andthecorrespondingimageoftheinverseoftheslopeofFig.8a(d). Theinverseoftheslopehasbeen... InordertoovercomethislimitationimposedbytheetchinginducedtopographyonthePFQNMmeasurementsof sampleHT250,weproceededtoapplythisprocedureontheHT350sample.Inthissample,weexpectedthatthe etchingwouldleadtoferriteetchedphaseswideenoughtoberidoftopographicalcontaminationonthe determinationofthecorrespondingEvalues.Figure9a,bshowcharacteristictopographicalandEmaps,taken simultaneouslyinthePFQNMmode,ontheHT350system.Inthetopographicalimagestheaustenitedomainsare clearlyvisibleastheycorrespondtothehigher(brighter)locationsthatdisplayaflatplateaulikemorphology,with aroughnessof1.5 nm.Inaddition,deepcrevices,asdeepas100 nm,havebeenproducedbytheetching procedure,leadingtoageneralsurfaceroughnesscloseto38 nm.Insomecases,thebottomofthesedeepcrevices isquiteflatandwide(i.e.largerthan100 nm).Thisismoreevidentforothershallowercrevicesoretchedstructures thatarearound20 nmdeep(i.e.,attherightbottomandtoppartsoftheimage,forinstance).Inthesecases,the bottomregionsarealsoatleast100 nmwideandflat(i.e.witharoughnessof3 nm).Thus,inprincipleweshould notexpectstrongartifactscomingfromthemorphologyatthesebottomcrevicesoftheferritephase.Thisis confirmedwhenweanalyzethecorrespondingEmap(Fig.9b).Oncemore,thesharpperimetersofthecrevices displayadarkcontrast,i.e.lowEvalues,inagreementwiththeobservationsrealizedinsampleHT250.However, thistimemostofthecrevicebottomspresentahigher(i.e.,largerEvalues)andhomogeneouscontrast.Infact,the correspondingnormalizedEdistribution(Fig.9c)showsnowaclearpeakat185 GPawithalongtaildowntothe lowEvalueregion.ThistailisduetothedarkEregionscorrespondingtolargeslopedmorphologicalregions.A carefulanalysisofthewiderferriteexposeddomainsrevealsaslightlylowerEvaluethanthatfoundonthehigher flataustenitedomains.Morespecifically,theseferritezonesshowanEvaluecloseto165 GPawhereasthe austeniteonespresentvaluesinthe185 GParange. Figure9 Thetopographicalmap(a)andEmap(b)ofsampleHT350withetched surface,measuredbyPFQNM;togetherwithitsnormalizedEdistribution (c). Therefore,byrevealingtheferriteandaustenitephasesonasamplewithwiderdomains,wehavebeenableto measuretheEvalueofeachphaseunderthesameconditions,andassessthattheyarequitesimilarwithintheerror ofthePFQNMmeasurementmode.Onefinalwaytoconfirmthisresultistomeasurebythesametechniquethe polishedsampleswithoutfurtheretching.Inthisway,ononehand,wewilldealwithflatsurfaces.Therefore,the measurementswillbefreefromstrongmorphologicalinducedartifactsasthosediscussedabovefortheetched samples.Ontheotherhand,wewillnotbeabletoidentifyeachphase. InFig.10a,baredisplayedthetopographicalandEmapsoftheHT250polishedsample.Clearly,thesurfaceis muchsmoother,witharoughnessbelow1 nm,thanintheetchedones.Still,somestructuresarevisiblewith heightsinthe10 nmrange.ThecorrespondingEmap,incontrast,isquitehomogeneous,althoughitcanbenoticed thatthestepsoftheabovementionedstructuresstillgivelowerEvalues,inagreementwithpreviousresults. However,theaverageEvalueresultstobe184 ± 35 GPa.ThecorrespondingnormalizedEdistributionisplottedin Fig.10c.ThecorrespondingdatafortheHT350polishedsampleareshowninFig.11a–c.Again,thesurfacestill presentssomemorphologicalfeatureswithheightdifferencescloseto10 nm,butasurfaceroughnessbelow1 nm( 3/3/2016 Nanomechanicalcharacterizationofnanostructuredbainiticsteel:PeakForceMicroscopyandNanoindentationwithAFM http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4658537/ 9/14 Goto: Fig.11a).ThecorrespondingEmapis,inthiscase,quitehomogeneous(Fig.11b)witharelativelynarrowE distribution(Fig.11c).TheaverageEvalueis178 ± 30 GPa.TheEvaluesobtainedinbothpolishedsamplesare quiteconsistentwiththosemeasuredontheausteniteandferritephasesinthecorrespondingetchedsamples.The factthattheEmapsofthepolishedsamplesarequitehomogeneousimpliesthatbothphases,indeedpresentatthe surface,dohavesimilarEvalues,inagreementwithourresultsobtainedontheetchedsamplesbyPFQNMand onetchedHT350samplebynanoindentation. Figure10 Thetopographicalmap(a)andEmap(b)ofsampleHT250withpolished surface,measuredbyPFQNM;togetherwithitsnormalizedEdistribution (c). Figure11 Thetopographicalmap(a)andEmap(b)ofsampleHT350withpolished surface,measuredbyPFQNM;togetherwithitsnormalizedEdistribution (c). Discussion TheresultsevidencetheadvantagesandlimitsofperformingnanomechanicalmeasurementswiththeAFMbased modesemployedinthiswork.RegardingtheEanalysisfromAFMindentationcurves,itisclearthattheir interpretationishamperedbothbythesamplemorphologyandbytheexperimentalprocedureitself.Inthefirst case,aswehaveseenalsointhePFQNMcase,thesurfaceroughnessandtheambiguityinperformingthecurves onwelldefinednanophasesarethemainsourceoferrors.ThesetwoeffectsusuallyleadtoEvaluessmallerthan expected,justasitoccursinPFQNMexperimentsbuteventoalargerextent.Inaddition,theproblemscoming fromeventualcontaminationofthetipduringtheindentationarrayexperimentscanalsoresultinlargedeviationsin theexperimentalresults. However,despitetheselimitationsandproblems,theAFMindentationdatacanstillbefurtheranalyzedinother ways.AsnotedpreviouslythecrossoverpointatwhichtheHertzianbehaviorfailsisknownastheyieldingpoint. Itisrelatedtotheonsetoftheplasticbehaviorofthegivendomain.Therefore,wecaninvestigatewhetherthe correspondingyieldingforcedependsonthenatureofthedomain.Thisinformationcouldbevaluableduetothe complexityofthemicrostructure. Thus,thestressatwhicheachphaseyieldscanbeestimateddirectlyfromthenanoindentationforcecurves,andfor thispurposethecorrespondingEvalueswillbeused.Itisknownthatthemaximumshearstress(τ )ofthe elasticregimeinananoindentationreadsasequation4 : whereP isthemaximumloadappliedattheelasticregime,i.e.,themaximumdeflectionmultipliedbyK,the springconstant.Obviously,afterthepreviousanalysisonYoung’smodulus,itisreasonabletoapplyeq.3onlyto curveswhichseemtobenotaffectedbytopography,i.e.,curves4,5and6ofHT350,forwhichcalculatedE valuesaremorereliable.Theforcevalueatwhichyieldingstartsisassumedtobethepointatwhichforcecurve presentsanabruptchangeoftrend,gettingawayfromtheelastictheoreticalsolution.P andτ areshownin Table2. Table2 Maximumforceandthecorrespondingmaximumshearstressofelastic regimeincurvesofFigure6. max 48 max max max