TECHNI CALARTI CLEATransmissionElectronMicroscopyInvestigationofReaustenitized-and-CooledHSLA-100SteelS.W.ThompsonReceived:3July2012 / Accepted:13August2012 / Publishedonline:17October2012SpringerScience+BusinessMedia, LLCandASMInternational2012Abstract Acontinuous-cooling-transformation diagramfor 0.06% C HSLA-100 steel (austenite grain size of about10 lm) was reexamined in combination with hardness dataandlight-microscopeimages.Datasuggestedthepresenceof a martensitic microstructure over a wide range ofcooling rates. A specimen cooled at approximately 17 C/swas chosen for detailed microstructural investigation.Transmissionelectronmicroscopyrevealedamicrostruc-tureconsistingprimarilyof martensitelathswithamod-erate-to-high dislocation density. About 5%of retainedaustenite was present as interlath lms. An orientationrelationship close to KurdjumovSachs was veried formartensiteandaustenite, andadjacent laths exhibitedanaxis/angle pair consistent withtwovariants of this rela-tionship. Some regions of twinned martensite wereobserved, and multi-variant cementite precipitation wasobservedwithinlargeautotemperedmartensitelaths. Allresults are consistent with this steel in its as-receivedmill-processedcondition.Keywords HSLA-100 TEM Martensite Retainedaustenite ElectrondiffractionIntroductionWilson et al. [1] examined the microstructures and mechani-cal properties of several low-carbon copper-precipitation-strengthenedalloys, includingHSLA-80, HSLA-100, andtwoalloys of intermediate compositions.1This class of steelswas developedtoreplacesteels withsimilar strengths, butnotablyhigher carboncontent [2, 3]. Akeyreasonfor thedevelopmentofthisclassofsteelswastoachieveimprovedweldabilitybyloweringthecarboncontent but maintaininghigh strength through copper-precipitation hardening. ThesteelsstudiedbyWilsonetal.[1]wereprocessedinacom-mercial mill by rolling to plate followed by a reaustenitze-and-quench step. Eachsteel was microalloyedwith niobium, and allsteels possessedanequiaxedne-grainedausteniticmicro-structure (e.g., about 10 lm) prior to the nal mill quench.The microstructure of HSLA-80 steel was shown tocontain primarily ne-grained polygonal ferrite over a widerange of cooling rates. Transformation of austenite topolygonal ferriteoccurredat relativelyhightemperatures(i.e., betweenabout550and800 C, dependingoncool-ingrate), withnal ferritegrainsizesof 5 lmor larger.Subsequent work [4] revealeda lowdislocation densitywithinthegrainsof polygonal ferrite, andsecondarymi-croconstituents that included carbon-enriched islands(a few micrometers in size) of bainite with coarse carbides,martensite, andsomeaustenite.Itwas concludedthatanHSLA-100steelwith0.06%Cexhibited a martensitic microstructure over a wide range ofcoolingrates. Twoalloysofintermediatecompositionsdisplayed a microstructure referred to as granular bainiteoverawiderangeofcoolingrates. Thelattermicrostruc-tureconsistedprimarilyofsmall crystalsofferrite(afewmicrometersinsize)withbothequiaxedandnon-equiaxedmorphologiesandmoderatedislocationdensity[5]. Addi-tionally, themicrostructurealsocontainedasmallamountS. W.Thompson(&)DepartmentofMetallurgicalandMaterialsEngineering,ColoradoSchoolofMines, Golden, CO, USAe-mail:sthompso@mines.edu1HSLAsteel referstohigh-strengthlow-alloysteel. Thenumericaldesignationsindicatetheminimumacceptableyieldstrengthinunitsofksi.1 3Metallogr. Microstruct. Anal. (2012)1:131141DOI10.1007/s13632-012-0023-yofcarbon-enriched, sub-micron-sizedblockyislandsofmartensiteandretainedaustenite(M/Aconstituent).The term granular bainite was used based on work byHabrakenandEconomopoulos [6] publishedinthesym-posiumproceedingsTransformationandHardenabilityinSteels. Sincethisarticlewaspublished, other termshavebeensuggestedforgranularbainitesuchascarbide-freebainite, acicularferrite, granularferrite, andothers. Intheremainderofthisarticle, thismicrostructurewillberefer-redtoasacicularferrite[5].Wilsonet al. [1] demonstratedthat theyield-strength/Charpy-impact-toughnessbehavioroftheHSLA-100steelwith a martensitic microstructure was superior to that of theHSLA-80 steel with a polygonal ferrite microstructure. Thestrength/toughness behaviors of the two steels withintermediate compositions (andacicular ferrite micro-structures) wereintermediate tothoseof the HSLA-80 and100steels. Inaddition, asecondHSLA-100steel withacomparatively low-carbon content of 0.04% C exhibited anacicularferritemicrostructurein29-mm-thickplates,withconcurrentlypoorer strength/toughnessbehavior, ascom-paredwiththe0.06%CHSLA-100steel[1].While the microstructures of the HSLA-80 steel and one ofthe intermediate compositions studied by Wilson et al. [1]have been documented in detail [4, 5] via transmission elec-tron microscopy,themicrostructure of the0.06%CHSLA-100 steel has not, until recently. The original conclusion thatthe latter alloy possessed a martensitic microstructure over awide range of cooling rates was strongly supported by severalimportant observations. Light microscopyshowedthat themicrostructure of 0.06% C HSLA-100 steel was notably dif-ferentthanthatoftheintermediatealloys(aswellasthe0.04%C HSLA-100 steel), thereby suggesting that thestructure was something other than acicular ferrite with M/A.The hardness of the 0.06%CHSLA-100steel was close tothatexpected for as-quenched martensite based on carbon content[7], unlike 0.04% C HSLA-100 steel and the intermediatealloys. Also, a continuous-cooling-transformation (CCT)diagram(fromthe 0.06%CHSLA-100steel, see Fig.1)showedawell-denedmartensite-start temperature(Ms)ofabout 425 Cover a wide range of cooling rates [1] that agreeswith the calculated value froman empirical equation ofAndrews [8]. A recent study via transmission electronmicroscopyshowedthatthe0.06%CHSLA-100steelpro-cessedexclusivelyinaplatemillpossessedafullymar-tensitic microstructure [9].IntwostudiespublishedaftertheworkbyWilsonetal.[1], themajor thrust wastoexaminenemicrostructuralfeatures via transmission electron microscopy of specimensthat came directly from CCT studies [4, 5]. Cylindrical testspecimens were reaustenitized to 905 Cand cooled atvariousratesinaGleeble1500thermo-mechanical sim-ulator[10].Hence,thepurpose ofthisstudyis,inessence,tocomplete a trilogy of studies that includes HSLA-80steel, an intermediate A710-type steel, and 0.06%CHSLA-100 steel. In each of the earlier papers [4, 5],transmissionelectronmicroscopyresults emphasizedthemost-prevalentmicrostructure,i.e., themicrostructurethatformed over a wide range of industrially important coolingrates.Collectively,thesepapersshowaprogressionofthepredominant microconstituent from polygonal ferrite(HSLA-80), toacicularferrite(intermediateA710-typecompositions), to martensite (HSLA-100) as alloy levelincreases but with a common austenite grain structure(equiaxedgrainsizeabout10 lm).Low-carbon copper-containing steels discussed above areput into service after a high-temperature (usually above600 C) agingheat treatment [1]. Thistreatment producessteel with an excellent combination of strength and toughness.While a high temperature is needed to induce copper precip-itation, other processes also occur, most notably the recoveryof dislocation substructure and elimination of low-angleboundaries. As discussed above, martensite has superiormechanical properties compared to acicular ferrite, and thesemicroconstituents alsohave important differences inage-hardeningbehavior. Thus, thepurposeof thisarticleistodocument the ne-scale microstructural features of a 0.06%CHSLA-100steelintheas-quenchedconditionthatwaspro-duced from undeformed, ne-grained austenite. A key aspectof this study will be to assess whether or not the observationsagree with generally accepted features of lath martensite.ExperimentalProceduresThereportedchemical compositionofthesteel examinedwas(in wt.%):0.06C,0.83Mn,0.010P,0.002S,0.37Si,1.66Cu, 3.48Ni, 0.58Cr, 0.59Mo, 0.028Nb, 0.028Al,balance Fe. The steel was received in the formof aFig.1 Continuous-cooling-transformation diagram for 0.06% CHSLA-100 steel. Prior austenite grain size approximately 10 lm.Labels areA(austenite), M(martensite), AF(acicular ferrite), GF(granularferrite), andPF(polygonalferrite)132 Metallogr. Microstruct. Anal. (2012)1:1311411 329-mm-thick (i.e., 1 and 1/8 inch) plate that had beenreaustenitizedandquenched. This is the same steel thatwasstudiedinRef. [1,9, 10].Test specimens were machined to a length of about102mmand a diameter slightly larger than 6mm, andpossessed a 12-mm-long reduced section of 3mmindiameter, with a 30-lmsurface polished nish. Thesespecimenswereheatedfromroomtemperatureatarateof3 C/s to905 C, heldfor 300s, andcooledat variousrates in a Gleeble1500 thermo-mechanical simulator. Thecooling programs were of a Newtonian nature [10, 11], anda simple estimate of cooling rate was determined bydividing300 C(i.e., 800500 C) bythetotal timethatthe specimencooledinthis 300 C-temperature interval(referredtoas Dt800?500).To determine the temperature range over which aus-tenite forms during reheating, a heating rate of about0.03 C/s (i.e., 2 C/min) was employed. The temperaturesat which austenite begins to form (Ac1) and at which a fullyausteniticmicrostructureisobtained(Ac3)duringheatingweredeterminedfromplotsofspecimendimensionversustemperature. Measured values of 680 and 825 C comparedfavorablywithcalculatedvalues(668and842 C)[8].Cylindrical test specimens were sectioned near thepositionofthethermo-couplethathadbeenweldedtothesurfacepriortoheat treatment. Specimensweremounted,polished, and etched using conventional metallographictechniques. After nal polishing, specimens were etchedwith 2% nitric acid in ethanol (2% nital), rinsed, and dried.Thesespecimenswereobservedvialightmicroscopy.Specimens for transmissionelectronmicroscopywerepreparedbycuttingthinwafersfromthemid-positionofthespecimenincloseproximitytotheattachedthermo-couples. Thethinwafer was groundtoabout 0.1mminthickness. Discs of 3mmindiameter were groundtoathickness of approximately 0.08mm. Specimens wereelectro-polishedina jet polisher at 40Vusinganelec-trolyte of CrO3, acetic acid, and water. Transmissionelectronmicroscopywasconductedat120kV.Bright-eld (BF) and centered-dark-eld (CDF) imagingtechniques were employed during transmission electronmicroscopyinvestigations. Additionally, critical informa-tion was obtained fromelectron-diffraction patterns, acommontechniquethat hasnot beenappliedinsufcientdetail to the microstructures observed. At 120kV, thecalculated electron wavelength is 0.00335nm, and thenominal cameralengthusedwas940mm. Apurealumi-numstandardwasusedtodetermineacalibratedcameraconstant of 3.26910-12m2(i.e., 32.6mm A). Based onnumerous analyses of electron-diffraction patterns from themicroscopeusedinthisstudyandthecommonlyselectedoperation parameters, the following uncertainties havebeen estimated: interplanar spacing 0.003nm, e.g.,0.1800.003nm,interplanarangles2,e.g.,712,and5for orientationrelationships(ORs) or axis/anglepairs, e.g., 1

10/555.Forresultspresentedinthisarticle, iftheuncertaintieslisted above were exceeded, the data were considered to beof a dubious nature, and alternate interpretations werepursued. Insomecases, satisfactoryanswerswithintheseuncertaintyrangeswerenot obtained, andinthosecases,theanalysisofthoseresultsisnotprovided.ResultsandDiscussionCCTDiagram,Hardness, andLightMicroscopyPrevious work[1] on0.06%CHSLA-100steel showedthat themicrostructureof amill-processedplatethat hadbeen reaustenitized and quenched was predominantlymartensitic, basedonlight-microscopeimages. Also, themicrohardness (Vickers hardness, HV) of about 365HVwas consistent withas-quenchedmartensitefor acarboncontentof0.06%[9].Furtherevidenceofapredominantlylow-carbon martensitic microstructure was obtained from aCCTdiagram[1], andFig. 1provides aslightlyrevisedversionofthisdiagram.This diagramincludes14different coolingcurves, andeach curve includes at least two points that reveal the start andcompletionof austenitetransformation. Thesepointsweredetermined fromcomputer-generated plots of specimendilationas a functionof temperature. For eight of these curves,numbersat thebottomofthegureindicatethemeasuredmicrohardness values (HV) obtained fromthe as-cooledspecimens. Additionally, toassist theinterpretationof thetransformation characteristics of austenite, Fig.1 includes thelabels A (austenite), M (martensite), AF (acicular ferrite), GF(granular ferrite), and PF (polygonal ferrite). Figure2 showsmicrohardnessasafunctionofthetimeintervalassociatedwithcoolingfrom800to500 C. Incomparisonwithanintermediate A710-type composition [5], 0.06% C HSLA-100 steel shows a much more gradual decrease in hardness ascooling rate decreases and, concurrently, as martensitetransitions to acicular ferrite with M/A constituent.Light micrographs fromspecimens cooledat approxi-mately 100 and 1 C/s (i.e., Dt800?500values of 3 and315s, respectively) are shown in Fig. 3. The dominantmicroconstituentsaremartensiteatthehighercoolingrate(100 C/s)andacicularferritewithM/Aatthelowerrate(1 C/s). In Fig.3(b), the features that are most-readilyevident are small islands or pools of martensite and/oraustenite(theM/Aconstituent). Thesefeaturesarepartic-ularlyevident becausetheinterfaces betweencrystals ofacicularferriteandpoolsofmedium-carbonM/AareveryMetallogr. Microstruct. Anal. (2012)1:131141 1331 3effectivesitesforetchingwithnital. However, crystalsofacicularferriteadjacenttooneanotherfrequentlyhavethesame crystallographic orientation, and so these adjacentcrystals etch at a similar rate. Even in cases where there is asignicant crystallographicmisorientationbetweenferritecrystals, the etching effect is weak as compared withacicularferriteM/A interfaces. Features labeled A, B, andC in Fig. 3(b) show different morphologies of M/A islands,wheretheislandmorphologyisdeterminedbythegrowthcharacteristics of the adjacent crystals of acicular ferrite. Incomparison, Fig. 3(a) reveals noobviousmedium-carbonM/Aislands,consistent with the diffusionless formation oflow-carbonmartensite. Forthismicrostructure, martensitepacket boundaries appear to be the sites readily revealed byetchingwithnital.Twoeasilydiscernibleadjacentpacketsof martensite are indicatedat A and B in this gure. Thesesame etching effects have been describedina previouspaperconcerninga similarsteelof lower alloy content [5].Despite the clear differences in the microstructuresrevealed by Fig.3, many features are very small andworthy of more-detailed examination via transmissionelectron microscopy. Additionally, electron diffractionprovides keycrystallographicinformationthat cannot berevealedbylight microscopy. Acicular ferrite withM/Ahas beenexaminedviatransmissionelectronmicroscopyfor alow-carboncopper-containingalloyof leaner com-positionthanthatofHSLA-100,andthereaderisreferredtothatpaperasbeingtypicalofthemicrostructureshownin Fig. 3(b) [5]. However, the microstructure inFig. 3(a) for a low-carbon copper-containing alloy hasbeen documented in much less detail. In a recent paper [9],transmissionelectronmicroscopywasusedtoexamineamill-processed 0.06%CHSLA-100 steel that had beenreaustenitizedandquenched. Intheas-receivedcondition,its hardness was approximately365HV. The data fromFigs. 1and2showthat aspecimenwithsimilarhardness(362HV) wasassociatedwithaDt800?500valueof 18s,andthusacoolingratecalculatedat 17 C/s. Thisinfor-mation was used to predict that the reaustenitized-and-quenchedplatehadbeencooledinthemill at about15 C/s, and it was also used to select the laboratoryspecimen cooled at 17 C/s for detailed transmissionelectronmicroscopyexamination. This coolingrateis inthe middle of the martensitic range and the associatedmicrostructureis believed to be generallyrepresentative ofthat obtained at all the cooling rates within this range(labeledMinFig.1).TransmissionElectronMicroscopyAnimportantaspectofthisarticleistopresentheretoforeunpublished transmission electron microscopy results fromlaboratoryprocessed0.06%CHSLA-100steel. Figure4shows atypical transmissionelectronmicroscope(TEM)Fig.2 Hardnessversustimeintervaltocoolfrom800to500 CFig.3 a Martensite (100 C/s) and b acicular ferrite with M/Aconstituent(1 C/s). Lightmicrographs, 2pct. nitaletch134 Metallogr. Microstruct. Anal. (2012)1:1311411 3image of martensite formed during cooling at a rate of17 C/s. The individual crystals of martensite possess alath-likemorphology[12], andtheapparentlathwidth(inthis particular region at this tilt condition) is approximately0.1 lm. The crystals of martensite show a moderate-to-high dislocation density in the as-cooled martensitestructureandadislocationdensityof approximately1015m-2(i.e., 1011cm-2) is a reasonable estimate [13, 14].This observation is consistent with volume accommodationeffectsduringmartensiteformation[15, 16].The intertwined nature of the laths in Fig. 4 is consistentwith two or more variants of, for example, the KurdjumovSachsOR[17]betweentheparentaustenitephaseandtheproduct martensite phase. An austenite grain-boundaryappears tohavebeenpresent (prior tomartensiteforma-tion) alongtheslightlycurvedlinear featurethat extendsfromAatthetopofthemicrographtoAatthebottomofthemicrograph.Figure5(a) is alow-magnicationBFTEMimageofmartensite laths, most of which are contained within asingleprioraustenitegrain.Figure5(b)isaCDFimageofthe same region produced from a (002) austenite reection.It appears that the crystal of austenite inthe upper leftportionofFig5(b)alsoshowsinterlathretainedaustenite,apparently because this particular spot was common tobothcrystalsof austenite. Thisstatement doesnot implythatthetwoaustenitegrainshadthesamecrystallographicorientation, andinfact wouldbeanunlikelyscenarioforequiaxed adjacent grains of austenite formed during asimplereaustenitizationstep.At higher magnication, individual lathsof martensiteappearedtobewhollycontained withinasinglepriorgrainof austenite, as expected[18]. Consistent withthiscomment, thearrowsnearthebottomofFig. 5(b)showacontinuousportionofretainedaustenitethatdelineatestheboundary between this grain of austenite and the one belowit.Similarly,arrowsattherightsideshowanotherslightlyless continuous grain-boundary lm. These featureswouldbeuncommonorabsent altogetheriflathsofmar-tensitecouldgrowfromonegrainintoanotherofrandomcrystallographicmisorientation.Figure6 shows the selected-area electron-diffractionpatternfromthe central portionof the regionshowninFig. 5. At least four zonesof martensitehavebeeniden-tied; however, only two zones have been indexed inFig. 6(b). The magnitudes of the associated ghkl vectors andthe angular relationships betweenthese vectors are con-sistent with a body-centered-cubic (BCC) phase with aFig.4 Lath martensite; bright-eld TEM micrograph. A prioraustenitegrain-boundaryappearstoextendfromAat thetoptoAatthebottomFig.5 Martensite laths within a single austenite grain. TEMmicrographs:aBFimageandbCDFimageofretainedausteniteMetallogr. Microstruct. Anal. (2012)1:131141 1351 3lattice parameter of approximately 0.287nm. Althoughmartensite is a body-centered-tetragonal phase, the ratio oflatticeparameters(thec/aratio)issosmall,especiallyforlow-carboncontents, that typical electron-diffractionpat-ternsdonot reveal thetetragonal natureofthemartensitephase.BFTEMimagesgeneratedat adifferent tilt conditionshowed that about half the laths in this prior austenite grainwas verybright andthe other half was verydark, sug-gestingthepresenceoftwodifferentpacketsofmartensiteand concurrently two crystallographic orientations ofmartensite. The two zones of martensite indexed inFig. 6(b) have beamdirections of

311 and

211: Theanglebetweenthesetwozonesis10, suggestingthat thespots for both the zones may be from the same packet, andthat theactual zoneaxisfor onepacket isbetween

311and

211:Therefore,themajorityoftheotherspotsinthepatternareprobablyfromasecondpacket of martensitewithahigh-indexzoneaxisthatwasnotdetermined.Unfortunately, the complexity of this pattern did notreveal sufcient informationtoprovethepresenceof theaustenitephase; however, the(002) austenitespot (AinFig. 6a) has theexpectedinterplanar spacingof approxi-mately0.180nm. Thecloseproximityof the(002) aus-tenitespotandthe 01

1martensitespot(MinFig. 6a)isreadily apparent. Unfortunately, the electron-diffractionpatterns obtained from this region at this tilt condition wereoverexposedwhenrecordedonlm; however, thesetwospotswereeasilydistinguishedontheuorescentviewingscreen of the TEM. In addition, a very small objectiveaperture was used to record the CDF image shown inFig. 5(b), thereby showing austenite very brightly, andsomeregionsofmartensitetoalesserextent. Cleariden-ticationof similar interlathlms as theaustenitephasewill be presented later in this article, though this example isFig.6 SADPof thecentral portionof themicrostructureshowninFig.5. a Original pattern and b facsimile of this pattern with indexedspotsfromtwozonesofmartensite.Electronbeamdirection=

311for one zone (grid made from solid lines) and beam direction=

211forasecondzone(gridmadefromdashedlines)Fig.7 Martensite laths within a single austenite grain. TEMmicrographs:aBFimageandbCDFimageofretainedaustenite136 Metallogr. Microstruct. Anal. (2012)1:1311411 3presented rst because the shape of this prior austenitegrainappearstobeeasilyenvisioned. Excludingtheillu-minatedregions of austenite inthe upper left corner ofFig. 5(b), the equiaxed prior austenite grain in the center ofthis gure has a maximumchord lengthof about eightmicrometers.Figure7showsasecondBF/CDFpair of micrographsrevealingmartensitecontainedwithinwhat againappearsto have been another single grain of austenite; although, itsapparent shape in two dimensions seems less conven-tional than that of Fig. 5(b). Nonetheless, its largest chordlength in this eld of view was measured to be about 7 lm.While this pair of micrographs is in some ways lessrevealingthanthoseofFig. 5, theassociatedelectron-dif-fraction pattern was interpreted in a more-complete fashionthanthatofFig. 6.Figure8 shows an SADPfromthe central region ofFig. 7. Specically, Fig.8(a) is the original pattern, (b) is afacsimile of the patternthat includes labels for indexedzones of martensitespots andaustenitespots, and(c) isanother facsimile with a second indexed zone of martensitespots. FromFig. 8, themagnitudesof theassociatedghklvectors and the angular relationships between these vectorsare consistent with a BCC phase with a lattice parameter ofapproximately0.287nmandaface-centered-cubic(FCC)phase with a lattice parameter of about 0.360nm, i.e.,consistent with values expected for martensite and aus-tenite. Detailed analysis of this pattern revealed an ORbetweenmartensite(M)andaustenite(A)phasesthatwaswithin5of

110M==

111A

1

1

1M==

10

1A

1

12M==

1

21Aandalsowithin5of001M==0

11A

110M==

111A

1

10M==

2

1

1ATherstrelationshipisavariantoftheKurdjumovSachsOR[17], while the second is a nearby variant of theNishiyamaWassermann OR [19, 20]. These ORs arerelatedby a 5.26 rotationaboutthecommonparallel axesFig.8 SADPof thecentral portionof themicrostructureshowninFig.7. a Original pattern and b facsimile of this pattern with indexedspotsfromonezoneof martensite(gridofsolidlinesandnumberswithout highlights) andone zone of spots fromaustenite (gridofdashedlines withnumbers ingrayovals), andcasecondzoneofmartensite. Electronbeamdirection=

2

10 for therst martensitezone, beamdirection=

5

31 for the secondmartensite zone, andbeamdirection=

4

1

1fortheaustenitezonecMetallogr. Microstruct. Anal. (2012)1:131141 1371 3of

110M==

111A:Basedontheestimateduncertainty,itisnotpossibletodeterminewhichORttheexperimentaldata the closest. As the maingoals of this studyare tocharacterizekeymicrostructural featuresviaimagingandto identify important crystallographic information viaelectron diffraction, the quoted accuracy is sufcient toprovethepoint.Specically,thephenomenologicaltheoryof martensitecrystallographyrequiresareproducibleORFig.9 Martensite laths within a single austenite grain. TEMmicrographs:aBFimageandbCDFimageofretainedausteniteFig.10 SADPofthecentralportionofthemicrostructureshowninFig.9. a Original pattern and b facsimile of this pattern with indexedspotsfromonezoneof martensite(gridofsolidlinesandnumberswithout highlights) andtwospotsfromaustenite(blacknumbersinwhite ovals), and c a second zone of martensite. Electron beamdirection= 11

1 for the rst martensite zone, and beamdirec-tion= 00

1forthesecondmartensitezonec138 Metallogr. Microstruct. Anal. (2012)1:1311411 3betweentheparentandtheproductphases[18],andthosereported by KurdjumovSachs and NishiyamaWasser-mann [17, 19, 20] are most often observed for carbon steel.Anal example of martensite laths formed within asinglegrainof austeniteis showninFig. 9. TheamountofretainedausteniteshowninFigs. 5(b),7(b),and9(b)issimilar topreviouswork[9], andaconservativeestimateisabout 5vol.%. Measurement oftheamount ofretainedausteniteviax-raydiffractionforthissteelwashamperedbythelowintensity, owingtothesmall amount of aus-tenite, andpeakbroadening, owingtothe small size ofthe interlathlms. Thus, the amount of austenite inthemicrostructure was underestimated by x-ray techniquesand is not reported. Additionally, a maximum chordlengthfromthisgrainofaustenitewasmeasuredat about9 lm.Theassociatedelectron-diffractionpatternisshowninFig. 10. Facsimiles of the original pattern reveal two zonesof martensite, but only a single rowof spots fromtheaustenite phase was identied with certainty (similar toFig. 6). Akeyaspect of themartensitepatterns was theaxis/angle pairofapproximately 1

10/55. Thisaxisanglepair isclosetotwodifferent variant pairsof theKurdju-movSachs OR. This andother variant pairs have beenexaminedindetail Moritoet al. [21, 22], andthevariantpair observedhereisconsistent withtheminimizationoftheoverallshapestrainformartensitecrystalsformedinagrainofaustenite.In summary, the complexity of electron diffractioninformation presented in this study wassuch thatno singleexample (i.e., Figs. 6, 8, 10) was identied that couldreveal all the key diffraction information at once. However,thethreeexamples collectivelyreveal datathat arecon-sistent with(i) theexpectedphases (martensiteandaus-tenite),(ii)afrequentlyreportedORforthesephases,and(iii) variant pairing that permits minimization of the overallshape strain associated with two variants of martensiteformedinthesamegrainofaustenite.So far in this article, evidence has been presented that isconsistent withlow-carbonlathmartensitemicrostructurewithabout5%ofretainedaustenite.Other,less-frequentlyobserved features will be described next. Although themajorityofthemartensitelathshadmoderate-to-highdis-locationdensities, best shownbyFig. 1, somelathscon-tained twinned regions. Twins in martensite are evidence ofstrain accommodation during martensite formation fromausteniteandarenot observedwithincrystalsof acicularferrite.Figure11showsanexampleof twinswithinalathofmartensite in both BF and dark-eld conditions. Theassociated selected-area electron-diffraction pattern fromthisregion, Fig. 12, showstwozonesof martensitespotsthatrevealthetwinrelationshipbetweenBCCcrystals. Intheoriginalpattern,numerousdouble-diffractionspotsareevident, andthearrows point toseveral suchspots. Thecoarsenatureof thetwins at AinFig. 11aresimilar tothose reported previously for this steel [9]. Finer, but more-difcult-to-seetwinsareevidentatC.Anadditionalnefeatureoccupyingasmallportionofthe microstructure is revealed by Fig. 13. In Fig. 13(a),darkplatelets of asecondphasearerevealed, andtheassociated selected-area electron-diffraction pattern isprovided in Fig. 13(b). The rotation between the twoimageshasbeenaccountedfor intheplacement of thesegures.Diffractionspotsfromthesedarkregionswerenotobtained; however, the shapes of these features andtheapproximate habit planes of {101} with respect to themartensitestronglysuggest thepresenceof cementite, asobserved elsewhere [23]. This type of multi-variantcementiteprecipitationischaracteristicof temperedmar-tensite, andinthiscaseitformedduringthequench.Owingtothelow-carbonlevelofthissteel,thevolumefraction of cementite is rather low, and combined withFig.11 Twinswithincrystalsoflathmartensite.TEMmicrographs:a BF image and b CDF image of twins. Twins at A and B seem to berathercoarse, whilethoseatCareveryneMetallogr. Microstruct. Anal. (2012)1:131141 1391 3different variants of thecementite, it is not entirelysur-prisingthat electron-diffractionspotsfromthisphasearenotevident. Thesefeaturesandotherprecipitatesmakeup\1%of the as-cooledmicrostructure for this steel. Thecementiteplateletsweretypicallyobservedinthelargestmartensitelaths. Therewaslittleifanyevidenceofothercarbide phases or copper-rich precipitates; however, theintent ofthisstudywastodocument thebasemicrostruc-ture andits keyfeatures, andnot toexamine ne-scaleprecipitation. It issuggestedthat thesecementiteprecipi-tates areformedduringautotempering, andthat theyareobservedinthelargestorrst-formedmartensitelathsthatformnearesttotheMstemperatureuponcooling.Finally, whilethelargemajorityoftheregionsinvesti-gatedviathin-foil transmissionelectronmicroscopydis-played characteristics consistent with a martensiticmicrostructure, a verylimitedamount of acicular ferriteappeared to be present, although evidence will not bepresented here. The difference between martensite andacicularferritewasmost-readilyrevealedbythepresenceofblockyislandsofretainedaustenitebetween adjacentcrystalsofacicularferriteascomparedwithinterlaththinlmsof austenitebetweenlathsof martensite. Thesedif-ferencesarebeinginvestigatedfurther,butasofthistime,the differences have not been documented in sufcientdetail.Fig.12 SADPfromtheregionnearAinFig.11.aOriginalpatternand b facsimile of this pattern with indexed spots from the martensitematrix(M) indicatedbysolidblacklinesandnumbers, andspotsfromthetwins (T) withdashedblacklines withblacknumbers inwhiteovals.Beamdirection= 1

13M==

3

11TFig.13 At least three variants of autotempered cementite precipitateswithinalargecrystal of lathmartensite. aBright-eldTEMimageand b indexed EDP from the martensite crystal. Beam direc-tion= 1

1

1: The image rotation between these images has beenaccountedforintheirplacementabove140 Metallogr. Microstruct. Anal. (2012)1:1311411 3SummaryandConclusionsA CCT diagram from a steel with chemical composition (inwt.%) of: 0.06 C, 0.83 Mn, 0.010 P, 0.002 S, 1.66 Cu, 3.48Ni, 0.58 Cr, 0.59 Mo, 0.37 Si, 0.028 Nb, 0.028 Al, balanceFe was reexamined and, in combination with hardness dataandlight-microscopeimages, themicrostructureobservedover a wide range of cooling rates was martensitic. Alaboratory processed specimen cooled at approximately17 C/s was chosen for detailed examination via trans-missionelectronmicroscopy.Conventional BF images were complemented with CDFimages andelectrondiffractiontoreveal microstructuralandcrystallographicfeaturesthatareconsistentwithlow-carbonlathmartensite. Thesefeatures includecrystalsof alath-likenature, withamoderate-to-highdislocationcon-tent. Crystals of martensitewerecontainedwithinsingleprioraustenitegrains,andoccasionallyregionsoftwinnedmartensite were observed. About 5vol.%of interlathretained austenite was present. The martensite/austeniteOR was close to that reported by KurdjumovSachs.Adjacent regions of martensite possessed two differentcrystallographic variants of the KurdjumovSachs OR.Someof thelargest laths wereautotempered, displayingmultiplevariantsofcementiteplatelets.Theseresultsareconsistent withpreviousworkonthesame steel that had been mill processed with a nalreaustenitization-and-quenchstep. Inbothcases, thehard-ness was about 360 HVand the prior austenite grainstructurewasequiaxedwithagrainsizeofabout10 lm.Acknowledgments ThisresearchwassupportedbytheAdvancedSteel ProcessingandProducts ResearchCenter withinTheGeorgeS. Ansell Department ofMetallurgical andMaterialsEngineeringatthe Colorado School of Mines (CSM) in Golden, CO, USA. G. Kraussisthankedfor enthusiasticdiscussions of martensiteandsteel. Theheat-treatedspecimenusedfor theTEManalyseswasproducedbyD. J.ColvinaspartofhisM.S.thesisinvestigationatCSM.References1. A.D. Wilson, E.G. Hamburg, D.J. Colvin, S.W. Thompson, G.Krauss, Properties andMicrostructures of Copper PrecipitationAgedPlateSteels, inMicroalloyedHSLASteels(ASMInterna-tional, MetalsPark, OH, 1988), pp. 2592752. T.W. Montemarano, B.P. Sack, J.P. Gudas, M.G. Vassilaros,H.H. Vanderveldt, High strength low alloy steels in navalconstruction. J. ShipProd.3,145162(1986)3. E.J. Czyryca, R.E. Link, J. Wong, D.A. Aylor, T.W. Monte-marano, J.P. Gudas, Development and certication of HSLA-100steelfornavalshipconstruction.Nav.Eng.J. 102,6382(1990)4. S.W. Thompson, D.J. Colvin, G. Krauss, Austenitedecomposi-tion during continuous cooling of an HSLA-80 plate steel. Metall.Mater. Trans. A27A,15571571(1996)5. S.W. Thompson, D.J. Colvin, G. Krauss, Continuous coolingtransformations and microstructures in a low-carbon, high-strengthlow-alloyplatesteel. Metall. Trans. A21A, 14931507(1990)6. L.J. Habraken, M. Economopoulos, Bainiticmicrostructures inlow-carbon alloy steels and their mechanical properties, inTransformationandHardenabilityinSteels (ClimaxMolybde-numCo., AnnArbor,1967), pp.691077. G. Krauss, Martensitictransformations, structureandpropertiesin hardenable steels, in Hardenability Concepts with Applicationsto Steel, ed. by D.V. Doane, J.S. Kirkaldy (TMS-AIME,Warrendale, 1978), pp.2292488. K.W. Andrews, Empirical formulaeforthecalculationofsometransformationtemperatures. J.IronSteelInst.203,721(1965)9. S.W. Thompson, Microstructural characterization of anas-quenched HSLA-100 plate steel via transmission electronmicroscopy. Mater. Charact. 2012(submitted)10. D.J. Colvin, The physical metallurgy of some low-carbon,Cu-bearingHSLAplatesteels, M.S. Thesis, ColoradoSchoolofMines, Golden,CO, 198811. G.T.Eldis,Acriticalreviewofdatasourcesforisothermaltrans-formation and continuous cooling transformation diagrams, inHardenabilityConcepts withApplications toSteel, ed. byD.V.Doane, J.S. Kirkaldy (TMS-AIME, Warrendale, 1978), pp. 12615712. G. Krauss, A.R. Marder, Themorphologyof martensiteinironalloys. Metall. Trans.2,2343(1971)13. P.B. Hirsch, A. Howie, R.B. Nicholson, D.W. Pashley, M.J.Whelan, Electron Microscopy of Thin Crystals (Butterworths,Washington, DC, 1965), p.42214. W.C. Leslie, The Physical Metallurgy of Steels (Hemisphere Pub.Co., Washington, DC, 1980),p. 21515. G. Krauss, Martensite in steel: strengthand structure.Mater. Sci.Eng.A273275,40(1999)16. B.P.J. Sandvik, C.M. Wayman, Characteristics of lath martensite:partI. crystallographicandsubstructuralfeatures. Metall. Trans.A14A,809(1983)17. G. Kurdjumov, G. Sachs, Uber denmechanismus der stahlhar-tung.Z. Phys.64,325(1930)18. H.K.D.H. Bhadeshia, Worked Examples in the Geometry ofCrystals(TheInstituteofMetals,Brookeld, 1987), pp. 576419. Z. Nishiyama, Sci. Rep.TohokuUniv.23, 637(1934)20. G. Wassermann, Mitt. K-W-IEisenforsch.17,149(1935)21. S. Morito, H. Tanaka, R. Konishi, T. Furuhara, T. Maki, Themorphology and crystallography of lath martensite in FeCalloys. ActaMater.51,17891799(2003)22. S. Morito, X. Huang, T. Furuhara, T. Maki, N. Hansen, Themorphology and crystallography of lath martensite in alloy steels.ActaMater.54,53235331(2006)23. H.K.D.H. Bhadeshia, Bainite in Steels (The Institute of Materials,London,1992), p. 65Metallogr. Microstruct. Anal. (2012)1:131141 1411 3