TurkJPhys25(2001),455{466.ITAKProductionandStructureofRapidlySolidied - PDF document

TurkJPhys25(2001),455{466.ITAKProductionandStructureofRapidlySolidiedAl-SiAlloysOrhanUZUNDepartmentofPhysics,GaziosmanpasaUniversity,60110Tokat-TURKEYTuncayKARAASLANDepartmentofPhysics,ErciyesUniversity,38039Kayseri-TURKEYMustafaKESKDepartmentofPhysics,ErciyesUniversity,38039Kayseri-TURKEYReceived10.10.AbstractAl-Sialloyswithcompositions8,12,16wt%Siweremanufacturedbychillcastingandmelt-spinning(MS)methods.Theresultingribbonsampleshavebeencharac-terizedbyopticalandscanningelectronmicroscopy(SEM)andX-raydiractometryandcomparedwiththeiringotcounterparts.Microstructuralexaminationsrevealedthatmicrostructuresofthemelt-spunribbonswerequitenecomparedtotheirin-gotcounterparts.BothXRDanalysesandmicrostructuralexaminationsindicatedthatthesolubilityofSiin-Alphasewasnearlycomplete.1.IntroductionIn1960,Duwezandhisco-workersdevelopedanoveltechniquetoextendsolidsolu-bility[1]andtoproducemetastablecrystalline[2]oramorphous[3]solidphaseinsomesimplebinaryeutecticalloysystems.Infact,RussianresearcherSalliwasthersttoapplyamethodsimilartooneabovein1958[4].Bothresearchersemployeda\guntech-nique"inwhichasmalldropletofmoltenmetalimpactedonachillsurfacewithhighvelocity,resultinginanon-uniform\splat"ofsolidiedmaterial.Thecoolingratesforthisprocesswereestimatedwithintherangeof10to10,muchfasterthantheconventionalsolidicationrateswhichare10orless[5,6].SincetheintroductionofrapidquenchingofmetallicmeltsbyDuwez,agreatvarietyoftechniqueshavebeendevelopedtoobtainalloysproducedbyrapidsolidication.Although,theeectofrapidsolidicationvarywidelyfromsystemtosystem,themajoreectsare:(i)decreasedingrainsize,(ii)increasedinchemicalhomogeneity,(iii)extensionofsolidsolubilitylimits,(iv)creationofmetastablecrystallinephases,and(v)formationofmetallicglasses[7, UZUN,KARAASLAN,KESKThedensityofsilicon(Si)is2.3gr/cmandofaluminium(Al)is2.7gr/cmsothatSiisoneofthefewelementswhichmaybeaddedtoAlwithoutlossofweightadvantage.Additionally,therelativelylowcoecientofthermalexpansion,highwearresistanceand uiditythatarecharacteristicofAl-Sialloyshavereceivedconsiderableinterestinthisalloyfamilyascandidatematerialsforautomotiveandaerospaceapplications[9,10].Therefore,Al-Sialloysaresoughtinalargenumberofapplicationsintheautomobileindustries,e.g.forpistons,cylinderblocksandliners[6,11].However,underthecon-ventionalsolidicationconditions,theSiphaseincastAl-Sialloysoftenexhibitacoarsemicrostructurethatleadstopoormechanicalproperties[6,10].RecentresultshavedemonstratedthatrapidsolidicationprocessesmaysubstantiallymodifythemorphologyoftheSiphaseascomparedtothosefoundinconventionallyprocessedmaterials[6,12-14].Therefore,variousrapidsolidicationtechniqueshavebeenappliedtowiderangeofcompositionofAl-Sialloys.Amongthemsplatquenching[15,16],atomization[17-19],andmeltspinning[11,19-23]resultedinanincreaseinthesolidsolubilityofSiin-Alwithoutforminganyintermetalliccompounds.Whilesomeoftheseattempts[15-18]focusedonmicrostructuralcharacterizationofrapidlysolidiedAl-Sialloys,others[19-23]concentratedonndingoptimalvaluefordeterminingtheretained-Siamountin-AlbyX-raydiractionmethods,dependingonlatticeparameterchangeofAlsolidsolution.However,thesestudieshavesuggestedmanydierentvalues,rangingfrom3.7x10nmto1.23x10nm,foreachincreasing1at%retained-Siamount-Al.Inthiswork,theAl-Sisystemhasbeenchosenessentiallyduetoitstechnologicalim-portance.Therefore,thepresentstudyfocusedonstructuralcharacterizationofrapidlysolidiedAl-Sialloyscomparativelywiththeirconventionallyproduced(ingot)counter-parts.AlthoughrapidsolidicationmayincreasetheequilibriumsolidsolubilityofSiin-Al,howtheamountofretained-SiinAlsolidsolutioncanbedeterminedisstillcon-troversial.Hence,themainpurposeofthepresentstudywastodeterminetheretainedSiamountinthe-Alphaseindetail.2.Melt-SpinningProcessingAlargenumberofdevicesdesignedforproducingrapidsolidicationmaterialhavebeenreportedoverthelastthreedecadesandthesecanbeclassiedinmanydierentways.Jones[24]classiesthemintothreecategories:(i)spraymethods,(ii)surfacemethods,and(iii)chillmethods.Spraymethodsinvolvewiththeproductionofaliquidmetalspray,whichislatertoberapidlysolidied.Thevariousdevicesinspraymethodsdierintwoaspectsas:Thoseonesusedforproductionoftheliquidsprayandthoseforquenching.Thespraymethodsincludesprayatomizationanddeposition,andtheguntechnique.Themostsignicantcharacteristicoftheweldmethodisthatbothmeltingandso-lidicationoccurinsituatthechillsurface,whichmayitselfbepartiallymeltedintheprocesses[24].Theprincipleinvolvedwithchillmethodsistheproductionofathinsectionofliquid UZUN,KARAASLAN,KESKmetalthatiscooledbyalargerchillblock.Amongthesetypesofmethodsthemeltspinningisthemostwidelyusedonebecauseofitsrelativelyhighcoolingratesupto,enablingproductionofthemetastablecrystallineoramorphouscontinuousstrips.Inadditiontothat,thehighpracticalityofthetechniquefavourshighvolumeindustrialmanufacturing[25,26].Infact,thedevelopmentofthistechniquehasbeenprimarilyresponsiblefortheacceleratedprogressofrapidsolidicationtechnologysincetheseventies.Thenameofthemelt-spinningprocessderivesfromitsinvolvementwiththeextrusionofmoltenmetalinawaywhichisakintothatusedtomanufacturesynthetictextilebres.Themelt-spinningcanbecarriedouteitherbyextrudingaliquidstreamintoacoolingmedium,calledthefree ightmelt-spinning(FFMS),orbyallowingamoltenjettoimpingeonarotatingchillblock,calledchillblockmeltspinning(CBMS).Theformermethodiswellknowntopolymertechnologistswhilethelaterdatesbackto1908andoftenusedtoproviderapidlysolidiedmaterials.Inthepresentwork,wealsousedtheCBMStechniquetoproducemelt-spunribbonsofAl-Sialloys.3.Experimental3.1.SpecimenPreparationThechillblockmelt-spinningdeviceisschematicallyshowninFig.1.Itbasicallyconsistsoftwoparts:(i)themeltingsystemand(ii)therotatingdisc.Themeltingsystemconsistsofafusedsilicatube65mmindiameterwithcompleteopeningsatbothends.Aclose-ttinggraphitecrucible,withoneendopentheotherrestrictedwithanozzle2mmindiameter,wasplacedintoafusedsilicatube.Thegraphitecrucibleisusedtobothheatthealuminumalloyandtoholdtheresultingmelt.Thefusedsilicatubewascappedwithagastightcaphavinganinletinthecentertoletargongasin.Therotatingdisc,22cmindiameter,wasmachinedfromabrassblock.Temperaturewasmeasuredbyplacingacromel-alumelthermocoupleatthecentralregionofcruciblewall.Firstly,Al-Sialloyswithnominalcompositionsof8,12,16wt%Siweremanufacturediningotmannerbymelting99.9%pureAland99.995%Simetalsinagraphitecru-cibleundervacuum.Afterthat,theseingotmaterialswerecutintopiecesfromwhichmelt-spunribbonswillbeproducedbytheMeltSpinning(MS)system.Secondly,thesesectionedingotmaterialswerere-meltedinagraphitecrucibleandspunintoribbonsintheMSsystem.Themoltenalloywithan850Cwasejectedfromthecrucibleunderanargonpressureofapproximately9,807.10Paontothebrasswheelrotatingatasurfacevelocityof40ms.Thus,therapidlysolidiedribbonswereproducedrangingfrom15mto150minthickness,from1mmto4mminwidthandfrom5cmto300cminlength.Asmentionedabove,wecannotgivecertainvaluesfortheaveragethickness,widthandlength.Butthesampleswiththeclosestvaluesforwidth,lengthandthicknesswerechosentoobtainthegreatestpossibleuniformity. UZUN,KARAASLAN,KESK compressed argonFigure1.Schematicofmelt-spinningprocess.3.2.MetallographyMicrostructuralexaminationsofbothingotandribbonsampleswerecarriedoutwithanOlympusBH2opticalmicroscope(OM)andtheJEOLJSM5400scanningelectronmicroscope(SEM).WheelandairsidesofribbonsandingotsampleswereexaminedaftermechanicalpolishingandetchinginaKeller-Wilcoxreagent.Afterthespecimenswerecoatedwithagoldlayertoenhancecontrast,SEMmicrographswereusuallytakenatanacceleratingvoltageof20kV.3.3.X-raydiractionForX-raydiractionanalyses,eachingotandribbonwereproducedthenexaminedwithoutapplicationofanytreatment.X-raydiractionstudiesforphaseanalysesandthelatticeparametermeasurementswereperformedusingaRigakuD/max-IIICDiractome-ter.X-raydiractionpatternswereobtainedutilizingCuKradiationwithawavelengthof1.5406A.Forphaseidentication,measurementswerescannedforawiderangeofdiractionangles(2)rangingfrom20to100degreeswithascanningrateof5deg/min.Thelatticeparametersweredeterminedfromtheindexplanesre ectionsofthealuminium-basedsolidsolution.ToenhancetheaccuracyofX-raymeasurementsforthelatticeparameterof-Al,bothscanningintervalandscanningrateweredecreasedto0.004degreeand0.2deg/min,respectively.4.ResultsandDiscussionMicrostructuralanalyseswereconductedbyopticalandscanningelectronmicroscopy.Themicrostructuresofingotsamples,knownasconventionallycastalloys,aregivenin UZUN,KARAASLAN,KESKFigs.2and3.OpticalmicrographsofingotAl-8wt%SialloyaregiveninFig.2a-b.Bothmicrographs,showedsimilarmicrostructureswithanonuniformdistributionofneedle-likeSiparticlesinthematrixof-Al.ThemicrographsofAl-12wt%Sialloy,giveninFig.2c-d,alsoshowedsimilarmicrostructureasAl-8wt%Si,exceptwithincreasingSiconcentration.Similarresultshavebeenreportedelsewhere[27-29].Al-16wt%Sialloyexhibitsnotonlyneedle-likeeutecticSiphasebutalsolargefacetedmassiveprimarySicrystalsthatsignifyahypereutecticmicrostructure(Fig.2e-f).WhenweconsidertheequilibriumphasediagramofAl-Sialloy,theaboveresultswouldbeexpected. a) Cross-sectionb) Longitudinal-sectionc) Cross-sectiond) Longitudinal-sectionf) Longitudinal-section Figure2.Opticmicrographsofingot(a,b)Al-8wt%Si(c,d)Al-12wt%Si(e,f)Al-16wt%Sialloys. UZUN,KARAASLAN,KESK b) Longitudinal-sectionc) Cross-sectiond) Longitudinal-sectionf) Longitudinal-sectionFigure3.SEMmicrographsofingot(a,b)Al-8wt%Si(c,d)Al-12wt%Si(e,f)Al-16wt%Sialloys.SEMmicrographsofthesameAl-8,12and16wt%SiingotalloysareshowninFig.3a-e.Similarly,thosemicrographs,giveninFig.3a-d,alsoshowednonuniformdistributionofneedle-likeSiphaseinthe-Almatrixforboth8and12wt%Sicomposition,whiletheAl-16wt%Simicrographs,giveninFig.3e-f,exhibitahypereutecticmicrostructure. a) Air sideb) Wheel sidec) Air Side d) Wheel sidee) Air sidef) Wheel sideFigure4.SEMmicrographsofmelt-spun(a,b)Al-8wt%Si(c,d)Al-12wt%Si(e,f)Al-16wt%Sialloys. UZUN,KARAASLAN,KESKTheSEManalysesoftheas-quenchedribbonswerecarriedoutonbothsurfacestodeterminethecoolingeectofthesubstrate.Whenweexaminetheairandwheelsidesofribbons,inFig.4,wenoticethattheairsidesgenerallyshowanebrancheddendriticmicrostructure,whilewheelsidescommonlyappeartopossessacellularmicrostructureforallmelt-spunAl-8,12,16wt%Sialloys.Thecellularstructureofribboncontaining16wt%Sishowedaslightdierencefromthoseofothersamples,indicatingthepresenceofahighSiratio.Microstructuralexaminationsrevealedthatthestructureofthespecimenspreparedbythemelt-spinningmethodwerequitedierentfromthoseofingotspecimens.NoneoftheprimarySicrystalwasobservedintheSEManalysisofmelt-spunribbonsincontrasttothatofingotcounterparts.Thisresulthasrevealedthat,undertheequilibriumcon-ditions,ahypereutecticAl-16wt%Sialloycanhaveahypoeutecticstructureduetotheexposureofundercoolingthroughrapidquenching.Thiseecthasbeenpreviouslyiden-tiedineutecticmetalsystems[19].Dependingonthefastestgrowingphaseatagivenundercoolingcondition,astructurewhichconsistsofbothprimaryandeutecticphasesoronlythelaterdevelops.Inthealuminium-siliconsystem,theprimaryaluminiumrichphasegrowsdendritically,whereasthesiliconrichphasegrowsinafacettedmanner.Comparingwithdendriticalphase,thefacetedgrowssloweratundercooling.Previously,JacobsonandMcKittrick[30]reportedthatrelativelylowundercoolingapplicationonanalloyofagivencompositioncausesdendriticalgrowthinwhichoneofthephasesgrowsrapidlyandtheotherphasesolidiesbetweendendriteswhilehigherundercoolingleadstocellulargrowth.Basedontheabovestatements,weconcludethatthedierencebetweenmicrostructuresofbothsidesresultedfromdierentcoolingrates.Thatwasexpectedduetothefactthatsolidicationstartsatthewheelside.Additionally,thelackofanyobservationofSicrystalisaresultofsuppressionofsiliconnucleation.XRDpatternsforallingotsandribbons,illustratedinFigs.5-7,showsimilarityinwhichallthepossible-AlandSiphasere ectionsareexpectedlypresent.Thosepatternsforquenchedribbons,showninthesamegureswiththoseofingotspecimens,alsoexhibitedallthepossiblere ectionsofAl-solidsolution,however,onlyveryweakandbroadre ectionsofSiwererecognized.Incontrasttotheingotcase,thelackofSire ectionsforquenchedribbonsintheXRDanalysisrevealedthatasubstantialamountofSihasbeenretainedinthesolidsolution.Todeterminetheamountofretained-Siuponrapidquenching,thelatticeparameterof-AlphasewasdeterminedusingBragg'sformulaforbothingotandas-quenchedribbons.ThelatticeparametersforallSicontentsaregiveninTable1.Intheory,thelatticeparameterof-AlphasedecreaseswithincreasingSicontentinthesolidsolution[11].Thequantityofdecreasehasbeendetectedviatheclosestdistanceofapproachoftheconstitutionalatoms[11].Theamountofdecreaseinthelatticeparametersforthe8,12,16wt%Sicontentsweredeterminedas
a=6.4x10nm,
a=8.3x10nm,and
a=9.3x10nm,respectively,ascalculatedinearlierstudies[19-21].Thevariabilityinthe-AllatticeparametersresultedfromextensionofSisolidsolubilityinAl-Sialloyshasearlierbeenreportedbyseveralresearchers[19-23]. UZUN,KARAASLAN,KESKTable.Latticeparametersandlatticeparameterchangesof-Alforeachmelt-spunandingotalloys Methodof Lattice LatticeParameter Sample Alloy Parameter Change Production (a;nm (
a;nm0.0005) Al-8wt%Si Ingot 4 Melt-Spun Al-12wt%Si Ingot 4 Melt-Spun Al-16wt%Si Ingot 4 Melt-Spun Figure5.TheXRDpatternsofAl-8wt%Si,a)Ribbons,b)Ingotalloys. UZUN,KARAASLAN,KESK Figure6.TheXRDpatternsofAl-12wt%Si,a)Ribbons,b)Ingotalloys.Thesevalues,forchangeinlatticeparameters(
),rangingfrom3.7x10nmto1.23x10nm,werereported,anddeningeachincreasing1at%retained-Siin-Alsolidsolution.Theamountsofretained-Siin-Alcalculatedfrombothofthelowestandhighest
valueswouldbequitedierent.Forinstance,whilethehighestvaluetakenasthechangeinlatticeparameter,theretained-Siamountwouldbe1%,thatresponds33%retained-Sibasedontheassumptioninwhichthelowestvalueacceptedasthechangeinlatticeparameter.Contrarytoingotcounterparts,XRDanalysesofmelt-spunAl-8,12,16wt%SialloysdidnotexhibitanystrongandsharpSire ection(Figs.5-7).Thereisonlyaveryweakandbroad(111)re ection.Therefore,wemaysaythatthealladdedSiamountwerenearlysolvedin-Alforeachmelt-spunAl-Sialloys.However,becauseoftheveryweakandbroadSire ectionsinthemelt-spunribbons,acertainSiamountcouldbelostinthe UZUN,KARAASLAN,KESKbackgroundwithoutbeingobserved.InourXRDanalysesofallmelt-spunalloys,noneofthepossiblere ectionsfromthe(220)indexplaneswith55%relativeintensitywereobserved.Similarly,thepossiblere ectionswiththelowerrelativeintensitiesfromtheotherindexplanes[(311),(400),(331),(422)]werenotobserved.Hence,wecanconcludethatatleast55%oftheaddedSiamountshavebeensolvedin-Alforallmelt-spunalloys.Thesevaluesare4.4wt%for8wt%Si;6.6wt%Sifor12wt%Si;8.8wt%Sifor16wt%Sicontents.Thus,wecansaythattheapplicationofmelt-spinningmethodinproducingmelt-spunalloyssignicantlyenhancestheequilibriumsolidsolubilitylimitofSiin-Al,whichwasreportedas0.02wt%Siatroomtemperature[10,31]. Figure7.TheXRDpatternsofAl-16wt%Si,a)Ribbons,b)Ingotalloys.Ontheotherhand,XRDanalysesshowthateachmelt-spunribbonhadonesharpandhighAlre ectioncontrarytoingotcounterpartspossessingnone.However,noneofotherpossibleAlre ectionswasdetected.Thisremarkableobservationresulted UZUN,KARAASLAN,KESKfromtheparallelorientationofoxidelmtotheribbonsurfacewithoutpresenceofanyintermetallicphase.Thelackofanyintermetallicphaseinmicrostructuralexaminationssupportsthesuggestionstatedabove.TheoxidelmontheribbonsurfaceresultedbecausevacuumwasnotusedduringtheMS.5.Conclusions1.Dierentcoolingratesresultindierentmicrostructuresonbothairandwheelsidesofribbons.2.ThesuppressionofsiliconnucleationpreventstheformationofSicrystalontheAl-basedmatrix.3.BothXRDandmicrostructuralanalysesshowthattheMSextendsthesolidsolu-bilityofSiin-Alofingotalloys.4.Theuseofrelativeintensitiesofthediractionlinesgivesmoreaccuratevaluesforretained-Siamountin-Alcomparedwiththeuseofthelatticeparameterschange.AcknowledgementsWewouldliketoexpressourgratitudetoTITAKfortheirnancialsupport(GrantNo:TBAG-AY/75).ThenancialsupportsfromResearchFoundationofGOU(GrantNo:95-31,96-20and00-08)andErciyesUniversity(GrantNo:96-051-1)arealsograte-fullyacknowledged.References[1]P.Duwez,R.H.WillensandW.Klement,J.Appl.Phys.(1960)1136.[2]P.Duwez,R.H.WillensandW.Klement,J.Appl.Phys.(1960)1500.[3]W.Klement,R.H.WillensandP.Duwez,Nature(1960)869.[4]I.V.Salli,LiteinoeProizvodstvo(1958)22.[5]T.R.AnantharamanandC.Suryanarayana,Rapidlysolidiedmetals,(TransTechPub.,Switzerland1987)p.5.[6]S.Das,A.H.YegnesvaranandP.K.Rohatgi,J.Mat.Sci.(1987)3173.[7]H.Jones,Mat.Sci.Eng.A133(1991)33.[8]B.Cantor,W.T.Kim,B.P.BewleyandA.G.Gillen,J.Mat.Sci.(1991)1266.[9]B.K.Prasad,J.Mat.Sci.(1991)867.[10]O.Uzun,Ph.D.Thesis,DepartmentofPhysics,GaziUniversityInstituteofScienceandTechnology,Turkey,1998. UZUN,KARAASLAN,KESK[11]S.Su,X.Liang,A.MoranandE.J.Lavernia,Int.J.Rap.Sol.(1994)161.[12]N.Fat-Halla,J.Mat.Sci.(1987)1013.[13]M.Rooyen,N.M.VanDerPers,Th.H.KeijserandE.J.Mittemeijer,Mat.Sci.Eng.,(1987)17.[14]B.K.Yen,J.Mat.Sci.,(1997)821.[15]M.Itagaki,B.C.GiessenandN.J.Grant,Trans.Quart.ASM.,(1968)330.[16]H.Matja,K.C.Russel,B.C.GiessenandN.J.Grant,Metall.Trans.A,6(1975)2249.[17]I.Yamauchi,I.Ohnaka,S.KawamotoandT.Fukusako,Trans.Jap.Ins.Met.,27,3(1986)[18]S.J.Hong,T.S.Kim,W.T.KimandB.S.Chun,Mat.Sci.Eng.A226(1997)878.[19]P.Todeshini,G.ChampierandF.H.Samuel,J.Mat.Sci.,(1992)3551.[20]S.K.BoseandR.Kumar,J.Mat.Sci(1973)1795.[21]A.Bendijk,R.Delhez,L.Katgerman,Th.H.DeKeijsher,E.J.MittemeijerandN.M.VanDerPers,J.Mat.Sci.,(1986)2803.[22]M.Rooyen,N.M.VanDerPers,Th.H.KeijserandE.J.Mittemeijer,Mat.Sci.Eng.,(1987)17.[23]A.M.BastawrosandM.Z.Said,J.Mat.Sci.,(1993)1143.[24]H.Jones,Mat.Sci.Eng.,(1984)145[25]R.W.Cahn,K.D.Laridjani,M.GreenholtzandR.Hill,Mat.Sci.Eng.,(1976)83.[26]R.B.Pond,R.E.MaringerandC.E.Mobley,ASMseminarseries,(1976)20.[27]F.YlmazandR.Elliott,J.Mat.Sci.,(1989)2065.[28]N.Fat-Halla,J.Mat.Sci.,(1989)2488.[29]D.R.Allen,M.GremaudandJ.H.Perepezko,Mat.Sci.Eng.,A226(1997)173.[30]L.A.JacobsonandJ.McKittrick,Mat.Sci.Eng.,(1994)355.[31]T.B.Massalski,Binaryalloysphasediagrams,(AmericanSoc.forMet.,Ohio1986)p.165.