Ag对Cu-Fe原位复合材料显微组织和性能的影响(英文原文+中文翻译)

www.actamat-journals.com

EffectofAgonthemicrostructureandproperties

ofCu–Feinsitucomposites

HaiyanGao,JunWang,DaShu,BaodeSun

*TheStateKeyLaboratoryofMetalMatrixComposites,ShanghaiJiaoTongUniversity,Shanghai200030,PRChina

Received11July2005;receivedinrevisedform24July2005;accepted26July2005

Availableonline19August2005

Abstract

MicrostructureandpropertiesofthreeCu–Fe–Aginsitucompositeswereinvestigated.DistributionofFeandAginthematrixwasanalyzedbyenergydispersiveX-rayspectroscopy.ThepresenceofAgreducedthesolutionofFeinthematrixathightempe-rature,andthestrengthandconductivityofCu–Fe–Agcompositesincreasedatthesametime.Ó2005ActaMaterialiaInc.PublishedbyElsevierLtd.Allrightsreserved.

Keywords:Cu–Fe–Ag;Insitucomposite;Microstructure;Mechanicalproperties;Electricalproperties

1.Introduction

Binarycopper-basemetalmatrixcompositescontain-ingabody-centeredcubic(bcc)transitionmetalsuchasNb,Fe,Cr,preparedbymechanicalprocessing,havebeenthesubjectofextensiveresearchinthepasttwodecades[1–6].Uponsolidificationofthesealloys,bccdendritesforminthecoppermatrixandsubsequentmechanicaldeformationreducesthedendritestoalignedfilamentswitharibbon-likecross-section[3,4,7].Thelimitedsolubilityofthesebccmetalsincopperallowsthematrixtoretainahighconductivity,makingthisinsitucompositeinterestingforapplicationswherehighconductivityandhighstrengtharerequired.Investiga-tionsfollowingtheselineswerealsoconductedonbinaryCu–Agcomposites[8–11].

TheCu–Fesystemisofparticularinterestbecauseofthelowcostofironcomparedtootherpossiblemetals.PureNbandAgarecostlymetals,whichimpedesthe

Correspondingauthor.Tel.:+862162932914;fax:+862162932870.

E-mailaddresses:gaohaiyan@sjtu.edu.cn(H.Gao),bdsun@sjtu.edu.cn(B.Sun).

*largescaleapplicationofsuchnewmaterialsindevicesotherthanhighfieldmagnets[12].However,theCu–Fesystemshavealowerconductivitythanothercopper-baseinsitucomposites.Thereasonliesin(1)therelativelyhighersolubilityofironincopperathightemperatures,coupledwithslowkineticsofironprecip-itationatlowertemperatures;and(2)theparticularlyharmfuleffectontheconductivityofironatomsinsolidsolution,namely9.2lXcm/wt.%Fe[13].Therefore,itisimportanttoremoveasmuchironfromsolidsolutioninthecopperaspossible.HansenandAnderko[14]reportedthefollowingsolutioncorrelation,x=2540·exp(À8800/T),whichpredictsanequilibriumsolubilityoflessthan1ppmFeat253°C.Ifthepredictedsolubil-ityat253°Ccouldbeachieved,theconductivityofthecoppermatrixwouldbereducedbyonly1%IACS(InternationalAnnealedCopperStandard,17.241nXmisdefinedas100%IACS)usingtheaboveresisti-vitydecrement.Thermal–mechanicaltreatmentshavebeenemployedtooptimizethestrengthandconducti-vity[13].

Intherecentyears,effortshavebeenmadetowidenthespectrumofinsitucompositestowardsternarycopper-basealloys[15–18].Thedevelopmentofternary

1359-6462/$-seefrontmatterÓ2005ActaMaterialiaInc.PublishedbyElsevierLtd.Allrightsreserved.doi:10.1016/j.scriptamat.2005.07.028

1106H.Gaoetal./ScriptaMaterialia53(2005)1105–1109

Cu–Fecompositesusuallyaimsatfurtherimprovingthestrengthandconductivityofconventionalbinaryalloys,andreducingthecostsaswell.Addingathirdelementallowstheexploitationofalargervarietyofpossiblekineticpathsforattainingacertainflowstress–conduc-tivityprofile.Forexample,Cu–15wt.%Fe–0.1wt.%Mghasatensilestrengthof1080MPawithaconductivityof56%IACS[13].Songetal.[19]reportedthatCu–9wt.%Fe–1.2wt.%Aghasastrength/conductivitycombinationof939MPa/56.2%IACSthroughthermal–mechanicaltreatment.Inthisstudy,theeffectsofAgonthemicrostructureandpropertiesinthecastanddrawnCu–Fe–Aginsitucompositeswereexamined.

2.Experimentaldetails

Threecompositionswerechosenforinvestigation,namely,Cu–12wt.%Fe–1wt.%Ag(denotedCu–12-Fe–1Ag),Cu–14wt.%Fe–3wt.%Ag(denotedCu–14-Fe–3Ag)andCu–11wt.%Fe–6wt.%Ag(denotedCu–11Fe–6Ag).Semi-sphereingotsofabout20mmindiameterwereseparatelypreparedfromelectrolyticCu,commercialFeandAgwithatleast99.9wt.%purityusingtungstenarc-meltingwithelectromagneticstirring.Theingotswerehotforgedintheopenairandthenfac-ing-machinedtoasectionaldimensionof14·14mmtoremovetheoxidationlayerandthesurfacedefects.Compositewireswereproducedthroughcoldrollingandsubsequentdrawingthroughvarioussuccessivedrawingdies.Duringrolling,thesectionaldimensionofthesampleswasreducedto5·5mminaseriesofsteps.Duringdrawingthesamplediameterwasfurtherreducedtoaminimumdiameterof0.4mm.Forcom-parisonpurposes,abinaryCu–12wt.%FecompositewasalsopreparedalongwithCu–Fe–Agcompositesusingthesameprocess.Thecoldworkstrainwasdefinedbyg=ln(A0/Af),whereA0istheinitialsectionalareaobtainedafterhotforgingandAfisthefinalsectionalarea.

TheevolutionofthemicrostructureatdifferentstrainswasinvestigatedusingaSirion200fieldemissionscanningelectronmicroscopy(FESEM).Thedistribu-tionofFeandAginthecoppermatrixwereanalyzedbyanenergydispersivespectrometer(EDS)inconjunc-tionwiththeFESEM.TensiletestsondeformedwirespecimensatroomtemperaturewereconductedusingaZwick/Roellmachineequippedwithanextensometerforaccuratestrainmeasurements.Thedeformationratewasfixedat2.5·10À4sÀ1attheelasticstageand5.0·10À4sÀ1attheplasticstage.Theultimatetensilestress(UTS)wastakenasameasureofthestrengthforcomparisonpurposesbecauseitwasveryreproduc-ibleandwelldefinedforsimilarspecimens.Specimensweretestedwithoutanyreductioninthegaugediameterduetotheirrelativelysmalldiameter.Atleastthree

specimensweretestedateachwiresizeandthereproducibilitywaswithin2%.ElectricalresistivitiesofspecimensweremeasuredusingaZY9858digitalmicro-ohmmeteratroomtemperature.

3.Resultsanddiscussion

Thefouralloysexhibitsimilarcastmicrostructure,i.e.,theFedendritesareembeddedinthecoppermatrixandrandomlyorientedwithrespecttotheingotaxis.ThedendritearmsofCu–Fe–Agcompositesareabout2–4lmindiameter,smallerthanthatofbinaryCu–Fecomposite(4–6lm),whichisinagreementwiththeobservationsofSong[15].Inaddition,thegreatertheamountofAgadded,thesmallertheFedendrites.High-ermagnificationobservationrevealedthatfewprecipi-tatescouldbeobservedinthematrixduetochillcoolingemployedintheexperiment.However,besidesFedendrites,someCu–AgeutecticcouldalsobeseendistributingalongthecopperboundaryinCu–11Fe–6Ag,asshowninFig.1.

TheequilibriumphasediagramofCuandFerevealsthatCuandFehavesmallmutualsolubilitythatdecaystonearzeroattheroomtemperature.ThemaximumsolubilityofFeinCuis4.1wt.%atperitectictempera-ture1096°C,andthatofCuina-Feis2.1wt.%ateutec-toidtemperature850°C.WithoutAgaddition,theamountofFeinthe‘‘matrix’’(whichinthispaperisusedtorefertothemicrostructureexcludingtheFeden-drites)shouldbethesameforthefouralloysbecausethetotalamountsofFeinthefouralloysareallfarbeyondthemaximumsolubilityofFeinCu.CuandAghavelimitedmutualsolubility,theirequilibriumphasediagramrevealsthemaximumsolubilityofAginCuis7.9wt.%attheeutectictemperature780°Canddowntozeroatroomtemperature.Duringnon-equilibriumsolidification,asAgcontentisapproaching7.9wt.%,

Fig.1.Microstructureofas-castCu–11Fe–6Agalloy.Cu–Ageutecticcouldbeseenclearlydistributedalongthecopperboundaries,theupper-leftcornerofthepictureisahighermagnificationimage(25,000·)oftheeutectic.

H.Gaoetal./ScriptaMaterialia53(2005)1105–1109

1107

Table1

DistributionofFeandAgofcomposites

Cu–12Fe

NominalcompositionFewt.%12.0Agwt.%–Matrix

Fewt.%4.5Agwt.%

–Feinformofdendrites,wt.%

7.5

aAmountofAgmeasuredisgiveninarange.

e.g.,theweightratioofAg/Cuis6.7wt.%inCu–11Fe–6Ag,theeutecticreactionislikelytotakeplaceduringchillcooling.

EDSanalysiswasusedfor‘‘matrix’’compositionmeasurements.Toimprovemeasurementprecision,selectedareasof10lm·5lminthe‘‘matrix’’wereanalyzedathighermagnificationobservationsof4000·andeightmeasurementswererecordedforeachspecimen,theaverageresultsareshowninTable1.MostoftheFeformsintodendritesduringsolidificationandtherestdissolvesinthematrixintheformofsolidsolution.However,Agexistsinthe‘‘matrix’’,eitherintheformofsolidsolutionorCu–Ageutectic.ItshouldbepointedoutthatthesmalldeviationofAgfromthenominalcompositionoftheCu–Fe–Agalloysisattri-butabletothemicro-non-homogeneityofthematerial.ForCu–11Fe–6Ag,inparticular,theamountofAginthematrixisnotaconstant,whichmayhaverelevancetotheformationofCu–Ageutecticinthe‘‘matrix’’.Furthermore,theamountofFein‘‘matrix’’decreasesobviouslywithanincreasingamountofAg,asshowninthefourthlineinTable1,whichindicatesthatthepresenceofAgpromotestheprecipitationofc-FefromliquidCu,i.e.,reducesthemaximumsolutionofFeinCuathightemperature.Thiscanbeexplainedby‘‘selec-tivesolution’’ofcopper.ComparedwithFe,Agatomshaveprioritytodissolveincopperbecauseofitssimilaratomradius,electronicstructureandcrystalstructuretocopper.Duringsolidification,asthemelttemperaturedropsneartheliquiduscurve,e.g.,20°Cabovetheliq-uidus,a‘‘quasi-solidphase’’hasformedinthemeltsduetocomponentfluctuation.Atthismoment,oncesuper-saturatedAgandFeexistincoppermeltsatthesametime,Agatoms,withthehelpoftheirpreferentialsolu-tionproperties,willoccupymanyadvantageousposi-tionsandimpedefurthersolutionofFeinCuaswell.Asaresult,greateramountsofFeatomsareforcedtoprecipitateintheformofprimaryphase.

MicrostructuresofdeformedCu–Fe–Agcompositesaresimilartothosereportedinpreviousstudies.Inthelongitudinaldirection,initiallyrandomlydistributeddendritestransformedintoalignedfibersgradually,whilethefilamentstakeonribbon-likemorphologyotherthancircularinthetransversedirection.Inmanypreviousstudies[3,4,7,20],ithadbeenreportedthat

Cu–14Fe–1AgCu–14Fe–3AgCu–11Fe–6Ag14.314.110.61.12.85.64.13.82.5

1.12.82.8–8.3a10.2

10.3

8.1

theirregularshapeoftheironphaseresultingfromtheh110ifibertexturesformedduringdeformation,whichpromotesplanestraindeformationratherthanaxiallysymmetricflow.However,thecoppermatrixdoesde-forminanaxiallysymmetricmannerduringthewiredrawing,andtheironfilamentsareconstrainedandforcedtofoldortwistaboutthewireaxistomaintaincompatibilitywiththematrixresultingintheirregularcross-sectionalshapes,asshowninFig.2.Furtherwiredrawingresultsinhomogeneityandrefinementoffila-ments.ItisnotedthatthefinerfilamentswithagreycontrastwithdarkerFeribbonsobservedinFig.2shouldbeAgfilaments,whichweredeformedfromtheas-castCu–Ageutectics.Asdiscussedlater,thesefinerAgfilamentsalsocontributetothestrengthofthecomposites.

Fig.3comparestheeffectsofthedrawingstrainpriortothetensiletestontheultimatestressforCu–14Fe–3Ag,Cu–14Fe–1Ag,Cu–11Fe–6AgandCu–12Fe.Stress–straincurvesshowexponentialdependenceforeachcompositeingeneral,whichisinaccordancewithpreviousresearch[3,4,21].Becausenon-machinedwireswereusedforthetensiletests,failurebreakingalwaysappearsnearthegripholderbecauseofstressconcentra-tion,andthemeasurementsweresomewhatlowerthantheactualtensilestresses.Atahigherdrawratio,whilethesizeofspecimensbecamesmallerthan0.5mmindiameter,thestressconcentrationeffectbecamestrongerandsometimessmalldecrementsinmeasurementswere

Fig.2.ThetransversesectionofdeformedCu–11Fe–6Agcomposite,atg=6.9.

1108H.Gaoetal./ScriptaMaterialia53(2005)1105–1109

mationofCu–Ageutectic,whichisingoodagreementwiththeobservedUTSdifferencesbetweenthealloys.Inaddition,forCu–11Fe–6Ag,asmallamountofCu–Ageutecticmayalsocontributetothestrength,especiallyasCu–Ageutecticcandeformtosmallfibersduringthesubsequentwiredrawingandfurtherstrengthenthecomposite.

Table2givesthemeasuredelectricalresistivityofthespecimensatadrawratioof2.5.TheresistivitiesofthecompositesdecreaseobviouslywithAgadditionintothecomposites,i.e.,thepresenceofAgincreasestheconductivityofbinaryCu–Fecomposite,whichmaybeduetotheloweramountofdissolvedFeinCu.Resis-tivitiesofCu–Fecompositescanbeevaluatedusingaparallel-circuitmodel[22]:

Fig.3.Effectofdrawratiogontheultimatetensilestrengthofthefourcomposites.

111¼fCuþfFe;qCqCuqFe

wherefCuandfFearevolumefractionsofCuandFe,respectively.BecauseofminordifferencesinthetotalamountoffilamentsplusthehigherresistivityofFe,thedifferenceincompositeresistivityresultingfromtheamountofironfilamentisnegligible.Therefore,themaindifferenceliesintheresistivityofcoppermatrix,whichcanbepartitionedintothecontributionoffourmainscatteringmechanisms[22]:qC=qpho+qdis+qint+qimp,whereqphoistheresistivitycontribu-tionfromthephononscattering,qdisthedislocationresistivity,qinttheinterfacescattering,andqimptheimpurityscattering.Compositeswiththesamedrawratiohavesimilarqphoandqdis;itisapparentthatqimpandqintaretheresistivitycontrolitems.Becauseoftherelativelysmalldrawratio,theinterfacebetweenmatrixandironfibersissimilar,andthedifferenceinqintbetweencompositesisalsosmall.ThemaindifferencesinthemeasuredresistivitiescomefromtheamountofFeandAgdissolvedinCu.Manypreviousinvestiga-tions[7,13,19]reportedthattheconductivitydifferencesbetweenCu–FeandCu–NbcompositeswiththesamevolumefractionofFeorNbwerealsocausedbythedis-solvedFeinthecoppermatrix.Asmentionedintheintroductionsection,every1wt.%Fedissolvedwillin-creasetheresistivityofCuby9.2nXm,while1wt.%Agleadstoanincreaseofonly1nXm.Resistivitydiffer-encesresultingfromdifferentdissolvedconcentrationsofFeandAgbetweenCu–Fe–AgandCu–Fecouldbecalculatedbasedupontheaboveresistivitydecrement;theresultsarelistedinTable2consideringallAgadded

observed.Intherange25,thestrengthofCu–14Fe–3Agincreasesfas-ter,andatadrawratioof6.0,theultimatetensilestressofCu–14Fe–1Ag,Cu–14Fe–3AgandCu–11Fe–6Agcompositesreach1119,1578and1357MPa,respec-tively,whilethatofCu–12Feisonly978MPa.

Previousstudiesrevealedthatduringtheinitialstageofdeformation,initiallyrandomlyorienteddendritestransformedintoalignedfibers,andthestrengthofcopper-baseinsitucompositeinthisstageobeystheruleofmixturesandHall–Petchcorrelationsthereafter.Hence,thestrengthsofCu–14Fe–3AgandCu–11Fe–6Agarecloseintherange25duetothehigherFefibercontent.TernaryCu–Fe–AgalloyshaveadditionalAgdissolvedinthecoppermatrix,leadingtosolutionstrengtheningofthematrix.Asaresult,evenwithsimilarironfibercontent,thestrengthofCu–14Fe–3AgisalwayshigherthanthatofCu–14Fe–1AgandsodoesbetweenCu–11Fe–6AgandCu–12Fe.HongandHillsuggested[9]thatthestrengtheningcomponentduetoalloyinginCu–8wt.%Agwas300MPa.IfthestrengtheningeffectduetoAgprecipitationispropor-tionaltoC1/2,whereCisthesilvercontent,thefrictionstressduetoAgprecipitationinCu–14Fe–1Ag,Cu–14Fe–3AgandCu–11Fe–6Agwillbe106,184and260MPa,respectively,withoutconsiderationofthefor-Table2

Resistivitiesofcomposites

Cu–12Fe

Resistivitymeasured,nXm

Resistivitydifferencescalculated,nXm

DuetoFeinmatrixDuetoAginmatrixTotal

41.2–––

Cu–14Fe–1Ag38.4À3.7+1.1À2.6

Cu–14Fe–3Ag37.3À5.7+2.8À2.9

Cu–11Fe–6Ag31.5À18.4+5.6À12.8

H.Gaoetal./ScriptaMaterialia53(2005)1105–11091109

intheformofsolidsolution.Itisfoundthatthecalcu-lateddifferencesagreewellwiththemeasurementdiffer-encesforthreeCu–Fe–Agcomposites.

4.Conclusion

TernaryCu–Fe–Aginsitucompositeswereinvesti-gated.TheeffectsofAginCu–Fe–Agcompositescanbesummarizedasfollows:(1)refiningtheprimaryFedendrites;(2)reducingthesolubilityofFeinCuathightemperature,andimprovingtheconductivityofbinaryCu–Fecomposites;and(3)strengtheningcompositesintheformofsolidsolutionorCu–Ageutecticanditsfibersduringsubsequentdeformation.SotheadditionofAgtoCu–Fegivesbothbetterstrengthandconduc-tivityoftheinsitucomposites.Ifproperheattreatmentsareusedduringdeformation,thestrengthandconduc-tivityofthecompositeswillbefurtherimproved.

Acknowledgments

TheauthorsacknowledgeM.Y.XiaandB.ChenofShanghaiNonferrousInstituteformaterialsprepa-rations.

References

[1]BevkJ,HarbisonJP,BellJL.JApplPhys1978;49:6031.

[2]FunkenbuschPD,CourtneyTH,KubischDG.ScriptaMetall1984;18:1099.

[3]FunkenbuschPD,CourtneyTH.ScriptaMetall1981;15:1349.[4]SpitzigWA,PelronAR,LaarsFC.ActaMetall1985;35:2427.[5]SnoeckE,LecouturierF,ThillyL,etal.ScriptaMater1998;38:1643.

[6]RaabeD,HangenU.ComputMaterSci1996;5:195.[7]BiselliIC,MorrisDG.ActaMetallMater1994;42:163.[8]BenghalemA,MorrisDG.ActaMetallMater1997;45:397.[9]HongSI,HillMA.ActaMater1998;46:4111.

[10]HongSI,HillMA.MaterSciEng1999;A246:151.

[11]HanK,VasquezAA,XinY,KaluPN.ActaMater2003;51:767.[12]RaabeD,GeJ.ScriptaMater2004;51:915.

[13]VerhoevenJD,ChuenSC,GibsonED.JMaterSci1989;24:1748.[14]StarkeMR,HoltDL.ActaMetall1972;20:569.

[15]SongJS,HongSI,KimHS.JMaterProcTechnol2001;113:610.[16]SongJS,KimHS,LeeCT,HongSI.JMaterProcTechnol

2002;130–131:272.

[17]RaabeD,MattissenD.ActaMater1998;46:5973.

[18]RaabeD,MiyakeK,TakaharaH.MaterSciEng2000;A291:186.[19]SongJS,HongSI,ParkYG.JAlloysCompd2005;388:69.[20]GoYS,SpitzigWA.JMaterSci1991;26:163–71.[21]SpitzigWA.ActaMetallMater1991;39:1085.

[22]VerhoevenJD,DowningHL,ChumbleyLS,GibsonED.JAppl

Phys1989;65:1293.

文 献 翻 译

论文名称Ag对Cu-Fe原位复合材料显微组织和性能的影响

学 院 材料科学与工程学院 班 级 105090102 学生姓名 杨海龙 学 号 10509010232 指导老师 周志明

1

Ag对Cu-Fe原位复合材料显微组织和性能的影响

高海燕，王军，舒达，孙宝德

上海交通大学金属基复合材料重点实验室，上海 200030, 中国

摘要：本文研究了3种Cu-Fe-Ag原位复合材料的显微组织与性能。通过采用X射线

衍射仪分析了Fe和Ag在Cu基体中的分布。高温下，Ag的存在降低了Fe在合金中的固溶度，同时使Cu-Fe-Ag合金的强度和电导率增加。

关键词：Cu-Fe-Ag合金，原位复合材料，显微组织，力学性能，电性能。 1. 前言

在过去的二十年间，二元体心立方（bcc）Cu基金属复合材料的金属，如Nb，Fe，Cr等Cu基合金的机械加工，一直受到广泛的研究[1-6]。根据合金的凝固作用，Cu基体中的bcc枝晶和随后的机械变形减少枝晶，并且变形后枝晶在横截面上上形成板条状的纤维结构[3,4,7]。这些体心金属在Cu基体中固溶度有限，使这些合金保持了高的导电性，这种原位复合材料在高强高导方面得到广泛的应用。在这些合金中，二元Cu-Ag复合材料也进行了研究。

Cu-Fe合金系列由于成本比别的材料更加低廉，因此引起了大家广泛的研究。纯Nb和Ag都是昂贵的金属，它阻碍了这种新材料在设备上的大规模应用和高强磁Fe的推广。然而，Cu-Fe系合金的导电性比其他Cu基原位复合材料低，原因在于Cu-Fe合金在高温下有相对较高的固溶度，而且Fe在低温时析出动力缓慢；并且固溶的Fe原子对导电性能尤为有害，即9.2μΩcm/wt.%Fe[13]。所以降低Fe在Cu中的固溶度至关重要。Hansen和Artore报道了固溶关系。x = 2540 × exp(-8800/T)，这意味着Fe在253 ℃下的平衡固溶度小于1 ppm。如果预测的固溶度在253℃可实现的，那么Cu基体的电导率将只减少1％IACS（国际退火Cu标准， 17.241μΩm的定义为100％IACS）。目前已使用热机械处理来优化合金强度和电导率[13]。

2

近年来，已作出努力拓宽三元Cu基合金原位复合材料范围。发展三元Cu-Fe复合材料通常旨在进一步提高常规二元合金的强度和电导率，并降低成本。为了使合金的流变应力-电导率达到一定的要求，加入第三种元素，使大量不同可能的动力学途径加以开发，以实现一定的流变强度和电导率综合性能。举例来说，Cu-15wt％Fe-0.1wt％Mg[13]，获得了1080MPa的抗拉强度与56％IACS的电导率。Song等[19] 发现Cu–9Fe–1.2Ag合金通过热处理后具有强度/电导率的良好结合，分别为939MPa/56.2％IACS的综合性能。本文研究了Ag对Cu-Fe-Ag原位复合材料铸态和变形后的显微组织与性能的影响。

2．实验方法

分别选定三种成分研究，即Cu-12wt％fFe-1wt％Ag（标注为Cu-12-Fe-1Ag） ，Cu-14wt％Fe-3wt％Ag（标注为Cu-14Fe-3Ag）和Cu-11wt％Fe-6wt％Ag（标注为Cu-11Fe-6Ag）。分别将纯度至少是99.9wt％的电解Cu、商用Fe和Ag采用带电磁搅拌的钨电极电弧熔炼成直径约为20毫米半球铸锭。将熔炼得到的铸锭在室温进行热锻，然后再加工到去除了氧化层和表面缺陷的一个断面尺寸为14×14毫米块体。复合材料通过冷轧及随后的拉伸成丝状。轧制过程中，样品的截面尺寸通过一系列步骤减少到5 ×5毫米。在拉伸样品时，直径进一步降低到最小直径为0.4毫米。为了便于比较，将二元Cu-12wt％Fe复合材料采用与Cu-Fe-Ag复合材料相同的过程。冷加其中A0是热锻后初始截面积，Af是最后的截面积。 工应变的定义是由η=ln(A0Af)，

采用Sirion 200场发射扫描电子显微镜（FESEM）观察显微组织的演变，采用FESEM自带的能谱仪（EDS）对Cu基体中的Fe和Ag的分布进行了分析。在室温下，采用Zwick/Roell机对拉伸标本进行拉伸试验，并配备了伸缩仪以便准确测量应变。

−4−1在弹性阶段5.0×10−4s−1和塑性阶段，变形率定为2.5 ×10s。最终的拉伸强度（UTS）

由于重复性非常好而为类似的标本作为强调对比的方法。由于其相对较小的直径没有任何削减衡量直径。至少对三个样本进行了测试，每个拉伸丝大小和可重复性是在2 ％左右时，在室温下电阻率的测定采用ZY9858数字微欧计。

3 .结果与讨论

3

四种合金有相似的铸态显微组织，即Fe枝晶是嵌入在Cu基体中和随机取向与分布。。Cu-Fe-Ag复合材料枝晶的直径大约为2-4μm，这小于二元Cu－Fe复合材料的枝晶的直径(4-6μm)，这与Song[15]所观测的一致。此外，Ag的加入越多， Fe枝晶越小。采用高倍率显微镜观察到少量的沉淀物在Cu基体内，主要是由于实验中的急冷。但是如图1所示，除了Fe枝晶外，也可以看到一些Cu－Ag共晶分布在Cu-11Fe-6Ag合金Cu的边界上。

图 1铸造Cu-11Fe – 6Ag合金的显微组织。

可以清楚地看到Cu－Ag共晶分布在Cu的边界，左上角的图片是共晶的一个高放大倍率的图象（25000 ×）。平衡相图中显示Cu和Fe有较小相互固溶度和在室温下接近于零。在包晶反应温度1096℃时Fe在Cu中最大的固溶度是为4.1wt％，在共晶温度850℃时Cu在α－Fe中最大的固溶度的为2.1wt％。在无Ag条件下，由于四种合金中都远远超越了平衡状态下Fe在Cu中最高的固溶度，4种合金中Fe在基体中的含量差不多（本文中指包括Fe枝晶的显微组织）。Cu和Ag都有限的相互固溶度，其平衡相图显示在共晶温度780℃时Ag在Cu中的最大固溶度的是7.9wt％，在室温下下降到零。而在非平衡凝固条件下，Ag含量接近7.9wt％。例如，Cu-11Fe-6Ag合金中Ag/Cu重量比是6.7wt％时，在激冷时可能发生共晶反应。

表1 Fe和Ag在合金中的分布

4

EDS分析显示基体的成分分布。为了提高了测量的精度，选定基体区域的10μm× 5μm进行了高放大倍率观察（4000×）分析，并对每个标本在记录八次测量结果，其平均值如表1所示。凝固过程中，大部分的Fe以枝晶的形式进入而其余的固溶在基体之中。然而Ag在基体中或者以固溶形式存在，或者以Cu－Ag共晶形式存在。而且如表1第4行所示，Ag的加入越来越多，Fe枝晶越来越少，这表明Ag的加入促进了γ-Fe从液Cu中的沉淀析出，也就是说，降低高温下Fe在Cu的最大固溶度。这也可以解释为Cu的选择性固溶。由于类似的原子半径，电子结构和晶体结构，Ag原子比Fe优先溶解在Cu中。在凝固过程中，由于熔体温度高于液相线近20℃。由于成分波动熔体形成“准固相”。一旦过饱和的Ag和Fe共存在于的Cu熔体中， Ag原子将借助它们的优先溶解性质，占用许多有利位置，并阻碍了Fe在Cu中进一步的固溶等。因此，大量Fe原子被迫沉淀出来形成初生相。

图2 η=6.9变形Cu-11Fe-6Ag合金横截面的显微组织

以往的研究结果表明Cu-Fe-Ag复合材料变形后的显微组织相似。在纵向，最初随机分布的枝晶渐渐地转变为纤维排列，而长的纤维带与横向的不一样，以板条状形式存在。许多以往的研究中[3,4,7,20]报告说不规则形状的Fe相由于在变形过程中<110>纤维织构形成的，促进了平面应变变形，而不是轴对称变形。然而，如图2所示，Cu基体在拉丝过程中以轴向对称的方式变形，Fe枝晶受到约束并被迫扭转来保持与基体产生不规则截面形状的均匀性和细小纤维的相容性。需指出的是如图2中所

5

示，与深色Fe板条对比的是这些细小的灰色纤维是Ag纤维，主要是铸造CuAg共晶发生变形得到的。如后面讨论的，这些细小的Ag纤维也有助于复合材料的强度提高。

图3 四种合金的最终拉伸强度和应变曲线

图3，对Cu-14Fe-3Ag，Cu-14Fe-1Ag，Cu-11Fe-6Ag和Cu-12Fe合金的极限应力和应变对比进行拉伸测量。总的来说，每种合金的复合应力应变曲线呈指数关系，这与以往的研究一致[3,4,21]。用于拉伸试验非加工丝拉伸试验经常失败，是因为应力集中打破了节点，测量的拉应力低于实际拉应力。在较高的拉伸系数，测量观察到小于0.5毫米直径大小的样品应力集中效应更强，并可以观察到一些碎片。在2<η<5时，Cu-14Fe-3Ag的强度与Cu-11Fe–6Ag类似和它们的强度比Cu-14Fe-1Ag合金高80-100MPa，比Cu－12Fe合金高150-250MPa。当η>5时，Cu-14Fe-3Ag的强度增加更快，在拉伸倍数为6.0 时，Cu-14Fe-1Ag，Cu-14Fe-3Ag和Cu-11Fe-6Ag合金最终的拉伸强度分别达到1119，1578和1357 MPa，而Cu-12Fe合金只有978MPa。

以往的研究显示在变形的最初阶段，开始随机方向枝晶转化为排列的纤维，原位复合材料中Cu基体的强度在这个阶段遵循混合物规律和Hall – Petch关系。因此，在2 <η< 5时，Cu-14Fe-3Ag和Cu-11Fe–6Ag的强度接近；在η>5时，由于Fe纤维的含量较高强度速度迅速提高。三元Cu-Fe-Ag合金具有额外的Ag固溶在Cu基体中，从而导致基体固溶的强化。因此，即使Fe纤维含量相差不大，Cu-14Fe-3Ag的强度总是高于Cu-14Fe-1Ag，Cu-11Fe-6Ag和Cu-12Fe的强度。Hong和Hill[9]等认为C u-8wt％Ag合金中组元的强化作用为300MPa。如果Ag沉淀析出强化效果正比与是C1 / 2，其中C是Ag的含量。而加强了在下合金C结构的强化效应。那么由于Ag

6

的沉淀，Cu-14Fe-1Ag，Cu-14Fe-3Ag和Cu-11Fe-6Ag摩擦应力将分别为106，184和260MPa。因为没有考虑到形成Cu－Ag共晶，与实测合金最终拉伸强度吻合良好。此外，少量Ag共晶也可能对Cu-11Fe-6Ag强度也有影响，特别是作为Cu－Ag共晶在随后的拉伸变形中成为小纤维会进一步强化。

表2 合金的电阻

表2给出了拉伸率为2.5的测量的样品电阻。由于Ag的加入，复合材料的电阻明显降低，即Ag的加入提高了二元Cu-Fe复合材料电导率，这可能是由于Fe在Cu中较低的固溶度。Cu-F复合材料的电阻率，可以平行电路模型[22]：111=fCUu+fFe ρcρcuρFe

其中

f

和

Cu

f

分别是Cu及Fe体积分数的等。因为细丝总额只是稍有差别，

Fe

再加上Fe的高电阻率，复合材料由于Fe纤维数量的不同在总电阻所产生的差别是微不足道的。因此，电阻率主要的差别在于Cu基体。可分为四个主要散射机制：ρC= ρ

pho

+ ρdis + ρint +ρimp ，其中ρpho是声子对电阻率的贡献，ρdis位错电阻，ρint

界面散射，ρimp杂质散射。具有相同的拉伸倍数复合材料也有类似ρpho和ρdis，很明显ρimp和ρint是控制电阻率的主要条件。由于相对较小的拉伸倍数，基体和Fe纤维界面之间相似，在复合材料中ρint之间的相差也很小。电阻率的主要不同来自Fe和Ag固溶于Cu基体的数量。许多先前的研究表明[7,13,19]，具有相同体积分数的Fe和Nb的Cu-Fe和Cu-Nb复合材料电导率之间的差别在于Fe固溶在Cu基体中。如引言所述，每1wt.％Fe固溶会增加Cu的电阻率9.2nΩm，而1wt.％Ag导致增加的只有1nΩm。Cu-Fe-Ag和Cu-Fe之间电阻率的差异是Fe和Ag的不同固溶度所造成的。考虑到所有Ag加入形成固溶体，根据上述电阻率递减计算出

f

Fe

结果如表2所示。

结果发现经Cu-Fe –Ag合金经计算的电阻与实际测量结果十分接近。

7

4. 结论

本文研究了三元原位Cu-Fe-Ag复合材料。Ag对Cu-Fe-Ag复合材料性能的影响归纳如下：（1）细化初生的Fe枝晶;（2）减少在高温下Fe在Cu中的固溶度，提高二元Cu-Fe复合材料的电导率;（3）复合材料由于形成固溶体或CuAg共晶及后面的变形成为纤维而强化。因此，Ag的加入可以使Cu-Fe原位复合材料获得更好的强度和电导率。如果变形过程中采用适当的热处理，复合材料的强度和电导率将得到进一步提高。

参考文献

[1] Bevk J, Harbison JP, Bell JL. J Appl Phys 1978, 49:6031.

[2] Funkenbusch PD, Courtney TH, Kubisch DG. Scripta Metall.1984; 18:1099. [3] Funkenbusch PD, Courtney TH. Scripta Metall 1981; 15:1349. [4] Spitzig WA, Pelron AR, Laars FC. Acta Metall 1985; 35:2427. [5] Snoeck E, Lecouturier F, Thilly L, et al. Scripta Mater 1998; 38: 1643. [6] Raabe D, Hangen U. Comput Mater Sci 1996; 5:195. [7] BiselliI C, Morris DG. Acta Metall Mater 1994; 42:163. [8] Benghalem A, Morris DG. Acta Metall Mater 1997; 45:397. [9] Hong SI, Hill MA. Acta Mater 1998; 46:4111. [10] Hong SI, Hill MA. Mater Sci Eng 1999; A246:151.

[11] Han K, Vasquez AA, Xin Y, Kalu PN. Acta Mater 2003; 51:767. [12] Raabe D, Ge J. Scripta Mater 2004; 51:915.

[13] Verhoeven JD, Chuen SC, Gibson ED. J Mater Sci 1989; 24:1748. [14] Starke MR, Holt DL. Acta Metall 1972; 20:569.

8

[15] Song JS, Hong SI, Kim HS. J Mater Proc Technol 2001; 113:610.

[16] Song JS, Kim HS, Lee CT, Hong SI. J Mater Proc Technol 2002; 130–131:272. [17] Raabe D, Mattissen D. Acta Mater 1998; 46:5973.

[18] Raabe D, Miyake K, Takahara H. Mater Sci Eng 2000; A291:186. [19] Song JS, Hong SI, Park YG. J Alloys Compd 2005; 388:69. [20] Go YS, Spitzig WA. J Mater Sci 1991; 26: 163–71. [21] Spitzig WA. Acta Metall Mater 1991; 39:1085.

[22] Verhoeven JD, Downing HL, Chumbley LS, Gibson ED. J Appl Phys 1989; 65:1293.

9