RepresentativeXRDdataforas-sinteredandpolishedsurfacesofCCTOceramicspreparedinairat11151CareshowninFig.3.Allre ectionsfromtheas-sinteredpelletsurfacewiththeexceptionofasmallpeakatB36.51correspondtothepatternreportedintheliterature10andICDDcardnumber75-2188forCaCutheas-sintered3Ti4O12.TheadditionalminorpeakobservedatB36.51inpelletsurface(seetheinsetinFig.3)wasremovedonpolishingthepelletsurfacesandmaycorre-spondtoCu2OprecipitatesatthesurfaceasthemostintensepeakintheXRDpatternforCuthe(111)re ectionat36.5212O(ICDDcardnumber77-199)is1.Unfortunately,thesecondmostintensepeak(200re ection)forCu2Ooccursat42.4231,whichisclosetothe(111)re ectionforCCTO;seetheinsetinoFig.3.AllremainingpeaksforCu2Ophase,30%,whichandgivenwaslimitedthesmallonlyvolumetotheas-sinteredfractionhaverelativeofintensitypellettheimpuritysurface,
Fig.3.TypicalX-raydiffractionpatternsobtainedfromanas-sinteredpelletsurface(bottomtrace)andthen,sequentially,fromlayer-by-layerpolishedpelletsurfaces(upper
traces).
itisnotpossibletoattributeunequivocallytheextrare ectiontoCuLow-magni cation2O.
BEIimagesof3and24hsamplesheattreatedinN2for6hat10001CareshowninFig.4.TheimagecontrastrevealstwocompositionallydistinctlayerssurroundingthebulkCaCu3Ti4O12materialinbothsamples;theselayersappeartoforma‘‘decompositionzone,’’Fig.4(a)and(b).Ahighermagni cationimagefora3hsampleshowstheinterfacebetweentheoutertwolayers,Fig.4(c)(asimilarimage,notshown,wasobtainedfora24hsample)andtheinterfacebetweentheinnerlayerandbulkmaterial,Figs.4(d)and(e).Theoutermostlayerwasahighlyporous(B50%dense),‘‘coral-like’’structureofB1mmdiameterinterconnectedstrands.Thisstructureformsthematrixofaninnerlayerinwhichtheporesare lledwithaphasethatappearsbrighterthanthematrix.Inbothsamples,thethicknessofthelayersvariesgreatlysuchthat,for3hsamples,theouterlayercanextendB100mmintothesampleor,conversely,theinnerlayercanextendfromthebulkinterface,throughtheouterlayerandtothepelletsurface.Theoverallthicknessofthedecompositionzonewasrelativelyconstantinbothsamples.Forthe3hsample,thedecompositionzonethicknesswasB150mmatthemajorpelletfaces,increasingtoB200–300mmalongthepelletcircumfer-ence.B100–150ThedecompositionTypicalEDSmm.
zonewasthinnerinthe24hsample,dataforthebrightprecipitatephaseandmatrixphaseobservedinthedecompositionzoneareshowninFigs.5(a)and(b),respectively.Thebrightphasecorrespondstoacopper-richphase,whichiscon rmedbyXRD(seelater)tobeCu2O.ThepresenceofCaandTiintheEDSdatamay,asbe-fore,arisefromtheirincorporationintotheCu2Olatticeand/orfrombeaminteractionswithsurroundingmaterial.ThedarkmatrixcontainsCa,Ti,andSi,andXRD(seelater)isusedtoshowevidenceofseveralphases,includingCaTiO3andTiO2.Forbothsamples,theinterfacebetweentheinnerlayerandthebulkmaterial,Figs.4(d)–(e),isclearlyde nedbya netex-tureofCuand,below2OstriationsoriginatingfromtheCaCutheinterface,themicrostructureiscomparable3Ti4O12grainsbothintermsofgrainsize,porosity,anddistributionofsecondaryphase(s)tothatoftheas-sinteredsamplesinFig.1.Thetextureisclearlylessstriatedinthe3hsample,suchthatthestriationsextendfor5–10mmbeforecoalescingintoamore‘‘globular’’morphology,Fig.4(d).Theinnerlayerofthe24hsample,how-ever,wasentirelystriatedbetweentheinterfacewiththeouterlayerandtheinterfacewiththebulkmaterial,Fig.4(e).
ThecrystallinephasescontainedwithinthedecompositionzoneandthebulkmaterialforsamplesheattreatedinN1CwerecharacterizedbyXRD;layer-by-layerand
typical
2at1000
2836JournaloftheAmericanCeramicSociety—Adamsetal.Vol.89,No.9
Fig.4.Low-magni cationbackscatteredelectronimagesscanningelectronmicroscopy(BEISEM)of(a)3hand(b)24hsamplesincrosssectionafterheattreatmentat10001CinN2.BEISEMshowinginterfacebetweentheouterlayerandtheinnerlayerofthedecompositionzonein(c)3hsampleafterheattreatmentat10001CinN2.BEISEMshowinginterfacebetweenthebulkphaseandtheinnerlayerofthedecompositionzoneof(d)3hand(e)24hsamples.
resultsareshowninFig.6.As-sinteredCaCu3Ti4O12pelletswereblackincolor,whereasthoseheattreatedat10001CinN2werewhite.Lightpolishingofthepelletsurfacesoftheheat-treatedsamplesresultedinasequenceofcolorchangesfromwhitetoorange,thendarkbrown,and nallyblack.Thesechangesaremacroscopicevidencefortheexistenceofthetwolayersinthedecompositionzoneandthebulkmaterialbelow,asobservedbySEM.XRDdatawereobtainedfromthevariouscoloredpelletsurfaces,assummarizedinFig.6.ThebottomtraceinFig.6showstheXRDdatafortheoriginalwhitesur-face,andsubsequentdatasetsarefortheorangesurface(layer1),darkbrownsurface(layer2),and nallytheblacksurface(layer3andbulk).
Re ectionsobservedfromthewhitesurfacecorrespondtoTiO2(rutile),(ICDDcard21-1276)andCaTiO3(ICDDcard22-153),inagreementwithEDSdataobtainedfromtheouterlayerandinnerlayermatrix.ThereisnoevidenceofCaCu3Ti4O12oranycrystallineCu-containingoxides.Inaddition,therewasnoevidenceforspheneorcrystallineSiO2intheXRDdata,sug-gestingthattheSidetectedbyEDSwaseitherpresentinanamorphoussilicatephaseorincorporatedintotheTiO2and/orCaTiO3lattice,orinsuchsmallquantitiesthatitcouldnotbedetectedbylaboratoryXRD.Thechangeinpelletsurfacecolorfromwhitetoorangecoincideswiththeobservationofextrare ectionscorrespondingtoCu2O(Fig.6,layer1).Thisismostclearlyobservedbytheappearanceofthe(311)re ectionatB73.71forCu2O;unfortunately,themostintensepeakforCu2O,atB36.51,the(111)re ection,occursinthevicinityofthesecondmostintensepeak,(101)re ection,forrutile.Thiscon rmsthemajorprecipitatephaseoftheinnerlayertobeCu2O,andtheouterlayer‘‘coral’’structureofTiO2andCaTiO3alsoformsthematrixoftheinnerlayer.Furtherpolishing,untilthechangeincolorfromorangetodarkbrown,yieldedre ec-tionsintheXRDpatterncorrespondingtoCaCu3Ti4O12,Cu2O,TiO2,andCaTiO3(Fig.6,layer2).Afterfurtherpolishing,allre ectionswerefoundtocorrespondtoCaCu3Ti4O12(Fig.6,layer3andbulk).