钙铝酸盐基质的制备3

Preparationofcalciumaluminatematrixcompositesbycombustionsynthesis

´H.C.YI,J.Y.GUIGNE

Guigne´InternationalLtd,Paradise,NF,CanadaA1L1C1,CanadaE-mail:hcyi@guigne.com

J.J.MOORE,F.D.SCHOWENGERDT

CenterforCommercialApplicationsofCombustioninSpace(CCACS),ColoradoSchoolofMines,Golden,CO80401,USA

L.A.ROBINSON,A.R.MANERBINOGuigne´InternationalLtd,Paradise,NF,CanadaA1L1C1,Canada

CaO–Al2O3–TiB2compositeshavebeenproducedbytheCombustionSynthesistechnique.Thesematerialshavematricesbasedonbinarycalcium-aluminatecompounds,i.e.,Ca3Al2O6(C3A),Ca12Al14O33(C12A7),CaAl2O4(CA),CaAl4O7(CA2)andCaAl12O19(CA6).

ExceptforsampleswiththematrixcompositionofC3A,thecombustionsynthesisreactionscanbecharacterizedasstableself-propagatingwaveswithcombustiontemperatures

rangingfrom2125Kto2717Kandcombustionwavevelocityfrom4.0mm/sto10.6mm/s.ForsampleswithamatrixcompositionofC12A7,CA,andCA2,predominantlyequilibriumcompoundphasewasformed,whileforsampleswithamatrixcompositionofC3A,non-equilibriumphaseswerealsopresent.TherewasnoevidenceofCA6formationforsampleswithamatrixcompositioncorrespondingtoCA6.󰀃C2002KluwerAcademicPublishers

1.Introduction

Calciumaluminates(CaO–Al2O3)areattractivece-ramicmaterialsforhightemperaturerefractoryappli-cations.Theyalsoplayimportantrolesinthesteelin-dustryasmetallurgicalslagsandincementtechnologyashydraulicmaterials.Inrecentyears,therehasbeenrenewedinterestintheircapabilityofglassformationbothforpracticalandscientificreasons.Calciumalu-minateglasseshaveinfrared(IR)transmissionsupto6µm,whichissuperiortoordinaryoxideglasses[1].Theyarescientificallyimportanttoosincetheydonotcontainanyoftheusualglass-formingions,andthere-forethetraditionalnetworkglass-formingtheoriesarenotapplicable.

ThephasediagramofthesystemisshowninFig.1[2,3].Therearefivecalciumaluminatecompoundsinthesystem,namely,Ca3Al2O6(C3A),Ca12Al14O33(C12A7),CaAl2O4(CA),CaAl2O7(CA2),andCaAl12O19(CA6).PropertiesofthesecompoundsarelistedinTableI.Calcia(CaO)hasameltingpointof3173K.Whenaluminaisfirstaddedin,theC3Acompoundisprecipitatedoutwhilethemeltingtem-peratureofthemixturedropscontinuouslyinthepro-cess.ThepureC3Acompoundmeltsincongruentlyat1817K.Furtherincreaseofaluminaleadstothefur-therdecreaseofmeltingpointsandformationofanewcompound,C12A7,whichhasameltingtemperatureof1722K.Initially,itwassuggestedthatthiscompoundwasunstableintheanhydrousCaO–Al2O3system

C2002KluwerAcademicPublishers0022–2461󰀃

[4,5].However,thereseemstobenodoubtofitsex-istenceanditwassynthesizedinmanystudiesinre-centyears[3,6–8].Ifmorealuminaisaddedtothe

mixturecontinuously,themeltingpointsstarttoriseandthecompoundsofCA,CA2andCA6formrespec-tively.TheCA,CA2compoundsmeltcongruentlyat1873Kand2048K,respectively,whiletheCA6com-poundmeltsincongruentlyat2156K.Thepurealumina(α-Al2O3)meltsat2327K.Thethermodynamicprop-ertiesofthesecompoundswereevaluatedbyHallstedt[3],andbyErikssonandPelton[9].Thestandardheatofformation(󰀃Hf298)dataforthesecompoundsshowninTableI,relativetothepureCaOandAl2O3oxides,arefromHallstedt[3].

ThecrystalstructuresofthesecompoundsarealsolistedinTableI.TheyareusedasreferencestoindextheX-raydiffractionspectrainthepresentwork.

Thetraditionaltechniqueofpreparingthecalciumaluminatecompoundsisbysolidstatereactions(sin-tering)betweencalcia(CaO),orcalciumcarbonate(CaCO3)andalumina(Al2O3)powdersattemperaturesinexcessof1673Kforalongperiodoftimeinordertoobtainthesinglephasedesired.Singhetal.[6]usedthistechniquetostudythereactionkineticsofC3A,CA,andCA2compounds.Itwasfoundthatwhenthepowdermixturewasheatedtothetemperaturerange1473–1673K,allofthethermodynamicallystablecom-poundswereinitiallyformed.Thesingleequilibriumphasecouldbeobtainedonlyafteraprolongedreaction

4537

TABLEIPropertiesofcalcium-aluminatecompounds

󰀃HCompoundsTm(K)Crystallography(kJmolf298

−1)aCa3Al2O6(C3A)

1817

Cubic,Space−5.2groupPa3[10]Ca12Al14O33(C12A7)1718Cubic,Space

−76.2groupI-43d[11]CaAl2O4(CA)1873Monoclinic,Space−15.5groupP21/n[12]CaAl2O7(CA2)2048Monoclinic,Space−11.8groupC2/c[12]CaAl12O19(CA6)

2156

Hexagonal,Space

−22.2

groupP63/mmc[13]

aFrom

Hallstedt[3].

Figure1PhasediagramoftheCaO–Al2O3system.

time.Ontheotherhand,chemicalsynthesisusingthesol-geltechnologyhasalsobeenusedtoproducethesematerials[7,8,14,15].Insteadofusingsolidoxidesasreactants,solutionscontainingcalciaandaluminaarefirstmade(gel)andthenheatedatrelativelylowtemperatures(e.g.,<1173K).Thistechniquewasnor-mallyusedtoproducepowdersorthinfilmofthesecompoundswhicharealmostalwaysintheamorphousstateatlowtemperatureswhenpreparedbythetech-nique.Crystallizationwasthenachievedbysubsequentheatingathighertemperatures.Tas[8]usedasimilartechniquetopreparetheliquidmixture,whichwasthenplacedinamufflefurnaceat783Katwhichtemper-aturethemixtureundergoesboiling,dehydration,thenglowstoincandescenceleadingtoaslowcombustion(takesabout15minutes).Theproductwasstillamor-phousandsubsequentsinteringathightemperatureswasrequiredinordertoobtaincrystallinecompounds.Inthiswork,theSelf-PropagatingHighTemperatureSynthesis(SHS),orCombustionSynthesis[16–18]isusedtoproducecompositeshavingmatricesofthesecompounds.AtypicalSHSprocessinvolvesforma-tionofagreenpelletfromreactantpowdersandig-nitionofthepelletusinganexternalheatsourcetogenerateaself-propagatingcombustionreaction.TheSHSprocesscanberealizedbytwomodes,i.e.,prop-agation(orcombustion)modeandsimultaneous(orthermalexplosion)combustionmode.Inthepropa-gationmode,thereactantsareignitedbyanexternal

4538

heatsource.Onceignited,thehighlyexothermicre-actionignitesthenextadjacentreactantlayerbyit-selftherebygeneratingaself-sustainingwavepropa-gatingtowardtheunreactedportionofthesample.Inthesimultaneouscombustionmode,thereactantsareheateduniformlyuntilthecombustionreactionisini-tiatedsimultaneouslythroughoutthewholepellet.Acombustionsynthesisreactionischaracterizedbyfourparameters:theignitiontemperature,combustiontem-perature,combustionwavevelocityandthecharacter-isticsofthecombustionwave.Theignitiontemperatureisthetemperatureatwhichthereactionratebecomesappreciableleadingtoself-propagatingreactions.Thecombustiontemperatureisthemaximumtemperatureachieved,whichislowerthantheadiabatictemperature.Thecombustionwavevelocityistheoverallcombus-tionrate,andthecombustionwavecharacteristicsde-scribethenatureofthecombustionwave,i.e.,whetherthecombustionwaveisstableorunstable.Allthesepa-rametersareaffectedsignificantlybythepropertiesofgreenpellets(i.e.,particlesize,greendensity,reactionenvironment,etc.)whichhaveaprofoundinfluenceonthecombustion.IntheSHStechnology,thereactantsarealsosolidpowdersbutareusuallydifferentthanthoseusedinthetraditionaltechniquesuchassinter-ing.Obviously,thetechniqueusedbyTas[8]wasdif-ferentfromtheSHSprocessdescribedabove,althoughtheauthorusedtheterminologyof“Self-propagatingCombustionSynthesis.”

2.TheSHSreactions

Thecombustionsynthesisreactionsareshownbytheequation:

Ti+B2O3+Al+αCaO→TiB2+Al2O3+αCaO+󰀃HTf0

(1)

Theself-propagatingcombustionreactionissustainedbyaluminiumreducingB2O3atthecombustionfrontformingthealuminiumoxide(Al(B)atomsthenreactwithtitanium2O3),andthefreeboron(Ti)form-ingtitaniumdiboride(TiB2).Thisisathermite-typereactionwhichreleasesalargeamountofheat(󰀃HTenoughtofusetheAlandCaOtogetherformingf0),amatrixconsistingof2OTiB32,Al2O3andCaOphases.Clearly,inordertoformthecalcium-aluminatecom-pounds,thereactionsrepresentedbyEquation1mustcontinueas:

TiB2+Al2O3+αCaO+󰀃HTf0→TiB2+CaαAl2O3+α+󰀃HTTf0+󰀃Hf0(c)

(2)

where󰀃HTf0(c)istheheatreleasedintheprocessofcompoundformationandthevaluesatroomtem-peratureareshowninTableIforallcompoundsinthesystem.Calcium-aluminatecompoundsareformedbyproperselectionoftheαvaluesinEquation2,thusformingcompositeswithcalcium-aluminatecom-poundsasthematrixandTiBrespectiveα2asthesecondphase.TableIIshowsthevaluestoformdiffer-entcalcium-aluminatecompoundsasthematrixphasesandtheoverallcompositioninthefinalproduct.Forex-ample,toproduceasamplewithamatrixcomposition

TABLEIIAdiabatictemperatureandcompositionoffinalproduct(wt%)producedbyreaction(1)Matrix

TadCompositionsofproducts(wt%)(CaO–Al2O3)α(K)Al2O3CaOTiB2C3A3.0216030.049.520.5C12A71.7232738.135.926.0CA1.0243744.824.630.6CA20.5265051.114.134.8CA6

0.17

2819

56.4

5.2

38.4

correspondingtotheCeverymole12A7compound,1.7molesofCaOisaddedtoofAltobeused,andthefinalproductisacompositecomposedof26wt%TiBand74wt%C212A7.

Theadiabatictemperature(Tad)isthemaximumpos-sibletemperaturewithoutheatloss.Itcanbedeter-minedbythefollowingequation[17–19]:

T󰀃HT󰀃f0

=

ad󰀂

C(pi)dT(3)

T0

i

where󰀃HT,Cf0istheheatofformationatthetemper-atureT0(Pi)istheheatcapacityfortheithprod-uct,andTadistheadiabatictemperature.Iftheig-nitionoccursatroomtemperature,thentheheatofformationcanbedeterminedfromstandardheatofformations󰀁ofproductsandreactantsby󰀃Hf298i[󰀃Hf298=(Pi)−󰀃H298

fromBarinf(Ri)].Usingthethermody-namicdataetal.[20],theadiabatictem-peraturesforthecombustionreactionsrepresentedbyEquation1forvariouscompositionsarecalculatedus-ingEquation3andtheresultsarealsolistedinTableII.Itisnoticedthattheadiabatictemperaturesforallcom-positionslistedinTableIIarehigherthantheirmeltingpointsshowninTableI.

3.Experimentalmethods

FinereactantpowderswereusedinthepresentworkandtheirspecificationsareshowninTableIII.ThesepowderswerefirstweighedaccordingtocompositionsdeterminedbyEquation1andthenthoroughlymixedbyballmillingforatleast4hours.Cylindricalgreenpelletswerethenpreparedbyuniaxialpressingtoarelativedensityof58.6±0.7%ofitstheoreticaldensity.Eachpellethadadiameterof12.7mm(0.5inches)andweighed2.5grams.

Thedetailsoftheexperimentalset-upweredescribedinapreviouswork[21].Briefly,theset-upincludedacombustionchamberandafullyautomatedignitionanddataacquisitionsystem.Afteragreenpelletwasloaded

TABLEIIISpecificationsofthereactantpowders

ParticlesizeImpurityReactants

(µm)(%)VendorsAluminum(Al)

<45<0.5AlfaAesaerBoronOxide(B2O3)<45<0.02AlfaAesaerCalciumOxide(CaO)<45<0.01CeracTitanium(Ti)

<45

<0.02

Cerac

intothecombustionchamberbyplacingitdirectlyontopoftheignitioncoil,ignitionwasachievedusingacomputer-controlledpowerinputintothetungstencoilinanultra-highpurityargonatmosphere.Thecontroloftheignitionpowerwasconsideredtobecrucialinob-tainingreliablecombustionparametersandpre-heatingtothegreenpelletbeforeignitionwascontrolledinapreciseandrepeatablemanner.Thedataacquisitionsystemrecordedignitionpower,temperaturehistory,andpressurechangeduringthewholecombustionsyn-thesisprocess.TwoC-typethermocouples(W-5%Re/W-26%Re)of0.005inchesindiameter(weldedunderflowingargonatmosphere)wereusedfortemperaturemeasurements.Finally,avideorecordingsystem,con-sistingofadigitalcamera,wasusedtorecordthewholecombustionprocessfromwhichthecombustionwavevelocitywasdetermined.

Allofthecombustionreactionswereignitedfromthebottomofthesamplestotakeadvantageoftheconvec-tioneffectofargon[22].Thecombustionwavevelocitywasdeterminedbylinearlyfittingthewavefrontpo-sition(obtainedfromframe-by-framemeasurementoftherecordedcombustionwave)versustimeplot.4.Resultsanddiscussion

4.1.Combustioncharacteristics

Atypicalcombustionsynthesisprocessinvolvesigni-tionofagreenpellet,formationofaself-propagatingcombustionwave,andcoolingdown.Inthecurrentwork,ignitionwasachievedbyresistanceheatingofaW-coilwithhighelectricalcurrent.Sinceboththecombustionparameters(combustiontemperature,wavevelocity,andnatureofthecombustionwaves)andmi-crostructureoftheproductdependstronglyonthechar-acteristicsoftheignitionpower[21],itisimportanttospecifytheignitionpowerprofilebeingused.AtypicalignitionpowerprofileforthepresentworkisshowninFig.2forasamplewitharelativegreendensityof60%oftheoreticalandamatrixcompositionofCcoilhadaspiralconfigurationwitha12Atyp-7.Theignitionicalcoldresistanceof58m󰀇.Atthetimeti,apre-setpower(voltageandcurrent)wasappliedtothecoil.Asitwasheatedup,theresistanceofthecoilincreased,resultinginmorepowerinputtotheignitioncoil.Thepowerinputtotheignitioncoilwasterminatedatthe

Figure2AtypicalignitionpowerprofileforasamplewithamatrixcompositionofC12A7.

4539

Figure3ImagesofatypicalcombustionsynthesisprocessforasamplewithamatrixcompositionofC12A7.

pre-determinedignitiontimet0.Theignitionpowerwascontrolledpreciselytoallowarepeatablepowerpro-fileforignition,thusallowingbothrepeatableandpre-dictablepre-heatingtothegreenpellets.

Sincetheignitionpowerprofilewasexactlythesameforallsamples,onlyignitionenergyhadtobecon-trolledinordertohaverepeatableeffectsonthegreenpellet.TheignitionenergyistheareaunderthecurveinFig.2,whichwas1773Joulesforthisparticularsam-ple.Theignitionenergyforallofthesamplesinthepresentworkwascontrolledtobe1793±52Joules.Usingtheignitionprofileandenergy,initiationofcom-bustionwaveswasfairlyeasyforallcompositionsandaself-propagatingcombustionwavecouldbeestablishedforallcompositionsexcepttheonewithamatrixcom-positioncorrespondingtoC3A.Forthiscomposition,althoughignitioncouldreadilybeachieved,thecom-bustionwavestopped(quenchedout)inallsamplesandself-propagatingcombustionreactionscouldonlybeachievedbyhigherignitionenergycausingsubstan-tialpreheatingtothegreenpellet.TypicalimagesofacombustionsynthesisprocessareshowninFig.3forasamplewithamatrixcompositionofC12A7.Aftertheignition,thecombustionsynthesisreactionwavewasinitiatedinasmalllayerofthepelletclosesttotheignitioncoil.Sincethecombustionsynthesiswasanexothermicchemicalreactionprocess,theheatreleasedduringreactionofthesmalllayerwouldignitetheadja-4540

Figure4TemperaturechangeduringthecombustionsynthesisprocessforasamplewithamatrixcompositionofC12A7.

centlayerthusgeneratingaself-propagatingcombus-tionwavepropagatingtowardtheunreactedpartofthepellet.Thecombustionwavepropagatedalongthelon-gitudinaldirectionofthecylindricalpellet.AtypicaltemperatureprofileisshowninFig.4.Itcanbeseenthatbeforethecombustionsynthesisreaction,thepelletwaspre-heatedtothetemperatureTp.Thecombustionzoneisaverynarrowregion,typicallybeing0.3sec-onds(thetimefromTptothecombustiontemperatureTc)andtheproductwasformedinthisnarrowzone.Afterward,thereactedproductcooleddown.

Figure5Theeffectsofcompositionontheadiabatictemperature(Tad),combustiontemperature(Tc),andcombustionwavevelocity(V).

Combustiontemperature(Tc)andwavevelocity(V)forallothercompositionsareshowninFig.5.ForallcompositionsexceptCstablemanner3A,thecombustionwavesprop-agatedinawithrelativelyhighspeed,from4.0–10.6mm/s.BoththeTcandVincreasewiththeincreaseoftheamountofAl2O3duetoincreasedexothermicity.ItcanalsobeseenthatthecombustiontemperaturesforthesampleswithmatrixcompositionsofCAandCA2wereveryclosetotheiradiabatictem-peraturesasshowninFig.5(dashedline).Thiscouldbeattributedeithertothenegligibleheatloss,orthefactthattheformationofcompoundsfromindividualoxides(Equation3)releasesextraheatwhichwasnotincludedintheadiabatictemperaturecalculations,orboth.Thisisparticularlytrueforthesampleswithama-trixcompositionofCAwas63K2sincetheaveragecombustiontemperaturehigherthantheadiabatictem-perature.ExceptforthecompositionofCcompositionshadcombustiontemperatures3A,allotherhigherthanthemeltingpoints(Tm)oftheirrespectivecompounds,asshowninTableIV.Therefore,amoltenproducthadformedinthecombustionfrontforthesereactions.FromFig.4,itcanbeseenthatthereisachangeincoolingrateofthereactedmoltenproductatthetimetwherethetemperaturecurvestartstodeviatefromthefbaseline.Thedeviationisbelievedtobecausedbyso-lidificationandcrystallizationoftheCsincethecrystallizationreleasesheat12Acausing7compoundaslowdownincoolingrate.Thisismoreobviousfromtheplotofcoolingrate(dT/dt)versustimeasshowninFig.6forthesamecurveshowninFig.4.Themax-imumcrystallizationrateoccursatthetimetthecorrespondingtemperatureTfmandattimetthefreezingfm.Ifthetemperatureatthefisdefinedastemperature(Tthenthevalueof󰀃T=Tf),m−Tfrepresentsundercool-TABLEIVLiquidformation(Tc−Tm>0)andundercooling(Tm−Tf)ofcalcium-aluminatecompoundsduringthecombustionsyn-thesisprocessMatrix

compoundsC3AC12Al7CACA2CA6

Tc−Tm󰀃T

−—53±407133±±661438024±±1153766582±±1987056273±±171129Tm−Tfm

—191±3345±24134±65232±117

Figure6CoolingrateofreactedproductforthesamesampleshowninFig.4.

Figure7TheXRDspectraforasamplewithamatrixcompositionofC3A.

ing.The󰀃TvaluesforallcompositionsinthecurrentworkareshowninTableIV.Ascanbeseen,theunder-coolingisrelativelylargerfortheCofthisonthetendency12Aof7compound.Theimplicationglassforma-tionwillbediscussedinanotherwork.

4.2.Microstructures

TheX-raydiffraction(XRD)spectraforthesamplecorrespondingtothematrixcompositionofCSinceaself-propagatingcombustion3AisshowninFig.7.wavecouldnotbeestablishedforthiscompositionandthewavewasquenchedoutduringburning,thespectrawerefromasamplethatwentthroughrepeatedignition(substantialpre-heatingtothepelletoccurredbeforeachievingthefinalcompletecombustion).Theequi-libriumphase(C3A)hasclearlybeenformed,buttheCThe12ACaO7phasehasalsoformedinsubstantialamount.phaseseemstobepresentaswell.Therefore,itcanbeconcludedthatforthisparticularcomposition,Equation1hasproceededtocompletion(sincenoele-mentalaluminiumexistsinthespectra)butthesubse-quentreactiontoformtheCoccurredpartially.The3Acompound(Equation3)onlyXRDspectraforthesam-plewithamatrixcompositioncorrespondingtoCisshowninFig.8.Thepredominantphaseinthe12Ama-7trixofthesampleistheequilibriumCexists.Itseems12AsmallamountofCAalsothat7phaseandatherealsoexistsasmallamountofanotherphasesincethereareextrapeaksthatcannotbeassignedtoeithertheCtheCAphase.ThisphaseshouldbeCaO-richwhen

12Aor74541

Figure8TheXRDspectraforasamplewithamatrixcompositionofC12A7.

Figure9TheXRDspectraforasamplewithamatrixcompositionofCA.

Figure10TheXRDspectraforasamplewithamatrixcompositionofCA2.

massbalanceisconsidered.TheXRDspectraforthesamplewithamatrixcompositionofCAisshowninFig.9.Clearly,almostpureCAcompoundhasformedinthematrix.ThesameconclusioncanbereachedforthesamplewithamatrixcompositionofCAinFig.10.InFig.11,theXRDspectrafor2,asshownthesamplewithamatrixcompositionofCAinthematrixareCAand6isshown.Thephasesformedformationof2α-AltheCA2O3andthereisnoevidenceof6compound.

FromthemicrostructuresindicatedbytheseXRDspectra,itcanbeconcludedthatreaction(1)proceededtocompletionforallcompositions.However,there-actionsrepresentedbyEquation2toformequilibriumpurecompoundphaseoccurringattheirrespectivecom-4542

Figure11TheXRDspectraforasamplewithamatrixcompositionofCA6.

bustiontemperature(Tc)insideaverynarrowcombus-tionfrontshouldbemodifiedto:

3C+A17−→K

C3A+C12A7+C(m)12C+7A2125−→KC12A7+CA(m)C+A2428−→K

CA+C12A7(m)C+2A2713−→KCA2+CA(m)C+6A2718−→KCA2+A

(4)

Here,theminbracketsrepresentsminorphase.Exceptforthe3C+Areaction,allotherreactionsshowninEquation4occurredintheliquidstate.Theliquidphaseisionicinnatureconsistingofcations(Ca+2,Al+3)andanions(O−2)[23].Obviously,thereactionsrepresentedbyEquation4involveexchangeofelectronsbetweenions.

Ingeneral,achemicalreactionisdictatedbyther-modynamicsandkinetics.Accordingtothethermody-namicdata[3,9],allofthebinarycompoundformationreactionsrepresentedbyEquation3arethermodynam-icallyfavourable.However,thereactionrateiscon-trolledbythekinetics,i.e.,theactivationenergyorthediffusivityofthespeciesinvolved.Clearly,boththeCreaction3AandCAratewith6compoundsshowkineticsofslowCA6beingtheslowestsincetherewasnoevidenceofthisphaseformationinthepresentwork,evenataveryhighcombustiontemperature(Tservedc=2718fortheK).CThelowreactionkineticswasalsoob-reaction3AcompoundwhenitwaspreparedviathesolidbysinteringCaOandAlduringcrystallizationofgelpowders2O3mixtures[6]and[7].AsfortheCA6compound,althoughpreviousworkshowedthatitwasdifficulttoformbelow1573K[3,24],itisstillsurprisingthatitisnotformedatallat2718Kinthepresentwork.

5.Conclusions

1.Thecombustionsynthesisreactionsforsynthesiz-ingcompositeswithmatricescorrespondingtobinarycalcium-aluminatecompoundswerehighlyexothermicwithhighcombustiontemperatures.Asaresult,there-actedproductswereinliquidstatesforallcompositions

excepttheonecorrespondingtothematrixcompositionofC3A.

2.Thecoolingofliquidproductischaracterizedbyrelativelylargeundercoolingunderthecurrentexperi-mentalconditions.

3.AlmostpurecompoundmatrixwasformedforsampleswithamatrixcompositionofCtheotherhand,incompleteformation12A7,CA,andCA2.OnofCformationofCA3AandnoC6wasobserved.Therefore,theformationkineticsof3AandCA6wereslowwithformationofCA6theslowest.

Acknowledgements

ThisworkwassupportedbytheNASASpaceProductDevelopmentProgramthroughtheCenterforCom-mercialApplicationsofCombustioninSpaceattheColoradoSchoolofMinesunderNASACooperativeAgreementNumberNCCW-0096,andbytheNASAMicrogravityResearchDivisionunderCooperativeAgreementNumberNCC3-659.AdditionalfundingwasprovidedbytheColoradoCommissiononHigherEducation,theColoradoSchoolofMinesandCCACSIndustrialPartnersCoorsTek,GuigneInternationalLtd,Hewlett-Packard,ITNEnergySystemsandSulzerOrthopedicsBiologics.ExperimentswereassistedbyJonathanCurlettandAndreaParrell.References

1.J.E.SHELBY,C.

M.SHAWandM.S.SPESS,J.Appl.

Phys.66(3)(19)1149.

2.G.A.RANKINandF.E.WRIGHT,Am.J.Sci.39(1915)1.3.B.HALLSTEDT,J.Amer.Ceram.Soc.73(1)(1990)15.

4.R.W.NURSE,J.H.WELCHandA.J.MAJUMDAR,Trans.Br.Ceram.Soc.(1965)323.5.Idem.,ibid.(1965)409.

6.V.K.SINGH,M.M.ALIandU.K.MANDAL,J.Amer.Ceram.Soc.73(4)(1990)872.

7.A.A.GOKTASandM.C.WEINBERG,ibid.74(5)(1991)1066.

8.A.C.TAS,ibid.81(11)(1998)2853.

9.G.ERIKSSONandA.D.PELTON,Metall.Trans.B24B(1993)807.

10.W.WONG-NG,H.MCMURDIE,B.PARETZKIN,

C.HUBBARDandA.DRAGOO,PowderDiffractionFile,38-1429.

11.Natl.Bur.Stan.,Circ.539(9)(1960)20,PowderDiffractionFile,

09-0413.

12.P.J.BALDOCK,A.PARKERandI.SLADDIN,J.Appl.

Cryst.3(1970)188.

13.L.KELLER,J.RASKandP.BUSECK,PowderDiffraction

File38-0470.

14.M.UBEROIandS.H.RISBUD,J.Amer.Ceram.Soc.73(6)

(1990)1768.

15.M.A.GULGUN,O.O.POPOOLAandW.M.KRIVEN,

ibid.77(2)(1994)531.

16.A.MUNIRandU.ANSELMI-TAMBURINI,Mater.Sci.Rep.

3(19)277.

17.J.J.MOOREandH.J.FENG,Prog.Mater.Sci.39(1995)243.18.H.C.YIandJ..J.MOORE,J.Mater.Sci.25(1990)1159.19.HC.YI,H.J.FENG,J.J.MOORE,A.PETRICandGUIGN´J.Y.

E

,InternationalJournalofSHS5(1)(1996)39.20.I.BARIN,O.KNACKEandO.KUBASCHEWSKI,“Ther-mochemicalPropertiesofInorganicSubstances”(SpringerVerlag,1977).

21.H.C.YI,J.Y.GUIGN´E

,J.J.MOORE,A..R.MANERBINO,L.A.ROBINSONandFD.SCHOWENGERDT,“CombustionSynthesisofGlass(B2O3-Al2O3-MgO)-Ceramic(TiB2)Composites,”J.Mat.SynthesisProcessing,inpress.

22.H.C.YI,T.C.WOODGER,J.J.MOOREandGUIGN´J.Y.

E

,Metall.Mater.Trans.29B(1998)8.23.M.HILLERT,B.JANSSON,B.SUNDMANandJ.AGREN,

ibid.16A(1985)261.

24.H.VERWEIJandC.M.P.S.SARIS,J.Amer.Ceram.Soc.

69(2)(1986)94.

Received30January

andaccepted20June2002

4543