Oxidation-reduction equilibria of vanadium in CaO-SiO2, CaO-Al2O3-SiO2 and CaO-MgO-SiO2 melts

Oxidation-reductionequilibriaofvanadiuminCaO-SiO2,CaO-Al2O3-SiO2andCaO-MgO-SiO2melts

H.FARAH∗,†

DepartmentofMechanicalandIndustrialEngineering,ConcordiaUniversity,Montreal,QCH3G1M8,Canada

E-mail:farah1@me.concordia.caM.BRUNGS

SchoolofChemicalEngineeringandIndustrialChemistry,UniversityofNewSouthWales,Sydney2052,Australia

AseriesofredoxstudiesofvanadiumhavebeencarriedoutinCaO-SiO2,CaO-MgO-SiO2andCaO-Al2O3-SiO2melts/slagsequilibratedoveroxygenpartialpressures(pO2)range10−2–10−9atmat1600◦C.V2O5levelwasvariedfrom1–5mol%.Threedifferentmelt

basicitiesandaluminacontentswereinvestigated.Magnesiacontentwasvariedbetween3.5and4.9wt%.Anewlydevelopedanalyticaltechniquebaseduponelectron

paramagneticresonance(EPR)spectroscopywassuccessfullyappliedtothesemelts.TworedoxequilibriacorrespondingtoV3+/V4+andV4+/V5+pairsfollowedtheO-typeredoxreactionovertheoxygenpartialpressurerangeinvestigated.Higheroxidationstatesofvanadiumwerestabilizedwithincreasingbasicityofslags.Tworedoxpairscoexisted

withinoxygenpressureregion10−4–10−6atm.However,redoxratiosdidnotindicatecleartrendswithincreasingV2O5contentinCaO-SiO2-V2O5system.InCaO-SiO2-Al2O3-V2O5slags,aslightincreaseinredoxratios(V3+/V4+)wasobviouswhenaluminaquantitywaschangedfrom3.22to5.44%atabasicityratio∼1.3,indicatinganincreaseinslagacidity.CaO-MgO-SiO2-V2O5slagsshowedasharpdecreaseinredoxratios(V4+/V5+)between10−2–10−6atmwithadditionof3.5wt%MgO,duetoincreasingfreeoxygenionsinslags.󰀃C2003KluwerAcademicPublishers

1.Introduction

Inourpreviousstudy[1,2],vanadiumredoxequilib-riumhasbeendescribedinquenchedsodiumsilicates.Howeverthefundamentalredoxdataofvanadiuminironfreeandironbearingcalciumsilicatesisunavail-ableand/orlimited.Asstatedearlierinourwork[1–3],vanadiumredoxchemistryisnotyetcompletelyunder-stoodbecauseoftheexistenceofthemultipleoxidationstates.Previouswork[4–6]hasindicatedthatoneofthemajordifficultiesininterpretationofthevanadiumredoxrectioninmeltsisthedependenceonconven-tionalanalysis,whichseriouslyaffectstheaccuracyoftheresults.VanadiumcanexiteitherasV3+/V4+orV4+/V5+redoxpairdependingonslagcomposition,althoughitispossiblethatoveracertainrangeofoxy-genpotential,theycouldcoexistattheexperimentaltemperatureintheseslags.Whenthisoccurs,determi-nationofvariousvanadiumspeciesviaanalysisinaque-oussolutionpresentsdifficultiesinthatthefollowing

reactionoccurs:

+5+4+V3(s)+V(s)⇔2V(soln)

(1)

Inaddition,theironintheslagscouldinterferewith

vanadiumanalysisbythefollowingreaction[7,8]:

V5++Fe2+⇔V4++Fe3+

(2)

Thusbeforewetchemicalanalysis,itisessentialtodetermineaccuratelytheconcentrationofatleastoneofvanadiumspeciestoenablewetchemicalresultstobecorrectedforreaction1or2.Thedetailshavebeendiscussedinourworkelsewhere[1–3,9,10].

MittelstadtandSchwerdtfeger[4]studiedcalciumsilicateswithdifferentbasicityratiosat1600◦CoverarangeofpO2(10−1–10−8atm)and10mol%V2O5con-tent.Howevervanadiumvolatilizationwasreportedun-derreducedoxygenpressures,whichwascompensatedbyreoxidisingthesampleafterequilibration;

∗Correspondingaddress:C.P.47527,1550deMaisonneuveWestMontreal,QCH3G2V7,Canada.

†Humera

FarahandMichaelBrungsarefacultymembersattheirrespectivedepartments.

C2003KluwerAcademicPublishers0022–2461󰀃1885

neverthelessthisapproachcouldhaveaffectedtheaccuracyoftheresults.Theauthorscommentedthatthethreevanadiumvalenciescouldn’tbeanalysedinaqueoussolutiontogether,owingtothereaction(1).TheCaO-MgO-FeOx-SiO2steelmakingslagsstud-iedbyInoueandSuito[11]wereequilibratedwithliquidironat1550–1650◦Candwerethenchemi-callyanalysedinaqueoussolution,assumingthattheFe3+/Fe2+ratioissameintheseslags,asintheslagswithoutvanadiumatafixedcomposition.Thesefac-torscontributedtotheerrorindeterminationofthetruevalencyofvanadium.Theauthoralsoadmitsthattheassumption,thatonlytwovanadiumspeciescancoex-ist,couldaffecttheactualvanadiumvalency.VanadiumdistributionbetweentheslagandmetalincreasedastheCaO/SiO2ratiowasraised,asexpected.

Fig.1presentsacomparisonofthevalencyratios(V4+/V5+)versus(CaO+MgO)/SiO2molarratiosatdifferenttemperatures(1550–1650◦C)usingtabulateddatafromInoueandSuito’swork.ThepresenceoftheV4+/V5+pairatthisreducedpressuredisagreeswiththatofMittelstadtandSchwerdtfeger[4]whodeter-minedanoverallvalanceof1.8,correspondingtoa(V3+/V4+)redoxpairunderoxygenpotential≤10−8atm.HoweverInoueandSuitoadmitthatitispossiblethatironfreeCaO-SiO2slagwouldcontainV3+ionatthisoxygenpressuresinceV3+ionisevidentat10−8–10−9atminV-Osystem.Thereforeitislikelythatironhasaffectedthetruevanadiumvalencyintheseslags.Aneffectofincreasingtemperatureontheredoxratiosisalsonotevident.Whiledeterminingthephosphorousdistributionbe-tweenmagnesiasaturatedsteelmakingslag(CaO-MgO-FeOx-SiO2)andliquidironat1600◦C,Selin[12,13]foundtheoverallcompositionofthemeltcorre-spondedtooxidationstateV4+.Sincethevanadiumvalancestateswerenotdetermined,thisfindingcannotindicatethetrueredoxpairofvanadiumpresentatthereducedoxygenpartialpressure.

InanotherstudymadebyTsukihashietal.[5],vana-diumvalenciesweremeasuredinCaO-CaF2-Na2O-SiO2at1300◦C.V4+andV5+ionswerefounddomi-nantintheseslagsandthevalencyincreasedwithanin-creaseinsodiumoxidecontentfrom0–3wt%.WermeandAstron[14–16]studiedthefundamentalsofthevanadiumredoxionsduringthedistributionexperi-mentsbetweencarbonsaturatedironandCaO-Na2O-SiO2(TiO2,Al2O3,MgO)andFeO(satd)-SiO2slagsat1300◦Cwith∼1.7%vanadium.VanadiumwasfoundasmonomerV3+andVO+ionsinsilicasaturatedslagsandasVO−2ioninFeOrichslags.Sincewetchemicalanalysiswasused,thepredictionofthetruevanadiumvalencywasaffectedandinadditiontheironintheslagscouldhaveinterferedwiththevanadiumanalysis.Contradictorybehaviorregardingthevanadiumdis-tributionwasalsoreportedbyWerme[14,15]whenheexaminedtheadditionofCaO,MgO,Na2O,Al2O3andTiO2,toironandsilicasaturatedFeO-SiO2meltsandfoundthatdistributionratioofvanadiumbetweenslagandirondecreasedwithincreasingadditionofCaO,MgO,Na2OandAl2O3underSiO2saturation,whichisincontrasttohispreviousfindings.

Figure1Variationof(V4+/V5+)ratioswith(CaO+MgO)/SiO2molarratiosforCaO-MgO-FeOx-SiO2-V2O5slagsat1550–1650◦C,TabulateddatafromInoue&Suito[11].

1886

Freeoxygenions(O2−)presentintheslagledtotheconceptofslagbasicity,whichcanbedefinedbaseduponionictheory,astheactivityofthefreea2O−oxygen[17]assumingbasiccomponentsproduceO

2−

ions,ionswhileacidcomponentsconsumethem.Redoxequilib-riumofanelementwithmultipleoxidationstatesuponintroductionintoameltcanberepresentdbyamassexpressionnamedanO-typereaction[4,18–20].Ma++

(2x−1)2O2−+14

O2⇔MO(ax−2x+1)

(3)

foroneelectronchangebetweenthevalencies,wherexisthenumberofoxygenatomsassociatedwiththeoxidizedredoxioninthemelttoformanoxo-anion.Thisexpressionaccountsfortheactivityoffreegenion,a2O−

oxy-anddescribetheequilibriaoftheionsformingoxoionse.g.,CrO24−,VO−redox3.AsecondtypeofredoxexpressionnamedasaR-typeredoxre-action[19]hasbeenusedtodescribethe3+redoxequilib-riaofanumberofionsi.e.,Co2+/Co,Mn2+R-type/Mn3+.ThepracticaldifferencebetweenOandredoxexpressionsisthattheypredicttheoppositeeffectofmeltbasicityupontheconcentrationredoxratio[21].Thatis,fromexpression3,anincreaseinmeltbasic-itymovestheequilibriuminfavourofformationoftheloweroxidationstate,whereastheoppositeeffectispredictedbyR-typereaction.WrightandJahanshahi[19]reviewedthevanadiumredoxbehaviorinvariousmeltsandconcludedthatthevanadiumfollowsanO-typeredoxreaction.Howeverinsomestudies,eitherorbothOandR-typeequationshavebeenusedtodescribethevanadiumredoxreaction,whichismisleadingandcontradictory[4,6,11,24].

Inpreviousstudies,theredoxdataofvanadiumismostlygeneratedusingchemicalanalysisandisavail-ableoveralimitedrangeofoxygenpotential,temper-atureandmeltcomposition.Theaimofthispresentworkwastoutilizethenewlydevelopedanalyticaltech-niquebasedonElectronParamagneticResonancespec-troscopy(EPR)inordertoremove:(1)oneofthema-jordifficultiesininterpretationofthevanadiumredoxreaction,thatisdependenceonconventionalanalysis,whichoftenseriouslyaffectstheaccuracyoftheresults,(2)thecontradictionsobserved,andalsotoproducere-liablefundamentaldataforavarietyofcalciumsilicabasedvanadiummelts.Thisdataisapplicabletounder-standingofvanadiumbehaviorinironandsteelmakingslags.

2.Experimentalprocedure

50gofmastermeltswerepreparedfromanalyticalgradeCaO,MgO,AlCfor2Oweredriedat120◦3,SiO24hrs,2andVthencarefully2O5.Allreagentsweighedandmixed.Theseweremeltedinaplatinumcrucibleat1600◦C,crushedinatungstenmillandanalysedbyXRF.CompositionsandidentificationsofslagsaregiveninTableI.Thetimetoapproachequilibrationtimewasdeterminedinaseparatekineticstudy.

Amolybdenumwoundelectricresistancetubefur-nacewasusedfortheequilibriumexperiments(Fig.2).

TABLEICompositionofslagsamplesCaO/SiO2V2O5Al2O3MgOS.no.(molarratio)(mol%)(wt%)(wt%)A

0.75––B0.95––C1.235––D0.71––E0.91––F1.213––G1.2353.22–H1.2255.44–I1.2257.87–J1.575–3.50K

1.51

5

–

4.87

Figure2Verticalsectionofthemolybdenumwoundelectricfurance:(A)Tocontrolthermocouple,(B)gasdeliverytube,(C)tosecondthermo-couple,(D)brassfittings,(E)wateroutlet,(F)gasoutlet,(G)aluminareactiontube,(H)aluminaoutertube,(I)heatingelement,(J)copperdrum,(K)crucible,(L)aluminafilling,(M)insulatingwool,(N)gasinlet,(O)waterinlet,and(P)quenchingchamber.

TemperatureofthefurnacewascontrolledbyanEu-rothermtemperaturecontroller(±5◦C).1–2gofsam-pleswere◦placedinplatinumcrucibleandequilibratedat1600Cin−air2andoverarangeofoxygenpartialpressures(10–10−9atm.).Thesampleswerethenquenchedinthebottomwater-cooledquenchingcham-berinthesamegaseousatmosphere.Ar,CO,COusedforthepreparationof2andAr-O2gasmixtureswerevar-iouscompositionsusingprecalibratedmassflowcon-trollers.Theequilibriumstudieswerecarriedoutforseriesofmasterslags(A-KinTableI)at1600◦Cinplat-inumcrucibles.Thecompositionvariablesinvestigated

1887

werealuminaandmagnesiacontentandbasicityratiosovertheabovementionedrangeofoxygenpartialpres-sures.

ThecombinedEPR/chemicalanalysismethodde-tailedelsewhere[3]wasemployedforallslags.AVar-ianE-linespectrometeroperatinginXbandat9.48GHzwasusedfortheanalysisofvanadium(IV).Allspec-trawererecordedover2000Gatroomtemperatureonquenchedslags.DetailsofEPRanalysisarealsogivenearlier[2,3].

3.Resultsanddiscussion

3.1.Vanadiumredoxreactionandoxygen

partialpressures

Thetwo4+/redoxV5+equilibriacorrespondingtoV3+pairsfollowedtheO-typeredox/reac-V4+andVtion[3]overtheexperimental2oxygenpartialpres-suresandtemperature(10−–10−9atm,1600◦C)withV2O5(5mol%),alumina(3–8wt%)andmagnesia(3–5wt%)contentsinvestigated.Theredoxratiosde-creasedwithincreaseinoxygenpressure,whichisinaccordancewithlimitedpreviousworkonsodiumandcalciumvanadiumsilicates[4,6];sincedataontheironfreeCaO-SiOandCaO-MgO-SiO-V2-V2Oslags5,CaO-SiOarenot2-Alavailable2O3-Vfrom2O5theliteratureunder2the2Oexperimental5conditionsinves-tigatedhere.Ourpreviousworkonsodaslags[2,3],isalsoinaccordancewiththefindingsforcalciumsilicatebasedslags,althoughthesodaslagsweremostlyacidic(Na2O/SiO2:0.5).

Generally4thetworedoxpairscoexistedatoxygenpo-tentials,10−–10−6atm.ThisregionisreadilyobservedinFig.3aandbwherethefractionsofthethreevana-diumspeciesareplottedatdifferentoxygenpressuresforCaO-SiO2-V2O5system(slagsBandC,TableI).Theoxygenpressureregionforcoexistenceoftwore-doxpairswasexpandedwithanincreaseinbasicity(B:0.9toC:1.23),aspresentedelsewhere[6].

3.2.Effectofbasicityonvanadiumredox

equilibria

To3fi+nd/Vthe4+effectofchangeinslagbasicity,redoxratios(V)ofslagsBandC(CaO-SiOarecomparedinFig.4a,whichreveals2-Vthat2O5system)theratiosdecreasedwithincreasingbasicity(0.9to1.23).Theseplotsindicateagoodlinearity.+ThesectionofoxygenpartialpressureswhereV4andV5+ionsexist,wasalsoextendedfrom10−4atmatabasicity:0.9,to10−6atm,byincreasingthebasicityto1.23+[10].

Fig.4bpresentsaplotof(V4+/V5)ratiosversusCaO/SiO2(molarratio)atpO210−2atmforslagsB,CandF-K(TableI).Itindicatesadecreaseinthere-doxratioswithincreasingCaO/SiOwasobtainedincorporating2ratioandfollow-inglinearequationalltheseslags.

V4+

logCaO2V5+=−1.413

SiO2

−1.386R=0.718(4)

Itisevidentfromtheresultsthatanincreaseinba-1888

sicitystabilizestheproportionofvanadiuminhigher

oxidationstatesintheslagsinvestigated,hencevana-diumfollowsanO-typeredoxreaction[18–20],whichagreeswithourpreviousconclusionsonsodiumsili-cates[2].

AstudymadebyTsukihashi◦etal.[5]onCaO-CaFSiO(VOCtodeterminetheeffect2-ofaddition2slagofx3%)at1300Nametal2OadditiononvanadiumvalenciesduringthehottreatmentrevealsthatatapO10−18atm,theV4+whereasnoneofthe/V5+pairwaspresentinthemelt2:presentslags(A-K,TableI)con-tains−9thispairevenatahigheroxygenpressure(i.e.,10atm)at1600◦C.Theirobservationmaybeat-tributedtothepresenceoftheseveraltransitionmetalsandtracesofironmetal,affectingthetruevalencyofvanadiumintheirslags.

Also,theadditionofsodaincreasedthevalencyofvanadiumintheCaFfindings,despitethe2slagswhichsupportsthepresentdifferenceinmeltcomposition,

thatanincreaseintheoxygenionactivity(a2−

vanadiumoxidationstates.Inaddition,O)in-creasedthetheincreaseinthevanadiumdistributionratiobetweentheslagandthehotmetalincreasedbyanorderofmagni-tudeonadding2wt%sodaintheCaFbyTsukihashietal.whichsupportsthe2slaginvestigatedearlierobserva-tionsindicatingtheeffectofincreasigbasicityinsodaslags[1,2,9,10].

Acomparisonofx,thenonstoichiometryofvana-diumoxide,betweenslagsBandC(Vbasicity:0.9,1.23respectively)◦andCaO-SiO2O5:5mol%,slagsmeltedat1600CoverpO2-V2O52range:10−1–10−8atmfromMittelstadtandSchwerdtfeger[4]ispresentedinFig.5.Herethetabulateddataatbasicityratios:1.0and1.36(Vysis[4]2Ois5plotted.∼10mol%)Itindicatesobtainedthatfromthechemical“x”remainedanal-lowerforslagsBandC,comparedtothoseofMit-telstadtandSchwerdtfeger,particularlyathigheroxy-genpartialpressures(≥10−5atm).Thoughthismaybeattributedtotheexperimentalerroranddifferenceinanalyticaltechniquesused,asourresultsshowlessuncertaintiesbyemploymentofnewlydevelopedana-lyticaltechnique.Neverthelesslowbasicity(0.9,1.23)reducesthe‘x”tosmallervaluescomparedtothatof1.0and1.36respectively,asexpected.

3.3.Dependenceofvalancestatesof

vanadiumonaluminacontent

Althoughthealuminacontentsinvestigatedproducedaslightincreaseintheredoxratios(V3+/V4+)byen-hancingtheacidityofthemelts,cleartrendshavenotbeen+observed.ThisisillustratedinFig.6mina/levelsV4+wherelog(V3)isplottedagainstlogpOfortheslagsG-I(CaO-SiO2atdifferentalu-aluminafreeslagC(CaO-SiO2-Al2O3-V2Osystem)andthe5system)(TableI).Aslightincreaseintheredox2-V2Ora-5tiosisobviouswhenthealuminaconcentrationwaschangedfrom3.22to5.44%atabasicityratio∼1.3.Aluminaisgenerallyconsideredamphoteric.Sincetheconcentrationofaluminainvestigatedwaslow,itmaybeinsufficienttoproduceasignificantchangeinthere-doxratios.Theauthorshavealsopreviouslyobserved

Figure3Dependenceoffractionofredoxspeciesofvanadium(CaO-SiO2-V2O5system,1600◦C)onoxygenpartialpressuresatCaO/SiO2:(a)0.9(slagB)and(b)1.23(slagC)SeeTableIfordetails.

thattheadditionofdifferentaluminacontentshadnoeffectontheoxygenpartialpressureregionwheretworedoxpairscoexistatmeasurableconcentrations[3].3.4.Effectofmagnesiaonvanadium

oxidationstates

Magnesia,abasicoxide,actsasanetworkmod-iferandincreasestheactivityoffreeO2−ionsin

1889

Figure4(a)ChangeinV3+/V4+ratios(CaO-SiO2-V2O5system,1600◦C)withlogpO2atdifferentbasicityratios(slagB:0.9&slagC:1.23)and(b)ChangeinV4+/V5+ratiosasafunctionofCaO/SiO2ratioforslagsB,C,E-K,SeeTableIfordetails.

1890

Figure5Comparisonofoveralloxidationstate,xinVOx,inCaO-SiO2-V2O5systemasafunctionofoxygenpartialpressuresatbasicity0.9&1.23(V2O5=5mol%)withthoseofMittelstadtandSchwerdtfeger[4]atbasicity1.0&1.36(V2O5=10mol%)at1600◦C.

Figure6Changeinredoxratios(V3+/V4+)(CaO-SiO2-Al2O3-V2O5system)withrespecttooxygenpressuresatdifferentaluminalevels(1600◦C).

themelt.Asexpected,theratiooftheredoxspecies(V4+/V5+)between10−2–10−6atmdecreasedsharplywiththe3.5wt%additionofMgO(CaO-MgO-SiO2-V2O5system)(Fig.7).Thereductionintheva-lencyratiowaslesswithafurtheradditionofMgOto4.87wt%.Thistrendgenerallyagreeswithpre-viouswork[11].Howeveritissurprisingthatad-ditionofmagnesiacausedtheoverlappingregionofthethreevalancestatestodiminish.Withintheexperimentalerror,therewasnooxygenpotential

1891

Figure7EffectofMgOadditiononredoxratios(V4+/V5+)overdifferentoxygenpressures(CaO-MgO-SiO2-V2O5system)(1600◦C).

atwhichallthevalenciescoexistedintheseslags[3].

Fig.8presentsvariationofvanadiumvalencyratios(V4+/V5+)versus(CaO+MgO)/SiO2molarratiosforslagC(CaO-SiO2-V2O5system),JandK(CaO-MgO-SiO2-V2O5system)atanoxygenpotentialof10−4atm.andthoseofCaO-MgO-FeO-SiO2slagsstudiedbyIn-oueandSuito[11]equilibratedwithliquidiron(V∼1.5wt%).Thedatapointsaretakenfromthetables,atanoxygenpressureof10−9atmat1600◦C.Itisdifficulttocomparetheseironbearingslagswiththepresentworkduetothedifferenceincompositionandunavailabilityofthedataovervariousoxygenpotentials.HoweverthebothsetsofdatashowadecreaseintheV4+/V5+ratiowithincreasingMgOcontentand/orthebasicityratio.InInoueandSuito’swork,FeOhasactedasaba-sicoxideandcontributedtowardsstabilizationofV5+;althoughthepresenceofV4+/V5+pairatthereducedpressure(10−9atm)disagreeswiththecurrentwork,wherenoneoftheslagscontainedthisredoxpairatthisoxygenpotential.

Othermagnesiabearingironsilicateslagswithvana-dium(wt%)∼1.7–2.3at1300◦ChavebeenstudiedbyWerme[15,16]andanoverallvalencyasV3+waspresentatpO2∼10−12–10−15atm.Itsupportsthecur-rentresultswheretheV4+/V4+pairwaspresentun-derloweroxygenpotentials(at10−8and10−9atm).Howeverotherexperimentaldataisunavailableoverthepresentoxygenpressurerangeforcomparison.Itshouldbenotedthatadditionofbasicoxides(CaO,MgO,Na2O)changedtheactivityofFeOandtheequi-libriumoxygenpotentialinWerme’sironbearingmelts,

1892

thereforevanadiumvalencydidnotincreasewithanin-creaseinbasicity.

3.5.EffectofV2O5content

onvandiumvalencies

TheredoxratiosV3+/V4+andV4+/V5+forslagsFandCwhichhavealmostidenticalbasicity(CaO-SiO2-V2O5system)containing3and5mol%V2O5respectivelyareplottedinFig.9.Fromacloseex-amination,itappearsthattheredoxratiosV4+/V5+didnotindicatecleartrends.TheratiosV3+/V4+areslightlylowerforslagF.Thereforetheredoxequi-libriaareindependentofthechangeinV2O5contentstudied.

Ithasbeenfoundearlieringlasses[22,23]thattheredoxratiosV5+/V4+increasedwiththeadditionofthisoxide,whichindicatesitsbasiccharacterornet-workbreakingordepolymerisingability.Incontrast,MittelstadtandSchwerdtfeger[4]havereportedade-creaseinvanadiumoxidationstatewithadecreaseinNa2O/V2O5molarratios(1.0to0.2)inNa2O-V2O5melts,indicatingV2O5asanetworkformer.Thedis-crepancymaybeattributedtothedifferencesintheoxygenpressureandpossiblytheglasscompositionstudied,sinceallglassinvestigationsreportedintheliteraturewerecarriedoutunderhighlyoxidizedcon-ditionsi.e.,inoxygenorair,whilethoseofMittel-stadtandSchwerdtfegerwereovertheoxygenpres-suresrangeof10−3and10−9–10−16atm.TheoveralloxidationstateinMittelstadtandSchwerdtfeger’sworkcorrespondedtoV5+at10−3atmatdifferentV2O5

Figure8Comparisonofredoxratiosofvanadiumbetweenpresentwork(slagsC(CaO-SiO2-V2O5),J&K(CaO-MgO-SiO2-V2O5system)andthoseofInoue&Suito[11](1600◦C).SeeTableIfordetails.

Figure9EffectofvariationoftotalvanadiumcontentonvanadiumredoxratiosinCaO-SiO2-V2O5system(V2O5=3mol%forslagF&5mol%forslagC).SeeTableIfordetails.

1893

contents.Below10−9atm,itdecreasedwithanincreaseinV2O5content.Hencetheeffectofreducedoxygenpressureontheoxidationstate,cannotbedisregarded.Thereforethisoxideactsasanetworkformer.Inthepresentinvestigation,withamaximumof5mol%in-vestigated,V2O5actedasaneutraloxide,whencom-paredtotheconcentrations>10mol%asfoundinotherinvestigations.

4.Conclusions

Thenewlydevelopedmethod,whichenablesquantita-tivedeterminationofthethreeoxidationstatesofvana-dium,wassuccessfullyevaluatedandappliedtocal-ciumsilicatesandcomplexmelts.Incalciumsilicates,thetworedoxequilibriacorrespondingtotheV3+V4+oxygen/V5+pairsfollowedO-typeredox/reac-V4+andthetionoverpartialpressures10−2–10−9atmat1600◦CwithV2O5(1–5mol%),alumina(3–8wt%)andmagnesia(3–5wt%)contentsinvestigated.Thisre-sultcontradictsofpreviousworkers[11,24]whousedbothOandR-typeredoxexpressionstodescribethevanadiumbehavior.Generallythetworedoxpairsco-existedinmeasurableconcentrationoveroxygenpo-tentials10−4–10−6atm.

InAl2Oincreasein3-CaO-SiOtheredoxratios,2-V2Owhen5slags,therewasaslightaluminaquantitywaschangedfrom3.22to5.44%atabasicityratio∼1.3.Howeveroverall,thereisnoobviouschangeintheav-eragevanadiumvalencyformeltscontainingbetween3and8wt%alumina.

InCaO-MgO-SiO4+3.5/V5+)decreased2-V2O5melts,theratioofredoxspecies(Vsharplywiththeaddi-tionof−2wt%−MgO6overtheoxygenpartialpressurerange10–10atmat1600◦C.AlthoughthesametrendwasobservedwhentheMgOcontentwasfurtherincreasedto∼5%,thereductioninthevalencyratiowassmall.Thistrendgenerallyagreeswithpreviouswork[4].Withintheexperimentalerror,therewasnooxygenpotentialatwhichallthevalenciescoexistedintheseslags.

TheredoxratiosV3+/V4+andV4+notindicateclear/trendsV5+(CaO-SiO2-V2O5system)didwithachangeinV2O5contentfrom3to5mol%.Thereforetheredoxequilibriaandtheoverallvalencyofvana-diumareindependentofthechangeinVstudied.

2O5content1894

Acknowledgment

Theauthorsgratefullyacknowledgethefinancialsup-portforthisprojectfromBHPResearch,CenterforMetallurgyandResourceProcessing,Newcastle,AustraliaandaccesstoEPRfacilitiesfromSchoolofPhysics,UniversityofNewSouthWales,Sydney,Australia.References1.H.FARAH,

M.P.BRUNGS,D.J.MILLERandG.R.

BELTON,Phy.Chem.Glasses39(1998)318.

2.H.CHAUDARY,M.P.BRUNGS,D.M.MILLERandG.R.BELTON,inProceedingsofthe5thInternationalConfer-

enceonMoltenSlags,SaltsandFluxes,Sydney,January1997(ISS,Warrendale,1997)p.493.

3.H.FARAH,Ph.D.Dissertation,UniversityofNewSouthWales,Australia,4..1999.

RMITTELSTADTandK.SCHWERDTFEGER,Met.Trans.B21(1990)111.

5.F.TSUKIHASHI,T.TAGAYAandN.SANO,Trans.I.S.I.J.28(1988)164.

6.W.D.JOHNSTON7..,J.Amer.Ceram.Soc.48(1965)184.HD.SCHREIBER,J.Non-Cryst.Solids42(1980)75.

8.Idem.,“AdvancesinMaterialsCharacterization”(Plenum,NewYork,9..1983)p.647.

HFARAH10..,J.Mater.Sci.38(2003)727.

HFARAHandM.P.BRUNGS11..,J.Mater.Chem.(2002).RINOUEandH.SUITO,Trans.I.S.I.J.22(1982)705.

12.R.SELIN,Ph.Dthesis,TheRoyalInstituteofTechnology,Stockholm,1987.

13.Idem.,inProceedingsofInternationalConferenceonMoltenSlagsandFluxes(Inst.ofMetals,Glasgow,1989)p.317.14.A.WERME15..,SteelResearch59(1988)6.

AWERME16..andC.ASTRON,Scan.J.Met.15(1986)273.AWERME,Ph.Dthesis,TheRoyalInstituteofTechnology,Stockholm,1987.

17.SCHLACKENATLAS“VerlagStahleisenM.B.H.,Dusseldorf,198118..Vol.1,p.2.

HD.SCHREIBER,J.Non-Cryst.Solids84(1986)129.

19.S.WRIGHTandS.JAHANSHAHI,(Apcche-Chemeca,Melbourne,20..1993)Vol.2,p.359.

SB.HOLMQUIST,J.Amer.Ceram.Soc.49(1965)228.

21.E.T.TURKDOGAN,“PhysicalChemistryofHighTemperatureTechnology”(AcademicPress,22..NewYork,1980)p.145.

E.J.FRIEABLE,L.KWILSONandD.L.KINSER,J.Amer.Ceram.Soc.55(1972)164.

23.S.GUPTA,N.KHANIJOandA.MANSINGH,J.Non-Cryst.Solids24.

.181(1995)58.

MNAGANO,A.KATO,I.MOCHIDAandT.SEIYAMA,J.Ceram.Soc.Japan78(1970)401.

Received7August2002

andaccepted26February2003