MnO2 机理

ChargeStorageMechanismofMnO2ElectrodeUsedin

AqueousElectrochemicalCapacitor

MathieuToupin,†ThierryBrousse,*,†,‡andDanielBe´langer*,†

De´partementdeChimie,Universite´duQue´beca`Montre´al,CasePostale8888,

succursaleCentre-Ville,Montre´al,Que´becH3C3P8,Canada,andLaboratoiredeGe´niedes

Mate´riaux,EcolePolytechniquedel’Universite´deNantes,LaChantrerie,

rueChristianPauc,BP50609,44306NantesCedex3,France

ReceivedMarch2,2004.RevisedManuscriptReceivedJune2,2004

ThechargestoragemechanisminMnO2electrode,usedinaqueouselectrolyte,wasinvestigatedbycyclicvoltammetryandX-rayphotoelectronspectroscopy.ThinMnO2filmsdepositedonaplatinumsubstrateandthickMnO2compositeelectrodeswereused.First,thecyclicvoltammetrydataestablishedthatonlyathinlayerofMnO2isinvolvedintheredoxprocessandelectrochemicallyactive.Second,theX-rayphotoelectronspectroscopydatarevealedthatthemanganeseoxidationstatewasvaryingfromIIItoIVforthereducedandoxidizedformsofthinfilmelectrodes,respectively,duringthecharge/dischargeprocess.TheX-rayphotoelectronspectroscopydataalsoshowthatNa+cationsfromtheelectrolytewereinvolvedinthechargestorageprocessofMnO2thinfilmelectrodes.However,theNa/MnratioforthereducedelectrodewasmuchlowerthanwhatwasanticipatedforchargecompensationdominatedbyNa+,thussuggestingtheinvolvementofprotonsinthepseudofaradaicmechanism.Animportantfindingofthisworkisthat,unlikethinfilmelectrodes,nochangeofthemanganeseoxidationstatewasdetectedforathickercompositeelectrodebecauseonlyaverythinlayerisinvolvedinthechargestorageprocess.

Introduction

Electrochemicalsupercapacitorsarecurrentlyinves-tigatedinvariousacademicandindustriallaboratoriesbecausetheycanbeusedascomplementarychargestoragedevicestoconventionalbatteriesinvariousapplicationsthatrequirepeakpowerpulses.1,2Intheseelectrochemicalsupercapacitors,theenergybeingstorediseithercapacitiveorpseudocapacitiveinnature.Thecapacitiveornonfaradaicprocessisbasedonchargeseparationattheelectrode/solutioninterface,whereasthepseudocapacitiveprocessconsistsoffaradaicredoxreactionsthatoccurwithintheactiveelectrodemateri-als.Themostwidelyusedactiveelectrodematerialsarecarbon,3,4conductingpolymers,5,6andbothnoble7-9andtransition-metaloxides.10-29

Themainmotivationfortheuseoftransition-metaloxidesliesintheirlowcostcomparedtonoblemetaloxidessuchasruthenium7,8andiridium9oxides.Theinitialstudieswereperformedonnickeloxide10andcobaltoxide11butmorerecentlyiron12-15andmanga-neseoxides14,16-29wereinvestigated.Theresearchef-fortsfocusedoncompoundsprovidinghighcyclabilityandcapacitance.Ontheotherhand,thechargestoragemechanismofMnO2hasnotbeeninvestigatedindetail.Untilnow,twomechanismswereproposedtoexplain

*Towhomcorrespondenceshouldbeaddressed.E-mail:thierry.brousse@polytech.univ-nantes.fr(T.B.)andbelanger.daniel@uqam.ca(D.B.).†Universite´duQue´beca`Montre´al.‡EcolePolytechniquedel’Universite´deNantes.

(1)Conway,B.E.ElectrochemicalSupercapacitors,ScientificFun-damentalsandTechnologicalapplications;KluwerAcademic/PlenumPress:NewYork,1999.

(2)Burke,A.J.PowerSources2000,91,37.

theMnO2chargestoragebehavior.Thefirstoneimpliestheintercalationofprotons(H+)oralkalimetalcations(C+)suchasLi+inthebulkofthematerialuponreductionfollowedbydeintercalationuponoxidation.17

MnO2+H++e-SMnOOH

or

(1)

MnO2+C++e-SMnOOC

(2)

Thesecondmechanismisbasedonthesurfaceadsorptionofelectrolytecations(C+)onMnO216

(MnO2)surface+C++e-S(MnO2-C+)surface

(3)

whereC+)Na+,K+,Li+.ThismechanismwasproposedfollowingtheobservationofsignificantdifferenceofthecyclicvoltammogramandthecapacitanceofMnO2inthepresenceofvariousmetalalkalicationsintheelectrolyte.16Itshouldbenoticedthatbothproposed

(3)Frackowiak,E.;Be´guin,F.Carbon2001,39,937.

(4)Lin,C.;Popov,B.N.;Ploehn,H.J.J.Electrochem.Soc.2002,149,A167.

(5)Villers,D.;Jobin,D.;Soucy,C.;Cossement,D.;Chahine,R.;Breau,L.;Be´langer,D.J.Electrochem.Soc.2003,150,A747.(6)Fusalba,F.;ElMehdi,N.;Breau,L.;Be´langer,D.Chem.Mater.2000,12,2581.

(7)Zheng,J.P.;Jow,T.R.J.Electrochem.Soc.1995,142,L6.(8)Soudan,P.;Gaudet,J.;Guay,D.;Be´langer,D.;Schulz,R.Chem.Mater.2002,14,1210.

(9)Conway,B.E.;Birss,V.;Wojtowicz,J.J.PowerSources1997,66,1.

(10)Nelson,P.A.;Owen,J.R.J.Electrochem.Soc.2003,150,A1313.

10.1021/cm049649jCCC:$27.50©2004AmericanChemicalSociety

PublishedonWeb07/16/2004

ChargeStorageMechanismofMnO2ElectrodemechanismsinvolvedaredoxreactionbetweentheIIIandIVoxidationstatesofMn.

Themechanismbasedonthesolid-statediffusionofprotonsinthebulkofthematerialissimilartothatproposedforRuO2.7However,onlyalimitedfractionoftheMnO2compositeiselectrochemicallyactive,thussuggestingthattheprotonicdiffusioninthebulkoftheMnO2compoundmightnotbeasfastasinthecaseofRuO2.21Subsequently,thechargestoragemightonlyinvolvedthesurfaceatomsoftheMnO2crystallitesoraverythinlayer.Thenitmightbeplausibletoassumethationsfromtheelectrolytewouldparticipateinthechargecompensationprocess.Ontheotherhand,thereportedcapacitancerangingbetween150and200F/gforcompositeelectrodecannotbeonlyassociatedtotheformationoftheclassicaldoublelayer.21Hence,thenatureofthechargestoragemechanismmustbepseudocapacitive.

Thisworkaimedatgettingabetterunderstandingofthechargestoragemechanisminmanganesedioxideelectrodeswhencycledinaqueouselectrolyte.TheelectrodeswerecharacterizedbycyclicvoltammetryandX-rayphotoelectronspectroscopyinordertodetermineachangeofthemanganesevalenceuponcharge/discharge.Additionally,theexperimentalresultswereusedtodeterminewhetherthechargestorageprocesswaslimitedtothesurfaceoftheoxideorifitoccurredinsidethebulkofthematerial.

ExperimentalSection

PreparationoftheMnO2Powder.TheMnO2powderwassynthesizedbycoprecipitation.21Briefly,KMnO4andMnSO4‚H2Oweremixedina2:3molarratio,leadingtoadarkbrownprecipitate.Theamorphousnatureoftheas-synthesizedMnO2powderwasconfirmedbytheX-raydiffraction(XRD)spectrum(FigureSI1),whichshowsbroadpeaksrelatedtoapoorlycrystallizedcompound.Fromchemicalanalysis,thestoichiometryforthepowderwasdeterminedtobeK0.02-MnO2H0.33,0.53H2O.Thereafter,thecompoundwillbenamed“MnO2”despitethatitdoesnotreflecttheexactcompositionofthesample.

(11)Lin,C.;Ritter,J.A.;Popov,B.N.J.Electrochem.Soc.1998,145,4097.

(12)Wu,N.-L.;Wang,S.-Y.;Han,C.-Y.;Wu,D.-S.;Shiue,L.-R.J.PowerSources2003,113,173.

(13)Wu,N.L.Mater.Chem.Phys.2002,75,6.(14)Brousse,T.;Be´langer,D.Electrochem.Solid-StateLett.2003,6,A244.

(15)Brousse,T.;Delahaye,T.;Be´langer,D.Inpreparation.

(16)Lee,H.Y.;Goodenough,J.B.J.SolidStateChem.1999,144,220.Seealsoamoredetailedversionofthisstudyinthefollowing:Lee,H.Y.;Manivannan,V.;Goodenough,J.B.C.R.Acad.Sci.Paris1999,t.2,Se´rieIIc,565.

(17)Pang,S.C.;Anderson,M.A.;Chapman,T.W.J.Electrochem.Soc.2000,147,444.

(18)Lee,H.Y.;Kim,S.W.;Lee,H.Y.Electrochem.Solid-StateLett.2001,4,A19.

(19)Hu,C.C.;Tsou,T.W.Electrochem.Comm.2002,4,105.(20)Chin,S.F.;Pang,S.C.;Anderson,M.A.J.Electrochem.Soc.2002,149,A379.

(21)Toupin,M.;Brousse,T.;Be´langer,D.Chem.Mater.2002,14,3946.

(22)Brousse,T.;Toupin,M.;Be´langer,D.J.Electrochem.Soc.2004,151,A614.

(23)Jiang,J.;Kucernak,A.Electrochim.Acta2002,47,2381.(24)Hu,C.C.;Tsou,T.W.Electrochim.Acta2002,47,3523.

(25)Jeong,Y.U.;Manthiram,A.J.Electrochem.Soc.2002,149,A1419.

(26)Broughton,J.N.;Brett,M.J.Electrochem.Solid-StateLett.2002,5,A279.

(27)Kim,H.;Popov,B.N.J.Electrochem.Soc.2003,150,D56.(28)Hu,C.-C.;Wang,C.-C.J.Electrochem.Soc.2003,150,A1079.(29)Chang,J.-K.;Tsai,W.-T.J.Electrochem.Soc.2003,150,A1333.

Chem.Mater.,Vol.16,No.16,20043185

Scanningelectronmicrographsrevealedthattheas-synthesizedR-MnO2powderismadeofsphericalgrains(FigureSI2).Thelengthscaleissystematicallyindicatedasawhitebaronthebottomleftcornerofthemicrographs.Astatisticalanalysisofthegraindiameterperformedovermorethan100particlesyieldedaGaussiandistributioncenteredat420nmwithastandarddeviationof190nm.Eachgrainseemstoresultfromtheagglomerationofsmallerparticles(FigureSI3).Usingthegeometricsurfaceofsphericalgrains(420nmdiameterassumingadensityof4.8g/cm3)thespecificsurfacewasestimatedtoavaluecloseto3m2/g.ThespecificsurfaceareadeterminedfromBETmeasurements(160(3m2/g)islargerthanthisvaluethusindicatingthatporesandvoidsexistinsidethegrainsexaminedbyscanningelectronmicroscopy.

PreparationoftheElectrodes.Toinvestigatetheinflu-enceofboththethicknessandthecompositionoftheelec-trodes,thickfilm(≈100µm)andthinfilm(<5µm)electrodeswereprepared.Forthethickfilmsamples,acompositewastypicallypreparedbymixing80%ofactivepowder,7.5%ofgraphite(AlfaAesar),7.5%ofacetyleneblack(AlfaAesar),and5%ofPTFE(poly(tetrafluoroethylene),Dupont)inethanol(Fisher).Then,coldrollingoftheobtainedpasteresultedinablackshinyfilm,whichwaspressedinaplyofastainlesssteelmesh(AlfaAesar,200mesh)currentcollectorunderapressureof9metrictons(seeFigureSI4aforaphotomicrographoftheseelectrodes).Typically,a2mgsquarepieceoffilmwasusedfortheelectrode,referredtoas“composite”electrodesfromnowon.

Alternatively,thinfilmswerepreparedbydispersinganappropriateamountofMnO2inasolutionofpolyvinylidenedifluoride-hexafluoropropylene(copolymerPVdF-HFP,Ky-narflex)inN-methylpyrolydinone(NMP,Fisher)toobtainafinalconcentrationof1mg(MnO2)/mLwith10%w/wofpolymer.Thedispersionwasleftinaultrasonicbathfor30min.Aplatinumfoil(1.5×0.5cm;thickness0.1mm;AlfaAesar)wasusedasthesubstrate(andcurrentcollector)andcoatedwiththeMnO2slurrybyadding5µLdropsofthedispersionwithamicropipet(Eppendorf)(seeFigureSI4bforaphotomicrographoftheseelectrodes).Theelectrodewasdriedinaovenat65°Cfor30minbetweeneachdrop.Typicallytheareaofplatinumcoveredwiththeslurrywas0.75cm×0.5cm.Themassofthedepositedmaterialwasdeducedbyweighingtheelectrodebeforeandafterthecoating.Thereafter,theseelectrodeswillbereferredtoas“thinfilm”electrodes.However,thisdenominationisnottotallycorrectsincethe“film”ismostlikelyseenasaggregatesofMnO2particles(thicknessislessthan2µm)ontheplatinumsubstrate(seeFigureSI4b).

ElectrochemicalMeasurements.Thecyclicvoltammetryandpolarizationexperimentswerecarriedoutwitha1470multipotentiostat(Solartron,Mobrey)usingtheCorrwaresoftware(ScribnerAssociates,version2.6).Abeakertypecellcontaininga0.1MNa2SO4electrolytesolutionwasusedforalltheelectrochemicalmeasurements.Thecyclicvoltammetryexperimentswereperformedbetween0and0.9VvsAg/AgCl(3MNaCl)atascanrateof5mV/s.

Thespecificcapacitance,Ccv,wascalculatedusingthevoltammetricchargeintegratedfromthecyclicvoltammogramaccordingtothefollowingequation

Ccv)

Q∆E×m

(4)

whereCcvisthespecificcapacitance(inF/g),Qisthecharge(inC),∆Eisthepotentialwindow(inV),andmisthemassofactivematerial(ing).

SurfaceCharacterizationoftheElectrodes.Afterpolarization,theelectrodesweredriedoutinavacuumovenatambienttemperaturefor1hour.TheXPSstudieswereconductedwithaVGEscalab220i-XLinstrumentequippedwithahemisphericalanalyzerandusinganaluminumanode(monochromaticKRX-raysat1486.6eV)asasource(at12-14kVand10-20mA).TheXPSspectrawereanalyzedandfittedusingCasaXPSsoftware(version2.2.27).TheC1sregion

3186Chem.Mater.,Vol.16,No.16,2004Figure1.Cyclicvoltamogramsin0.1MNa2SO4at5mV/sof(A)acompositeelectrodecomposedof80%MnO2,7.5%graphite,7.5%acetyleneblack,and5%Teflonand(B)a90%MnO2and10%PVdF-HFPthinfilmelectrodesupportedonaPtfoil.

wasusedasareferenceforsurfacechargingandwassetat284.9eV.AmixtureofGaussian(70%)andLorentzian(30%)functionswasusedfortheleast-squarescurvefittingproce-dure.

ThemanganeseoxidationstatewasdeterminedfromtheMn3sandO1scorelevelspectra.TheprocedureusedtoanalyzetheMn3sspectrahasbeendescribedpreviously.21,30,31InthecaseoftheO1sdata,theaveragemanganeseoxidationstateforthethreeelectrodescanbecomputedfromtheintensitiesoftheMn-O-MnandMn-OHcomponentsac-cordingto

Ox‚State)

(IV*(SMn-O-Mn-SMn-OH))+(III*SMn-OH)

S(5)

Mn-O-Mn

whereSstandsforsignalofthedifferentcomponentsoftheO1sspectra.Sinceallmanganeseatomsarebondedtoanoxygenatom,theMn-O-Mnsignalshouldrepresentthecontributionoftwospecies:MnOOHandMnO2.Hence,theXPSsignalrelatedtotheMn(IV)speciescanbecomputedbysubtractingthecontributionofthehydroxylgroup(Mn-OH)fromtheMn-O-Mnsignal.BindingenergiesandmanganeseoxidationstatesofauthenticsamplescanbefoundinTable1ofref30.

ResultsandDiscussion

ElectrochemicalBehavioroftheMnO2Compos-iteElectrode.Figure1Ashowsatypicalcyclicvolta-mmogramforacompositefilmelectrodein0.1MNa2SO4atascanrateof5mV/s.Thecyclicvoltammetryresponsewhenthenegativeandpositivepotentiallimitsarerestrictedto0and0.9V,respectively,ischaracter-isticofapseudocapacitiveelectrodematerial,butitisnotperfectlyrectangularduetopolarizationresistance.Thiseffectisnoticeablymoresignificantatthelesspositivepotentiallimitcomparedtothepositivelimit,andthisspecificpointwillbediscussedlater.ThecyclicvoltammogramissimilartothatpreviouslyreportedforMnO2-basedcompositeelectrode,andacapacitanceof150F/gcanbecomputedforthiselectrode.14,21,22

(30)Chigane,M.;Ishikawa,M.J.Electrochem.Soc.2000,147,2246.(31)Chigane,M.;Ishikawa,M.;Izaki,M.J.Electrochem.Soc.2001,148,D96.

Toupinetal.

Figure2.XPSsurveyforanoxidizedandareducedcompositeandthinfilmelectrodes.

XPSSurfaceAnalysisoftheCompositeElec-trodes.Toobservethechangeinoxidationstateofmanganesewhentheelectrodeiscycledbetween0and0.9Vandifionicspeciesareinvolvedinthechargestoragemechanism,thecompositeelectrodeswerecharacterizedbyX-rayphotoelectronspectroscopy(XPS).PriortotheXPSmeasurements,thecompositeelec-trodeswerepolarizedat0or0.9Vuntilthechargepassedwasequaltothechargeintegratedfromthecyclicvoltammogram.ThesurveyspectrapresentedinFigure2foroxidizedandreducedMnO2-basedcompos-iteelectrodesshowMn2p(642eV),Mn3s(84eV),andO1s(530eV)peaksattributedtomanganesedioxide.TheC1s(285and293eV)peaksandF1s(685eV)areassociatedwiththepresenceofacetyleneblack,graph-ite,andPTFE.ThetwospectraarealmostidenticalwiththeexceptionoftheNa1speakthatisalmostabsentfortheoxidizedcompositeelectrodes(videinfra).

TheMn3s,Mn2p,andO1scorelevelspectracanbeusedtoassessthechangeinoxidationstateofmanga-nesefortheoxidizedandreducedMnO2electrodes(seeExperimentalSection).TheMn3scorelevelspectrashouldusuallyshowapeaksplittingandadoubletduetotheparallelspincouplingofthe3selectronwiththe3delectronduringthephotoelectronejection.30,33,34Theenergyseparationbetweenthetwopeaksisrelatedtothemeanmanganeseoxidationstate.Sincealowervalenceimpliesmoreelectronsinthe3dorbital,moreinteractioncanoccuruponphotoelectronejection.Con-sequently,theenergyseparationbetweenthetwocomponentsoftheMn3smultipletwillincrease.30Theinversetrendwillbeobservedwhenthemanganesevalencyincreases.

TheMn3scorelevelspectrawererecorded(FigureSI5)foroxidizedandreducedMnO2-basedcompositeelectrodes,andtherelevantdataareincludedinTable1.ThedataofTable1revealedthatthepeaksplittingofthedoubletoftheMn3scorelevelspectraisalmost

(32)Long,J.W.;Young,A.L.;RolisonD.R.J.Electrochem.Soc.2003,150,A1161.

(33)Moulder,J.F.;Strickle,W.F.,Sobol,P.E.;Bomben,K.D.HandbookofX-rayPhotoelectronSpectroscopy;Perkin-ElmerCorpora-tion:PhysicalElectronicsDivision:EdenPrairie,MN,1992.

(34)Briggs,D.;Seah,M.P.PracticalSurfaceAnalysis,2nded,;JohnWiley&Sons:1996;Volume1.

ChargeStorageMechanismofMnO2ElectrodeChem.Mater.,Vol.16,No.16,20043187

Table1.DataObtainedfromtheXPSSpectra

thinfilmoxidizedas-preparedreduced

0E(V)0.90

Mn3s(eV)peak1peak288.8888.7288.98

84.1083.8083.65

∆eVd4.784.925.33

Mn2p(eV)

3/2∆BEMn-Oa642.6642.4642.3

112.8112.7112.4

oxidationstate

Mn3s/O1sb

4.0/4.03.6/3.72.9/3.1

Mn-O-MnMn-OHH-O-HMn-O-MnMn-OHH-O-HMn-O-MnMn-OHH-O-H

O1s(eV)c

BE(eV)

529.8531.3532.4529.7531.0532.4529.9531.1532.6O1s(eV)c

BE(eV)

Mn-O-MnMn-OHH-O-HSO42-Mn-O-MnMn-OHH-O-HSO42-Mn-O-MnMn-OHH-O-HSO42-Mn-O-MnMn-OHH-O-HSO42-530.2531.0533.5532.0530.3531.0533.3532.0530.3531.6533.8532.0530.2531.0533.4532.0

area%80.62.916.564.820.714.548.643.48.0

compositeelectrodeoxidized

E(V)1.25

Mn3s(eV)peak1peak288.74

83.65

∆eVd5.09

Mn2p(eV)

3/2∆BEMn-Oa642.4

112.7

oxidationstate

Mn3s/O1sb

3.5/3.5

area%50.024.98.516.753.226.99.810.160.330.46.52.849.925.38.616.2

0.989.3984.414.98643.1112.73.5/3.5

reduced088.5883.604.98643.0112.73.5/3.5

-0.6589.4184.395.02643.1112.73.5/3.5

DifferenceinbindingenergybetweentheMn2p3/2andO1s[Mn-O-Mn]peaks.bThefirstentryisobtainedbytheMn3speakshiftandthesecondaftertheslashbytherelativeareacalculationoftheO1scomponents.cForthecompositeelectrode,thepresenceofsulfatewasshownbyaS2ppeakat168.6eV).33Hence,acomponentat532eVwasaddedtofittheO1speakenvelopeinordertotakeintoaccountthecontributionoxygenatomsofthesulfatespecieswhencomputingtheMnredoxstate.dDifferencebetweenthebindingenergiesofpeak1andpeak2.

a

thesameforalltheelectrodes.Thisvaluecloseto5.00eViscomparedto5.79,5.50,5.41,and4.78eVforreferencesampleofMnO,Mn3O4,Mn2O3,andMnO2,respectively.30Hence,themanganeseoxidationstateremainedatabout3.5.Similarfindingswereobtainedforcompositeelectrodespolarizedatmorepositive(1.25V)andmorenegative(-0.65V)potential.TheabsenceofachangeofthemanganeseoxidationstateispuzzlinginlightofpreviousreportsonelectrodepositedMnO230,31andbirnessiteMnOxambigelfilms32inaqueouselec-trolytes.EvenwhentheXPSmeasurementswereperformedwithatakeoffangleof30°or45°,incondi-tionswhereathinnersurfacelayerisprobed,33,34attemptstoobserveachangeofthemanganeseoxida-tionstatefailed.Then,itwassuspectedthattheoxidationstateoftheelectrodechangedduringthedryingstepandexposuretoair.Themeasurementoftheopencircuitpotential(OCP)forbothreducedoroxidizedelectrodes,monitoredbeforeandafterthedryingstep,revealedadriftoftheOCPtoanaveragepotentialof0.45Vfollowingthedryingstep.ThisobservationsuggeststhatonlythesurfaceMnO2maybeinvolvedintheredoxpseudocapacitivereaction.Inthiscase,thesurfaceofthefilmcouldbebroughtbackveryclosetotheoxidationstateofOCPconditionsbyaredoxreactiondrivenbythechemicalpotentialbetweenthesurfaceandthebulkofthematerial.

ElectrochemicalBehavioroftheMnO2ThinFilmElectrode.Toavoidthisphenomenon,thinfilmelectrodeswereusedtoensurethatamoresignificantfractionoftheMnO2filmwouldbeinvolvedintheelectrochemicalreactionduringtheoxidationandre-ductionsteps.ThiswasaccomplishedbyusingMnO2thinfilmsupportedonplatinumelectrodesasdescribedintheExperimentalSection.First,experimentswereperformedwithelectrodesofdifferentfilmthicknesstoestimatetheelectrochemicallyactivefractionoftheMnO2film.ThemassandthicknesswerecontrolledbyaddingbetweenoneandfivedropsofaMnO2-PVdF-HFPmixturewithadryingstepbetweeneachaddition(seeTable2).Figure1BshowsarepresentativecyclicvoltammogramfortheMnO2powdersupportedonplatinumfoil,whichdisplaysthecharacteristiccapaci-tivebehaviorbetween0and0.9V.Thespecificcapaci-tanceof1380F/gobtainedforthiselectrodeisclosetothetheoreticalvalueof1370F/gexpectedforaredoxprocessinvolvingoneelectronpermanganeseatom.ThelargerpolarizationofthePt-MnO2electroderelativetothecompositeelectrode(Figure1A),demonstratedbythemorepronouncedcurvatureofthecyclicvolta-mmogramnearthepotentiallimits,isduetotheabsenceoftheconductivecarboninthethinfilmelectrode.Table2showsthevariationofthevoltam-metricchargemeasuredforeachelectrodeasafunctionoftheMnO2mass.Forthethinnerfilm(electrodeA),alltheMnO2materialistakingpartintheelectro-chemicalredoxprocess,duringthecyclicvoltammetryexperimentperformedat5mV/s,sincetheCoulombicefficiencyisabout100%.TheCoulombicefficiencyiscalculatedfromthemeasuredvoltammetricchargeofthecyclicvoltammogramandthetheoreticalcalculatedchargebyassumingthetransferofoneelectronperMnatomandthatthewholeMnO2massiselectrochemi-callyactive.Table2indicatesthattheCoulombic

3188Chem.Mater.,Vol.16,No.16,2004Toupinetal.

Table2.ElectrochemicalandCompositionDataforMnO2ThinFilmElectrodeswithDifferentLoadings

mass(µg)5.0(0.310.0(0.815.0(1.520.0(2.525.0(3.8

amountofMnO2(moles)5.75×10-81.15×10-71.73×10-72.30×10-72.88×10-7

voltammetricchargea(C)/(C/g)0.0056/12500.0106/11900.0148/11000.0156/8750.0186/835

calculatedcharge(C)b0.00550.01110.01660.02220.0277

coulombicefficiency(%)c

10195897067

specificcapacitance

(F/g)

138013201230970930

electrode

ABCDE

a

CalculatedbytakingintoaccountthemassofMnO2inthesample(about89.1%).ThefirstvalueisinC,whereasthesecondisthespecificvoltammetricchargeinC/g.bThecalculatedchargewasobtainedfromtheamountofMnO2ontheelectrodebyassumingthetransferofoneelectronperMnatom.cCoulombicefficiency(%))(voltammetriccharge/calculatedcharge)×100.

efficiencydecreasedwhenthefilmthicknessincreasedandthatitreachedonly67%for25µgofMnO2.TheseresultsdemonstrateclearlythatasignificantfractionofMnO2wasnotelectrochemicallyaddressablewhenthefilmthicknessincreased.Incontrast,inthecaseofarapidprotonicdiffusioninthebulkoftheactivematerial,thechargewouldincreaselinearlywiththemassoftheelectrode.17ThissuggeststhatslowerionictransportisoccurringwithintheactivematerialorthatprotonscannotdiffusefreelyacrossthethicknessoftheMnO2particles(videinfra).Thisissupportedbytherelativelylowdiffusioncoefficient(6×10-10cm2/s)forprotonsinmanganesedioxide.35

SomeinsightintothedecreaseoftheCoulombicefficiencywithanincreaseofthefilmthicknesscouldbeobtainedbycalculatingthesurfaceoftheplatinumcurrentcollectorcoveredbyMnO2particules.Bycon-sideringsphericalparticleswithameandiameterof420nm,theformationofa“monolayer”oftheseMnO2particleswillrequire33µg.However,asdepictedinFigureSI4b,theMnO2particlestendtoagglomerateratherthanformingamonolayer.TheMnO2depositedontheplatinumsubstrateleadstoalargernumberofclusters,whichreducetheareaofmaterialexposedtotheelectrolyte.IftheelectrochemicalprocessistakingplaceonlyatthesurfaceoftheMnO2exposedtotheelectrolyte,thechargewilldecreaseastheweightofMnO2isincreased.Thiscanexplainwhythinfilmsusuallyexhibithighercapacitancevaluesthanbulkcompositeelectrodes.17Inthisstudy,thegravimetricchargeofourcompositeelectrodes(100µmthick)islimitedto135C/g.Ontheotherhand,whentheweightofMnO2ismuchlowerasforthePt-MnO2samples,thechargeincreasedupto1250C/gforverysmallamountofMnO2.ThisCoulombicefficiency,closeto100%,impliesthatprotonsoralkalicationscandiffusethrougha“monolayer”ofMnO2particles.Thisisexpectedbyconsideringtheparticlesize(<420nmascanbeseeninFigureSI3),thecharge/dischargetime(180sforascanrateof5mV/s),andthediffusioncoefficientforprotonsinmanganesedioxide.35WhenalargeramountisMnO2orwhenathickerfilmisused,thediffusionofactivecationsisclearlyhindered.

XPSSurfaceAnalysisoftheMnO2ThinFilmElectrodes.ThesurveyspectrafortheoxidizedandreducedthinfilmelectrodesaredepictedinFigure2.Thesespectradifferslightlyfromthoserecordedforthecorrespondingcompositeelectrode(alsoshowninFigure2).Thesedifferencesbecomemoreevidentonthehigherresolutionspectraasitwillbedemonstratedbelow.

(35)Ruetschi,P.J.Electrochem.Soc.1984,131,2737.

Figure3.Mn3scorelevelspectraforreduced,as-prepared,andoxidizedthinfilmelectrodes.Thepeakseparationbetweenthetwopeaksisindicatedandcanbeusedtodeterminetheoxidationstateofmanganese.Therawdataarerepresentedbythedots,andthefitteddataarerepresentedbythelines.

Figure3depictstheMn3scorelevelspectraforthereduced,as-prepared,andoxidizedMnO2thinfilmelectrodes.Theseparationofpeakenergies(∆Eb)oftheMn3scomponentsincreasedfrom4.78eVfortheoxidizedfilmto5.33eVforthereducedfilm(seealsoTable1).Inthecaseoftheas-preparedthinfilm,anintermediatevalueof4.91eVwasfound.TheEbvaluesareinagreementwiththoseexpectedforMn4+andMn3+oxides,whichshouldhaveapeakseparationofabout4.7and5.4eV,respectively.30,36Inaddition,itwaspreviouslyshownthatalinearrelationexistsbetweentheenergyseparationoftheMn3speaksandtheoxidationstateofmanganeseintheoxide.21,30,31Fromthisrelationship,themeanmanganeseoxidationstatecanbeestablishedat4.0,2.9,and3.6forthe

(36)Audi,A.A.;Sherwood,M.A.Surf.InterfaceAnal.2002,33,274.

ChargeStorageMechanismofMnO2ElectrodeFigure4.O1scorelevelspectraforreduced,as-prepared,andoxidizedthinfilmelectrodes.Therawdataarerepresentedbythedots,andthefitteddataarerepresentedbythelines.

oxidized,reduced,andas-preparedelectrodes,respec-tively.Thus,incontrasttothethickercompositeelectrode,achangeoftheoxidationstateofmanganesecanbeobservedforthethinnerfilmelectrodeonplatinumsubstrate.Thischangeisalsoaccompaniedbyacolorchangeuponredoxswitching.

TheO1scorelevelspectrawerealsousedtoconfirmthechangeofmanganeseoxidationstateduringredoxswitching.Figure4showsasignificantdifferenceoftheO1senvelopebetweenthereduced,as-prepared,andoxidizedthinfilmelectrodes.Indeed,thereducedfilmshowsadistincthigh-intensityshoulderonthehigherbindingsideofthemainpeak.Togetsomeinsightintothechemicalmodificationthatareoccurringuponredoxswitching,theO1sspectrawereanalyzedbycurvefitting.Figure4(seealsoTable1)showsthatthespectracanbefittedwiththreecomponentswhicharerelatedtotheMn-O-Mnbond(529.8(0.1eV)forthetetravalentoxide,theMn-OHbond(531.1(0.2eV)foranhydratedtrivalentoxide,andfinallytoaH-O-Hbond(532.5(0.1eV)forresidualstructurewater.28,37,38So,thechangeoftheshapeoftheO1senvelopeiscausedbythevariationoftheMn-O-MnandMn-OHcontributions(Table1).Fortheoxidizedelectrode,theMn-O-Mncomponentcontributestoabout81%oftheO1speak,whereasforthereducedelectrode,theintensityoftheMn-O-MnandMn-OHsignalsisalmostidentical.ThevariationoftherelativeintensityoftheO1scomponentsindicatesachangeofthemanganeseoxideoxidationstatebetweentheoxidized

(37)Banerjee,D.;Nesbitt,H.W.Geochim.Cosmochim.Acta1999,63,3025.

(38)Banerjee,D.;Nesbitt,H.W.Geochim.Cosmochim.Acta2001,65,1703.

Chem.Mater.,Vol.16,No.16,20043189

Figure5.Mn2pcorelevelspectraforreduced,as-preparedandoxidizedthinfilmelectrodes.Therawdataarerepresentedbythedots,andthefitteddataarerepresentedbythelines.

andreducedstates.Table1indicatesthatthemeanmanganeseoxidationstate,obtainedfromeq5,isequalto4fortheoxidizedelectrode,whereas,forthereducedelectrode,anoxidationlevelof3.1isfound.Inthecaseoftheas-preparedelectrode,anintermediateoxidationstateof3.7isdetermined.ThesevaluesareinexcellentagreementwiththosecomputedfromtheMn3sdata(videsupra).

Figure5showsMn2pspectraofthinfilmsamples.SuchMn2pcorelevelspectrahavebeenrecentlyusedtodeterminethecontributionofvariousMnspecies.28However,theevaluationoftheMn(II),Mn(III),andMn-(IV)contributionsrelyonacurvefittingprocedure,whichcanbearbitrarysincetheshapeofthespectraoftheelectrodesdoesnotappeartochangedrastically(Figure5).Nevertheless,thebindingenergyseparation(∆EMn-O)betweentheMn2p3/2andO1s[Mn-O-Mn]peakshasbeenfoundtochangeslightlywhentheelectrodeisoxidizedorreduced.30,33AsshowninTable1,the∆EMn-Ovalueswerefoundtobeequalto112.4and112.8eVforthereducedandoxidizedfilmelec-trodes,respectively.Thelarger∆EMn-Ofortheoxidizedthinfilmisinagreementwithliteraturedata31despitethefactthattheabsolutereportedvaluewasslightlylarger.Thus,alltheXPSdataareconsistentwithachangeofmanganeseoxidationstateuponswitchingbetween0and0.9V.TheredoxswitchingofMnO2electrodesmightinvolveionicspeciesfromtheelectro-lytesolution(0.1MNa2SO4),andtheappropriateXPScorelevelspectra(Na1sandS2p)weremeasuredtofurthercharacterizethechargestoragemechanism.Figure6showsthattheintensityoftheNa1speakislargerforthereducedthinfilm(thinfilm,0V)incomparisontotheoxidizedelectrode(thinfilm,0.9V).ThisisconsistentwithachargecompensationofMnOO-byNa+forthereducedfilm(eq3).Inaddition,Figure6demonstratesthatthesulfateanionsarenotinvolvedintheredoxprocess,sincetheS2pspectraarefeature-less.Thisistobecontrastedwiththesignificant

3190Chem.Mater.,Vol.16,No.16,2004Figure6.Na1sandS2pcorelevelspectraforcompositeandthinfilmelectrodes.

differenceinthevariationoftheNa1sandS2pspectraforthecompositeelectrode.Whenthecompositeelec-trodeisswitchedfrom0to0.9V,theNa1ssignaldecreased,whereastheS2psignalincreased.Obviously,somesaltseemstobetrappedsincebothNa+andSO42-arefoundintheelectrodes.Theoxidizedcompositeelectrodecontainsanexcessofsulfate(ratioS/Mn)0.064),whereasthereducedelectrodehasanexcessofNa+(ratioNa/Mn)0.105).ThevariationoftheNa+andSO42-concentrationsforthecompositeelectrodemayappearpuzzlingconsideringthatthevalenceofmanganesedoesnotappeartochangebetweentheoxidizedandreducedstates.TheexcessofNa+andSO42-forthereducedandoxidizedelectrodes,respec-tively,canbeattributedtothepresenceofcarboninthecompositeelectrodes.Thus,thisbehaviorisconsis-tentwiththatexpectedforaclassicaldoublelayerchargestorageprocess.1Theinvarianceofthemanga-neseoxidationstateforthecompositeelectrodeswhichcontrastswiththechangeobservedforthethinfilmelectrodessuggeststhatonlyathinlayerofMnO2ontheelectrodeisinvolvedintheredoxinterconversion.Thisisalsoconfirmedbythehighcapacitancerecordedforthethinnerfilms,whichalmostreachedthetheo-reticalvalues(videsupra).

ThevariationoftheNa/Mnratio,forthethinfilmelectrode,uponpotentialswitchingdeservessomecom-ments.Asmentionedabove,thehigherNa/Mnratioforthereducedelectrodeincomparisontotheoxidizedelectrode-isconsistentwithachargecompensationofMnOObyNa+.Ontheotherhand,theNa/MnismuchlowerthanthatexpectedforacompletecompensationbyNa+.TheseresultsclearlydemonstratethatprotonsareinvolvedintheredoxprocessofMnO2.DespitetheseexperimentalXPSresults,theexactmechanismfortheuptakeofNa+isunclear.Recentelectrochemicalmea-surementsoftheintercalationofalkalimetalionsintobirnessitemanganesedioxideinaqueousmediasug-gestedthatprotonsaredirectlyintercalatedbutnotthealkalimetalcations.39Thepresenceofthealkalimetal

(39)Kanoh,H.;Tang,W.;Makita,Y.;Ooi,K.Langmuir1997,13,6845.

Toupinetal.

cationsintheoxidematrixwasexplainedbyanion-exchangereactionbetweenthecationsandH+.Accord-ingly,asimilarmechanismcannotberuledoutcom-pletelyforourelectrodes.

Conclusion

Inthiswork,thechargestoragemechanisminMnO2electrode,usedinaqueouselectrolyte,wasinvestigatedbycyclicvoltammetryandX-rayphotoelectronspec-troscopy.Themainobjectivewastodeterminewhetherthemanganeseoxidationstatechangedduringpotentialswitchingbetween0and0.9VvsAg/AgCl.Tothisend,thinMnO2filmsdepositedonaplatinumsubstrateandthickerMnO2compositeelectrodeswereused.X-rayphotoelectronspectroscopy(XPS)measurements(Mn3sandO1s)withthethickcompositeelectrodesdidnotrevealanychangethatcouldbeassignedtoavariationofthemanganesevalency,and,atthispoint,thechargestoragemechanismcouldbebasedonelectrostaticeffectsonly.Infact,thechargestoragewouldbesimilartothatobservedforcarbonelectrodes.1Ontheotherhand,acompletelydifferentXPSbehaviorwasnoticedforthethinfilmelectrodes.BoththeMn3sandO1sspectrawereconsistentwithmanganeseoxidationstatesof+3and+4forthereducedandoxidizedforms,respectively.TheXPSdataalsoshowthatNa+cationsfromtheelectrolyteareinvolvedinthechargestorageprocessofMnO2thinfilmelectrodes.TheNa/MnratioforthereducedelectrodeismuchlowerthanwhatisanticipatedforchargecompensationdominatedbyNa+andsuggeststheinvolvementofprotons.TheapparentdiscrepancybetweentheXPSdata(Mn3sandO1sspectra)canbeexplainedbythecyclicvoltammetrydataofthethinfilmelectrodeswhichestablishedthatonlyathinlayerofMnO2isinvolvedintheredoxprocessandiselectrochemicallyactive.Presumably,thisthinsurfacelayercannotbeprobedforthecompositeelec-trodebecausethisregionisbroughtbacktothechemical(oxidation)stateofthebulkbyinternalredoxintercon-version.

Acknowledgment.ThefinancialsupportoftheNaturalScienceandEngineeringResearchCouncil(NSERC)andtheCanadianFoundationforInnovation(CFI)isacknowledged.Oneoftheauthors(T.B.)wouldliketothank“l’Universite´deNantes”forgivinghimtheopportunitytoworkinUQAMandUQAMforwelcom-inghimasavisitingprofessor.The“Ministe`reFranc¸aisdesAffairesEtrange`res”andthe“Ministe`redesRela-tionsInternationalesduQue´bec”arealsogreatlyac-knowledgedforsupportingthisworkwithintheframe-workofthe“CommissionPermanentedeCoope´rationFranco-Que´be´coise”(project#59-102).

SupportingInformationAvailable:X-raydiffractionpatternoftheas-synthesizedMnO2powder(FigureSI1),scanningelectronmicrographoftheas-synthesizedMnO2powder(FiguresSI2andSI3),scanningelectronmicrographsofaMnO2compositeelectrode(FigureSI4a),scanningelectronmicrographofMnO2powder-PVdF-HFPcoatedonPt(FigureSI4b),andMn3scorelevelspectra(FigureSI5).ThismaterialisavailablefreeofchargeviatheInternetathttp://pubs.acs.org.CM049649J