photoactivity issues problematic of enhancementfactors Novel synthesis for WO approaches -TiO /MWCNT compositephotocatalysts- Catalysis Today

(1)

factors

Enik ˝o Bárdos

^a

, Gábor Kovács

^a,b,c

, Tamás Gyulavári

^a

, Krisztián Németh

^a

, Egon Kecsenovity

^a

, Péter Berki

^a

, Lucian Baia

^b,c

, Zsolt Pap

^b,c,d

, Klára Hernádi

^a,∗

aDepartmentofAppliedandEnvironmentalChemistryUniversityofSzeged,RerrichBélatér1,HU-6720,Szeged,Hungary

bFacultyofPhysics,Babes¸–BolyaiUniversity,M.Kog˘alniceanu1,RO-400084Cluj–Napoca,Romania

cInstituteforInterdisciplinaryResearchonBio-Nano-Sciences,Babes¸–BolyaiUniversity,TreboniuLaurian42,RO-400271Cluj-Napoca,Romania

dInstituteofEnvironmentalScienceandTechnology,UniversityofSzeged,TiszaLajoskrt.103,HU-6720Szeged,Hungary

a r t i c l e i n f o

Articlehistory:

Received30September2016

Receivedinrevisedform25February2017 Accepted15March2017

Availableonline22March2017

Keywords:

WO3nanocrystallites WO3-TiO2nanocomposites Photocatalyticactivity Carbonnanotubes Photocatalysis

a b s t r a c t

The “build-up” methodology, the importance of the order of the semiconductor layers in WO3- TiO2/MWCNT composite materials was studied in terms of the applied synthesis pathway, morpho-structuralparameters(meancrystallitesize,crystalphasecomposition,morphology)andpho- tocatalyticefficiency(usingoxalicacidasmodelpollutant).Theappearance ofTiWOxphaseinthe compositescontributedtotheenhancementofthephotocatalyticefficiencies,asdifferentsynthesis approachesledtodifferentcrystalphasecompositions.Although,itwasproventhatabeneficialphase’s presencecanbehinderedifanexcessofMWCNTorWO3wasapplied.Astheratioofthementioned materialswasreduced,activecompositeswereobtained,butthepreviouslynoticedTiWOxdisappeared.

Therefore,itwasproven,thatinthecaseofWO3-TiO2/MWCNTnanocompositesystemseveralphoto- catalyticactivityenhancementfactorscanbeintroduced,butnotsimultaneously(thedisappearanceof TiWOxatlowMWCNTandWO3contentsandtheappearanceofhighlycrystallineanatase).

1. Introduction

Catalysts are of great importance in the synthesis of petro- chemicals, plastics, medical agents or fuelcell materials. Their applicability extends also for environmental applications, for examplethereduction ofemissionsand degradationof various organicpollutants[1].Thecatalyticactivityandselectivitydepend stronglyonseveralparameters, suchascrystal phase and stoi- chiometriccomposition,particlesizeandmorphology,crystallinity gradeofthecatalystortheshapeofthecatalystnanocrystals[2,3].

Amongthecatalyticprocesses,photocatalysisisaveryimportant alternativeinthedegradationoforganicpollutants,anditisbased ontheirradiationofasemiconductorwithphotonsofgreateror equalenergythanitsbandgap.Titaniumdioxide(TiO₂)isasemi- conductormetaloxide,whichhasmanyadvantageousproperties thatmakesitalmostidealforphotocatalyticapplications[4,5],itis photostable,biocompatible,non-toxic,relativelycheapandavail- ableinlargequantity[6].UnderUVirradiation,itisanexcellent

∗Correspondingauthor.

E-mailaddress:hernadi@chem.u-szeged.hu(K.Hernádi).

materialforthedegradationofvariousinorganicandorganiccom- pounds[7].However,becauseofitslargebandgapenergy,TiO₂ absorbslightonlyintheUV-A,nearvisibleregion.Thisisthemain drawbackofitswiderapplicationsinceonly4–5%ofthesolaremis- sionspectrumfallsinthisrange[8,9].Toovercomethisproblem there aremany differentapproaches inthe scientiﬁcliterature.

Themostimportantmethodstoincreasethephotocatalyticactivity containcatalystmodificationbydoping,metaldecoration,surface sensitization,orbydesigningmoreefficientcomposites(e.g.mixed semiconductoroxides)[10],thelatterpossibilityseemstoberather promising.Efficientchargeseparationcanbeobtainedbythecom- binationoftwosemiconductorparticlesthathavedifferentenergy levels[11]. Byproperly tailoringtheproperties ofthe compos- itecomponents,metaloxideheterojunctionscancertainlybean interestingapproachtomodifythephotocatalyticperformancesof individualoxides[12].Everyoxidecomponentcandisplaydifferent chemicalinteractionwiththecontaminanttodecompose.Surpris- ingly,inmanycasesoxidesystemsusedforcatalysisshowmany similaritieswiththeonesinvestigatedforsensing[13].

In the last few years, tungsten trioxide − titanium dioxide (WO₃/TiO₂) composite photocatalysts have drawn researchers’

attention since they possess interesting and unique photocat- http://dx.doi.org/10.1016/j.cattod.2017.03.019

(2)

alyticfeatures[14–16].WO₃showsamoderatelysmallbandgap energy,strongabsorptionwithinthevisiblespectrum,long-term performanceand highchemical stabilityin a widepHrange of aqueoussolutionsunderoxidativecircumstances[17,18].Itiswell- knownin theliteraturethat theconnectionbetweenWO3 and TiO₂enablesimprovedchargeseparationandvisiblelightresponse [14–16,18–22]. Furthermore, WO₃ based composites possess a strongabsorptioninvisiblelight,buttheirphotocatalyticactivities areverylow.

Comparingbandgapsofthetwocomponents(whilerelatively small−2.4–2.8eV−forWO3andhigh−3.0–3.2eV−forTiO2)itis obviousthatthetransferofelectronsfrombulkTiO2toWO3might furtherminimizetherecombinationlosses.

Besides theabove-mentioned composite formation of semiconductoroxides,theadditionofcarbonnanotubes(CNT)canbe alsoa feasible methodto elongate thelifetime of photogener- atede⁻/h⁺ pairs and preventtheirrecombination [23].Because oftheiruniquechemical,physicalandelectronicproperties,CNTs arewidelyinvestigated[24,25]inordertopromoteapplications intheﬁeldofcompositematerialsandphotocatalysts[23].Since CNTs canact as anelectron reservoir tostore temporarily and transferthephotogeneratedelectronsfromTiO₂[25],withadding CNTstotheoxide,thephotocatalyticactivityofTiO₂canbefur- therimproved[26–28]and possiblyconfervisible lightactivity [29].The combinationof thebeneﬁcialproperties ofWO3 with theadvantagesofmultiwalledcarbonnanotubes(MWCNTs)can also result in a synergic effect tocreate materialswith differ- entproperties for targetedapplications.Thus, researchers have studiedWO₃/MWCNTcompositesinpreviousworksinorderto investigatetheirgassensing [30,31]andelectrochromic performances[32].Differentimpregnationmethodswerealsopresented toprepareWO₃/MWCNTcomposites:eitherasimpleone-stepfab- ricationmethodorcombinedwithprecipitationbasedsynthesis techniques[33]. Accordingtoourknowledge, just a few publi- cationscanbefoundregarding thephotocatalytic propertiesof ternarycompositescontainingtitaniumandtungstenoxideswith MWCNT. Literaturedata are rather inconsistentconcerningthe build-upmethodology(synthesisapproachesandoxidelayercom- position)oftheabove-discussednanocomposites.

Moresynthesis approaches werealready developedin order toinvestigatetheavailablepossibilitiestoobtaindifferentcon- tactorderofthecompositecomponents.Twopotentialmethods wereidentiﬁed:modiﬁed/slowhydrolysis(MSH)byRétietal.[26], andimpregnation(I)methodwhichwasinvestigatedbyVassetal.

[33].Thesetwoapproaches(combinedaMSHI)providedreliable resultsconcerningtheuniformcoverageofMWCNTswithdifferent semiconductoroxides(In2O3,SnO2,Al2O3,ZnO,SnO2-In2O3).The advantagesanddrawbacksofthesesynthesismethodsaresumma- rizedinFig.1.

Inthispaper,ourinterestistoclarifytheeffectoftheheattreat- mentonthecrystallinityandmorphologyofcompositematerials thatwere obtainedby themethodslistedabove, whilevarying

Fig.1.Thedifferentpreparation methodadvantagesand disadvantages(slow hydrolysisandimpregnation(SHI),amodiﬁedslowhydrolysisandimpregnation, (MSHI)andasimultaneoussynthesis(SS)).

theorderofthedifferentsemiconductorlayersandattemptingto depositsimultaneouslythetwooxides(simultaneousslowhydrol- ysisandimpregnation−denotedasSSmethod).

2. Materialsandmethods 2.1. Chemicals

Allchemicalsusedwereofanalyticalgrade.TheMWCNTsfrom Nanothinx(NTX1type)titaniumisopropoxideTi(iPrO)₄andtung- sten(VI)hexachloride(WCl6)fromSigma-Aldrich,P25fromEvonik Aeroxide(Germany)wereusedwithoutfurtherpuriﬁcationforthe preparationofthecompositematerials(secondarycomposites,for detailspleaseconsultSections2.2and2.3)

2.2. Characterizationmethods

FormeasuringtheDRS(DiffuseReﬂectanceSpectroscopy)spectra of the samples, a JASCO-V650 spectrophotometer with an integrationsphere(ILV-724)wasused(␭=250–800nm).

Theoxidelayerthicknessandcoverageratiooftheoxidesonthe MWCNTwereveriﬁedwithFEITechnaiG2X-TWINTEM(200kV).

Thesampleswerepreparedasfollows:asmallamountoftheexam- inedcompositewassonicatedin1.25cm³ofethanol.Afewdrops fromthissuspensionweredepositedanddriedontothesurfaceof thegrid(CF200CuTEMgrid).

TheSEMmicrographswererecordedonaHitachiS-4700Type IIFE-SEMscanningelectronmicroscopeequippedwithacoldﬁeld emissionsourceoperatingintherangeof5–15kV.Thesamples weremountedonaconductivecarbontape.

PowderXRDanalysiswascarriedoutonaRigakuMini-ﬂex- IIDiffractometer(angle-range:2␪=20–80^◦,␭=0,15418nm)using characteristicX-ray(CuK_␣)radiation.

TheRamanspectroscopymeasurementswerecarried outon aThermoScientiﬁcDXRRamanmicroscopewitha532nmlaser (5mW).

A photoreactor system with 2 LighTech 40W ﬂuorescent lamps (␭max≈365nm, irradiation distance=5cm, irradiation time=2h) was used to measure the photocatalytic activities.

ThenanocompositesweregrindedtogetherwithP25(secondary composites)in ordertodifferentiatetheirimpactonthephoto- catalyticactivity.Thephotocatalystsuspensioncontainingoxalic- acid (initial concentration of oxalic-acid c0,oxalic-acid=0.5mM;

c_suspension=20mg/20cm³=1.0gL⁻¹;total volumeofthesuspen- sion V_suspension=20cm³) was continuously stirred during the experiments.Theconcentrationdecrease oftheoxalicacidwas followed using a Merck-Hitachi L-4250 system with Merck- Hitachi L-7100 pump. The eluent wasH₂SO₄ (concentration of H2SO4=19.3mM),theappliedﬂowratewas0.8cm³min⁻¹andthe detectionwavelengthwas205nm.

2.3. Preparationofnanocomposites

The TiO₂-WO₃/MWCNT composites were prepared by slow hydrolysisandimpregnation(SHI),amodiﬁedslowhydrolysisand impregnation,(MSHI)andasimultaneoussynthesis(SS)method.

Thenomenclature ofthecomposites containstheused method abbreviation (listedabove)− ﬁrstoxide−calcinationtempera- ture− second oxide− calcination temperatureweight ratioof thecomponents(justforthosesamplesinwhichtheratioofthe componentswaschanged).Ifnoheattreatmentwasappliedfor thespeciﬁc oxide,thecalcination temperatureis denoted with 0. Examples:MSHI-TiO₂-400-WO₃-450, MSHI-TiO₂-0-WO₃-450.

In thecase ofSSseriesthejointheattreatmenttemperatureis includedinthenomenclature−e.g.SS-TiO2-WO3-700.

(3)

Fig.2.TheXRDpatternsoftheobtainedcompositematerials(−TiWOx,-monoclinicWO3,䊏−AnataseTiO2,䊐−RutileTiO2).

Theadditionofcarbonnanotubeswasdecidedasinthepho- tocatalytic degradation process is presumably responsible for enhancedchargeseparation,ascarbonnanotubesareefﬁcientelec- tronconductingmaterials,whichcanseparatetheelectronsfrom thesemiconductor,minimizingthechanceofrecombination.The compositesmadebySHIwerepreparedusinganalkoxideprecur- sor,titaniumisopropoxideTi(iPrO)₄ andethanol assolvent.The weightratioof thereactantswas1:10:15(MWCNT/TiO₂/WO₃).

The dispersion of 100mg MWCNTs and 150cm³ ethanol was stirredvigorouslyunderinertatmosphere(Ar)whiletheprecur- sor(3.71cm³)wasaddedfollowedbyahomogenizationperiodof 1h.Thesampleswerekeptatroomtemperatureforaweekuntil thesolventevaporatedcompletely.Thesampleswereseparated intotwoparts:theﬁrstpartwasannealedat400^◦Cfor3hinair andgrindedintoﬁnepowderinanagatemortar.Thesecondpart wasnotheattreated.Fortheimpregnationmethod,TiO₂/MWCNT wasappliedastherawmaterial(bothinannealedandnotannealed forms)andWCl6astheprecursor.Initially,160mgTiO2/MWCNT

wasaddedin200cm³acetoneandasuspensionwaspreparedvia sonicationfor45min.Duringthistime,thecalculatedamountof precursor(257mg)wasdissolvedinanotherportionofacetone (20cm³).Finally,thesolutionoftheprecursor wasaddeddrop- wiseintotheTiO2/MWCNT-suspensionandthenthemixturewas heatedto55–60^◦Ctoevaporatethesolventonaheatedmagnetic stirrer.Whenthesolventwascompletelyremoved,ablackpowder wasobtained.Thispowderwasdriedfurtherat90^◦Cfor24h.The sampleswereannealedat450^◦Cfor3hinair.Compositeswerealso preparedwhenthedepositionorderoftheoxideswasreversed.

The other TiO2-WO3/MWCNTcomposites were prepared by MSHImethod.Theweightratioforthepreparationwas1:10:15.

2.65cm³Ti(EtO)₄ wasaddedinto14.74cm³ethanolandhomog- enized using ultrasonication for 15min to form the precursor solution.ThedispersionofMWCNTswasstirredvigorouslywhile theprecursorsolutionwasaddeddropwise.Themolarratioswere n(TiO2):n(H2O):n(ethanol)=1:3:30.Aftertheadditionofthepre-

(4)

cursorsolution,thebeakerwascoveredwithplasticparaffinfilm andwasstirredforanadditional1h. Afterstirring,thesamples werekeptat80^◦Cuntilallthesolventevaporated.Sampleswere annealedat400^◦Cfor4hinairandgrindedintofinepowderinan agatemortar.Theimpregnationprocesswasthesamelikeinthe caseofSHImethod,butthesampleswereannealedat700^◦Cfor 3hininertatmosphere(Ar).

ThecompositeswerepreparedalsobytheSSmethod:100mg MWCNTwasaddedinto150cm³acetonehomogenizedbyultra- sonication for 45min. The dispersion of MWCNTs was stirred vigorouslyunderinertatmosphere(Ar)whiletheTi(iPrO)4precur- sor(3.71cm³)wasadded.WCl6precursor(257mg)wasdissolved inanotherportionofacetone(20cm³)andwasaddeddropwiseto theMWCNTdispersion.Then,thesuspensionwaslefttobestirred for1h.Afterstirring,thesampleswerekeptat23^◦Cforaweek untilallthesolventevaporated.Sampleswereannealedat700^◦C for3hininertatmosphere.

For the photodegradation of oxalic acid, the previously described MWCNT-based composites and Evonik Aeroxide P25 wereusedforthepreparationofthephotocatalyst(secondarycomposites).Ineachcase,aspeciﬁcratiowasestablishedbetweenthe composite components,30%MWCNT-based composite and 70%

TiO₂ (EvonikAeroxideP25).Thenanocompositeswereprepared viamechanicalmixinginanagatemortarfor3×5min.

3. Resultsanddiscussion

3.1. Crystallinestructureofnanocomposites

TheXRDanalysisoftheobtainedcompositematerialsstarted withthesamplesSHI-TiO₂-400-WO₃-0and SHI-TiO₂-400-WO₃- 450. Both of them were obtained starting from heat treated TiO2/MWCNT,whichcanbeobservedbythepresenceofthemain anatase(JCPDS21-1272)peakat25(2␪^◦).InthecaseofSHI-TiO₂- 400-WO₃-0,besidestheanatasepeak,onlyabaselinedistorting wasobserved,behaviorthatindicatesthatWO3isclearlypresent asamorphousphase.Whenasupplementaryheattreatmentwas appliedat450^◦CthesignalsofmonoclinicWO₃emerged(JCPDS 43-1035,themeancrystallitesizewasof20.2nm).Byapplyingthis supplementaryheattreatment,aslightincreaseofthemeancrys- tallitessizeofanataseTiO₂,from25nmto27nm,wasobserved (Fig.2).

Theabove-mentionedobservationschangeddrastically,when theobtained TiO₂/MWCNTwas not calcinated. Therefore, after impregnationwithWO₃precursornosignofanycrystallinemate- rialwasdetected(thecaseofsampleSHI-TiO2-0-WO3-0).However, ifapostcalcinationwasapplied,bothsignaturesofanataseTiO₂and monoclinicWO₃appearedintheXRDpattern(SHI-TiO₂-0-WO₃- 450),althoughasmalleramountofamorphousmaterialremained asthepatternshowsaslightincreaseinthemainbaseline.The meancrystallitesizeofTiO₂ andWO₃ was29.9nmand74.0nm, respectively.

Reversingthedepositionorder,anotherinterestingaspectof thesematerialsisrevealed.IftheWO₃/MWCNTinitialcomposite wasnotcalcined(compositeSHI-WO₃-0-TiO₂-0)andtheprecur- sorofTiO2wasdeposited,acompletelyamorphousmaterialwas obtained (although there are signs of really small amounts of crystallinematerial).Whenapostheattreatment(compositeSHI- WO3-0-TiO2-400)wasapplied,asmallanatasesignalappeared, whiletherestofthematerialremainedamorphous.Nocrystallite sizecalculationwaspossibleinthecaseofthesetwosamples.

ToinvestigatethebehaviorofTiO2,whichwassubsequently deposited,apriorheat-treatmenttotheWO₃/MWCNTcomposite wasapplied.After theimpregnation process,the material con- tainedonlycrystallinemonoclinicWO3,whilenosignofanyTiO2

Fig.3.TheRamanspectraofthecompositematerials(*−WO3relatedbands,䊏− AnataseTiO2,䊐−RutileTiO2).

diffractionpeakswasdetected(sampleSHI-WO₃-450-TiO₂-0).The mean crystal size of WO3 was28.4nm. Whenthe supplemen- taryheattreatmentwasapplied(sampleSHI-WO₃-450-TiO₂-400), theanatasepeakemerged,andthemeancrystallitesizeofWO₃ increasedslightlyto33.4nmandthenewlyformedanatasecrys- tals’sizewas34.6nm.

WhenTiO₂wasthefirstoxidedepositedonthesurfaceofMWC- NTs,crystallinematerialswereobtainedinthreecases.WhenWO₃ wasthefirstone,thefinalmaterialwascrystallineonlyintwocases.

Itseemslikethepre-depositionofWO₃inhibitsthecrystallization ofthesecondoxide.Thisphenomenonwaspreviouslyobservedin literature[34,35].Hence,theﬁrststrategywasfurtherdeveloped.

Asitwasalreadydetailedinthesectionfocusingonthesynthe- sisofthesematerials,themodiﬁedsol-gelandthesimultaneous precursorhydrolysisstrategieswereapplied.

ThecompositedenotedasMSHI-TiO₂-400-WO₃-0possessthe samecrystallineproperties asthesampleSHI-TiO₂-400-WO₃-0, namely theycontain anatasephase withthe same mean crys- tallite size. However, if a supplementary heat treatment was applied(thistimeat700^◦C,inordertofurtherincreasethecrys- tallinityofthesample),newdiffractionpeaksappearedat23.4and 33.1(2␪^◦).Thesepeaksreappearedinthecaseofthecomposite (SS-TiO₂-WO₃-700,whileSS-TiO₂-WO₃-0wascompletelyamor- phous),togetherwiththealreadymentionedanatasereﬂections andthenewonesattributedtorutilephase(themaindiffraction peakat27(2␪^◦),JCPDS21-1276).NeitherTiO₂ norWO₃ crystal phaseswereidentiﬁablebythepeaksmentionedearlier.Apossible explanationcouldbefoundintheworkofQureshietal.[36],who shownthepossibleformationofTiWO_x,arathersurprisingcom- poundobservedrarelyinthelastdecade.TiWOxappearedonlyin the1:10:15weightratiosamples.Thismeansthatarelativelyhigh amountofWO₃andMWCNTisneededtofavortheformationof TiWOx,Wheretheamountofthementionedtwocomponentswas loweredTiWOxwasnotpresentatall.Itshouldbealsonotedthat

(5)

Fig.4. Scanning(a,c)andtransmission(b,d)EMsofTiO2/MWCNT-basedcomposites(SHI-TiO2-0-WO3-450)andSHI-TiO2-400-WO3-450(c,d)).

Fig.5. Scanning(a,c)andtransmission(b,d)EMsofWO3/MWCNT-basedcompositesSHI-WO3-0-TiO2-400(a,b)andSHI-WO3-450-TiO2-400(c,d)).

(6)

Fig.6.Scanning(a,c)andtransmission(b,d)EMsofcompositesMSHI-TiO2-400-WO3-700(a,b)andSS-TiO2-WO3-700(c,d).

theheattreatmentwasalsocrucialtoobtaintheabovementioned compoundasitwasobtainedonlywhenacalcinationwasapplied at700^◦C.

Thephotocatalytic testsdidnotshow anypromising results asitwillbedetailedinSection2.4,althoughtwomethodswere involved(SSandSHImethods),whichafterfurtheroptimization couldleadtophotoactivecomposites.Thestrategyadoptedwas basedonthefact thattheCNTand theWO₃ ‘s relativecontent wasratherhighcomparedtoTiO2.Thus,theWO3 andCNTcon- tentswerelowered(newcomponentweightratios1:18:1,1:16:3).

WhentheXRDpatternswereanalyzedforthesecompositesitwas foundthatjustanatasecrystallizedfromthetwooxides,whilethe non-heattreatedsampleswerecompletelyamorphous.

Bycomparingthemeancrystallitesizeofthesecompositestwo importantaspectscanbepointedout.Thus,whentheratioofWO₃ decreased(relativetoTiO2)itwasfoundthatthecrystallitesizeof theformedanataseslightlyincreased(from16.3to17.4nm−sam- plesSHI-TiO₂-0-WO₃-4501:16:3vs.SHI-TiO₂-0-WO₃-4501:18:1), inthecaseofsampleseriesSHIandmoreintensively(from22.4nm to47.5nm−samplesSS-TiO₂-WO₃-7001:16:3vs.SS-TiO₂-WO₃- 7001:18:1) whentheSSmethodwasapplied.Thisobservation emphasizesagainthatthepresenceofWO3inhibitsthecrystalliza- tionofTiO₂.Furthermore,atthisratiovalues,WO₃doesnotappear atallintheXRDpatternsindicatingitspotentialdopantrole.

ItshouldbementionedthatinallthecaseswhereWO3wasnot observableintheXRD,amorphousW(OH)x ismostprobablythe primaryhydrolysisproduct,thereforenoXRDpeakswereexpected andsubsequentlyobtained.

TheFT-Ramanspectrarevealthepresenceofbandscharacter- isticforTiO₂ anataselocatedaround144, 197(these twowere notincludedintheFigure),394,512,638andrutilephaseat445, 610cm⁻¹(bothofthemoverlappedwithanatase),respectively.The

DandGbandsoftheMWCNTdidnotsufferanymodiﬁcation,and thereforetheywerenotincludedinFig.3.

TheWO3crystallinestructures(includingtungstates)giverise tocharacteristicRamanbands around710and 800cm⁻¹,while O W O bending modes are located at≈260cm⁻¹ and W O stretchingmodesat950–960cm⁻¹ [34,35].Mostoftheidentifi- cationsmadeabovecoincidewiththefindingsdeterminedfrom the XRD results, therefore the observations made for samples SHI-TiO2-400-WO3-0,SHI-TiO2-400-WO3-450andthosewithsim- ilarstructurescanbeconsideredvalid.However,itisinteresting that thebandassociatedwiththepresence ofW O unitsdoes not appear in thespectraof theinvestigated samplesmeaning thatcompletehydration(duringhydrolysis)ornon-stoichiometric specieswereformed.Thebandat330cm⁻¹,whichcorrespondsto W-OH₂ stretchingvibrations,suggestedthepreviouslyenounced possibility,asthesespeciesappearwhendefectiveWOxspeciesare present[34,35].ThiswasalsoconfirmedbytheformationofTiWOx

identiﬁedbyXRD.

Furthermore,iftheratioofthecompositecomponentsisshifted in the favor of titania (while lowering the MWCNT and WO₃ content−sampleSS-TiO₂-WO₃-7001:18:1)thepreviouslymen- tionedRamanbands,specifictoWO3,aremissing,whiletheones attributedtothepresenceofanataseTiO₂intensifiedsignificantly.

Thisinformationpointsoutincreasedcrystallinity,highermean crystallitesizeandpossibledopingofanatase/amorphousWO3.

3.2. Morphologicalinvestigationsusingscanningand transmissionelectronmicroscopy

BothSEMandTEMinvestigationsrevealedthatcarbonnano- tubeswerehomogeneouslycoveredbythebinaryoxideandthe thicknessofthelayerwasapproximately20–40nm(Fig.4a,b−

(7)

Fig.7. Scanning(a)andtransmission(b)EMsofchangedweightratiocompositesSS-TiO2-WO3-7001:18:1.

sampleSHI-TiO2-0-WO3-450).WhenTiO2washeattreatedprior WO₃deposition(Fig.4c,d),theSEMandTEMmicrographsshowed thattitaniumdioxideandtungstenoxideevenlycoveredthesur- faceof MWCNT,however,in manyregions both oxidesformed separateaggregatesaswell.Fromthemicrographsitcanbeseen thatthethicknessofthelayerwasabout20–40nmagainandthe sizeofdetachedoxideparticleswereintherangeof50–200nm.

ChangingthedepositionorderoftheoxidesontheMWCNT, itcanbeclearlyseenfromSEMandTEMmicrographsthatcarbon nanotubeswerenotcoveredhomogeneouslywithtitaniumdioxide andtungstenoxide(Fig.5a,b),asmentionedpreviously,butlarger aggregatesof100–500nmsizewereformedandattachedtothe surfaceofMWCNT.However,whenWO₃receivedaheattreatment at450^◦CbothSEMandTEMmicrographsconﬁrmedthatcarbon nanotubeswerehomogeneouslyandthicklycoveredbythebinary oxidelayerofapproximately100nm(Fig.5c,d).

Inthe case ofthe compositessynthesized bymodiﬁed slow hydrolysisandimpregnation(MSHI),SEMandTEMmeasurements provedthecoverageonthesurfaceofcarbonnanotubes(Fig.6a,b) Whilethehomogeneousregionofthelayerwascomposedoftung- stenoxide,thetitaniumdioxideformscrystallinenanoparticles insidethecoverage.Theaveragethicknessofthelayerwasofabout 50nm.

Whenthesampleswerepreparedwithsimultaneoussynthesis (SS)methodFig.6(c,d)itcannotbeseenhomogeneouslayeronthe surfaceofMWCNT.SEMandTEMimagesshowlargeroxideparti- clesofapproximately100nm,randomlyattachedontothecarbon nanotubes.Whentheweightratiowaschanged,thesurfaceofcar- bonnanotubewascoveredhomogeneouslywithtitaniumdioxide andtungstentrioxide.Theaveragethicknessofthelayerwasof 50–100nm(Fig.7a,b).

3.3. OpticalpropertiesoftheindividualWO₃andcomposites

Inordertoevaluatetheopticalpropertiesofthenanomaterials, diffusereﬂectancemeasurementswereperformedonthecompos- iteswiththebestphotocatalyticefﬁciencies.ItcanbeseeninFig.8, thatthematerialsabsorbvisiblelightinrelativelyhighratio(more than80%).ThisbehaviorcanbeattributedtothehigherMWCNT contentinthecomposites(5%).

Duetothesmalllight absorptionedge,theband-gapvalues cannotbeevaluatedinthetypicalway,namelybytheKubelka- Munkequation,becausethevalues obtaineddo not reﬂect the reality(values4.0–5.0eV).Therefore,theﬁrstorderderivativeof theDRSspectra(Fig.9)wascalculatedinordertogaininformation abouttheelectrontransitionsoccurringduringtheirradiationof thenanomaterials,resultingrealisticband-gapvalues.

Fig.8.ThereﬂectancespectraofselectedWO3-TiO2/MWCNTcomposites.

Analyzing the first order derivative of the reflectance data, mainlythreeelectrontransitionbandscanbeseen,bothrespon- siblefor theUV light absorptionof thecatalysts, at≈362, 392 and≈400nm,respectively.Thesepeaks,intermsofenergy,arecor- respondingtotheband-gapvaluesofanatase,inthefirstcase,to TiWO_xat392nm,andtomonoclinicWO₃for400nm.Inadditionto thequalitativepresenceofthesetransitionsitisclearlyvisiblethat thedifferentlysynthetizedcompositeshaveachangeintheratioof thesepeaks.Anotheraspect,whichcanbeobservedwhenWO₃is depositedonthesurfaceofthenon-calcinatedTiO₂/MWCNTsam- ples,consistsinthatthepeakofWO3cannotbeobservedonthe firstorderderivativeoftheDRSspectra.Thisabsenceofthesignal ismostprobablyduetothepresenceofnon-crystalline/amorphous WO₃,which“delays”thecrystallizationprocessofbothsemicon- ductorsinthenanocompositesandmaysuggestapossibledoping.

WhenWO₃wasdepositedonthethermallythreatedTiO₂-based material,thecharacteristicderivativepeakofWO₃appearsaround 400nm,whichmeansthatthecrystallizationprocessissuccess- fulinthiscase,thesemiconductor-oxideis“transformed”intoits monoclinicform(resultswereprovenalsobyXRDandRamanmea- surements,asdiscussedinSection3.1).

Astheratioofthecomponentswaschangedto1:18:1(sam- pleSS-TiO2-WO3-7001:18:1) anatasecrystallized(a clearlight absorptionedgewasvisibleintheDRSspectrum−Fig.8),while nosignsof WO₃ wasdetectedintheXRD patternsand Raman spectra.AsitcanbeseeninFig.9thederivativespectrumdoes notshowanyadditionalbands,exceptthecharacteristicbandof

(8)

Fig.9.FirstorderderivativespectraoftheWO3-TiO2/MWCNTcomposites.

anataseat365nm.ThisexplicitlypointsoutthattheWO₃present inthatsamplemustbeamorphous[35].

3.4. PhotocatalyticactivityofWO₃-TiO₂/MWCNT-based nanocomposites

Theinitialcompositeswere notactiveunder UVirradiation.

ThisresultcanbeattributedtotherelativelyhighamountofWO₃ presentinthecomposite,and,asitisknownfromtheliterature, WO₃haslowphotocatalyticactivityunderUVirradiation[37].In ordertogainfurtherinformationaboutthephotocatalyticpotential ofthecompositesawell-knowncommercialphotocatalyst(P25)in aratioof70%P25−30%compositewasprepared.

Aftertwohoursof photodegradation(Fig.10)of oxalicacid, itcanbeconcludedthatthebestperformingcompositewasthe

SHI-TiO₂-0-WO₃–450+P25,withadegradationyieldof83.3%and the composite made bysimultaneous synthesis (+P25) (75.1%).

TheSHI-WO3-450-TiO2-400andSHI-WO3-0-TiO2-400composites had a similarphotocatalytic efficiency toward oxalicacid, both degradinglessthan60%ofthemodelcontaminant(58.4and59.2%, respectively).Thelowestefficiencywasshownbythecomposite madebymodifiedsol-gelsynthesis,havingayieldlessthan40%.

Thecommercialcatalyst(P25)hadthehighestefﬁciency,decom- posing96.8%oftheoxalicacid.Thisexperimentpointedoutclearly whichoftheinvolvedsynthesismethodscouldleadtothecom- positematerialsthatmaypossessphotocatalyticactivity.Although theobtainedcompositesdeterioratedtheactivityofP25,theper- formedresearchpointedoutwhichofthemethodshaspotentialto yieldphotoactivenanocomposites.Thetwomostpromisingmeth- odswerechosen,SHI-TiO2-0-WO3-450andSS−TiO2−WO3-700

(9)

Fig.10.PhotocatalyticactivitiesofvariousTiO2-WO3/MWCNTcomposites(compositesobtainedbytheﬁrstsynthesisapproachandmixedwithP25−bluecolumns;

compositesobtainedinthesecondoptimizationstagewithoutP25–redcolumns).(Forinterpretationofthereferencestocolourinthisﬁgurelegend,thereaderisreferred tothewebversionofthisarticle.)

andtheircomponent contents furtheroptimized:from1:10:15 (MWCNT:TiO₂:WO₃)to1:16:3an1:18:1.

ThemostefficientsamplewastheSS-TiO2-WO3–700^◦C1:18:1 weight ratio with a degradation yield of 67.3%. The other composites showed low photocatalytic performances, the SHI- TiO₂-0-WO₃–450^◦C 1:16:3 degraded 19.7%, while the sample SHI-TiO2-0-WO3–450^◦C1:18:1removedonly11.9%,andfinallythe sampleTiO₂-WO₃–700^◦C1:16:3wasnearlyinefficient(Fig.10).

3.4.1. Searchingfortheoriginofthephotocatalyticactivityof WO3-TiO2/MWCNT-basednanocomposites

Mostofthesamplesshowedverylowphotocatalyticactivity, inwhichacrucialrolewasassignedtothecompositecomponents ratio.Oneofthechosenweightratiosofthecompositecomponents was1:10:15(MWCNT:TiO2:WO3),whichshowedarelativelyhigh amountofWO₃andMWCNT.AsWO₃hasverylowindividualpho- tocatalyticactivityunderUVirradiation[37],therefore,theactivity deteriorationcouldbeascribedwiththehighestpossibilitytothe relativelyhighcarboncontent.

In order toenlighten the possible reasons for theobtained photocatalyticactivitytrenditwasnecessarytocarryoutaddi- tional investigations. One of the key structural features which might beresponsible for theactivity is thesize of thespecific surfacearea.We expectedthatahighvalueofthespecificsur- face area resultshigh photocatalytic activity, but theobtained resultsdidnotsupportthisstatement.TheSHI-TiO₂-0-WO₃-450 (1:16:3)samplehadarelativelyhighsurfacearea(112.0m²g⁻¹) however,thementionedsampledegradedonly19.7%oftheoxalic acidcomparedtosampleSS-TiO₂-WO₃-700(1:18:1)whichpos- sessed88.5m²g⁻¹surfaceareaandshowedasignificantlyhigher degradationefficiencyof67.3%.AlsosampleSHI-WO3-0-TiO2-400 (1:10:15)shouldbepointed out,asit wasthesample withthe highestspecificsurfacearea,butnophotocatalyticactivityatall.

Therefore,weconcludedthat,thephotocatalyticactivitywasinde- pendentfrom thespeciﬁc surfaceareainourcase. Thespeciﬁc surfaceareavaluesofallsamplescanbefoundinthesupporting information(notallthevalueswereintroducedanddiscussedin detailasmostofthemshowedlowphotocatalyticactivity).

ThefourkeysampleswereexaminedwithXPSinordertoinves- tigatethesourceofthephotocatalyticactivity,asitwasevidentthat

thesurfacequalitywasthekeyfeatureinourcase.TheXPSspectra ofthesamplesrevealedthesamespeciesineachofthecomposites.

Themaindifferencewasintheirconcentration.

TheX-rayphotoelectronspectra(Ti2p,O1s,W4f)ofsampleSS- TiO₂-WO₃-7001:18:1arepresented inFig.11.In theTi2pXPS spectrumTi⁴⁺andaverysmallamountofTi³⁺(0.96at.%−peaks at457.3eVand463.4eV)alongwithTi⁴⁺(90.04at.%− peaksat 459.1eVand464.8eV)[38]wasdetected,whileintheO1sspec- trum,theusualcomponentsinusualratiowereidentiﬁed:oxide oxygenfromTiO2(530.3eV−74.82at.%),surfaceOHgroupoxy- gen(531.3eV− 21.78at.%), and oxygenfromH₂O(532.8eV − 3.40at.%).Furthermore,intheO1sspectrumofthematerialsno low binding energy oxygenspecies (528.8eV) appeared. Based onourrecent publications [38,39] this interesting oxygentype couldbeattributedtothepresenceofoxygendefects,ortooxy- genatomswhichareneighbouringtoreducedTisites(i.e.Ti³⁺).

InthesamplesSHI-TiO₂-0-WO₃-4501:16:3,SHI-TiO₂-0-WO₃-450 1:18:1, SS-TiO₂-WO₃-700 1:16:3 the amount of Ti³⁺ increased graduallyfrom0.96at.%(SS-TiO₂-WO₃-7001:18:1)to3.21at.%(SS- TiO2-WO3-7001:16:3),whiletheoxygenformsshowedthesame ratio.ThismeansthattheTi³⁺actedasrecombinationcenters,con- sumingmostprobablytheavailableholes,inhibitingtheoxalicacid degradationbydirectholeoxidation.

The W4f XPS spectrum of sample SS-TiO₂-WO₃-700 1:18:1 revealedtwosigniﬁcantWspecies(W⁶⁺andW⁴⁺−36.0/38.2eV and 35.8/37.8eV). W⁴⁺ presenceconferred ourmaterialsa pale bluishcolor,asitwaspresentin21.2at.%fromthetotalamountof W.AsWO₃wasnotpresentintheXRDpatternsofthesesamples, themostpossibleformofthisoxideisamorphouswithdangling, reducedW⁴⁺sites,similarwiththestructurereportedbyAkurati etal.[40].Astheratioofthesespecieswereunchangedintheother samples,emphasizingthatinthecaseofthesecompositesTi³⁺plays theactivitydeﬁningrole.

4. Conclusions

Itiswellknownthatduetoitscomplexityphotocatalysiscanbe inﬂuencedbymanydifferentparameters.Ourinvestigationwith aternarycompositesystem(WO₃-TiO₂/MWCNT)describedhere demonstratedthisstatementexemplarily.Itwasprovedthatnot

(10)

Fig.11.XPSspectra(Ti2p,O1sandW4f)ofsampleSS-TiO2-WO3-7001:18:1.

onlythesynthesismethodbutalsothesequenceofdepositingthe components(TiO₂followedbyWO₃),affectsignificantlythefinal propertiesofthenanocomposites.Wehavelearntthattheproper- tiesoftheoxidedepositedinthefirstinstancehasastronginfluence onthecrystallinityofthesecondphase (e.g.differentcrystallite sizeoftheindividualoxideparticles).Atthesametime,thehigher degreeofcrystallinitydoesnotnecessarilycoincideswiththehigh photocatalyticactivity(theexampleofSHI-TiO₂-0-WO₃-450).The appearanceofTiWOxcrystalphaseathighertemperatureisalso aninterestingandsomewhatunusualphenomenonandmightbe worthconsidering,sincesample SS-TiO₂-WO₃-700alsoshowed promisingphotocatalyticactivity.However,itcannotbestatedthat thisnewphaseisexclusivelyresponsiblefortheobservedphoto- catalyticperformances(seeMSHIsamples).Thiswasproven,when theratioofthecompositecomponentswaschanged,becausein thosesamplesTiWOxdisappearedandanatasecrystallizedwitha highcrystallitesize,makingitthemaindeterminingfactorforthe

photoactivity.Also,itisimportantthatatlowWO₃ andMWCNT contentWO₃doesnotappearatallsuggestingapossiblepresence onlyasamorphousWO3.Concerningthesequenceofdepositing theoxidesitcanbeconcludedthattheexistenceofMWCNT-TiO₂- WO₃ junctionismoreadvantageous.Thephotocatalystshowing thehighestactivityshowedthelowestamountofTi³⁺,emphasizing theroleofrecombinationcenterofthisspeciesinthesecompos- ites.Basedonliteraturedata[26]itcannotbeexcludedthatthe morphologyofthecompositecomponentsisalsoimportant.

Acknowledgments

ThisworkwassupportedbyagrantfromSwitzerlandthrough Swiss Contribution(SH/7/2/20). G.K.wantsto acknowledgethe ﬁnancialsupportofHungarianAcademyofSciences,whileZs.P acknowledgesthePremiumPost-Docprojectprovidedalsobythe HungarianAcademyofSciences.Theauthorsarealsoverygrate-

(11)

[6]A.Mills,S.J.LeHunte,Photochem.Photobiol.A108(1997)1–35.

[7]M.A.Fox,M.T.Dulay,Chem.Rev.93(1993)341–357.

[8]S.Banerjee,S.K.Mohapatra,P.P.Das,M.Misra,Chem.Mater.20(2008) 6784–6791.

[9]Q.Li,L.Chen,G.J.Lu,Phys.Chem.C111(2007)11494–11499.

[10]N.A.Ramos-Delgado,M.A.Gracia-Pinilla,L.Maya-Trevi ˜no,L.Hinojosa-Reyes, J.L.Guzman-Mar,A.J.Hernández-Ramírez,Hazard.Mater.263(Pt.1)(2013) 36–44.

[11]S.Malato,P.Fernández-Ibá ˜nez,M.I.Maldonado,J.Blanco,W.Gernjak,Catal.

Today147(2009)1–59.

[12]K.Zakrzewska,ThinSolidFilms391(2001)229–238.

[13]M.Epifani,E.Comini,R.Díaz,T.Andreu,A.Genc¸,J.Arbiol,P.Siciliano,G.

Faglia,J.R.Morante,ProcediaEng.87(2014)803–806.

[14]S.Higashimoto,Y.Ushiroda,M.Azuma,Top.Catal.47(2008)148–154.

[15]C.W.Lai,S.Sreekantan,Int.J.HydrogenEnergy38(2013)2156–2166.

[16]A.K.L.Sajjad,S.Shamaila,B.Tian,F.Chen,J.Zhang,J.Hazard.Mater.177 (2010)781–791.

[17]É.Karácsonyi,L.Baia,A.Dombi,V.Danciu,K.Mogyorósi,L.C.Pop,G.Kovács,V.

Cos¸oveanu,A.Vulpoi,S.Simon,ZsPap,Catal.Today208(2013)19–27.

[18]Y.-C.Nah,A.Ghicov,D.Kim,S.Berger,P.Schmuki,J.Am.Chem.Soc.130 (2008)16154–16155.

[29]L.Tian,L.Ye,K.Deng,L.J.Zan,SolidStateChem.184(2011)1465–1471.

[30]C.Balázsi,K.Sedlácková,E.Llobet,R.Ionescu,Sens.ActuatorsB133(2008) 151–155.

[31]C.Bittencourt,A.Felten,E.H.Espinosa,R.Ionescu,E.Llobet,X.Correig,J.J.

Pireaux,Sens.ActuatorsB115(2006)33–41.

[32]P.M.Kadam,N.L.Tarwal,S.S.Mali,H.P.Deshmukh,P.S.Patil,Electrochim.Acta 58(2011)556–561.

[33]A.Vass,P.Berki,Z.Nemeth,B.Reti,K.Hernadi,Phys.StatusSolidiB250 (2013)2554–2558.

[34]G.Kovács,ZsPap,C.Cotet,V.Cos¸oveanu,L.Baia,V.Danciu,Materials8(2015) 1059–1073.

[35]G.Kovács,L.Baia,A.Vulpoi,T.Radu,É.Karácsonyi,A.Dombi,K.Hernádi,V.

Danciu,S.Simon,ZsPap,Appl.Catal.B147(2014)508–517.

[36]M.Qureshi,PrakashGupta,J.Chromatogr.A62(1971)439–448.

[37]I.Székely,G.Kovács,L.Baia,V.Danciu,ZsPap,Materials9(2016)258.

[38]Zs.Pap,É.Karácsonyi,Zs.Cegléd,A.Dombi,V.Danciu,I.C.Popescu,L.Baia,A.

Oszkó,K.Mogyorósi,Appl.Catal.B111(2012)595–604.

[39]Zs.Pap,V.Danciu,Zs.Cegléd,Á.Kukovecz,A.Oszkó,A.Dombi,K.Mogyorósi, Appl.Catal.B101(2011)461–470.

[40]K.K.Akurati,A.Vital,J.-P.Dellemann,K.Michalow,T.Graule,D.Ferri,A.

Baiker,Appl.Catal.B79(2008)53–62.