Organic adsorbates have higher affinities to fluorographene than to graphene

ContentslistsavailableatScienceDirect

Applied

Materials

Today

jo u r n al h om ep a g e :w w w . e l s e v i e r . c o m / l o c a t e / a p m t

Organic

adsorbates

have

higher

affinities

to

fluorographene

than

to

graphene

Eva

Otyepková,

Petr

Lazar,

Klára ˇ

Cépe,

Ondˇrej

Tomanec,

Michal

Otyepka

∗

RegionalCentreofAdvancedTechnologiesandMaterials,DepartmentofPhysicalChemistry,FacultyofScience,Palack´yUniversityOlomouc,tˇr.17.Listopadu12,77146Olomouc, CzechRepublic

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received1September2016 Receivedinrevisedform 24September2016 Accepted24September2016 Keywords: Enthalpy Entropy IGC Graphene Fluorographene

a

b

s

t

r

a

c

t

Thelargesurfacesoftwo-dimensionalcarbon-basedmaterials,suchasgrapheneandfluorographene, areexposedtoanalytes,impuritiesandotherguestmolecules,soanunderstandingofthestrengthand natureofthemolecule–surfaceinteractionisessentialfortheirpracticalapplications.Usinginversegas chromatography,wedeterminedtheisostericadsorptionenthalpiesandentropiesofsixvolatileorganic molecules(benzene,toluene,cyclohexane,n-hexane,1,4-dioxane,andnitromethane)tographeneand graphitefluoridepowders.Theadsorptionentropiesofthemoleculesrangedfrom−17to−34cal/molK andthemaximumadsorption-inducedentropylossoccurredfornitromethaneatthehigh-energysites. Theenthalpiesofbulkieradsorbateswerealmostcoverage-independentonbothsurfacesandranged from−11to−14kcal/mol.Despitethefactthatfluorographenehaslowersurfaceenergythangraphene andgraphenerepresentsanidealsurfaceforthe␲–␲stacking,theadsorbateshadloweradsorption enthalpiestofluorographenethantographeneby∼9%.Thesefindingsimplythatbulkierairboneorganic contaminantsreadilyadsorbtotheinvestigatedsurfacesandcanmodifythemeasuredsurfaceproperties. ©2016TheAuthors.PublishedbyElsevierLtd.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).

1. Introduction

Graphene’scapacitytoadsorbsmallmolecules[1]isoneofits mostimportantpropertiesandiswidelyexploitedinitspractical applications[2,3].Becausesmallmoleculeadsorptionchangesthe electronicpropertiesofgraphene[4],devicesbasedon graphene-typematerialscouldbesensitivesensorsofguestmolecules[5–7]. Theadsorptionoforganicmoleculesaffectsalsothesurface prop-ertiesof graphene [8] suchas its hydrophobicity[9,10], which wouldinturnaffectitsuseandperformanceinallfield,wherethe surfaceofgrapheneisinvolved,e.g.,coating-technologies, electro-chemistry,catalysis,andlubrication.Grapheneisalsoregardedasa usefuladsorbentinapplicationssuchaspollutantremovalanddrug delivery[11–14].Noncovalentinteractionstographenealsoaffect itscolloidalstability,dispersabilityandprocessabilityofgraphene dispersions[15,16].Thementionedapplicationsfueledresearchof graphene’ssurfacepropertiesand affinityofguestmoleculesto graphene.

In2010,fluorographenewassynthesizedviathefluorinationof grapheneandmechanicalandchemicalexfoliationofgraphite fluo-ride[17–19].Itwasinitiallyconsideredtobearatherinertmaterial, analogoustoTeflon.However,recentexperimentshaverevealed thatitisreactiveunderambientconditionsand,consequently,can

∗ Correspondingauthor.

E-mailaddress:[email protected](M.Otyepka).

beutilizedasastartingmaterialforthepreparationofgraphene derivatives[20–23].Thesurfacepropertiesoffluorographenehave notbeenexploredsoextensivelyasthoseofgraphene,although variousinterestingfeatureshavebeenobserved.Graphitefluoride hasalowersurfaceenergythangraphite[24],andhasbeenused asanindustriallubricantforoverfortyyears[25].However, fluo-rinationofgrapheneyieldsmaterialsexhibitingnontrivialfriction behavior[26].Graphitefluoride,fluorographene,andpartially fluo-rinatedgrapheneshavebeensuccessfullyusedaselectrochemical detectors[27–32]whoseperformanceisdrivenbytheir conduc-tivitiesandtheanalyte’sadsorptiontotheelectrodesurface[28]. Forabroadapplicationoffluorinatedgraphenesinpractice,itis necessarytotakeintoaccountapotentialhazardforhumancells. Pioneeringstudiesoncelllines identifiedsomerisksassociated withinteractionof fluorinatedgraphenes withcells,which was dependentonsizeanddegreeoffluorination[33,34],ontheother hand,thetoxicityoftherespectivematerialsisstillsignificantly underexplored.

Bothgrapheneandfluorographeneare2Dhydrophobic mate-rials.Intheir3Danalogues,i.e.,graphiteandgraphitefluoride,the layersareboundbytheLondondispersiveforces[24,35].Both sur-faces,however,differ,asgrapheneisflatsurfaceofahoneycomb latticeofthesp2_{carbonatomswith␲-electroncloudsaboveand} belowthecarbonlayer[36],ideallysuitedfor␲–␲stacking interac-tions[37,38].Ontheotherhand,fluorographenehasthehexagonal latticemadeofthesp3_{carbonatoms,towhichthefluorineatoms} areattached.Inaddition,thecovalentC–Fbondisratherpolar[39] http://dx.doi.org/10.1016/j.apmt.2016.09.016

E.Otyepkováetal./AppliedMaterialsToday5(2016)142–149 143

imprintingsomeelectrostaticfiledacrossthefluorographeneplane [40].Thequestionishowthedistinctpropertiesofgrapheneand fluorographeneinfluencetheiradsorptionbehavior.

Inthispaper,westudytheadsorptionpropertiesofgraphene andfluorographeneusinginversegaschromatography(IGC).IGC measures the retention of probe molecules (adsorbates) in a columnloadedwiththeadsorbent, and thesurface-energyand adsorption enthalpy and entropy can be determined from the resulting retention curves [41–46]. IGC is dominantly used for measurementofthesurfaceenergyand itsabilitytodetermine theadsorptionenthalpyandentropy isoftenbeingoverlooked. IGChascoupleofadvantagesoverotherexperimentaltechniques, e.g.,itprovidesanaveragedresponseofthematerial,andenables tomanipulate withthetargetedcoverageoftheadsorbent.The targetedcoverageoftheadsorbentallowsdeterminingthe depen-denceoftheisostericadsorptionenthalpyonthecoverage,andto studymicroscopicprocessesrelatedtotheadsorption,identifythe high-energysites,etc.

UsingtheIGCtechnique,we comparedtheadsorptionofsix volatileorganicmoleculestographeneandgraphitefluoride pow-ders. For the measurements, we used nitromethaneas a small molecularprobe andfivesimilarlysized organicmolecules,one aliphatic(n-hexane),twoalicyclic(cyclohexaneand1,4-dioxane), and two aromatic (benzene and toluene). Applying the Lang-muiradsorptionmodelwedeterminedthecorrespondingisosteric adsorptionenthalpiesandentropiesasafunctionofthesurface coverage(rangingfrom0.5%to18%ofamonolayer).WeusedIGCto determineboththeisostericadsorptionenthalpiesandentropies. Theexperimentalfindingswereinterpretedwiththehelpof den-sityfunctionaltheory(DFT)calculations,whichmayprovideuseful insightsintostrength andnatureofinteractionto2Dmaterials [8,47].

2. Experimental 2.1. Materials

Weusedagraphenepowder(A01,GrapheneSupermarket)that has been extensively characterized using various experimental techniques in our previous studies [8,48]. It consists of 3-nm (height)grapheneflakeswithahighsurfacearea(915m2_/g)[48] and a diameter of only a few micrometers. The graphite fluo-ride powder (Sigma–Aldrich) had a surface area of 236.9m2_/g (see [24]) and consisted of flaky lamellar crystals, with a ter-race height of 10–15nm. We used volatile organic molecules asadsorbents; these included: n-hexane(Merck,LiChrosolv for LC,≥98%),toluene(Sigma–Aldrich,ChromasolvforHPLC,99.9%), 1,4-dioxane(Sigma–Aldrich,ChromasolvPlusforHPLC,≥99.5%), benzene(Sigma–Aldrich,ChromasolvforHPLC,≥96%), cyclohex-ane (Sigma–Aldrich, Chromasolv Plus for HPLC, ≥99.9%), and nitromethane(Sigma–Aldrich,ChromasolvforHPLC,99.9%). 2.2. Microscopicmeasurements

Samples were characterized via transmission electron microscopy(TEM;JEOL2100equippedwithaLaB6electrongun), atan accelerating voltageof 200kV. Thecorresponding images wereobtainedbyusingaTengra(EMSIS)camera.SamplesforTEM analyseswerepreparedbydispersinggraphitefluoride/graphene nanopowder in ethanol and sonicating for 5min. One drop of theresultingsolutionwasplacedonacoppergrid,coveredwith a holey carbon film, and dried at roomtemperature. Scanning electronmicroscope(SEM;SEMHitachiSU6600equippedwitha Schottkyelectronsource)imagesofthedrieddropwereobtained atanacceleratingvoltageof5kV,insecondaryelectron(SE)mode.

Samples for SEM analyses were affixed to conductive carbon tapeand placed inanaluminum holder.The surfacetopologies ofthesesampleswereevaluatedviascanningprobemicroscopy (SPM) in semi-contact atomic force microscopy (AFM; Ntegra, NT-MDT)mode,usinganHaNCprobe.Similarly,high-resolution TEM(HR-TEM, Titan,FEI) elementalmaps ofthesamples were obtained in scanning transmission electron microscopy (STEM) mode(80keV)usingaSuper-Xenergy-dispersivedetector(EDS; Bruker).

2.3. IGCsetupanddatafitting

IGCexperimentswereperformedusingasurfaceenergy anal-ysisinstrument(SEA;SMS-iGC2000equippedwithCirrus–SEA ControlSoftware,Version1.3.3.2,SurfaceMeasurementSystems Ltd.,UK2011);theresultingchromatogramswereallprocessed byCirrusPlus (SEA DataAnalysisSoftware,Version1.2.1.2, Sur-faceMeasurementSystemsLtd.,UK,2012).Fortheexperiments, 18.8mgofgraphenepowderor23.9mgofgraphitefluoride pow-derwereloadedinto30-cm-longsilylatedglasscolumnsandwere washedbyheliumgasfor2h(10sccmheliumcarriergas,at80◦C) priortheexperiments.Thepartialpressuresweredeterminedfrom adsorbatepeakmaxima,basedoninstrumentcalibration,andthe temperatureofthecolumnwascontrolledbyaninstrumentoven withadeclaredstabilityof±0.1◦_{C.Thecoveragewascontrolled} bytheinjectiontimeoftheadsorbatevapor,whichwascalculated fromthesurfaceareaandvaportensionoftheadsorbateat40◦C. Theisostericadsorptionenthalpiesandentropieswerefittedusing thefollowingequation[46]:

K=



(1−



)p/p0 =e−(Had−TSad )/RT)

whereK,



,p,p0_{,Had,Sad,R,andTaretheequilibrium} con-stant,coverage,pressure,standardpressure,isosteric(atthegiven



) adsorption enthalpy, isosteric adsorption entropy, universal gasconstant,andtemperature,respectively.Thechromatograms wereusedforfittingonlyfornetretentiontimes(vs.methane) of>0.26min(i.e.,greaterthan3ofthemethanepeakfittedbya Gaussiandistribution).

2.4. DFTcalculations

Theprojector-augmentedwavemethod,asimplementedinthe Viennaab-initiosimulationpackage(VASP)suite,wasusedforthe calculations[49,50].Theenergycutofffortheplane-wave expan-sionwassetto400eV.Moreover,theinteractionenergyandforces werecalculatedbyapplyingtheoptimizedvanderWaalsfunctional optB86b-vdW[51],whichencompassesbothlocalandnon-local electron–electron correlationeffects, suchas Londondispersive forces[8].Thegrapheneandfluorographenesheetsweremodeled bya6×6supercell(72carbonatoms,or72carbonand72fluorine atoms)andtheforcesbetweenthesurfaceandthemoleculewere fullyrelaxed.Inaddition,theperiodicallyrepeatedsheetswere sep-aratedbyatleast16 ˚Aofvacuum,anda3×3×1k-pointgridwas usedtosampletheBrillouinzone.Thethermalcorrectionsforthe enthalpyweretakenfromourpreviouswork[8].The polarizabil-itiesof moleculeswerecalculated byapplyingtheelectricfield (LCALCEPStagin VASP)astheratioof thechangeofthedipole momenttothemagnitudeofelectricfield(dP/dE).

3. Resultsanddiscussion

3.1. Structuralfeaturesofgrapheneandfluorographitepowders Bothnanomaterialswereanalyzedbyawiderangeof micro-scopicandspectroscopictechniquesaspartofourpreviousstudies

Fig.1. SEMandTEMimagesofthe(a,c)graphenenanopowderand(b,d)graphitefluorideflakes.

[8,24,46,48].The graphenenanopowder consistsof 3-nm-thick, crumpledflakes (seeFig.1aand c), withRamancharacteristics comparable tothoseof few-layer grapheneand a distinct few-lamellarstructure[48].Thegraphitefluoridealsohasalamellar structurewithsignificantlyexposedfluorographeneterraces(see Figs.1b,d,and2),andtheindividuallamellaeconsistof homoge-neouslydistributedcarbonandfluorineatoms(Fig.2).Thelateral sizeofindividualgraphitefluorideflakesasdeterminedbySEM amountsto∼15␮m×25␮m(Fig.S1inSupportingInformation) withverticalheightsof∼7␮m.

3.2. Adsorptionenthalpiesofgraphene

Wemeasuredtheisostericadsorptionenthalpiesofsixvolatile organicmoleculesatcoveragevaluesrangingfrom0.25%to20%of monolayertobothgrapheneandfluorographene.Itshouldbenoted thattheexperimentalsetupallowedamaximumcoveragevalue of3%and7%fornitromethaneongrapheneandfluorographene, respectively.Theadsorptionenthalpiesofbothmaterials exhib-ited a weak dependence (see Fig. 3) onthe coverage over the entirerangeofcoveragesinvestigated.However,theadsorption enthalpyofnitromethane decreasedsharplyat coverages rang-ingfrom0.25%to1.00%.Weassumethatthedecreaseoccurred duetotheadsorptionofnitromethaneintohigh-energyadsorption sitesinthegraphenepowder.Thisresultconcurswithour previ-ousfinding[48]thattheinvestigatedgraphenepowdercontained 0.24±0.03%high-energysites(inthatcase,stepsandcavities[48]), as determinedfrom adsorption experiments withacetone. The enthalpy-coverageprofilesforlargermoleculesdidnotindicate thepresenceofanyhigh-energysites.Inordertocorroboratethis hypothesis,weperformedDFTcalculationofadsorptionofbenzene

onthestep-likedefect.Thestep-likedefectwasveryenergetically favorablein thecase ofacetoneand ethanolmolecules [46,48]. TheDFTcalculationunveiledthatthebindingofbenzenetostep wasonlyslightly(by20%)morefavorablethanthebindingtothe graphenesurface.Suchamoderateenergeticgainandgreatersize ofthemoleculesuppresstheroleofhigh-energysites.

The adsorption enthalpiesreached a steady state at surface coverage values of >2% (Fig. 3). At this range of coverage, the adsorptionenthalpyofgrapheneandgraphitestemsfrom adsorp-tiontothegraphenesurface[8,46,48].Therefore,inthecurrent study,weattributeadsorptionenthalpiesatcoveragesof>2%to terrace/surface adsorption of both materials. The likelihood of adsorbate–adsorbateinteractionincreaseswithincreasing cover-age,asindicatedbythetransitionofadsorptionisothermsfromthe LangmuirtoBETregimes.Hence,weaveragedtheisosteric adsorp-tionenthalpiesmeasuredat2–10%ofmonolayercoverageandused theseaveragedvalues(seeTable1)thereafter.

The adsorption enthalpy of toluene to graphene was −13.8±0.4kcal/mol,which isvery closetotheIGC-determined valueof−13.5±0.3kcal/mol[8]andfallswithintherangeofthe TPD-determinedenthalpiestohighly-orientedpyrolyticgraphite (HOPG) (−16.6±1.7 and −13.2±1.4kcal/mol, for a monolayer (ML) and the fourth ML, respectively, as calculated from the desorption activation energy EA, Had=−EA−1/2RT at 323K) [52].Themeasuredadsorptionenthalpyofbenzenetographene was −11.9±0.3kcal/mol. Recent thermal-desorption studies of the interaction of polycyclic aromatic hydrocarbon molecules withthe basal plane of graphite yield, through theanalysis of desorption kinetics, activationenergy of 0.50eV (11.5kcal/mol) [53]atsubmonolayercoverages,inexcellentagreementwithour IGCresult.

E.Otyepkováetal./AppliedMaterialsToday5(2016)142–149 145

Fig.2. AFMimagesandcorrespondingheightprofile(upperrightpanel)revealthelamellarstructureofagraphitefluorideflakewithexposedfluorographeneterraces.The fluorineatomsarehomogeneouslydistributedoverthefluorographenesurface,asshownbytheelementalmap(lowerrightpanel),obtainedviaHRTEM/EDS.

Table1

Saturatedadsorptionenthalpiesandentropiestographeneandgraphitefluoridepowders.

Compound Graphene Fluorographene

H(kcal/mol) S(cal/molK) H(kcal/mol) S(cal/molK) Tmin–Tmax

(K) 1,4-Dioxane −10.8±0.1 −26.9±0.2 −11.7±0.2 −29.9±0.4 313–363 Benzene −11.9±0.3 −28.0±1.1 −12.7±0.2 −30.1±0.7 313–363 Cyclohexane −11.4±0.3 −28.0±0.7 −12.2±0.2 −30.4±1.3 313–363 Nitromethane −6.3±0.1 −16.8±1.1 −9.2±0.2 −25.9±0.8 313–363 n-Hexane −13.5±0.2 −31.2±0.5 −14.3±0.3 −34.1±0.6 313–363 Toluene −13.8±0.4 −28.9±1.5 −13.1±0.2 −27.8±0.8 313a_–363

a_{Thetemperatureintervalforgraphenepowderwas333–363K.}

The lower valueobtained for the adsorptionenthalpy of n-hexanetographenepowderrelativetoourpreviouslydetermined [8]value(−13.5±0.2vs. −12.2±0.2kcal/mol)canbeattributed tothe differencein fitting methods;in ourprevious study[8], weusedthelinearizedClausius–Clapeyronequationfordata fit-ting.Ontheotherhand,itshowsthattheerrorbarsassessedfrom thedatafittingcouldbeunderestimatedandthereal experimen-taluncertaintyoftheadsorptionenthalpiesdeterminedbyIGCis around1kcal/mol.Nonetheless,theclosecorrespondencebetween thetheoretically predictedandmeasuredadsorption enthalpies [8,48]indicatethattheIGCtechniquecanreliablydeterminethe adsorptionenthalpiesofvolatileorganiccompoundsto carbona-ceousmaterials.

The higher adsorption enthalpy of cyclohex-ane, (−11.4±0.3kcal/mol), compared to n-hexane, (−13.5±0.2kcal/mol), can be explained in terms of the larger conformational freedom of then-hexane moleculeand slightly

higher polarizability ˛ of n-hexane (˛=11.9×10−24cm3 _taken from Ref. [54] and 12.7×10−24cm3 _{from our DFT calculation,} refer to Table S1 for calculated polarizabilities of all studied molecules) with respect to cyclohexane (˛=11.0×10−24cm3 fromRef.[54] and10.9_×10−24cm3 _{fromDFTcalculation,Table} S1).Thehigherconformationalfreedomenabledbetterfittingof n-hexanetothegraphenesurface.Thehigherpolarizabilityleads tostrongerLondondispersiveforces,whichdrivetheadsorption of non-polar organicmolecules tographene[8].The difference betweencyclohexane and n-hexanewasconfirmed byourDFT calculation,whichyieldedadsorptionenthalpiesof−11.5kcal/mol and −13.6kcal/mol,respectively.Theseresultscorroboratedthe experimentaldata,therebyverifyingthatthedifferencein adsorp-tionenthalpystemsfromthedifferenceinstructureandcharacter ofcyclohexaneandn-hexane.

Theadsorptionenthalpyof1,4-dioxane,(−10.8±0.1kcal/mol), is slightly higher than the adsorptionenthalpy of cyclohexane,

Fig.3.IGC-determinedisostericadsorptionenthalpyandentropyofmolecularprobestographenepowderforcoveragesrangingfrom0.25%to10%ofamonolayer(forsome molecules,weextendedtheplotto20%coverage,seeFig.S2oftheSupportingInformation).

indicatingthattheenthalpychangesonlymodestlywithoxygen substitutionof twomethylene groups. Theadsorptionenthalpy ofaromaticbenzene,(−11.9±0.3kcal/mol),isbyafractionlower than that of its saturated counterpart cyclohexane. However, accordingtoexperimentalmeasurements,benzeneisless polar-izablethan cyclohexane [54], sothehigher affinity of benzene tographene cannot be attributed solely to theLondon disper-siveforces.Sincebenzenebearspermanentquadrupolemoment [55](whilequadrupolemomentofcyclohexaneisnegligible)and that electrostaticscontributes by∼30% tobinding tographene [8],thehigheraffinityofbenzenecanbeexplainedbythehigher contributionof electrostatics. This documents that the adsorp-tionenthalpytographeneisdrivenbyasubtleinterplayamong individualcontributionstothenoncovalentinteractions.OurDFT calculationyieldstheadsorption enthalpyof −11.7kcal/molfor benzene,lowerthanthatofcyclohexane(−11.5kcal/mol)inline withtheIGCresult.Nevertheless,eventhecalculationcannot ulti-matelydistinguish,whichofthecontributionstotheadsorption promotesadsorptionofbenzeneovercyclohexane.Toluene con-tainsonemoremethylgroupthanbenzeneand,accordingly,has higherpolarizability(˛=12.3×10−24cm3_{takenfromRef.[54]and}

12.6×10−24cm3 _{fromourDFTcalculations,TableS1)andlower} adsorptionenthalpy(_−13.8±0.4kcal/mol).Thecalculated adsorp-tionenthalpyoftolueneof−14.5kcal/molcorroboratesitsstrong adsorptionongraphene.

3.3. Adsorptionenthalpiestofluorographene

In the case of graphite fluoride, we obtainedrelatively sta-bleadsorptionenthalpiesevenforthesmallprobe,nitromethane (Fig.4).Thisisindicativeofeitheralowfractionofhigh-energy sitesorasimilaritybetweenthebindingenergytosteps/cavities and thebindingenergy tothefluorographene surface.In other words, these structural features, i.e., steps and/or cavities, do not createhigh-energy sitesonthis material for themolecules we used. Graphite fluorite consists of flaky lamellar crystals with exposed terraces composed of fluorographene layers (cf. Figs. 1 and 2). It is reasonable to assume that the adsorption processoccurspredominantlyontheseexposedterraces,i.e.,on thefluorographene.Theobservedadsorptionenthalpiesto fluo-rographenecanthereforebeattributedtoadsorptiontographite fluorideterracesFig.5showsoptimizedorientationsofbenzene

E.Otyepkováetal./AppliedMaterialsToday5(2016)142–149 147

Fig.4. IGC-determinedisostericadsorptionenthalpiesandentropiesofselectedmolecularprobestographitefluoridepowderforcoveragesrangingfrom0.25%to10%ofa monolayer(forsomemoleculesweextendedtheplotupto20%coverage,seeFig.S3oftheSupportingInformation).

on both graphene and fluorographene surfaces. The calculated adsorption enthalpy of cyclohexane, n-hexane and benzene to single-layerfluorographene was−11.0kcal/mol,−12.5kcal/mol, and−10.2kcal/mol,respectively.Thecalculationidentifiestoluene as the most strongly adsorbed molecule on fluorographene (−13.2kcal/mol). Interestingly, thecalculationascribes stronger bindingontofluorographenetocyclohexanethantobenzene,in contrasttothebindingongraphene(seeabove). TheaxialC–H bondsofcyclohexanefitwellintoC–Fpocketsonfluorographene, whichdiminishestheadvantageofflatgeometryofbenzene.

Comparing adsorption on fluorographene and graphene as seen in IGC experiment, the same relative trends were observed.Forexample,n-hexanehadloweradsorptionenthalpy, (−14.3±0.3kcal/mol), than cyclohexane (−12.2±0.2kcal/mol) and 1,4-dioxane (−11.7±0.2kcal/mol). Similarly, toluene had loweradsorptionenthalpy,(−13.1±0.2kcal/mol), thanbenzene (−12.7±0.2kcal/mol).Theonly exceptionwasthattoluene had thelowestadsorptionenthalpytographene,whilen-hexanewas the most strongly bound moleculeto the fluorographene. The differencesinadsorptionenthalpyweregovernedbydifferences inmolecularpolarizabilities,i.e.,Londondispersiveforces.More

importantly, the adsorption enthalpiesto fluorographene were slightly lower than those to graphene by ∼9% (except for the adsorption enthalpy of toluene). That is a somewhat surpris-ingresult,becauseflurographenehaslowersurfaceenergythan graphene[24].Itshouldbe,however,notedthattheadsorption enthalpiesarecontrolledbyadelicatebalanceamongattractive andrepulsiveforces,whichimpliesthatthereisnotanydirect rela-tionbetweenthecohesive(orsurface)energyandtheadsorption enthalpyofanadsorbate.Furthermore,theresultsindicatethatthe ␲electronsofgraphenedonotprovideanysignificantcontribution tothebindingofsmallneutralmolecules.Itshouldbementioned thatsimilarconclusionwasproposedbyGrimme,whostudiedthe interactionbetweenstackedaromaticunitsandsaturated (hydro-genated) ringsofthesamesize.Hefoundthattheeffectof the ␲systemisrathersubtle,andinteractionsbetweenaromaticand saturatedfragmentsofthesamesizearesimilarlystrong[56]. 3.4. Adsorptionentropiestographeneandfluorographene

Using Langmuir adsorption model, we also determined the adsorptionentropiesfromIGCexperimentsbyfittingtheprimary

Fig.5. Structuresofbenzeneongraphene(left)andfluorographene(right)surfacesshowthatbenzeneringisparalleltoandlies315and312pmabovethegrapheneand fluorographenesurfaces,respectively.Carbonatomsareingray,fluorineatomsinyellowandhydrogensinpalecyan.

data(cf.Methods).Tobestofourknowledge,IGChasnotbeen usedtodetermineandcomparetheadsorptionentropiesofseveral adsorbatesyet.Measuredadsorptionentropiesof∼−30cal/molK were obtained, although a higher value (−16.8±1.1cal/molK) resulted for nitromethane on graphene. The lower adsorption entropy,atlowcoverage,ofnitromethaneongrapheneconfirms thattheentropylossis highestatthehigh-energysites, where strongbondingrestricts molecularmotion[57]. Theadsorption entropies exhibited similartrends tothe adsorptionenthalpies (themean coefficientof determinations,r2_{,of isosteric} adsorp-tionenthalpiesvs.entropieswere0.78and0.69forgrapheneand fluorographene,respectively).Furthermore,theentropieson flu-orographenewereonly slightlylower (by∼10%)than thoseon graphene(withoneexception,toluene).Theadsorptionentropies follow the adsorption enthalpies, which might indicate that strongerthebindingthehighertheconfigurationalspace restric-tion.Itshouldbenotedthatthenumberofdeterminedentropiesis somewhatlimitedtomakegeneralconclusionsandshouldbe sub-jectoffurtherresearch.Itwassuggestedthatthesurfaceentropies followtheadsorbatesgasphaseentropies[58].Inourcase,the entropylosscorrespondedto∼40%oftheidealgasphaseentropy oftheorganicmoleculesSg,whichfollowsthetrendobservedfor metaloxidesurfaces[58].

4. Conclusion

UsingtheIGCtechnique,wedeterminedtheisosteric adsorp-tionenthalpiesand entropiesofsixvolatileorganiccompounds to graphene and fluorographene as a function of the sur-face coverage. The results obtained for the smallest molecule used,i.e.,nitromethane,indicatedthatgraphenecontained high-energy sites. On the other hand, the adsorption enthalpies and entropies of bulkier molecules were relatively coverage-independent.Nitromethanehadthehighestadsorptionenthalpies andentropiesonbothgrapheneandfluorographene.Thestrongest bindingwas observed ongraphene for toluene, and on fluoro-graphenefor n-hexane.The adsorption enthalpies were driven mainly by polarizabilities of individual molecules, i.e., London

dispersive forces. The molecules on average exhibited lower adsorption enthalpies tofluorographene than to graphene and thisfindingwascorroboratedbyDFTcalculations.Theadsorption entropiestographeneandfluorographenefollowedthe adsorp-tionenthalpiesandcorrespondedto∼40%ofnegativegasphase entropies. The adsorption-induced entropy loss was highest at high-energysites.Ingeneral,theusedprobeshaveratherstrong affinitiestobothmaterialscorrespondingto∼10–15%ofa com-mon covalentbondand the affinitiesincreased withmolecular sizeandpolarizability.Thehighaffinitiesexplainthefactthat air-borne contaminantsreadilyadsorbonthe surfacesand inturn modifysurfaceproperties asobserved experimentally[9,10].In addition,theorganicprobeshaveonaveragehigheraffinityto flu-orographenethantographene,whichimpliesaneffectiveusage offluorographeneinapplications,whereadirectcontactwiththe molecularprobeisrequired,e.g.,electrochemicalsensing[27–29]. Ontheotherhand,thehigh-affinitywarnsthatmeasuredsurface propertiesoffluorographenemightbeaffectedbyadsorbed air-bornecontaminants.

Acknowledgements

Financial support from the Ministry of Education, Youth and Sports of the Czech Republic, via projects LO1305 and CZ.1.05/2.1.00/19.0377,isgratefullyacknowledged.M.O.isfunded bytheEuropeanUnion’sHorizon2020researchandinnovation programmeunderERC-CoG-2015(grantNo.683024).

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,atdoi:10.1016/j.apmt.2016.09.016.

References

[1]K.S.Novoselov,etal.,Electricfieldeffectinatomicallythincarbonfilms, Science306(5696)(2004)666–669.

[2]K.S.Novoselov,etal.,Aroadmapforgraphene,Nature490(7419)(2012) 192–200.

E.Otyepkováetal./AppliedMaterialsToday5(2016)142–149 149

[3]V.Georgakilas,etal.,Functionalizationofgraphene:covalentand non-covalentapproaches,derivativesandapplications,Chem.Rev.112(11) (2012)6156–6214.

[4]F.Schwierz,Graphenetransistors,Nat.Nanotechnol.5(7)(2010)487–496.

[5]F.Schedin,etal.,Detectionofindividualgasmoleculesadsorbedongraphene, Nat.Mater.6(9)(2007)652–655.

[6]G.S.Kulkarni,etal.,Electricalprobingandtuningofmolecularphysisorption ongraphene,NanoLett.16(1)(2016)695–700.

[7]J.D.Fowler,etal.,Practicalchemicalsensorsfromchemicallyderived graphene,ACSNano3(2)(2009)301–306.

[8]P.Lazar,etal.,Adsorptionofsmallorganicmoleculesongraphene,J.Am. Chem.Soc.135(16)(2013)6372–6377.

[9]Z.T.Li,etal.,Effectofairbornecontaminantsonthewettabilityofsupported grapheneandgraphite,Nat.Mater.12(10)(2013)925–931.

[10]D.Martinez-Martin,etal.,Atmosphericcontaminantsongraphiticsurfaces, Carbon61(2013)33–39.

[11]G.K.Ramesha,etal.,Grapheneandgrapheneoxideaseffectiveadsorbents towardanionicandcationicdyes,J.ColloidInterfaceSci.361(1)(2011) 270–277.

[12]J.M.D.Tascon,Overviewofcarbonmaterialsinrelationtoadsorption,Adsorp. Carbons(2008)15–49.

[13]J.Xu,L.Wang,Y.F.Zhu,DecontaminationofbisphenolAfromaqueous solutionbygrapheneadsorption,Langmuir28(22)(2012)8418–8425.

[14]T.H.Liu,etal.,Adsorptionofmethylenebluefromaqueoussolutionby graphene,ColloidsSurf.BBiointerfaces90(2012)197–203.

[15]D.W.Johnson,B.P.Dobson,K.S.Coleman,Amanufacturingperspectiveon graphenedispersions,Curr.Opin.ColloidInterfaceSci.20(5–6)(2015) 367–382.

[16]A.Mohamed,etal.,Graphene-philicsurfactantsfornanocompositesinlatex technology,Adv.ColloidInterfaceSci230(2016)54–69.

[17]J.T.Robinson,etal.,Propertiesoffluorinatedgraphenefilms,NanoLett.10(8) (2010)3001–3005.

[18]R.R.Nair,etal.,Fluorographene:atwo-dimensionalcounterpartofteflon, Small6(24)(2010)2877–2884.

[19]R.Zboril,etal.,Graphenefluoride:astablestoichiometricgraphenederivative anditschemicalconversiontographene,Small6(24)(2010)2885–2891.

[20]M.Dubecky,etal.,Reactivityoffluorographene:afacilewaytowardgraphene derivatives,J.Phys.Chem.Lett.6(8)(2015)1430–1434.

[21]V.Urbanova,etal.,Thiofluorographene-hydrophilicgraphenederivativewith semiconductingandgenosensingproperties,Adv.Mater.27(14)(2015) 2305–2310.

[22]K.E.Whitener,etal.,Grapheneaselectrophile:reactionsofgraphenefluoride, J.Phys.Chem.C119(19)(2015)10507–10512.

[23]P.Lazar,etal.,Dichlorocarbene-functionalizedfluorographene:synthesisand reactionmechanism,Small11(31)(2015)3790–3796.

[24]P.Lazar,etal.,Thesurfaceandstructuralpropertiesofgraphitefluoride, Carbon94(2015)804–809.

[25]R.L.Fusaro,H.E.Sliney,Graphitefluoride(Cfx)N–newsolidlubricant,ASLE

Trans.13(1)(1970)56.

[26]S.Kwon,etal.,Enhancednanoscalefrictiononfluorinatedgraphene,Nano Lett.12(12)(2012)6043–6048.

[27]X.Y.Chia,etal.,Fluorographites(CFx)(n)exhibitimprovedheterogeneous electron-transferrateswithincreasingleveloffluorination:towardsthe sensingofbiomolecules,Chem.Eur.J.20(22)(2014)6665–6671.

[28]V.Urbanova,etal.,Fluorinatedgraphenesasadvancedbiosensors–effectof fluorinecoverageonelectrontransferpropertiesandadsorptionof biomolecules,Nanoscale8(24)(2016)12134–12142.

[29]K.K.Tadi,S.Pal,T.N.Narayanan,Fluorographenebasedultrasensitive ammoniasensor,Sci.Rep.6(2016).

[30]T.V.Vineesh,etal.,Synergisticeffectsofdopantsonthespindensityof catalyticactivecentresofN-dopedfluorinatedgrapheneforoxygenreduction reaction,Appl.Mater.Today1(2)(2015)74–79.

[31]H.L.Poh,etal.,Fluorographenesviathermalexfoliationofgraphiteoxidein SF6,SF4,MoF6atmospheres,J.Mater.Chem.C2(26)(2014)5198–5207.

[32]V.Mazanek,etal.,Tuningoffluorinecontentingraphene:towardslarge-scale productionofstoichiometricfluorographene,Nanoscale7(32)(2015) 13646–13655.

[33]W.Z.Teo,etal.,Cytotoxicityoffluorographene,RSCAdv.5(129)(2015) 107158–107165.

[34]W.Z.Teo,etal.,Fluorinatednanocarbonscytotoxicity,Chem.Eur.J.21(37) (2015)13020–13026.

[35]L.Spanu,S.Sorella,G.Galli,Natureandstrengthofinterlayerbindingin graphite,Phys.Rev.Lett.103(19)(2009).

[36]A.K.Geim,K.S.Novoselov,Theriseofgraphene,Nat.Mater.6(3)(2007) 183–191.

[37]E.M.Perez,N.Martin,pi–piinteractionsincarbonnanostructures,Chem.Soc. Rev.44(18)(2015)6425–6433.

[38]V.Georgakilas,etal.,Noncovalentfunctionalizationofgrapheneand grapheneoxideforenergymaterials,biosensing,catalytic,andbiomedical applications,Chem.Rev.116(9)(2016)5464–5519.

[39]Y.Sato,etal.,Ontheso-called“semi-ionic”C–Fbondcharacterin fluorine-GIC,Carbon42(15)(2004)3243–3249.

[40]F.Karlicky,M.Otyepka,Bandgapsandopticalspectraofchlorographene, fluorographeneandgraphanefromG(0)W(0),GW(0)andGWcalculationson topofPBEandHSE06orbitals,J.Chem.TheoryComput.9(9)(2013) 4155–4164.

[41]R.Ho,J.Y.Y.Heng,Areviewofinversegaschromatographyandits developmentasatooltocharacterizeanisotropicsurfacepropertiesof pharmaceuticalsolids,KonaPowderPart.J.(30)(2013)164–180.

[42]S.Mohammadi-Jam,K.E.Waters,Inversegaschromatographyapplications:a review,Adv.ColloidInterfaceSci.212(2014)21–44.

[43]H.Balard,etal.,Studybyinversegaschromatographyofthesurface propertiesofmilledgraphites,J.Chromatogr.A1198(2008)173–180.

[44]H.Balard,Estimationofthesurfaceenergeticheterogeneityofasolidby inversegaschromatography,Langmuir13(5)(1997)1260–1269.

[45]E.Papirer,etal.,Comparisonofthesurfacepropertiesofgraphite,carbon blackandfullerenesamples,measuredbyinversegaschromatography, Carbon37(8)(1999)1265–1274.

[46]F.Karlicky,etal.,Interplaybetweenethanoladsorptiontohigh-energysites clusteringongraphenegraphitealtersthemeasuredisostericadsorption enthalpies,J.Phys.Chem.C119(35)(2015)20535–20543.

[47]M.Pykal,etal.,Modellingofgraphenefunctionalization,Phys.Chem.Chem. Phys.18(9)(2016)6351–6372.

[48]P.Lazar,etal.,Thenatureofhighsurfaceenergysitesingrapheneand graphite,Carbon73(2014)448–453.

[49]P.E.Blochl,Projectoraugmented-wavemethod,Phys.Rev.B50(24)(1994) 17953–17979.

[50]G.Kresse,D.Joubert,Fromultrasoftpseudopotentialstotheprojector augmented-wavemethod,Phys.Rev.B59(3)(1999)1758–1775.

[51]J.Klimes,D.R.Bowler,A.Michaelides,VanderWaalsdensityfunctionals appliedtosolids,Phys.Rev.B83(19)(2011).

[52]H.Ulbricht,etal.,Thermaldesorptionofgasesandsolventsfromgraphiteand carbonnanotubesurfaces,Carbon44(14)(2006)2931–2942.

[53]R.Zacharia,H.Ulbricht,T.Hertel,Interlayercohesiveenergyofgraphitefrom thermaldesorptionofpolyaromatichydrocarbons,Phys.Rev.B69(15)(2004).

[54]W.M.Haynes,CRCHandbookofChemistryandPhysics,93rded.,CRCPress, 2016.

[55]J.Vrbancich,G.L.D.Ritchie,Quadrupole-momentsofbenzene,

hexafluorobenzeneandothernon-dipolararomatic-molecules,J.Chem.Soc. Farad.Trans.II76(1980)648–659.

[56]S.Grimme,Dospecialnoncovalent␲–␲stackinginteractionsreallyexist? Angew.Chem.Int.Ed.47(18)(2008)3430–3434.

[57]C.T.Campbell,J.R.V.Sellers,Theentropiesofadsorbedmolecules,J.Am. Chem.Soc.134(43)(2012)8109–18115.

[58]C.T.Campbell,J.R.V.Sellers,Enthalpiesandentropiesofadsorptionon well-definedoxidesurfaces:experimentalmeasurements,Chem.Rev.113(6) (2013)4106–4135.