An Artificially Lattice Mismatched Graphene/Metal Interface: Graphene/Ni/Ir(111)

0.82 MB·English

by  D. Pacilé

AnArtificiallyLatticeMismatchedGraphene/MetalInterface: Graphene/Ni/Ir(111)
Daniela Pacile´1,2∗, Philipp Leicht3, Marco Papagno1,2, Polina M. Sheverdyaeva2, Paolo
Moras2, Carlo Carbone 2, Konstantin Krausert3, Lukas Zielke3, Mikhail Fonin3, Yuriy S.
Dedkov4,†, Florian Mittendorfer5, Jo¨rg Doppler5, Andreas Garhofer5, and Josef Redinger5
1 DipartimentodiFisica,Universita`dellaCalabria,87036ArcavacatadiRende(CS),Italy
2 IstitutodiStrutturadellaMateria,ConsiglioNazionaledelleRicerche,Trieste,Italy
3 FachbereichPhysik,Universita¨tKonstanz,78457Konstanz,Germany
3 4 Fritz-Haber-InstitutderMax-PlanckGesellschaft,Faradayweg4-6,14159Berlin,Germany
1
5InstituteforAppliedPhysicsandCenterforComputationalMaterialsScience,ViennaUniversityofTechnology,1040Vienna,Austria
0
2

n
(Dated:December11,2013)
a
J
7 We report the structural and electronic properties of an artificial graphene/Ni(111) system obtained by the

intercalation of a monoatomic layer of Ni in graphene/Ir(111). Upon intercalation, Ni grows epitaxially on
]
i
c Ir(111),resultinginalatticemismatchedgraphene/Nisystem.ByperformingScanningTunnelingMicroscopy
s
- (STM)measurementsandDensityFunctionalTheory(DFT)calculations,weshowthattheintercalatedNilayer
l
r
leadstoapronouncedbucklingofthegraphenefilm. Atthesametimeanenhancedinteractionismeasuredby
t
m
Angle-ResolvedPhoto-EmissionSpectroscopy(ARPES),showingacleartransitionfromanearly-undisturbed
.
t
a to a strongly-hybridized graphene π-band. A comparison of the intercalation-like graphene system with flat
m
grapheneonbulkNi(111),andmildlycorrugatedgrapheneonIr(111),allowstodisentanglethetwokeyprop-
-
d ertieswhichleadtotheobservedincreasedinteraction,namelylatticematchingandelectronicinteraction. Al-
n
o thoughthelatterdeterminesthestrengthofthehybridization,wefindanimportantinfluenceofthelocalcarbon
c
configurationresultingfromthelatticemismatch.
[

1
v
9 I. INTRODUCTION
0
2
1 Metalsupportedgraphenehasreceivedrenewedinterestasitprovidesamodel-systemforstudyinggraphenemodificationson
.
1
well-definedlargeareasamples. RecentphotoemissionandScanningTunnelingMicroscopy(STM)studieshaveshownthatthe
0
3 carriermobility,chirality,andbandgapcanbetailoredbyaperiodicperturbationpotential[1],doping[2,3],intercalation[4],
1
: andhybridizationwiththesupportingsubstrate[5]. Althoughtherearecountlessstudiesongraphenegrownontransitionmetals
v
i
X [5–7],thevastlydifferinginteractionofgraphene(G)withtransitionmetals(Me)isnotfullyunderstoodonabasiclevel[8,9].
r Asanimportantcontributiontotheinteractionstemmsfromnon-local(van-der-Waalslike)interactions, thevariabilityofaG
a
film with the supporting metal is only partially explained by the so-called d band model [10]. That Pt and Ir interact more
weaklywithGthanNiorCoisnotsurprisingaccordingtothismodel,butchangesfromonemetaltothenextoneintheperiodic
tableareexpectedtobemoregradual. Instead, forneighboringelements, likePdandPt, andRhandIr, theG-Meinteraction
∗Emailaddress:daniela.pacile@fis.unical.it
†Presentaddress:SPECSSurfaceNanoAnalysisGmbH,Berlin,Germany
2
seems to abruptly switch from strong to weak [5, 6]. According to the experimental findings, the G-Me interaction has been
partitionedintothesetwomaincategories,wherefromanelectronicstructurepointofview,astrongoraweakinteractionmeans
aperturbedoranalmostunperturbedgrapheneπ-bandattheK-pointoftheBrillouinzone. Ontheotherhand,thetermstrong
appearsinappropriateifintendedforchemisorptionbetweengrapheneandtheunderlayingmetal. Indeed,forNi(111),whichis
consideredoneofthemetalsbelongingtothestrongcategory,onlyamoderateadsorptionenergyof67meVpercarbonatomhas
beenrecentlyevaluatedonthebasisofhigh-levelmany-bodycalculations(andupto160meV/Cusingsemi-empiricalforcefield
corrections), whichstillisintherangeoftypicalphysisorptionsystems[11–13]. Therefore, itshouldbenotedthatthestrong
interactionmainlyimpliesastronghybridizationbetweenthegrapheneπstatesandthesubstrate,butisnotnecessarilyreflected
intheadsorptionenergies. ArelatedquestionisapossiblecorrelationbetweentheG-Melatticemismatchandthestrengthof
interaction. Severalmetalsbelongingtothestrong(Rh,Ru)andweak(Ir,Pt,Cu)categoriesformmoire´ structures,comprising
differentinteractionswiththeGlayer.
Amongseveraltransitionmetals,Ni(111)hasbeenmoststudiedassubstratematerialforG-Meinterface,bothbytheoryand
experiment. The close lattice match between G and Ni allows the growth of a commensurate (1×1) graphene overlayer, with
carbonatomsatatopandfcc-hollowsites,separatedfromthesubstrateby2.11A˚ and2.16A˚,respectively[14]. Angle-resolved
photoemission(ARPES)datashowapronouncedenergygapattheK-pointoftheBrillouinzonebetweenπ andπ∗ [15],asa
resultofbrokensymmetryforthetwocarbonsublatticesaccompainedbystronghybridizationbetweenNi3dandgrapheneπ
states.
HerewereportanewG-Ni(111)system,obtainedbytheintercalationofasingleepitaxiallayerofNiingraphene/Ir(111).For
thissystem,thelatticemismatchbetweengrapheneandtheNilayerisincreased. AlthoughtheepitaxialgrowthofNionIr(111)
willleadtoadditionalelectroniceffects,suchasanarrowingoftheNidband,thelocalchemicalelectronicenvironmentisstill
similarenoughtoallowforacomparisionwithG/Ni(111). Theintercalationleadstoalocallyenhancedinteraction, resulting
in a strong corrugation of the graphene layer. We investigated the artificially mismatched G/Ni by STM, Density Functional
Theory(DFT)includingvanderWaascontributions(vdW-DF)andARPES,providingawidecharacterizationofelectronicand
structuralproperties.ThecomparisonbetweenG/Ni/Ir(111)andG/Ni(111)allowstoranktheinfluenceoftwoimportantfactors,
latticemismatchandchemicalinteraction,affectingtheG-Meadsorptionmechanism.
II. METHODS
Thepresentedstudieswereperformedintwodifferentexperimentalchambersunderidenticalexperimentalconditions,allow-
ing for a reproducible sample preparation. Photoemission experiments were carried out at the VUV-Photoemission beam line
oftheElettrasynchrotronradiationfacility(Trieste,Italy),usingaScientaR4000electronenergyanalyzeratabasepressureof
5×10−11mbar. Angle-resolvedphotoemissionspectrawerecollectedatroomtemperature(RT)usingaphotonenergyof80
eV,withtotalenergyresolutionof100meVandangularresolutionof0.1◦.
STMexperimentswerecarriedinultra-highvacuum(UHV)system(basepressure5×10−11mbar)equippedwithanOmicron
variabletemperaturescanningtunnelingmicroscope. AllSTMmeasurementswereperformedintheconstant-current-modeat
RTusingelectrochemicallyetchedpolycrystallinetungstentipscleanedinUHVbyflash-annealing. Thesignofthebiasvoltage
correspondstothevoltageappliedtothesample. TunnelingcurrentandvoltagearelabeledI andU ,respectively.
T T
3
FIG.1: (coloronline)(a)TopographicSTMoverviewshowingthemorphologyofgraphenewithapartiallyintercalatedNisub-monolayer.
Niaccumulatesatstepedges(Barea)showingincreasedmoire´corrugationinSTMascomparedtoG/Ir(111)(AandCareas)(70×70nm2;
U =0.65V;I =1.21nA).CorrespondingLEEDimageintheinset. (b)AreaswithNiintercalatedunderneathgraphene(Bareas)show
T T
reducedmeanapparentheightinthelineprofilesandthehistogram. Thehistogramshowsthefrequencyofapparentheightvaluesappearing
inthemagnificationdepictedin(c)(blackcurve)andwithinareasonterraceA,BorC(yellow,orangeandbrowncurves,respectively). (c)
Magnificationofthedottedsquarein(a)(46×8.6nm2).
Experimentally, the G/Ir(111) system was prepared by the procedure described in [1]. Intercalation of Ni underneath a
graphenelayerwasperformedviaannealingofthepre-depositedfilminthetemperaturerangeof670-800K.Startingfromthe
sub-monolayerregime,theNicoveragewasestimatedonbareIr(111)bymeasuringtheintensityratioofNi-3pandIr-4fcore
levels.
Thespin-polarizedDFTcalculationswereperformedwiththeViennaAb-initioSimulationPackage(VASP)[16,17],using
PAWpotentials[18,19]andanenergy-cutoffof400eV.AsGGAexchange-correlationfunctionalstendtoseverelyunderbind
the adsorption of graphene on Ni(111) [11], the calculations were performed using van der Waals DFT (vdW-DF) with the
opt86b functional [20, 21]. In the calculations, a (10×10) graphene sheet was adsorbed on a (9×9) Ir(111) substrate with
a lattice constant of 2.735 A˚, consisting of a three layer slab and an additional intercalated epitaxial Ni layer. A Γ-centered
3×3×1 k-point mesh was used to relax the structures keeping only the two bottom-most layers fixed. The C 1s core level
shiftswerecalculatedintheinitialstateapproximation. Forthegraphicalvisualisation,theresultingtotalcorelevelspectraare
displayedasasumoverGaussianfunctionswithastandarddeviationof0.25. TheSTMsimulationswereperformedusingthe
Tersoff-Hamannapproximation[22]usingtheintegratedchargedensitybetweenE andE +0.2eV.
F F
III. RESULTSANDDISCUSSION
InSTM,theG/Ir(111)surfacedisplayslargefullygraphenecoveredterraceswithseveralhundredsofnanometerswidthand
straight steps following the direction of the graphene moire´, consisting of distinct fcc-hcp carbon configurations and virtually
indistinguishabletop-hollow(top-fccandtop-hcp)sites. G/Ir(111)wasimagedhereinthedark-atop-contrast[23], whereele-
vatedfcc-hcpregionsappearasblackdepressionsinmiddleofbrightringsinSTMtopographies. UponNiintercalationstraight
4
FIG. 2: (color online) (a) Structural model for a single layer of graphene on Ni/Ir(111). The color coding indicates the height of the
corrugation∆hinthegraphenelayer. (b)Comparisonofthecorrugationintheoptimizedstructureofgraphene/Ni/Ir(111)(upperpanel)and
graphene/Ir(111)(lowerpanel). (c)SimulatedSTMimageforthestatesbetweenE and0.2eV.(d)AtomicallyresolvedSTMtopographyof
F
graphene/Ni/Ir(111)(8×8nm2; U =50.0mV;I =35.0nA).Theoretical(e)andexperimental(f)C1scorelevelshiftsofG/Ni/Ir(111)
T T
comparedtoG/Ir(111).
terracesbecomedisruptedandirregularlyextendedbyareaswithinvertedmoire´contrast(whiteprotrusionsondarkbackground)
ascomparedtopristineG/Ir(111)steps.Figure1adepictsadetailedSTMtopographywithgraphenecoveringIr(A,C)including
anintermediateareawithinvertedcontrast(B)andtworemainingNiclusters(inwhite)ontopofgraphenewithheightinthe
nanometer range. From the morphology of the sample after intercalation it becomes clear that areas A and C display pristine
grapheneonadjacentIrsubstratelevels,whereasareaBcorrespondstographeneonaNi-intercalatedregion. Toshedmorelight
on the overall as well as site specific graphene-substrate interaction we analyzed the area depicted in Figure 1c and evaluated
lineprofilesacrosstheterracesandhistogramsshowingthedistributionofapparentheightvalues(Figure1b).
Line profile 1 crosses the fcc-hcp sites of G/Ir(111) on terrace A, which appear as dark depressions, and continues across
terraceBwherebrightprotrusionsnowoccupytheformerfcc-hcpsites. Heightprofile2crossesthebrightprotrusionsofterrace
B.Alargepeak-to-valleyheightvariationof0.6A˚ ismeasuredonterraceBascomparedtothemuchshallowerpeak-to-valley
corrugationof0.25A˚ ofG/Ir(111)onterracesAandC.InthehistograminFigure1bthefrequencyofapparentheightvaluesis
shownforequallysizedareasonterraceA,BandC,respectively,aswellasforthecompleteareainFigure1c. ForG/Ir(111)on
terraceAandCthedistributionisnarrow(0.45A˚ peakwidth)andfeaturesadistributionmaximumreflectingthetop-hollowsites
andadistinctshoulderatlowerapparentheightcorrespondingtothefcc-hcpregions. ForNiintercalatedgrapheneonterrace
Bthedistributionismuchwider(0.8A˚ peakwidth)withamaximum0.6A˚ belowthemaximumofG/Ir(111)andashoulder
extendingfarintotheG/Ir(111)region. AssumingtheintercalatedNiatomsarrangepseudomorphicallyontheIr(111)surface
withcomparableinterplanedistance,themeasureddistancebetweenequivalentpointsontheterracesAandBreflectstoalarge
extentthedifferenceinthegraphene-metaldistance(seealsothediscussionoftheDFTresults),andleadstotheintriguingresult
that graphene on intercalated Ni is in average 0.6A˚ closer to the topmost substrate layer compared to the pristine G/Ir(111).
5
Thealmostunaffectedcontinuationofthegraphenemoire´ onintercalatedNi–albeitwithincreasedcorrugation–justifiesthe
assumption of pseudomorphic arrangement of the intercalated Ni atoms. Moreover, the experimental data do not show any
indicationfortheformationofasurfacealloy.
The DFT calculations allow to investigate the structural changes in the (10 × 10) graphene sheet adsorbed on a (9 × 9)
Ni/Ir(111)substrate[1MLNipseudomorphicallyarrangedonIr(111)]. Itshouldbenotedthatthestrengthofthehybridization
iscloselyrelatedtotheminimalgraphene-substratedistances[11]. Consequentlythestructuralanalysisallowstodeconvolute
the effects of the lattice mismatch between the graphene sheet and the substrate, and the chemical properties of the interface.
Figures2(a-b)displaythegeometryofthegraphenesheetaftertherelaxation. ThemodelshowsthattheintercalationoftheNi
layerleadstoapronouncedcorrugationof∆h=1.51A˚ inthegraphenelayer, significantlylargerthanforG/Ir(111)(Figure
2b). Yet it should be noted that more than 70% of the carbon atoms in the graphene layer are adsorbed at a close distance of
about2.0-2.2A˚ fromtheNilayer. Thecomparisonwiththeinter-planedistanceofGonNi(111)(2.1A˚)thereforehintsata
similarbindingofgraphenetothesubstrate,despitethelargeexperimentalstrainofroughly9%duetopseudomorphicgrowth
oftheNilatticeinG/1MLNi/Ir(111). Therefore,theG-Nibondingseemstobemainlyaffectedbytheelectroniccontribution,
whichalsodrivesthestrongcorrugationofthegraphenelayerinthemismatchedstructure.
Intheflatregions,themagneticmomentoftheNiatomsiscompletelyquenchedbytheinteractionwiththegraphenesheet,
whiletheNiatomsunderthegraphenebubblesyieldasmallmagneticmoment(<0.4µ ). Wefindthatnomagneticmoment
B
is induced in the graphene sheet that can be due to the small net magnetic moment of the underlying Ni film, opposite to the
graphene/Ni(111)systemshowinganinducedmagneticmomentofcarbonatoms[24].
Seen from an atomistic point of view, the close adsorption configuration of the graphene layer is reached not only for the
top-fccconfigurationpreferredonNi(111),butalsofortheadsorptioninabridge-likeconfiguration. Bothconfigurationsyielda
closeadsorptiondistancereachingvaluesaslowas1.94A˚.Incontrast,theweakinteractioninthefcc-hcpsites(greenregionsin
Figure2a)leadstotheformationoflocalprotrusions,withamaximaldistanceof3.45A˚ totheNilayercommonforphysisorbed
graphene. Nevertheless, thisdistanceisstillsmallerthanthecalculatedmaximal(vdW-DF)distance3.7A˚ (∆h=0.37A˚)for
theadsorptionofGonthebareIr(111)surface(Figure2b).
AdirectcomparisonoftheobtainedSTMdata(Figure2d)andasimulatedSTMimage(Figure2c)reflectsthestructureofthe
adsorbedgraphenesheet: theelevatedfcc-hcpregionsappearbrightest,whilethelow-lyingareaswithtop-hollowconfiguration
appearasadarkbackground. InagreementwiththestructurerecentlyreportedforG/Ru(0001)[25]andG/Rh(111)[26–28],the
regionswherethegraphenesheetisadsorbedinalocalbridgeconfigurationistheareaofthesmallestdistancetothesurface.
TheseareasappearasfaintlyvisibledepressionsinSTMtopographies. Apeak-to-valleycorrugationofupto1A˚ fitswellthe
corrugationof1.3A˚ inthesimulatedimage. Ontheatomiclevel,G/Ni/Ir(111)showsringsofcarbonatomseverywherewithin
themoire´ supercellinFigure1d, howeverwiththestronglyboundareasadifferenceintheintensitybetweentheneighboring
atomsisobservedindicatingabrokensublatticesymmetry.
Previousstudiesdemonstratethatthecorrugationandhybridizationofthegraphenelayerwiththemetallicsubstrateisstrongly
reflectedintheC1sline-shape[29,30]. Figure2fshowstheC1scorelevel(CL)takenat445eVofG/Ir(111)anditsevolution
during the intercalation of about one third of monolayer and one monolayer of Ni atoms. According to the existing literature
[29,31],intheG/Ir(111)systemtheC1sbindingenergyisfoundat(284.10±0.20)eV.Aftertheintercalationof0.33MLofNi
6
FIG.3: (coloronline)(a-d)ARPESdispersionsalongΓKasafunctionoftheamountofNiatomsintercalatedunderneathG/Ir(111).(a)and
(d)showextremecasesof0MLandthickNi,respectively.(e)Carbonprojectedbandstructureofa(1x1)modelsystemofgraphene/Ni/Ir(111).
ThebandstructurewasevaluatedataG-Nidistanceof2.0A˚ (brown-darkgraydots)and3.4A˚ (green-lightgraydots).
atomsasecondcomponentathigherbindingenergyisseen. Therelativeintensitiesofthetwocomponentsisfullyinvertedfor
1ML of Ni atoms intercalated. The main peak centered at 284.90 (±0.20) eV is close to the value found for graphene grown
onbareNi(111)[32], whereasinglepeakat284.7(±0.18)eVwasmeasured, withanintrinsiclinewidthof216meV.Inour
system,theC1sline-shapeexhibitsatotalwidthofabout840meVandastrongasymmetrytowardslowerbindingenergy,likely
convolutingdifferentcomponents.ThisisconfirmedbythecalculatedcorelevelstatesofG/Ni/Ir(111):althoughthecalculations
predictonlyaminorCLshiftforthecarbonatomsintheelevatedregionsofthemoire´ pattern,thestronglyinteractingCatoms
inthelowerregionsaredominantintheconvolutionconsideringallcontributions(Figure2e)andthusdonotexhibitadouble
C1speakasobservedforG/Re(0001),G/Rh(111)orG/Ru(0001)[29,30].
Toshed morelightonthe overallinteractionof graphenewiththemismatched Nilayers, wehave performedARPESmea-
surements mainly looking at the graphene π-band. Figures 3(a-d) show ARPES maps of the electronic band dispersion of (a)
G/Ir(111); (b) G/0.33MLNi/Ir(111); (c) G/1MLNi/Ir(111); (d) G/thickNi/Ir(111), all taken along the ΓK direction. The last
measurementwascollectedaftertheintercalationofseveralmonolayersofNi. Theπ stateofG/Ir(111)(Figure3a)exhibitsa
minimumattheΓ-pointat8.30eV,andapproachestheFermilevelattheK-pointat70meV,wheretheplanarσ statereaches
11.35eVofbindingenergy[1]. Accordingtotheliterature[1,3,29,33,34], replicasbandsofbothπ andσ statesduetothe
moire´ superpotential are seen close to the K-point. After the intercalation of about 1/3 ML of Ni atoms (Figure 3b), new π
andσ statesappearathigherbindingenergytogetherwiththedstatesofNi, weaklydispersingintherange0-2eVbelowthe
Fermi level. The co-existence of double π and σ states reflects a non-homogenous surface with clean areas of G/Ir(111) and
patcheswhereNiatomsareinbetween. WhenafullmonolayerofNiatomsisintercalatedviaannealing(Figure3c),thenew
bandstructureevolvesclearly. TheπstatenowexhibitsaminimumattheΓ-pointat10.03eVandreachesamaximumatabout
2.16eVattheK-point,whereitmergeswiththedstatesofNi. Theσstateisalsoshiftedtoahigherbindingenergycompared
toG/Ir(111), withaminimumattheK-pointatabout12.48eV.WhenseveralmonolayersofNi(above5)areintercalatedvia
annealing(Figure3d), theelectronicstatesofIrarenolongerdetected, andthebandstructurereflectsthatofG/Ni(111)[15].
7
Notably,whiletheσstateisunaffectedbythenumberofNilayersattheK-point,theπstateexhibitsaclearshifttowardslower
bindingenergies(upto2.16eVfor1MLintercalatedNi)withrespecttoGgrownonmultilayers(Figure3d)orbulknickel[15],
wheretheπ bandmaximumisfoundatabout2.65eV.ThisfindingisrelatedtothedifferencesinwidthoftheNi-dstatesofa
singleintercalatedNilayercomparedtoasurfaceofbulknickel:narrowingoftheNi3dbanduponthedecreasingoftheNilayer
thickness. Takingintoaccountthatthepositionofthegraphene-derivedπbandattheΓpointisthesameforboththickandthin
(1ML)intercalatedNilayers,wecanconcludethattheenergyshiftofπ bandwithrespecttofree-standinggrapheneispurely
definedbythechargetransferbetweenNiandCatomsattheclosestdistancethroughthedonation/back-donationmechanism
[9]. Atthesametimethepresenceaswellasthewidthofthebandgapbetweenπandπ∗graphene-derivedstatesisdetermined
by thebroken symmetryfor twocarbon atoms inthe grapheneunit cellin thissystem accompanied bya stronghybridization
betweenNi3dandgrapheneπstates.
For a comparison with experiment, we also evaluated the theoretical band structure. In order to avoid the back-folding
induced by the larger supercell, we have mimicked the local interactions by calculating the electronic structure for a smaller
(1×1) model of G/Ni/Ir(111) in a top-fcc configuration at an average distance of the flat regions (2.0 A˚) and at the maximal
heightof thebubbles (3.4 A˚).The resultingband structureis shownin Figure3e. Although theeffect ofthe latticemismatch
between the graphene sheet and the substrate is lost in this smaller model, the calculations clearly show that the interactions
at the elevated regions of the bubbles (green-light gray dots in Figure 3e) are rather weak, resulting in a nearly unperturbed
graphene band structure. On the other hand, a much stronger interaction can be expected for the dominant flat regions in the
vicinityofthesurface,leadingtoalargesplittingoftheπ bandattheDiracpoint(brown-darkgraydotsinFigure3e). These
findings agree with the experimentally observed opening of a band gap in the ARPES data. Furthermore, no π band splitting
isobservedexperimentallyinoursystem,duetothemetallicnatureofgraphene[35],incontrasttotheelectronicbehaviorof
h-BN grown on selected transition metals [36], where the dielectric nature of the overlayer allows to observe double σ and π
statescorrespondingtoupperandlowerregions.
IV. CONCLUSION
Inconclusion,wehaveshownthattheadsorptionofgrapheneonepitaxiallayersallowstostudytheinfluenceofthelattice
mismatch between the graphene layer and the support, while keeping the chemical environment similar. For graphene on
Ni/Ir(111), we find that the interaction is locally strongly enhanced for specific adsorption configurations. Consequently, in
contrast to G/Ir(111), the moire´ structure of G/Ni/Ir(111) exhibits a strong corrugation, with a modulation of about 1.5 A˚ and
aminimumG-Medistanceslightlysmallerthan2A˚ .ThegraphenebandstructureprobedbyARPESshowsacleartransition
from a nearly-free standing to a strongly-hybridized character of the π band, in analogy with graphene grown on bulk nickel.
ThehybridizationbetweenNidstatesandgrapheneπstatesisdirectlyrelatedtothestronglyinteractingtop-hollowandbridge
configurationsinthelowerpartsofthemoire´mesh,atadistanceofabout2.0-2.2A˚ fromtheNilayer.Incontrast,theinteraction
issignificantlyweakerforotherregions(fcc-hcpconfigurations)ofthemoire´ mesh,whereonlyavanderWaalslikebindingis
observed. Thereforewecanidentifytheroleoftwoimportantcontributionstotheadsorption: whiletheelectronicinteraction
dominatesinthestronglyinteractingregions, thelatticemismatchbetweengrapheneandthemetalsupportisdecisiveforthe
ratiobetweenstronglyandweaklyinteractingregions. Weexpectthatthisinterplayisoneofthekeyfeaturesformismatched
8
graphene-metalinteractions.
Acknowledgements
This work has been supported by the Italian Ministry of Education, University, and Research (FIRB Futuro in Ricerca,
PLASMOGRAPH Project), by the European Science Foundation (ESF) under the EUROCORES Program EuroGRAPHENE,
by the Austrian Science Fund (FWF) project I422-N16, and by the Research Center “UltraQuantum” of the University of
Konstanz(ExcellenceInitiative). TheViennaScientificCluster(VSC)isacknowledgedforCPUtime.
[1] S.Rusponi,M.Papagno,P.Moras,S.Vlaic,M.Etzkorn,P.M.Shverdyaeva,D.Pacile´,H.Brune,andC.Carbone,Phys.Rev.Lett.105,
246803(2010).
[2] A.Nagashima,N.Tejima,andC.Oshima,Phys.Rev.B50,17487(1994).
[3] M.Papagno,S.Rusponi,P.M.Sheverdyaeva,S.Vlaic,M.Etzkorn,D.Pacile´,P.Moras,C.Carbone,andH.Brune,ACSNano6,199
(2012).
[4] A.Varykhalov,J.Sa´nchez-Barriga,A.M.Shikin,C.Biswas,E.Vescovo,A.Rybkin,D.Marchenko,andO.Rader,Phys.Rev.Lett.101,
157601(2008).
[5] M.Batzill,Surf.Sci.Rep.67,83(2012).
[6] J.WintterlinandM.-L.Bocquet,Surf.Sci.603,1841(2009).
[7] Yu.S.Dedkov,K.Horn,A.Preobrajenskij,andM.Fonin,inGrapheneNanoelectronics(Springer,Berlin2012).
[8] E.VoloshinaandYu.Dedkov,Phys.Chem.Chem.Phys.14,13502(2012).
[9] S.M.Kozlov,F.Vin˜es,andA.Go¨rling,J.Phys.Chem.C116,7360(2012).
[10] B.HammerandJ.K.Nørskov,Adv.Catal.45,71(2000).
[11] F.Mittendorfer,A.Garhofer,J.Redinger,J.Klimesˇ,J.Harl,andG.Kresse,Phys.Rev.B84,201401(R)(2011).
[12] E.N.Voloshina,A.Generalov,M.Weser,S.Bo¨ttcher,K.Horn,andYu.S.Dedkov,NewJ.Phys.13,113028(2011).
[13] L.Adamska,Y.Lin,A.J.Ross,M.Batzill,andI.I.Oleynik,Phys.Rev.B85,195443(2012).
[14] Y.Gamo,A,Nagashima,M.Wakabayashi,M.Terai,andC.Oshima,Surf.Sci.374,61(1997).
[15] Yu.S.DedkovandM.Fonin,NewJ.Phys.12,125004(2010).
[16] G.KresseandJ.Furthmu¨ller,Comput.Mater.Sci.6,15(1996).
[17] G.KresseandJ.Hafner,Phys.Rev.B47,558(1993).
[18] P.E.Blo¨chl,Phys.Rev.B50,17953(1994).
[19] G.KresseandD.Joubert,Phys.Rev.B59,1758(1999).
[20] J.Klimes˘,D.R.Bowler,andA.Michaelides,J.Phys.:Condens.Matter.22,022201(2010).
[21] J.Klimes˘,D.R.Bowler,andA.Michaelides,Phys.Rev.B83,195131(2011).
[22] J.TersoffandD.R.Hamann,Phys.Rev.B31,805(1985).
[23] A.T.N’Diaye,J.Coraux,T.N.Plasa,C.Busse,andT.Michely,NewJ.Phys.10,043033(2008).
[24] M.Weser,Y.Rehder,K.Horn,M.Sicot,M.Fonin,A.B.Preobrajenski,E.N.Voloshina,E.Goering,andYu.S.Dedkov,Appl.Phys.
Lett.96,012504(2010).
[25] M.IannuzziandJ.Hutter,Surf.Sci605,1360(2011).
9
[26] B.Wang,M.Caffio,C.Bromley,H.Fru¨chtl,andR.Schaub,ACSNano4,5773(2010).
[27] M.Sicot,P.Leicht,A.Zusan,S.Bouvron,O.Zander,M.Weser,Yu.S.Dedkov,K.Horn,andM.Fonin,ACSNano6,151(2012).
[28] E.N.Voloshina,Yu.S.Dedkov,STorbru¨gge,A.Thissen,andM.Fonin,Appl.Phys.Lett.100,241606(2012).
[29] A.B.Preobrajenski,M.L.Ng,A.S.Vinogradov,andN.Ma˙rtensson,Phys.Rev.B78,073401(2008).
[30] E.Miniussi,M.Pozzo,A.Baraldi,E.Vesselli,R.R.Zhan,G.Comelli,T.O.Mentes¸,M.A.Nin˜o,A.Locatelli,S.Lizzit,andD.Alfe`,
Phys.Rev.Lett.106,216101(2011).
[31] C.Busse,P.Lazic´,R.Djemour,J.Coraux,T.Gerber,N.Atodiresei,V.Caciuc,R.Brako,A.T.N’Diaye,S.Blu¨gel,J.Zegenhagen,and
T.Michely,Phys.Rev.Lett.107,036101(2011).
[32] A.Gru¨neis,K.Kummer,andD.V.Vyalikh,NewJ.Phys.11,073050(2009).
[33] E.Starodub,A.Bostwick,L.Moreschini,S.Nie,F.ElGabaly,K.F.McCarty,andE.Rotenberg,Phys.Rev.B83,125428(2011).
[34] M.Kralj,I.Pletikosic´,M.Petrovic´,P.Pervan,M.Milun,A.T.N’Diaye,C.Busse,T.Michely,J.Fujii,andI.Vobornik,Phys.Rev.B84,
075427(2011).
[35] T.Brugger,S.Gu¨nther,B.Wang,J.H.Dil,M.-L.Bocquet,J.Ostewalder,J.Wintterlin,andT.Greber,Phys.Rev.B79,045407(2009).
[36] T.Greber,M.Corso,andJ.Ostewalder,Surf.Sci.603,1373(2009).