Chemical gating of epitaxial graphene through ultrathin oxide layers.

View Article Online View Journal

Nanoscale Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: R. Larciprete, P. Lacovig, F. Orlando, M. Dalmiglio, L. Omiciuolo, A. Baraldi and S. Lizzit, Nanoscale, 2015, DOI: 10.1039/C5NR02936H.

This is an Accepted Manuscript, which has been through the Royal Society of Chemistry peer review process and has been accepted for publication. Accepted Manuscripts are published online shortly after acceptance, before technical editing, formatting and proof reading. Using this free service, authors can make their results available to the community, in citable form, before we publish the edited article. We will replace this Accepted Manuscript with the edited and formatted Advance Article as soon as it is available. You can find more information about Accepted Manuscripts in the Information for Authors. Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains.

www.rsc.org/nanoscale

Page 1 of 18

Nanoscale

layers Rosanna Larciprete,1, ∗ Paolo Lacovig,2 Fabrizio Orlando,3, 4 Matteo Dalmiglio,2 Luca Omiciuolo,3 Alessandro Baraldi,2, 3, 5 and Silvano Lizzit2

Published on 15 June 2015. Downloaded by Ryerson Polytechnic University on 18/06/2015 10:24:28.

1

CNR-Institute for Complex Systems,

Via Fosso del Cavaliere 100, 00133 Roma, Italy 2

Elettra Sincrotrone Trieste S.C.p.A., AREA Science Park, S.S. 14 Km 163.5, 34149 Basovizza, Italy 3

Physics Department, University of Trieste, Via Valerio 2, 34127 Trieste, Italy

4

present address Laboratory of Radiochemistry and Environmental Chemistry, Paul Scherrer Institut, 5232 Villigen PSI, Switzerland 5

IOM-CNR Laboratorio TASC, Area Science Park, S.S.14 Km 163.5, 34149 Trieste, Italy

1


View Article Online Chemical gating of epitaxial graphene through ultrathin oxide DOI: 10.1039/C5NR02936H

Nanoscale

View Article Online

DOI: 10.1039/C5NR02936H


Abstract

Page 2 of 18

We achieve a controllable chemical gating of epitaxial graphene grown on metal

substrates by exploiting the electrostatic polarization of ultrathin SiO2 layers synthesized below it.

Intercalated oxygen diffusing through the SiO2 layer modifies

the metal-oxide work function hole doping graphene.

The graphene|oxide|metal


heterostructure behaves as a gated plane capacitor with the in-situ grown SiO2

layer acting as a homogeneous dielectric spacer, whose high capacity allows the

Fermi level of graphene to be shifted by a few hundreds meV when the oxygen coverage at the metal substrate is of the order of 0.5 monolayer. The hole doping can be finely tuned by controlling the amount of the interfacial oxygen, as well as

by adjusting the thickness of the oxide layer. After complete thermal desorption of

oxygen the intrinsic doping of SiO2 supported graphene is evaluated in the absence of contaminants and adventitious adsorbates. The demonstration that the charge state of graphene can be changed by chemically modifying the buried oxide/metal interface hints at the possibility of tuning level and sign of doping by the use of other

intercalants capable to diffuse through the ultrathin porous dielectric and reach the interface with the metal.

KEYWORDS: Graphene, ultrathin oxide, chemical gating, iridium, photoemission, NEXAFS

2

Page 3 of 18

Nanoscale

View Article Online

INTRODUCTION

DOI: 10.1039/C5NR02936H


I.

Ultrathin oxide layers on metal substrates represent a flexible class of supports for cat-

alysts and other reactive adsorbates as their characteristics can be manipulated to tailor the impact on chemical processes.1 The possibility, unavailable for bulk oxides, to adjust


the oxide thickness gives the opportunity to enable charge transfer and tune the electronic properties of adsorbates.2,3 Moreover, oxygen vacancies and defects in the oxide layer or at

the metal oxide-interface are effective means to modify the charge state and the equilibrium shape of metal nanoclusters,4 and to strongly enhance the chemical reactivity of the surface

oxide.5 Recently, an additional and remarkable tool to engineer ultrathin oxides has been envisaged in the introduction of dopant atoms at the metal oxide interface to adjust the electron flow through the dielectric.6–10 The availability of oxide-metal supports with tun-

able electronic properties gains a paramount importance in the context of graphene (Gr)

integration in technological nanodevices. In this case the controllable doping obtained by the

spontaneous charge transfer through the thin dielectric barrier, could enable new graphene-

based architectures and activate unexplored graphene functionalization mechanisms. Large area single layer graphene on ultrathin, metal supported oxides, would pave the way for the manipulation of the buried oxide/metal interface to tailor the doping level of graphene. In this context, though the possibility to synthesize ultrathin oxide layers below epitaxial

graphene has been proved,11,12 the demonstration that the Fermi level position of graphene can be tuned by chemically modifying the ultrathin oxide/metal support up to now has not been given.

Here we grow high quality ultrathin SiO2 layers with variable thickness at the graphene/Ir(111) interface by stepwise intercalation of silicon and oxygen.11 We demonstrate that by controlling the amount of oxygen which diffuses to the SiO2 /Ir(111) interface and the SiO2 layer

thickness, it is possible to modulate the doping of the graphene layer lying on top of the oxide. We have used photoemission spectroscopy as a tool to monitor the charge transfer

from and to graphene in the different steps of the process. Our results show that during

oxidation, the Ir(111) surface below the thin oxide layer gets covered by atomic oxygen

which significantly downshifts the Fermi level of graphene with respect to the Dirac point. The doping level of graphene can be regulated by controlling the thickness of the SiO2 spacer

and the amount of the interfacial oxygen as in an electrostatically gated architecture. The 3

Nanoscale

Page 4 of 18


View Article Online charge configuration, where the buried dopant is protected from the environment, could DOI: 10.1039/C5NR02936H

be exploited for selective and efficient molecular sensors. Complete oxygen elimination

allows the intrinsic doping of SiO2 supported graphene to be evaluated in the absence of contaminants and adventitious adsorbates. Our results pave the way to the use of ultrathin


oxide/metal supports to achieve chemical gating of graphene.

II.

EXPERIMENTAL DETAILS

The experiments were performed at the SuperESCA beamline of the synchrotron radiation source Elettra (Trieste, Italy). The Ir crystal was treated with the standard cleaning procedure, consisting of repeated sputtering cycles and oxygen treatments at high temperature with final flash annealing to 1370 K to completely desorb the residual oxygen from the surface. High quality graphene was grown by making cycles of ethylene dosing at 5×10−7

mbar onto the Ir(111) surface kept at 520 K followed by annealing to 1470 K. The sample quality was checked by means of low-energy electron diffraction (LEED) and photoemission

from the C1s and the Ir4f7/2 core levels. The graphene sample showed sharp LEED spots

with the characteristic moiré pattern typical of a single domain graphene on the incommen-

surate substrate. Silicon was evaporated at a rate of 0.02 ML/min by means of an electron beam evaporator, placed at around 60 mm from the sample surface that was kept at 720 K.

During Si dose, fast XPS spectra of the C1s and Si2p regions were collected simultaneously at a photon energy of 400 eV. The sample kept at 720 K, was placed almost in contact, in

front of a custom made O2 doser consisting of a Mo tube with a diameter of 6 mm. During oxidation the background O2 pressure was maintained at 5 × 10−4 mbar. With this setup we

estimate that the pressure at the sample surface was one order of magnitude higher. The thickness d of the Si oxide layer was evaluated by using the relation I = Io exp(−d/λSiO2 ), where the ratio I/I0 was calculated from the attenuation of the Ir 4f7/2 core level measured

at photon energy of 136 eV and at normal emission on the graphene-covered metal and

on the same interface with the SiO2 intralayer. The photoelectron escape depth of 70 eV

kinetic energy photoelectrons in SiO2 , λSiO2 , was taken to be 6.5 ± 0.7 ˚ A13 . With this value for the thickest SiO2 layer d resulted to be ∼ 1.6 nm, corresponding to 6.6 ± 0.7 ML of SiO2 14 , calculated by considering the Si atomic volume density in SiO2 NSiO2 = 2.3 × 1022

atoms/cm3 . Thus the amount of deposited Si turned out to be equivalent to 2.6 ± 0.3 ML (1 4

Page 5 of 18

Nanoscale


View Article Online ML = 1.4 ×1015 atoms/cm2 ). In this experiment we prepared samples with DOI: SiO10.1039/C5NR02936H 2 thickness

between 6.6 and 3.5 ML. Thermal annealing was performed by heating the sample in steps up to 1050 K. Each time the sample was rapidly brought to the desired temperature (in less

than 1 minute) and then cooled down. Thermal annealing was performed in UHV (base pressure 1×1010 mbar) and during heating the background pressure increased by less than


one order of magnitude. Annealing at temperature higher than 1050 K caused a massive decomposition of the SiO2 layer which strongly deteriorated the graphene layer as well.

High resolution Ir4f7/2 , C1s, Si2p and O1s core level spectra were measured at a photon energy of 136, 400, 150 and 650 eV, respectively, with an overall energy resolution ranging from 40 to 150 meV. For each spectrum, the binding energy was calibrated by measuring

the Fermi level position of the Ir substrate. The measurements were performed with the photon beam impinging at grazing incidence (70◦ ) while photoelectrons were collected at ˘ normal emission angle. The core level spectra were best fitted with Doniach-Sunji´ c functions

convoluted with Gaussians, and a linear background. The C K-edge NEXAFS spectra were measured in the Auger yield mode by revealing the photoelectrons at a kinetic energy of 260

eV corresponding to the C-KLL transition. Angular dependent spectra (see Supplementary Information) were taken as a function of the angle θ between the electric field E of the

photon beam (which was horizontally polarized) and the normal to the substrate plane (or between the x-ray beam and the substrate plane). The angle θ was varied between 20◦ (grazing incidence) and 90◦ (normal incidence) by rotating the samples.

III.

A.

RESULTS AND DISCUSSIONS

Growth of graphene|oxide|metal heterostructures

We start with a monolayer (ML) graphene grown on Ir(111) by thermal dissociation of ethylene molecules. The first step to grow the Gr|SiO2 |Ir(111) heterostructure is the

intercalation of a few monolayers Si below graphene, which is achieved by evaporating silicon on the graphene/metal surface held at 720 K.15,16 The bottom curves in Fig.1 show the C1s

and Ir4f7/2 spectra measured on the clean Gr/Ir(111) surface and after the intercalation of Si at 670 K. For the Gr/Ir(111) sample the C1s spectrum consists of a single narrow peak at 284.09 eV17 and the Ir4f7/2 spectrum exhibits the surface (S, 60.31 eV) and the 5

Nanoscale

Page 6 of 18


View Article Online bulk (B, 60.85 eV) components due to metal atoms in the first and in theDOI:lower crystal 10.1039/C5NR02936H

layers, respectively.18 After the intercalation of 2.6 ML of Si the C1s spectrum shows a

single component (C1 , 284.25 eV) which has a similar intensity and is shifted by +160 meV

with respect to the pristine spectrum. The stable C1s intensity indicates that the graphene

signal is not attenuated by Si atoms adsorbed on top of it and therefore demonstrates that


Si does not accumulate on graphene but diffuses completely underneath forming a layer on

the Ir surface. The C1s BE shift confirms that graphene supported on Si is n-doped.16 In the Ir4f7/2 core level the Si intercalation damps not only the S but also the B peak, while the

spectral line shape peaked at 60.70 eV indicates the appearance of new components. This demonstrate that the Si atoms do not simply chemisorb on the Ir surface but intermix with

the metal, intensely perturbing the upper crystal layers. Accordingly the Si2p spectrum

exhibits the presence of several doublet components due to the formation of different Si-Si and Si-Ir bonds.

The synthesis of the SiO2 layer is achieved by oxidizing the Si intercalated Gr/Ir(111) interface by exposing the sample held at 720 K to molecular oxygen. The Si2p, C1s, Ir4f7/2

and O1s core level spectra measured at increasing oxidation time tox are shown in the upper

part of Fig.1. The exposure to O2 transforms the Si2p spectrum into a broad peak centered at 102.58 eV,10 indicative of Si in a SiO2 environment.

During oxidation the intensity of the C1s peak (C1 ) due to graphene supported by the

intercalated Si is gradually transferred to the new component C2 , without measurable intensity loss (see inset in Fig.1). C2 , which arises from the graphene regions on top of SiO2 , is shifted with respect to C1 due to hole doping.11,19 With increasing oxidation the whole

C1s spectrum moves to lower BE and finally C2 is located 230 meV below the position

occupied in Gr/Ir(111). Such a sizable doping is due to the effect of the electronegative O-Ir(111) surface which forms underneath, as oxygen partly diffuses to the SiO2 /Ir interface and chemisorbs on the metal substrate9–11 . In fact, on the fully oxidized interface the

Ir4f7/2 spectrum, besides the bulk peak B, exhibits the IrA (60.75 eV) and IrB (61.12 eV)

components due to Ir atoms bonded to one and two oxygen atoms, respectively,19,20 whereas the small peak IrC (61.60 eV)21 can be tentatively assigned to Ir atoms bonded to three O atoms. The relative intensities of the IrA and IrB components with respect to B are consis-

tent with an O coverage of ∼ 0.5 ML,19 this estimation being only marginally influenced by the minute IrC component. Accordingly, the O1s spectrum, together with a main peak OA 6

Page 7 of 18

Nanoscale

View Article Online



DOI: 10.1039/C5NR02936H

FIG. 1. SiO2 synthesis below graphene. Si2p, C1s, Ir4f7/2 and O1s core levels spectra measured on the Gr/Ir(111) surface (black), after the intercalation of 2.6 ML of Si (violet) and during the oxidation by O intercalation (grey). In the latter case the spectra were measured after the exposure of the sample held at 720 K to O2 (p=4×10−4 mbar) for increasing time tox up to 260 minutes. The inset shows the intensity of the C1s components C1 and C2 as a function of the oxidation time. The top blue curves show the Si2p, C1s, Ir4f7/2 and O1s core levels spectra measured after the oxidation of the Gr/Ir(111) interface intercalated with 1.4 ML of Si; the blue numbers indicate the relative BE shifts measured with respect to the sample with the thicker oxide layer.

(531.82 eV) due to the SiO2 phase, shows a smaller OB (529.98 eV) component arising from the O atoms chemisorbed on Ir(111).19 The SiO2 polarization arising from the presence of the O layer chemisorbed on the Ir surface is responsible for the BE shift of ∼ -1 eV exhibited by the Si2p and O1s spectra with respect to the BE values usually measured for SiO2 /Si interfaces.13

The average thickness of the oxide SiO2 layer results to be ∼ 1.6 nm, equivalent to 6.6 ± 0.7 MLs.14 The lack in the LEED pattern of any diffraction spot ascribable to the SiO2

indicates that the layer is predominantly amorphous. For comparison the upper blue curves in Fig.1 represent the Si2p, C1s, Ir4f7/2 and O1s spectra measured after the oxidation of the

Gr/Ir(111) interface intercalated with a smaller amount of Si (1.4 ML). In this case, whereas

the Ir4f7/2 spectrum indicates an O coverage on the metal substrate similar to the previous case, the lower SiO2 thickness (∼ 3.5 ML) determines a different energy level alignment at 7

Nanoscale

Page 8 of 18


View Article Online the Gr/SiO2 /O-Ir(111) interfaces. In fact both the Si2p and O1s spectra show an extra BE DOI: 10.1039/C5NR02936H

shift of ∼ -280 meV because of the stronger electrostatic effect at the SiO2 surface which now is closer to the O-Ir(111) interface. This reflects in an extra BE shift of -130 meV for the C1s peak, giving a total shift of -360 meV with respect to Gr/Ir(111).


B.

Graphene doping

The C1s spectra shown in Fig.2a summarize the amount of doping measured when graphene is supported on Si/Ir(111) or SiO2 /O-Ir(111). For comparison it is shown also the spectrum measured on the O intercalated Gr/Ir(111) surface in the absence of the SiO2

spacer, which has a similar amount (∼ 0.5 ML) of O chemisorbed on the Ir surface. With

respect to Gr/Ir(111), Si intercalation induces an upward shift of 160 meV due to n-doping, whereas for the Gr/O-Ir(111) the C1s peak shifts by -500 meV due to p-doping.19 In the

presence of the SiO2 layer the effect of the O covered Ir surface is lower and therefore the C1s BE shift varies between -230 and -360 meV, depending on the thickness of the dielectric spacer.

FIG. 2. a) C1s core level spectra, b) NEXAFS spectra and c) valence band spectra measured on the Gr/Ir(111) surface before and after the intercalation of Si (2.6 ML) or O (0.5 ML), and after the synthesis of SiO2 (3.5 and 6.6 ML) by stepwise intercalation of Si and O. In a) the BE shifts with respect to the C1s position measured for the Gr/Ir(111) surface (black curve) are reported at the top of the figure. In c) the bottom and the top spectra show for comparison the VB spectra measured on the bare Ir(111) substrate and on the SiO2 /O-Ir(111) interface (thickness of the SiO2 layer 4.4 ML), the latter obtained by oxidizing the Si layer deposited on the bare Ir(111) surface. The NEXAFS spectra were aligned at the position of the σ ∗ resonance at 291.5 eV.

The effect of the SiO2 support on the electronic band structure of graphene is illustrated 8

Page 9 of 18

Nanoscale


Article Online by the X-ray absorption spectra at the C K-edge which probe the densityDOI: of10.1039/C5NR02936H theView empty

states above the Fermi level. Figure 2b compares the NEXAFS spectra measured at grazing incidence (θ=20◦ ) for the Gr/Ir(111) surface before and after the intercalation of Si,

for the Gr/SiO2 /O-Ir(111) structures with the 3.5 ML thick SiO2 intralayer and for the Gr/O-Ir(111) surface. The spectra are dominated by the sharp A peak due to the C1s → π ∗


transition, whereas the σ ∗ resonance at 291 eV is hardly visible, being strongly depleted in

this experimental geometry. The angular dependent spectra (see Supporting Information)

confirm that in all interfaces graphene remains flat on the substrate. For the Gr/Ir(111) surface the A′ feature appearing at 284 eV signals the interaction between graphene and Ir, resulting in a slight p-doping.17 The evidence that this feature vanishes in the Gr/Si/Ir(111) sample confirms that the intercalated Si n-dopes graphene.15 The NEXAFS spectrum mea-

sured on the Gr/SiO2 /O-Ir(111) surface maintains the spectral structure typical for pristine graphene indicating that the electronic energy bands are preserved. For this sample the

A′ feature moves away from the main peak due to the downward Fermi level shift after hole doping,22 in agreement with the behavior of the C1s core level. With respect to the Gr/Ir(111) surface the leading edge of the A′ feature for the Gr/SiO2 /O-Ir(111) sample is

shifted by 480 meV, a value of the same order of the BE shift of 360 meV measured be-

tween the corresponding C1s spectra (see Fig.2a). The comparison with the O intercalated Gr/Ir(111) confirms that without the SiO2 layer the hole doping is stronger and manifests by shifting the A′ feature by 640 meV.

The quality of the SiO2 layer below graphene can be appreciated by examining the VB spectra measured at normal emission reported in Fig.2c. The spectra taken on the Ir(111)

surface and on Gr/Ir(111) before and after the intercalation of 2.6 ML of Si are also shown. The graphene σ, π and σ ′ bands at 20.9, 7.8 and 2.7 eV, respectively, are preserved after Si intercalation and oxidation. In this latter case the π band is indiscernible because of the appearance of the typical SiO2 spectral structures, that are the O2s at ∼ 24 eV, the O2p

lone pairs in the 4-8.5 eV region and the features due to O2p-Si3p and O2p-Si3s hybridized states centered at 10 eV and 13.5 eV, respectively23 . The top curve shows the VB spectrum measured on the SiO2 /O-Ir(111) surface obtained by oxidizing a Si layer evaporated on the bare Ir(111) surface.

For the SiO2 valence band the intensity ratio between the non-bonding (O2p) and the bonding (O2p-Si3p and O2p-Si3s) features has been reported to vary with the Si-O coordi9

Nanoscale

Page 10 of 18


View Article Online nation in thin SiO2 films,23 and might be indicative of differences in the structural order, DOI: 10.1039/C5NR02936H

density and dielectric properties of the oxide material24 . The strong similarity between the spectral features measured for the SiO2 layers synthesized with and without graphene

excludes any influence on the SiO2 layer structure due to the presence of the graphene overlayer, in spite of the fact that on the bare Si/Ir(111) surface the oxidation rate is almost ten


times faster.

C.

Chemical gating of graphene

The possibility to tunably doping graphene by modifying the buried metal-oxide interface was explored by removing the O chemisorbed on the Ir surface by thermal annealing. After the desorption from the Ir surface the oxygen diffuses through the porous ultrathin oxide7,8,10 and evolves from the sample relieving the graphene doping.

The results obtained by stepwise heating up to 1050 K the Gr/SiO2 /O-Ir(111) structure

with the 3.5 ML oxide layer are summarized in Fig.3. Each time the sample was rapidly brought to the desired temperature (in less than 1 minute) and then cooled down. The

Si2p spectra measured while annealing the sample are reported in the 2D plot of Fig.3a. Above 800 K the Si2p peak starts shifting to higher BEs. Such shift involves all SiO2

spectral features, being equivalent for the Si2p and O1s spectra taken on the sample at

increasing temperature10 and shown in Figs.3b and 3e. In parallel, the component OB in

the O1s spectrum due to O atoms chemisorbed on the Ir(111) surface disappears, indicating the thermal desorption of the interfacial oxygen. The loss of oxygen manifests in the Ir4f7/2

spectra shown in Fig.3c where the IrA , IrB and IrC components arising from Ir atoms bonded

to O progressively decrease while the surface peak S emerges in the spectrum. In the sample

heated to 1050 K the high intensity of the S component, comparable with that of the pristine Gr/Ir(111) surface, demonstrates a negligible interaction between the metal surface and the

SiO2 layer. The oxygen atoms desorbing from the Ir surface, while passing through the graphene layer determine the removal of C atoms ( 18% after the annealing at 1050 K) which likely evolve as CO and CO2 .19 In this respect we can safely rule out the formation

of a high density of graphene defects such as single- and double-vacancies or Stone-Wales defects because, according to DFT calculations performed on graphite25 and graphene26 , the

presence of this kind of defects in a concentration even lower than 0.5% would clearly show 10

Page 11 of 18

Nanoscale

View Article Online



DOI: 10.1039/C5NR02936H

FIG. 3. Thermal annealing of the Gr/SiO2 /O-Ir(111) structure with a 3.5 ML thick SiO2 layer. a) 2D plot of the Si2p spectra measured during sample heating at rate of 1K/s. The red dots indicate the position of the spectral maxima. b) Si2p, c) Ir4f7/2 , d) C1s and e) O1s high resolution spectra measured on the sample heated at increasing temperature; in c) the bottom part shows the deconvolution of selected spectra. f) Thermal dependence of the BE shifts ∆ESi2p , ∆EO1s and ∆EC1s of the Si2p, O1s and C1s with respect to the corresponding BE values measured for the sample annealed at 1050 K.

up in the C1s spectra in the form of components at BEs lower than that of the main peak.

Therefore we believe that C removal occurs at the the edges of the graphene islands, which maintain a high crystalline structure.19

After the desorption of interfacial oxygen the O1s and Si2p peaks have both shifted by + 850 meV,10 approaching the BE position typical for SiO2 /Si.13

Figure 3d shows how the polarization of the SiO2 intralayer reflects in the doping state

of graphene. During sample annealing the C1s peak shifts in concert with the Si2p and O1s spectra by 280 meV. Figure 3f compares the shifts ∆ESi2p , ∆EO1s and ∆EC1s of the Si2p,

O1s and C1s peaks with respect to the BE values measured for the sample annealed at 1050 K as a function of the temperature.

The behavior of ∆ESi2p , ∆EO1s and ∆EC1s is better elucidated by considering the relation

which for a back gated graphene links the gate voltage Vg to the Fermi level shift ∆EF around 11

Nanoscale

Page 12 of 18


√ View Article Online the conical point, that is ∆EF ∝ sign(Vg ) |Vg |.27 For the Gr|SiO2 |Ir(111) structure ∆ESi2p DOI: 10.1039/C5NR02936H and ∆EO1s measure the local potential in the dielectric28 , i.e. quantifies Vg , whereas ∆EC1s

satisfactorily traces ∆EF (see Fig.4a), as it has been demonstrated for Gr/Ir(111)29 and √ the Gr/O/Ir(111) surfaces.19 Therefore ∆EC1s is expected to vary as the - |∆EO1s | and √ |∆ESi2p | and indeed the plot in Fig.4c shows a linear trend up to 970 K demonstrating that


the heterostructure behaves as a gated plane capacitor with the in-situ grown ultrathin SiO2

layer acting as homogeneous dielectric spacer. At 1050 K the capacitance of the structure decreases. We relate this behavior to morphological and structural changes occurring at

high temperature in the SiO2 layer30 which degrade the electrical response. It is worth mentioning that, due to the high capacity of the ultrathin oxide layer, the efficiency of the

electrostatic gating of graphene in this system is much higher than in conventional devices which generally use hundreds nm thick dielectric spacers. A similar ∆EC1s of 220 meV

has been measured for graphene supported on 100 nm thick Si3 N4 layer when applying an external voltage of ∼ 22 V.28

Our results thus show that the amount of electronegative oxygen chemisorbed on the metal acts as an efficient gate electrode on the oxide supported graphene. The hole doping can be tuned by controlling the amount of this interfacial oxygen which diffuses through the porous ultrathin oxide, as well as by adjusting the thickness of the SiO2 layer. Figure 4d shows that the ∆C1s values measured in samples with different SiO2 thickness31 exhibit a √ nearly linear dependence on 1/ d as expected in the planar capacitor model32 . After the

complete removal of the interfacial oxygen the C1s peak is found at 284.01 eV, which gives a shift of -80 meV with respect to the Gr/Ir(111) surface. Considering the intrinsic p-doping of -70 meV19 of the Ir(111) supported graphene, the resulting ∆EF can be estimated ∼-150 meV. The diagram in Fig.4b summarizes the doping of graphene supported on the 3.5 ML SiO2 layer in the different configurations.

The Fermi level shift of -150 meV, corresponding to a carrier density of ∼ 1.6 × 1012

cm−2 , has to be discussed in the context of the intrinsic doping of graphene supported on

SiO2 .33 According to the calculations, the interaction between SiO2 and graphene depends on the structure, polarity and termination of the oxide surface,34–36 but is weak both for the Si- and for the O-polar SiO2 surfaces.34,35 On the contrary, significant doping is usu-

ally observed in exfoliated SiO2 -supported samples,37 with the amount and sometimes even the sign of the carrier transfer38 changing after thermal annealing. Therefore the role of 12

Page 13 of 18

Nanoscale

View Article Online



DOI: 10.1039/C5NR02936H

FIG. 4. a) Schematic drawing illustrating the effect of the oxygen chemisorbed on the Ir surface on the interface potential. W and W ′ indicate the graphene work functions in the two structures. b) Summary of the doping level in graphene in the different configurations. The √ values of the Fermi level shift√ refers to the structure with a 3.5 ML SiO2 layer. c) ∆EC1s vs − |∆ESi2p | (red circles) and vs − |∆EO1s | (blue squares); the solid curves are the linear fits of the data in the low and high temperature regions. The dashed line extrapolates the behavior observed at low temperature to the complete removal of oxygen and the right scale evaluates the corresponding C1s BE shift ′ ∆EC1s . d) ∆C1s measured for Gr|SiO2 |O-Ir(111) samples with different SiO2 layer thickness d vs −1/2 d .

external factors is believed to be crucial. In our case since the presence of contaminants and adsorbates on the SiO2 surface and/or on graphene can be ruled out, we could be led

to the conclusion that the amount of p-doping we reveal results from the intrinsic charge transfer in the intimate graphene-SiO2 contact. However, since graphene acts as a donor in the presence of SiO2 oxygen defects and dangling bonds,34,36,39 we cannot exclude that

local structural inhomogeneities which emerge in the SiO2 layer above 970 K contribute to

the hole doping we measure. We could speculate that the effect of the oxide defects on the 13

Nanoscale

Page 14 of 18


View Article Online doping could be evaluated by extrapolating in Fig.4c (dashed line) the gating DOI: trend observed 10.1039/C5NR02936H

at low temperature for the intact oxide up to the complete desorption of interfacial oxygen ′ corresponding to ∆ESi2p = ∆EO1s = 0. By doing this, the effective ∆EC1s consequent to

the oxygen removal is estimated to be ∼ 440 meV. With this assumption the Fermi level of the SiO2 supported graphene would stay very close to the charge neutrality point, in


agreement with the recent calculations reporting that the interaction between graphene and SiO2 occurs without charge transfer to or from graphene.35

IV.

CONCLUSION

The strategy demonstrated here to gate graphene by chemically modifying the oxide/metal interface buried in the substrate allows the charge state of graphene to be defined

without the need to include doping species in the honeycomb lattice or to bind on top or

below it adsorbates which might also degrade the electrical conductivity. Being the active interface protected from the reactions taking place on the surface, the stability of the charge configuration is preserved. By using this approach different doping states can be achieved by intercalating below graphene chemical species able to diffuse through the ultrathin porous oxide layer and reach the interface with the metal. The potential of this approach relies also

on the fact that the synthesis of dielectric layers other than SiO2 , such as high-k oxides,

nitrides or oxynitrides is in principle affordable. Moreover the evidence that good quality graphene layers can be grown on Ir thin films40,41 discloses the possibility to extend this approach beyond single crystal substrates, hinting at Gr|oxide|metal heterostructures built using low-cost thin metallic layers.

V.

ACKNOWLEDGEMENTS

This work was supported by MIUR through the program ”Progetto Premiale 2012” -

Project ABNANOTECH and through the program PRIN 2010-2011 for the project entitled ‘GRAF. Frontiers in Graphene Research: understanding and controlling Advanced Functionalities (N.20105ZZT SE0 01)’. We acknowledge financial support from ”Università degli 14

Page 15 of 18

Nanoscale

∗

[email protected]

1

L. Giordano and G. Pacchioni, “Oxide films at the nanoscale: New structures, new functions, and new materials,” Acc. Chem. Res. 44, 1244–1252 (2011).



View Article Online Studi di Trieste-Finanziamento di Ateneo per progetti di ricerca scientifica- FRA 2014”. DOI: 10.1039/C5NR02936H

2

L. Giordano, G. Pacchioni, J. Goniakowski, N. Nilius, E. D. L. Rienks, and H.-J. Freund, “Charging of metal adatoms on ultrathin oxide films: Au and Pd on FeO/Pt(111),” Phys. Rev. Lett. 101, 026102 (2008).

3

M. Sterrer, T. Risse, U. Martinez Pozzoni, L. Giordano, M. Heyde, H.-P. Rust, G. Pacchioni, and H.-J. Freund, “Control of the charge state of metal atoms on thin MgO films,” Phys. Rev. Lett. 98, 096107 (2007).

4

J. Jung, H.-J. Shin, Y. Kim, and M. Kawai, “Activation of ultrathin oxide films for chemical reaction by interface defects,” J. Am. Chem. Soc. 133, 6142–6145 (2011).

5

B. Yoon, H. Häkkinen, U. Landman, A. S. Wörz, J.-M. Antonietti, S. Abbet, K. Judai, and U. Heiz, “Charging effects on bonding and catalyzed oxidation of co on au8 clusters on mgo,” Science 403-407, 16948 (2005).

6

U. Martinez, L. Giordano, and G. Pacchioni, “Tuning the work function of ultrathin oxide films on metals by adsorption of alkali atoms,” J. Phys. Chem. 128, 164707 (2008).

7

U. Martinez, J.-F. Jerratsch, N. Nilius, L. Giordano, and G. Pacchioni, “Tailoring the interaction strength between gold particles and silica thin films via work function control,” Phys. Rev. Lett. 103, 056801 (2009).

8

T. Jaouen, S. Tricot, G. Delhaye, B. Lépine, D. Sébilleau, G. Jézéquel, and P. Schieffer, “Layerresolved study of Mg atom incorporation at the MgO/Ag(001) buried interface,” Phys. Rev. Lett. 111, 027601 (2013).

9

D. Löffler, J. J. Uhlrich, M. Baron, B. Yang, X. Yu, L. Lichtenstein, L. Heinke, C. B¨ uchner,

M. Heyde, S. Shaikhutdinov, H.-J. Freund, R. Wlodarczyk, M. Sierka, and J. Sauer, “Growth and structure of crystalline silica sheet on Ru(0001),” Phys. Rev. Lett. 105, 146104 (2010). 10

R. Wlodarczyk, M. Sierka, J. Sauer, D. Löffler, J. J. Uhlrich, X. Yu, B. Yang, I. M. N. Groot,

S. Shaikhutdinov, and H.-J. Freund, “Tuning the electronic structure of ultrathin crystalline silica films on Ru(0001),” Phys. Rev. B 85, 085403 (2012).

15

Nanoscale

View Article Online S. Lizzit, R. Larciprete, P. Lacovig, M. Dalmiglio, F. Orlando, A Baraldi, L.DOI: Gammelgaard, 10.1039/C5NR02936H


11

Page 16 of 18

L. Barreto, M. Bianchi, E. Perkins, and P. Hofmann, “Transfer-free electrical insulation of epitaxial graphene from its metal substrate,” Nano Lett. 1, 4503–4507 (2013). 12

L. Omiciuolo, E. R. Hernández, E. Miniussi, F. Orlando, P. Lacovig, S. Lizzit, T. Onur, A. Lo-

catelli, R. Larciprete, M. Bianchi, S. Ulstrup, P. Hofmann, D.Alfè, and A. Baraldi, “Bottom -up


approach for the low-cost synthesis of graphene alumina nanosheet interfaces using bimetallic alloys,” Nat. Commun. 5, 5062 (2014). 13

F. J. Himpsel, F. R. McFeely, A. Taleb-Ibrahimi, J. A. Yarmoff, and G. Hollinger, “Microscopic structure of the SiO2 /Si interface,” Phys. Rev. B 38, 6084–6096 (1988).

14

M. P. Seah and et al., “Critical review of the current status of thickness measurements for

ultrathin SiO2 on Si part V: Results of a CCQM pilot study,” Surf. Interf. Analysis 36, 1269– 1303 (2004). 15

L. Meng, R. Wu, H. Zhou, G. Li, Y. Zhang, L. Li, Y. Wang, and H.-J. Gao, “Silicon intercalation at the interface of graphene and Ir(111),” Appl. Phys Lett. 100, 083101 (2012).

16

J. Mao, L. Huang, Y. Pan, M. Gao, J. He, H. Zhou, H. Guo, Y. Tian, Q. Zou, L. Zhang,

H. Zhangand Y. Wang, S. Du, X. Zhou, H. Castro Neto, and H.-J. Gao, “Silicon layer intercalation of centimeter-scale, epitaxially grown monolayer graphene on Ru(0001),” Appl. Phys. Lett. 100, 093101 (2012). 17

A. B. Preobrajenski, M. L. Ng, A. S. Vinogradov, and N. M˚ artensson, “Controlling graphene corrugation on lattice-mismatched substrates,” Phys. Rev. B 78, 073401 (2008).

18

P. Lacovig, M. Pozzo, D. Alfè, P. Vilmercati, A. Baraldi, and S. Lizzit, “Growth of domeshaped carbon nanoislands on Ir(111): The intermediate between carbidic clusters and quasifree-standing graphene,” Phys. Rev. Lett. 103, 166101 (2009).

19

R. Larciprete, S. Ulstrup, P. Lacovig, M. Dalmiglio, M. Bianchi, F. Mazzola, L. Hornekær,

F. Orlando, A. Baraldi, P. Hofmann, and S. Lizzit, “Oxygen switching of the epitaxial graphenemetal interaction,” ACS Nano 6, 9511–9518 (2012). 20

M. Bianchi, D. Cassese, A. Cavallin, R. Comin, F. Orlando, L. Postregna, E. Golfetto, S. Lizzit, and A. Baraldi, “Surface core level shifts of clean and oxygen covered Ir(111),” New J. Phys. 11, 063002 (2009).

21

W.-H. Chung, D.-S. Tsai, L.-J. Fan, Y.-W. Yang, and Y.-S. Huang, “Surface oxides of Ir(111) prepared by gas-phase oxygen atoms,” Surf. Sci. 606, 1965–1971 (2012).

16

Page 17 of 18

Nanoscale

View Article Online N. A. Vinogradov, K. A. Simonov, A. V. Generalov, A. S. Vinogradov, D. V. Vyalikh, C. LaubDOI: 10.1039/C5NR02936H


22

schat, N. M˚ artensson, and A. B. Preobrajenski, “Controllable p-doping of graphene on Ir(111) by chlorination with FeCl3 ,” J. Phys.: Condens. Matter 24, 314202 (2012). 23

G. Hollinger, S. J. Sferco, and M. Lannoo, “Deviation from the classical 4:2 coordination in very thin SiO2 films grown on silicon,” Phys. Rev. B 37, 7149–7152 (1988).


24

H. M. Tsai, S. C. Ray, C. W. Pao, J. W. Chiou, C. L. Huang, C. H. Du, W. F. Pong, M.-H. Tsai, A. Fukano, and H. Oyanagi, “Enhancement of Si-O hybridization in low-temperature grown ultraviolet photo-oxidized SiO2 film observed by x-ray absorption and photoemission spectroscopy,” J. Appl. Phys. 103, 013704 (2008).

25

A. Barinov, O. B. Malcioglu, S. Fabris, T. Sun, L. Gregoratti, M. Dalmiglio, and M. Kiskinova, “Initial stages of oxidation on graphitic surfaces: Photoemission study and density functional theory calculations,” J. Phys. Chem. C 113, 9009–9013 (2009).

26

T. Susi, M. Kaukonen, P. Havu, M. P. Ljungberg, P. Ayala, and E. I. Kauppinen, “Core level binding energies of functionalized and defective graphene,” Beilstein J. Nanotechnol 5, 121–132 (2014).

27

Y.-J. Yu, Y. Zhao, S. Ryu, L. E. Brus, K. S. Kim, and P. Kim, “Tuning the graphene work function by electric field effect,” Nano Lett. 9, 3430–3434 (2009).

28

M. Copuroglu, P. Aydogan, E. O. Polat, C. Kocabas, and S. S¨ uzer, “Gate-tunable photoemission from graphene transistors,” Nano Lett. 101, 2837–2842 (2014).

29

A. Dahal, R. Addou, H. Coy-Diaz, J. Lallo, and M. Batzill, “Charge doping of gaphene in metal/graphene/dielectric sandwich structures evaluated by C-1s core level photoemission spectroscopy,” APL Mater. 12, 042107 (2012).

30

K. Xue, J. B. Xu, and H. P. Ho, “Nanoscale in situ investigation of ultrathin silicon oxide thermal decomposition by high temperature scanning tunneling microscopy,” Nanotechnology 18, 485709 (2007).

31

Note1, The equivalent oxygen coverage of ∼ 0.5 ML in each case was evaluated from the ratio between the IrA and the IrB components of the Ir4f7/2 spectrum.

32

M. Bokdam, P. A. Khomyakov, G. Brock, Z. Zhong, and P. J. Kelly, “Electrostatic doping of graphene through ultrathin hexagonal boron nitride films,” Nano Lett. 11, 4631–4635 (2011).

33

K. Xu, C. Zeng, Q. Zhang, R. Yan, P. Ye, K. Wang, A. C. Seabaug, H. G. Xing, J. S. Suehle, C. A. Richter, D. J. Gundlach, and N. V. Nguyen, “Direct measurement of Dirac point energy

17

Nanoscale

34

View Article Online

DOI: 10.1039/C5NR02936H

X. F. Fan, W. T. Zheng, V. Chihaia, Z. X. Shen, and J.-L. Kuo, “Interaction between graphene and the surface of SiO2 ,” J. Phys.: Condens. Matter 24, 305004 (2012).

35

N. T. Cuong, M. Otani, and S. Okada, “Semiconducting electronic property of graphene adsorbed on (0001) surfaces of SiO2 ,” Phys. Rev. Lett. 106, 106801 (2011).


36


at the graphene/oxide interface,” Nano Lett. 13, 131–136 (2013).

Page 18 of 18

A. N. Rudenko, F. J. Keil, M. I. Katsnelson, and A. I. Lichtenstein, “Interfacial interactions between local defects in amorphous SiO2 and supported graphene,” Phys. Rev. B 84, 085438 (2011).

37

S. Ryu, L. Liu, S. Berciaud, Y.-J. Yu, H. Liu, P. Kim, G W. Flynn, and L. E. Brus, “Atmospheric oxygen binding and hole doping in deformed graphene on a SiO2 substrate,” Nano Lett. 10, 4944–4951 (2010).

38

H. E. Romero, N. Shen, P. Joshi, H. R. Gutierrez, S. A. Tadigadapa, J. O. Sofo, and P. C. Eklund, “n-Type behavior of graphene supported on Si/SiO2 substrates,” ACS Nano 2, 3430– 3434 (2008).

39

S. Wu, R. Yang, M. Cheng, W. Yang, G. Xie, P. Chen, D. Shi, and G. Zhang, “Defect-enhanced coupling between graphene and SiO2 substrate,” Appl. Phys. Lett. 91, 4944–4951 (2010).

40

C. Vo-Van, A. Kimouche, A. Reserbat-Plantey, O. Fruchart, P. Bayle-Guillemaud, N. Bendiab,

and J. Coraux, “Epitaxial graphene prepared by chemical vapor deposition on single crystal thin iridium films on sapphire,” Appl. Phys. Lett. 98, 181903 (2011). 41

C. Struzzi, N.I. Verbitskiy, A.V. Fedorovd, A. Nefedov, O. Frank, M. Kalbach, G. Di Santo,

M. Panighel, A. Goldoni, J. Gärtner, W. Weber, M. Weinl, M. Schreck, Ch. W¨ll, H. Sachdev, A. Gr¨ uneis, and L. Petaccia, “High-quality graphene on single crystal Ir(111) films on Si(111) wafers: Synthesis and multi-spectroscopic characterization,” Carbon 81, 167–173 (2015).

18

Tuneable pressure effects in graphene oxide layers.

Photoluminescence investigation about zinc oxide with graphene oxide & reduced graphene oxide buffer layers.

Growth of Epitaxial Oxide Thin Films on Graphene.

Ordered and ultrathin reduced graphene oxide LB films as hole injection layers for organic light-emitting diode.

oxide hexagonal ring-graphene hybrids through chemical etching of metal hydroxide platelets by graphene oxide: energy storage applications.

Chemical solution route to self-assembled epitaxial oxide nanostructures.

Effect of structure on the tribology of ultrathin graphene and graphene oxide films.

Structural and morphological modifications of thermally reduced cerium oxide ultrathin epitaxial films on Pt(111).

Structure, Morphology and Reducibility of Epitaxial Cerium Oxide Ultrathin Films and Nanostructures.

Synthesis of ultrathin polymer insulating layers by initiated chemical vapour deposition for low-power soft electronics.

reduced graphene oxide in human lung cells mediated through oxidative stress.

Atomistic understandings of reduced graphene oxide as an ultrathin-film nanoporous membrane for separations.

Perfluorodecyltrichlorosilane-based seed-layer for improved chemical vapour deposition of ultrathin hafnium dioxide films on graphene.

Ultrathin dendrimer-graphene oxide composite film for stable cycling lithium-sulfur batteries.

Single-molecule junctions with epitaxial graphene nanoelectrodes.

Exceptional ballistic transport in epitaxial graphene nanoribbons.

Microscopically-tuned band structure of epitaxial graphene through interface and stacking variations using Si substrate microfabrication.

Magnetoplasmons in quasineutral epitaxial graphene nanoribbons.

Highly efficient and ultra-broadband graphene oxide ultrathin lenses with three-dimensional subwavelength focusing.

Photoinduced charge separation in ordered self-assemblies of perylenediimide-graphene oxide hybrid layers.

Molecular Dynamics Simulations Reveal that Water Diffusion between Graphene Oxide Layers is Slow.

Precise and ultrafast molecular sieving through graphene oxide membranes.

Electro- and magneto-modulated ion transport through graphene oxide membranes.

Ultrathin Gas Permeable Oxide Membranes for Chemical Sensing: Nanoporous Ta₂O₅ Test Study.