SISTEMA DE BIBLIOTECAS DA UNICAMP
REPOSITÓRIO DA PRODUÇÃO CIENTIFICA E INTELECTUAL DA UNICAMP
Versão do arquivo anexado / Version of attached file:
Versão do Editor / Published Version
Mais informações no site da editora / Further information on publisher's website:
https://journals.aps.org/prb/abstract/10.1103/PhysRevB.87.081403
DOI: 10.1103/PhysRevB.87.081403
Direitos autorais / Publisher's copyright statement:
©2013 by American Physical Society. All rights reserved.
DIRETORIA DE TRATAMENTO DA INFORMAÇÃO Cidade Universitária Zeferino Vaz Barão Geraldo
CEP 13083-970 – Campinas SP Fone: (19) 3521-6493 http://www.repositorio.unicamp.br
Atomic surface structure of graphene and its buffer layer on SiC(0001): A chemical-specific
photoelectron diffraction approach
L. H. de Lima, A. de Siervo,*R. Landers, and G. A. Viana
Instituto de F´ısica Gleb Wataghin, Universidade Estadual de Campinas, Campinas 13083-859, S˜ao Paulo, Brazil A. M. B. Goncalves and R. G. Lacerda
Departamento de F´ısica, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil P. H¨aberle
Universidad T´ecnica Federico Santa Mar´ıa, Valpara´ıso, Chile
(Received 27 June 2012; published 7 February 2013)
We report a chemically specific x-ray photoelectron diffraction (XPD) investigation using synchrotron radiation of the thermally induced growth of epitaxial graphene on the 6H -SiC(0001). The XPD results show that the buffer layer on the SiC(0001) surface is formed by two domain regions rotated by 60◦with respect to each other. The experimental data supported by a comprehensive multiple scattering calculation approach indicates the existence of a long-range ripple due the (6√3× 6√3)R30◦ reconstruction, in addition to a local range buckling in the (0001) direction of the two sublattices that form the honeycomb structure of the buffer layer. This displacement supports the existence of an sp2-to-sp3 rehybridization in this layer. For the subsequent graphene layer this
displacement is absent, which can explain several differences between the electronic structures of graphene and the buffer layer.
DOI:10.1103/PhysRevB.87.081403 PACS number(s): 61.48.Gh, 61.05.js, 74.25.Jb, 68.37.Ef
Graphene, a single sheet of sp2-bonded carbons arranged in a honeycomb lattice, has attracted enormous attention due to its unique physical and electronic properties, as, for example, a relativistic dispersion relation near the Fermi level described by the massless free particle Dirac equation, room temperature quantum Hall effect, high carrier mobility, possible topological magnetic ordering, a tunable band gap, among others.1–3It is not a surprise that graphene is nowadays considered a perfect playground material to be explored in basic research from both theoretical and experimental points of view. Graphene also holds many promises for the development of industrial applications. Several of these new properties were investigated using exfoliated graphite samples,4but this way of obtaining graphene presents many difficulties for technological applications. It has been known since 1975 that the graphitization of SiC surfaces occurs when they are heated to temperatures above 1100◦C.5–7For the SiC(0001) surface in particular, the graphitization exhibits an epitaxial relationship with the substrate and is azimuthally ordered. By adjusting the growth parameters, it is possible to obtain large areas of graphene that are a single or few layers thick on the surface of SiC,8–14which could lead, in the near future, to a wafer-sized graphene sheet directly grown on an insulating SiC substrate. The initial stage of graphitization (first layer) is characterized by a (6√3× 6√3)R30◦ reconstruction, also known as the buffer layer (BL).15The BL is a layer of carbon atoms arranged in a honeycomb structure, whose lattice is mismatched in relation to the substrate to which some of the atoms are bonded covalently. Despite its topological similarity to graphene, the BL exhibits physical and electronic properties that are dramatically different from those of exfoliated graphene: For instance, it has no π bands around the K point15,16and it shows a gap opening.17 The BL also has an important influence on the subsequent layer grown, which is single-layer graphene
that has a conical dispersion of the π bands, but with a band gap at the Dirac point due to the presence of the BL substrate. The mechanism behind the gap opening and its relation to the atomic structure of the graphene/BL system remains an open issue, despite its importance.
The atomic structure of BL itself is a subject of controversy in the literature. The pioneering work of van Bommel5 suggests that the graphitelike layers are weakly bonded to the surface of SiC(0001) while other studies suggest a Si-rich interface.7Another controversy involves the 6× 6 hexagonal reconstruction of the BL observed in scanning tunneling microscopy (STM) images,6,8,10,18,19which do not agree with the (6√3× 6√3)R30◦reconstruction commonly observed in low-energy electron diffraction (LEED) patterns.6–8,10 These reports involved the indirect determination of the structure by combining STM and theoretical results. It is clear that a direct determination of the structure from experimental data is indicated.
In this Rapid Communication, we present a surface struc-ture determination of the graphene/buffer layer on SiC(0001). We have used the unique advantage of the chemical sensitivity of synchrotron-based, high-energy-resolution photoelectron diffraction to probe the local order of carbon atoms in each of the chemically different environments, determining separately the atomic structure of each contribution at a specific region, i.e., C or Si in the bulk of SiC(0001), C in BL/SiC (0001), and C in graphene/BL/SiC(0001). The approach used in this work represents a direct, element- and chemical-specific determination of the graphene/BL which has the ability to unambiguously differentiate the contributions of each layer. Based on our experimental data analysis, which includes a comparison with theory, the existence of the long-range ripple structure due the (6√3× 6√3)R30◦ reconstruction is confirmed. In addition, there are strong indications that the
L. H. DE LIMA et al. PHYSICAL REVIEW B 87, 081403(R) (2013)
sublattices that form the honeycomb structure are decoupled in the BL, but not in the graphene that is essentially flat. This breaking of the sixfold symmetry would explain, from a structural point of view, the peculiar differences observed in the electronic properties of the buffer layer compared to graphene.
The x-ray photoelectron diffraction (XPD) experiments were carried out at the SGM (in the case of the buffer layer) and PGM (in the case of graphene) beamlines of the Brazilian Synchrotron Light Laboratory (LNLS) using 450 eV photons. The base pressure in the ultrahigh-vacuum (UHV) chamber was below 1× 10−10mbar during the experiments. For this study, we used on-axis Si-terminated SiC crystals (n doped, N, 2–4× 1018cm−3). The samples were hydrogen etched in order to remove polishing damage and to chemically passivate the surface. In UHV, the sample was submitted to several sessions of annealing with the temperature precisely monitored by a infrared pyrometer with emissivity set to 95%. The sample was first degassed for 1 h at temperatures lower than 900◦C and finally it was heated for 9 min at 1100◦C and for 6 min at 1150◦C to produce the buffer layer. The pressure was always below 1.0× 10−9 mbar during heating. To produce the graphene layer the same procedure was used with one additional heating cycle for 6 min at 1160◦C. The measurements on both beamlines were done using the same end station and each stage of the experiments was monitored with x-ray photoelectron spectroscopy (XPS) and LEED (see also Supplemental Material20).
In the BL experiment we can clearly resolve two compo-nents in the C 1s line [Fig. 1(a)] with the highest binding energy corresponding to the C-C bonding (buffer layer) and the lower binding energy corresponding to the C-Si in SiC. In this experiment the BL is not completely covering the surface. There is a coexistence of the BL phase with some regions displaying the previous phase (√3×√3)R30◦, as indicated in the LEED pattern. However, its contribution to the BL component in the XPD pattern can be considered negligible, since the peak intensity contribution for the (√3×√3)R30◦ is much weaker compared to the BL intensity, as demonstrated in the Supplemental Material20 and previously by Johansson et al.21 For the graphene experiment, we could resolve one additional component related to the C-C bonding in a true graphene layer, as seen in Fig. 1(b). The relative binding energies and intensity ratio between each component are consistent with previous results for either 1 monolayer (ML) of buffer layer and 1 ML of graphene.15 These thicknesses were determined by XPS and confirmed using XPD. The angle scanned XPD experiments were performed with 450 eV photons, with the Si 2p and C 1s core levels being monitored. The photon energy was carefully chosen, so that for the C 1s electron the kinetic energy (KE) is about 165 eV where the multiple scattering and backscattering regime dominates, so that each C 1s component has a strong sensitivity to a particular chemical environment. On the other hand, for Si 2p the KE is 348.5 eV where the forward scattering regime starts to dominate, so that Si 2p emission is suitable to probe the stacking structure and number of graphite layers. A detailed description of the experimental setup, data acquisition, and analysis is presented in the Supplemental Material20 and elsewhere.22 The experimental patterns are compared with
Intensity (arb. units)
Binding Energy (eV)
290 289 288 287 286 285 284 283 282 281 280 hν = 450 eV
buffer layer SiC
1.7 ± 0.1 eV
100 eV (a)
buffer layer
Intensity (arb. units)
Binding Energy (eV)
288 287 286 285 284 283 282 hν = 450 eV 1.1 ± 0.1 eV 0.9 ± 0.1 eV buffer layer SiC graphene 98 eV (b) graphene
FIG. 1. (Color online) C 1s core-level spectra excited with photons of 450 eV and normal emission. (a) Situation with only the buffer layer and substrate. (b) Situation with the buffer layer and graphene. In each case the solid lines are the fitting envelope (black), SiC component (red), buffer-layer component (blue), and graphene component (green). LEED patterns are displayed in the insets.
simulations from theMSCDpackage developed by Chen and Van Hove,23 which includes a fast structural optimization based on genetic algorithm by M. L. Viana et al.24This code performs multiple-scattering cluster calculations based on the Rehr-Albers separable representation of the free propagator.23 The structure is determined in a fitting procedure that searches for the set of parameters that optimizes the agreement between the theoretical and experimental diffraction curves through minimization of the reliability factor Ra (a perfect agreement corresponds to Ra = 0 and a complete disagreement is given by Ra 1).22
Figure 2 shows the experimental (left column in blue) and simulated (right column in red) photoelectron diffraction patterns, for the best model and after structural relaxations. The patterns shown in Figs.2(a)–2(c)were obtained in the BL experiment [i.e., see Fig.1(a)] and are, respectively, for C 1s from BL, C 1s from SiC, and Si 2p from SiC. The patterns shown in Fig.2(d)are for the case of C 1s from the graphene
FIG. 2. (Color online) Experimental (left column) and simulated (right column) photoelectron diffraction patterns. (a) C 1s, buffer layer, (b) C 1s, SiC, (c) Si 2p, SiC, and (d) C 1s, graphene.
component [i.e., see Fig.1(b)]. The excellent agreement of the simulated patterns with the experimental ones is reflected in the very low Ra factors, i.e., Ra= 0.178 for the C 1s of the BL, Ra= 0.136 for the C 1s from SiC, Ra = 0.157 for Si 2p, and especially the Ra = 0.064 for the C 1s from graphene.
For BL (without graphene), as the first model, we include a flat single layer of carbon atoms on the surface of the SiC(0001) in a honeycomb structure. The first results of our
FIG. 3. (Color online) (a) The honeycomb structure is formed by two hexagonal lattices (A and B). In the relaxation process, we permitted these lattices to move in the z direction, forming a buckling in the short-range order and a ripple in the long-range order (details in text). (b) Distances between atomic layers when the surface contains only a buffer layer. (c) Distances between the atomic layers when the surface contains a buffer layer and graphene. In graphene the atoms are coplanar.
simulations only solved part of the observed structures in the diffraction patterns. According to previous studies,9 the procedure of etching the SiC in a H2flux leaves a surface with predominately three- and six-bilayer steps. Since the density of Si atoms in SiC is three times greater than the density of C atoms in the BL, it is expected that the evaporation of three layers of Si will form one graphene (buffer) layer. Therefore, the steps with three bilayers would be retained after the graphitization and the BL grown on different terraces could be oriented 60◦relative to each other. To include such a domain structure, we have done two calculations with clusters rotated by 60◦relative to each other, and the final theoretical pattern was a equally weighted combination of both domains. Such a model significantly decreased the Rafactor (∼50%) in comparison to the single domain model. The BL was treated in the same manner in the second experiment where graphene is present.
The best in-plane lattice parameters that adjusted the data were 3.081 ˚A for SiC and 2.461 ˚A for the BL and graphene. In order to determine the distance between the BL and the substrate, the atoms of the outermost surface layers were allowed to relax along the 0001 direction. To form a honeycomb structure, two hexagonal lattices were used as shown in Fig.3. In the optimization process of the structure,
L. H. DE LIMA et al. PHYSICAL REVIEW B 87, 081403(R) (2013)
TABLE I. Summary of XPD simulations for the C emitter in BL.
Model Single domain Two domains
Flat 0.434 0.428
Ripple 0.427 0.383
Buckling 0.402 0.201
Buckling + ripple 0.445 0.178
we permitted these lattices, A (green) and B (red), to move in the 0001 direction. For the case with only a BL, this relaxation produced a remarkable reduction of the Raby∼40% compared to the coplanar structure. For the case with 1 ML of graphene coverage, the displacement in BL almost disappears compared to the bare case. We did not find any displacement in the graphene layer.
For the case with only BL, we found a distance of 0.36 ˚A between the two lattices A (green) and B (red). Previous experimental results reporting such a buckling in the literature are lacking, however, a more complicate structure also including a long-range ripple was suggested in some theoretical works.25It could be claimed that the XPD technique is only sensitive to short-range order and cannot distinguish these superstructures. However, it is reasonable that C emitters sitting at different positions and heights relative to the Si layer below the BL can introduce particular features in the diffraction pattern. To try to include such a long-range ripple in our XPD simulations we added into our previous flat models a simple corrugation by changing the height (A) and width (B) of a Gaussian distribution (see Fig.3), in a way similar to the LEED work of Moritz et al.26 for graphene on Ru(0001). The best results after minimization are summarized in TableIand Fig.3. We can observe that only the long-range ripple did not improve the Ra factor significantly, which is expected since XPD is more sensitive to local order. However, including a long-range ripple in the buckling model did improve the Ra. In first approximation, such a simple model describes better some lower polar angles that are more sensitive to the interface through backscattering. It also improved all higher polar angles, demonstrating the existence of a pyramidization through a sp2-to-sp3rehybridization of the
C-C bonding, in agreement with a more sophisticated theoret-ical model suggested by Sclauzero et al.27 Any other kind of atomic arrangement, such as hexagon-pentagon-heptagon defects,28 could not be distinguished with the present study. However, further XPD investigations that can separate the S1 and S2 components15 in the BL signal could find valuable structural information to solve in detail the atomic structure of the BL as well as its relation to the Si-rich interface below it.
Results for the second atomic layer of the SiC substrate us-ing either the photoelectrons from the C 1s and Si 2p emitters are already bulk values (1.90 and 0.62 ˚A). Nevertheless there is a expansion of about 11% in the first Si-C bilayer and a distance of 1.97 ˚A between the BL and SiC. Such a distance is in good agreement with previous density functional theory (DFT) calculations and is attributed to the strong covalent bonding of some carbon atoms in the buffer layer with Si atoms at the rich Si surface of the SiC(0001) surface.29 For the case of the graphene experiment our simulations indicated that it is basically flat and well decoupled from the BL. The buckling displacement was not found at all in the graphene layer and the buckling in the BL was reduced to only 0.09 ˚A (much lower than in the previous case).
In summary, our approach of chemically resolved XPD demonstrated that while graphene is strictly flat, BL shows a more complex structure with a long-range ripple and buckling at the local range order that decouples the honeycomb sublat-tices that form the buffer layer in SiC(0001). The displacement of the sublattices is more intense when the buffer layer is not covered with graphene. This ripple-buckling model structure agrees with a sp2-to-sp3 rehybridization (pyramidalization) that has effects on the stability of this layer27 and explains several of the electronic differences between graphene and the buffer layer.
This work received financial supported from FAPESP, CNPq, CAPES, FAPEMIG, LNLS (SGM 8586 and SGM 10029) and INCT-Carbono from Brazil. The authors would like to thank the LNLS staff, especially G. Rodrigues, for their technical support during beamtime and George G. Kleiman for valuable discussions.
*Also as Joined Appointed Research of Laborat´orio Nacional de Luz
S´ıncrotron - LNLS, 13083-970 Campinas, S˜ao Paulo, Brazil; E-mail: asiervo@ifi.unicamp.br
1A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov, and
A. K. Geim,Rev. Mod. Phys. 81, 109 (2009).
2A. K. Geim and K. S. Novoselov, Nat. Mater. 6, 183 (2007).
3Y. Zhang, Y.-W. Tan, H. L. Stormer, and P. Kim,Nature (London)
438, 201 (2005).
4K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang,
S. V. Dubonos, I. V. Grigorieva, and A. A. Firsov,Science 306, 666 (2004).
5A. J. Van Bommel, J. E. Crombeen, and A. Van Tooren,Surf. Sci.
48, 463 (1975).
6W. Chen, H. Xu, L. Liu, X. Gao, D. Qi, G. Peng, S. C. Tan,
Y. Feng, K. P. Loh, and A. T. S. Wee, Surf. Sci. 596, 176 (2005).
7I. Forbeaux, J.-M. Themlin, and J.-M. Debever,Phys. Rev. B 58, 16396 (1998).
8C. Berger, Z. Song, X. Li, X. Wu, N. Brown, C. Naud, D. Mayou,
T. Li, J. Hass, A. N. Marchenkov, E. H. Conrad, P. N. First, and W. A. de Heer,Science 312, 1191 (2006).
9J. Hass, W. A. de Heer, and E. H. Conrad,J. Phys.: Condens. Matter
20, 323202 (2008).
10W. A. de Heer, C. Berger, X. Wu, P. N. First, E. H. Conrad, X. Li,
T. Li, M. Sprinkle, J. Hass, M. L. Sadowski, M. Potemski, and G. Martinez,Solid State Commun. 143, 92 (2007).
11J. L. Tedesco, G. G. Jernigan, J. C. Culbertson, J. K. Hite, Y. Yang,
K. M. Daniels, R. L. Myers-Ward, C. R. Eddy, Jr., J. A. Robinson, K. A. Trumbull, M. T. Wetherington, P. M. Campbell, and D. K. Gaskill,Appl. Phys. Lett. 96, 222103 (2010).
12C. Virojanadara, M. Syv¨ajarvi, R. Yakimova, L. I. Johansson,
A. A. Zakharov, and T. Balasubramanian,Phys. Rev. B 78, 245403 (2008).
13C. Virojanadara, R. Yakimova, J. R. Osiecki, M. Syv¨ajarvi, R. I. G.
Uhrberg, L. I. Johansson, and A. A. Zakharov,Surf. Sci. 603, L87 (2009).
14C. Virojanadara, R. Yakimova, A. A. Zakharov, and L. I. Johansson, J. Phys. D: Appl. Phys. 43, 374010 (2010).
15K. V. Emtsev, F. Speck, Th. Seyller, L. Ley, and J. D. Riley,Phys. Rev. B 77, 155303 (2008).
16C. Riedl, C. Coletti, T. Iwasaki, A. A. Zakharov, and U. Starke, Phys. Rev. Lett. 103, 246804 (2009).
17S. Kim, J. Ihm, H. J. Choi, and Y.-W. Son,Phys. Rev. Lett. 100, 176802 (2008).
18V. W. Brar, Y. Zhang, Y. Yayon, T. Ohta, J. L. McChesney, A.
Bostwick, E. Rotenberg, K. Horn, and M. F. Crommie,Appl. Phys. Lett. 91, 122102 (2007).
19P. Mallet, F. Varchon, C. Naud, L. Magaud, C. Berger, and J.-Y.
Veuillen,Phys. Rev. B 76, 041403 (2007).
20See Supplemental Material at http://link.aps.org/supplemental/ 10.1103/PhysRevB.87.081403 for a detailed description of the experimental setup, data acquisition, and theoretical simulations.
21L. I. Johansson, F. Owman, and P. M˚artensson,Phys. Rev. B 53, 13793 (1996).
22A. de Siervo, E. A. Soares, R. Landers, T. A. Fazan, J. Morais, and
G. G. Kleiman,Surf. Sci. 504, 215 (2002).
23Y. Chen, F. J. Garcia de Abajo, A. Chasse, R. X. Ynzunza, A. P.
Kaduwela, M. A. Van Hove, and C. S. Fadley,Phys. Rev. B 58, 13121 (1998).
24M. L. Viana, R. D. Muino, E. A. Soares, M. A. Van Hove, and
V. E. de Carvalho,J. Phys.: Condens. Matter 19, 446002 (2007). 25F. Varchon, P. Mallet, J.-Y. Veuillen, and L. Magaud,Phys. Rev. B
77, 235412 (2008).
26W. Moritz, B. Wang, M.-L. Bocquet, T. Brugger, T. Greber,
J. Wintterlin, and S. Gunther,Phys. Rev. Lett. 104, 136102 (2010). 27G. Sclauzero and A. Pasquarello, Phys. Rev. B 85, 161405(R)
(2012).
28Y. Qi, S. H. Rhim, G. F. Sun, M. Weinert, and L. Li,Phys. Rev. Lett. 105, 085502 (2010).
29A. Mattausch and O. Pankratov, Phys. Rev. Lett. 99, 076802 (2007).
