Article pubs.acs.org/JPCC
Interface Oxidative Structural Transitions in Graphene Growth on SiC (0001) Weijie Lu,*,†,⊥ Roland Barbosa,‡,⊥ Edwina Clarke,§,⊥ Kurt Eyink,† Lawrence Grazulis,† William C. Mitchel,† and John J. Boeckl† †
Materials and Manufacturing Directorate, AFRL/RXPS, Air Force Research Laboratory, Wright-Patterson Air Force Base, Ohio 45433, United States ‡ Chimie Physique des Materiaux (Catalyse-Tribologie), Université Libre de Bruxelles, Campus Plaine, CP 243, B-1050 Bruxelles, Belgium § Department of Biological Sciences, University of Alabama, Tuscaloosa, Alabama 35487, United States ABSTRACT: The structural transition from a three-dimensional SiC lattice to a twodimensional graphene sheet is a crucial element in the growth mechanism of graphene on SiC. An interfacial defective transition layer near the surface of the SiC substrate is believed to be an intermediate structure for graphene layer formation. The transition layer consists of SiOxCy, vacancies, and other defects in the SiC lattice, which result from Si evaporation via thermal degradation of the SiC lattice and oxidation reactions of residual oxygen and other oxygen containing molecules on the SiC surface at high temperatures. This partially oxidized and structurally degraded SiC lattice layer is formed at temperatures lower than the graphene growth temperature but then decomposes with increasing temperature, leading to graphene formation. Then, the growth mechanism for graphene on SiC (0001) in high vacuum consists of multiple steps, including Si removal by thermal decomposition and oxidation, collapsing of the near surface SiC lattice, conversion from sp3 to sp2 carbon, and an increase in the degree of low-dimensional graphitization. The proposed atomic scale mechanism is able to explain experimental phenomena in graphene/SiC structural growth, such as graphene coverage at step edges, growth environment effects, graphene domain size, and thickness variations.
■
INTRODUCTION Among the several techniques available for fabrication of graphene, growth on the Si-face SiC (0001) is favored from the standpoint of electronic device fabrication due to the simple growth process, good structural uniformity, and large planar surface morphology. Also, this graphene structure is produced on commercially available atomically flat chemical−mechanical polished (CMP) semi-insulting SiC wafers suitable for conventional wafer scale processing.1,2 Electronic devices based on the graphene/SiC structure are believed to have promising potential for future high-frequency applications.2 Many factors affect the graphene growth and quality on SiC. It is known that the interface structures and the electronic properties of graphene/SiC depend on the crystal orientation of the SiC surface it is grown on.3 Graphene grown on SiC (0001) has exhibited a more uniform topology and better control in graphene layer number than that on C-face SiC (000−1). However, the electronic transport properties of graphene on C-face SiC are superior compared to graphene on Si-face SiC. In addition, the growth environment is also known to affect the graphene properties. It is known that graphene produced in ultrahigh-vacuum (UHV) conditions exhibits poor surface morphology and low carrier mobility,4 whereas graphene grown in atmospheric pressure argon has improved carrier transport properties and surface morphology.5 Despite a large number of publications on the growth process of graphene/SiC (0001), a detailed growth mechanism © 2012 American Chemical Society
at the atomic scale for graphene/SiC structure is still not clear. The relationships between the graphene/SiC (0001) interface and low-dimensional confined graphitization on the surface, and the electronic properties are not understood. One of the main concerns from a material quality standpoint is that the carrier mobility for graphene/SiC (0001) is only about 1000− 2000 cm2/(V s),6 which is about 10 times lower than that usually observed for exfoliated graphene on SiO2. Understanding the nature of graphene/SiC (0001) growth is critical toward improving transport properties in the graphene films.7 SiC surface graphitization in ultrahigh vacuum (UHV) has been investigated for decades. Generally, heat treatment in vacuum at high temperature induces Si evaporation from the SiC lattice and results in formation of a sp2-graphitized layer on the surface.8−10 This simple process has been employed in many investigations to explain the growth mechanism of graphene/SiC. Forbeaux et al.9 established the well-known sequence of surface reconstructions for 6H-SiC (0001) in UHV as increasing temperature. A (3 × 3) reconstruction is seen at temperatures below 900 °C but converts to a (√3 × √3)R30° surface above this temperature. At around 1100 °C it converts to (6√3 × 6√3)R30° while a graphite (1 × 1) appears around 1400 °C. The temperature for each of these transitions varies Received: February 29, 2012 Revised: June 21, 2012 Published: June 25, 2012 15342
dx.doi.org/10.1021/jp301996h | J. Phys. Chem. C 2012, 116, 15342−15347
The Journal of Physical Chemistry C
Article
Figure 1. AFM images of SiC samples after annealing in 10−6−10−7 Torr for 30 min at (a) 1300, (b) 1400, and (c) 1500 °C.
and silicon vapor. However, many observed experimental phenomena in the graphene/SiC growth process are difficult to explain by these mechanisms. For example, graphene grown on SiC in atmospheric pressure Ar5 exhibits better morphology and higher carrier mobility than that grown in UHV.4 Our previous work20 has observed that graphene grows on oxidized SiC likely via silicon oxycarbides decompositions. In this study, we use AFM, HR-TEM, and X-ray photoelectron spectroscopy (XPS) to study growth of graphene on SiC (0001) and propose a growth mechanism in high vacuum which includes multiple interface structural transitions from the SiC lattice to graphene.
somewhat in the literature, but the sequence does not. While important, consideration of these surface reconstructions alone limits structural evolution on the SiC surface only to the topmost SiC bilayers.9,10 In general, two approaches have been used to investigate the growth mechanism of graphene on SiC: (a) from the evolution of surface morphology during the growth process and (b) from thermal decomposition of the SiC lattice and formation of graphene. As examples of (a), Borovikov et al.11 suggested a graphene growth mechanism based on step flow decomposition in which fingerlike structures observed during graphene growth are explained by step-edge instability. In this model, the Si desorption from steps is likely the main controlling process in graphene growth.12 The model explains the finger-like growth of graphene on SiC by postulating that graphene growth initiates at SiC step edges since the terrace step edges and other topological defects on SiC are low-energy sites for graphene nucleation.13,14 These proposed mechanisms were mainly based on the images of ex situ AFM (atomic force microscopy) and/ or STM (scanning tunneling microscopy) at different growth stages, and the graphene growth mechanisms were correlated with the surface morphological changes assuming that Si evaporates from the topmost SiC surface bilayers only. From the interface structural evolution approach of (b), the formation of one graphene layer requires carbon atoms from the decomposition of about three SiC bilayers, and a metastable transitional interface is formed when only one or two SiC bilayers decompose as observed by high-resolution transmission electron microscopy (HR-TEM).15 The measured spacing between the interface layer and the top SiC bilayer is smaller than the interlayer spacing in bulk graphite but larger than the covalent bond length between C atoms and the Si atoms of the top SiC bilayer. To accommodate this, an adatom model with an additional partial layer of adatoms between the SiC and graphene was proposed by Hass et al.3 A strong substrate−graphite bond is found in the first all-carbon layer by density functional theory calculations and X-ray diffraction for few graphene layers on SiC. This first layer is devoid of graphene electronic properties and acts as a buffer layer.16 Similarly, a bottom-up growth mechanism proposed by Huang et al.17 suggested thermal decomposition of one single SiC bilayer underneath the graphene layers causes the accumulation of carbon atoms to form a new graphene buffer layer at the graphene/SiC interface. Weng et al.18 found that the buried interface layer possesses a lower average areal density of carbon atoms than graphene, indicating that it is not a graphene-like sheet. Gao et al.19 reported the presence of Si vacancies as much as a few micrometers below the graphene layer. These published growth mechanisms have a common assumption: thermal decomposition of SiC results in graphene
■
EXPERIMENTAL METHODS Semi-insulating on-axis chemical mechanical polished (CMP) 6H SiC (0001) wafers from II-VI Inc. were used for this study. After cleaning with 20% HF solution and deionized water, the SiC sample (10 × 10 mm2) was placed in the growth chamber and heated to the specific temperature. Hydrogen pregrowth annealing was not performed. The base pressure at the growth chamber is at about 10−9 Torr, and the sample is placed on a small (1.0 in. diameter) BN resistance heater with 20 °C/min ramp rate from room temperature to the desired temperature. The pressure was maintained at 10−6−10−7 Torr during the growth at high temperature with a turbo pump. Using an FEI Strata DB-235, samples were prepared with a focused ion beam (FIB) for cross-sectional TEM analysis. The 30 keV Ga+ ion beam was then used to prepare thin foils that were subsequently broad-beam Ar+ ion polished at 1 keV to remove any Ga+ beam damage. These electron transparent samples were later examined using a Philips CM-200 field emission gun (FEG) transmission electron microscope (TEM) operated at 200 kV. The X-ray photoelectron spectroscopy (XPS) measurements were performed on a Kratos-800 electron spectrometer. Mg Kα (1253.6 eV) radiation was used. The data were analyzed using the Vision2 software. Nanoscope IIIa (Bruker Co.) was used for AFM (atomic force microscopy) measurements. Tapping-mode was applied for all AFM measurements, and at least three areas on each sample were measured.
■
RESULTS AND DISCUSSION Morphological and Interface Evolutions from SiC to Graphene. Chemical mechanical polished (CMP) on-axis SiC (0001) samples were annealed in an HV chamber at temperatures from 1200 to 1500 °C for 30 min at pressures ranging from 10−6 to 10−7 Torr. As shown in Figure 1a, at 1300 °C, irregular and wavy shaped surface features appear on the surface, which has an overall roughness of less than 1 nm. Continuing to increase the temperature to 1400 °C, Figure 1b 15343
dx.doi.org/10.1021/jp301996h | J. Phys. Chem. C 2012, 116, 15342−15347
The Journal of Physical Chemistry C
Article
Figure 2. Cross-sectional TEM images of SiC sample after annealing in 10−6−10−7 Torr for 30 min at (a) 1200 and (b) 1300 °C.
shows that the irregular-shaped surface features remain, and some stripes-like are formed. Finally, Figure 1c shows the large graphene domains with dimensions of several micrometers and atomic flatness that are formed after annealing to 1500 °C. Of interest, the SiC surface becomes rough at 1200 °C, well before the structural features of graphene which are formed at 1400 °C. The AFM images in Figure 1 show the morphological evolution during graphene growth in HV and are consistent with other published results.11−13 Figure 2 shows the cross-sectional TEM images of the samples annealed to 1200 and 1300 °C for 30 min in 10−6− 10−7 Torr. After annealing at 1200 °C (Figure 2a), a dark contrasted layer up to 10 nm thick appears the near SiC surface. The TEM image shows that the upper most SiC bilayers become disordered, contain defects, and some are discontinuous. After annealing to 1300 °C, shown in Figure 2b, the SiC lattice is still intact but changes in contrast. There is no evidence in the TEM images for the formation of graphene at 1200 °C. It is noted that cross-sectional TEM images have been made of unannealed CMP SiC wafers, and this damaged SiC layer was not observed. At 1300 °C, the distance between SiC bilayers is 2.3 Å (L1a) in the interface region and 2.5 Å (L1b) in the SiC bulk. The interlayer spacing in graphene on SiC is about 3.4−3.6 Å,15,18 and the bulk SiC bilayer spacing is 2.5 Å. Therefore, graphene is not formed on SiC at 1300 °C. The shorter lattice spacing interfacial SiC bilayers has been observed by Norimatsu et al.15 as the interfacial defects. The contrast changes in the TEM images are often considered to be a result of charged defects and/or an electronic charge density induced by chemical reactions with foreign atoms.21 The detailed interpretation and analysis of the contrast changes in TEM are complicated,21,22 and the structural and bonding nature of the defective layer is under further investigation. The darker contrasts in the cross-sectional TEM images likely indicate that structural defects and lattice distortions near the surface of the SiC take place after annealing in high vacuum at 1200−1300 °C, which is below the temperature for graphene growth. Figure 3 shows a cross-sectional TEM image of graphene/ SiC after annealing at 1500 °C at 10−6−10−7 Torr. Three layered graphene was grown on SiC with an interlayer spacing of 3.5 Å (L2a), and the bulk SiC bilayer spacing is 2.5 Å (L2b). The interfacial lattice defective SiC layer under the graphene still is observed, but the thickness is reduced to a few SiC bilayers. It is also shown a nonuniform amorphous carbon-like layer between the graphene and the defected SiC lattice with a thickness of ∼1 nm. Not all published reports show these layers between the SiC and the graphene,15 but Colby et al.23 recently
Figure 3. Cross-sectional TEM images of graphene/SiC sample after annealing in 10−6−10−7 Torr at 1500 °C for 30 min.
reported that an amorphous carbon layer is formed between graphene and C-face SiC substrate at the growth temperature of 1600 °C in atmospheric pressure Ar. It is observed that the amorphous layer is not uniformly present across the wafer and does not appear in samples annealed at 1500 °C. From the above discussion, it is concluded (a) the defected interfacial layer is strongly related to the annealing temperature and the thickness is significantly reduced after the formation of graphene at higher temperature and (b) the interfacial structure of graphene/SiC is not uniform. It is known that SiC surface structures are not ideal or identical, and the chemical reactivity differences of different surface areas likely result in some variation in the graphene structure and the interface. XPS measurements were made on the samples in Figure 1 to further investigate the nature of the defected SiC surface and amorphous layers observed via TEM. Figure 4a shows C1s XPS of the SiC sample after annealing at 1300 °C for 30 min at 10−6−10−7 Torr in the HV chamber. The binding energy of the C1s peak from SiC is at 282.7 eV, the C1s peak for silicon oxycarbide (SiCxOy) is at 283.6 eV, the sp3 carbon peak is at 285.0 eV, the sp2 carbon peak is at 284.5 eV, and a C−O group resulting from air exposure is at 286.7 eV.24,25 As shown in Figure 4a, after annealing at 1300 °C, a mixture of sp3 and sp2 carbon covers the surface of the SiC substrate. The AFM images in Figure 1 show that the stripe-shaped surface features are formed on the SiC surface at 1300 °C, but the graphene is not yet formed. It is well-known that only the sp2-graphitized C1s peak at 284.5 eV is observed in XPS on graphene well grown at higher temperatures.26 Figure 4b shows Si2p XPS of the sample prepared at 1300 °C in the HV chamber. The binding energy of the Si2p peak from SiC is at 100.2 eV, and a shoulder peak at 101.3 eV is from SiCxOy (oxygen atoms substitute some Si sites in the SiC 15344
dx.doi.org/10.1021/jp301996h | J. Phys. Chem. C 2012, 116, 15342−15347
The Journal of Physical Chemistry C
Article
graphene/SiC (0001) are based on the simple assumption that SiC thermally decomposes one layer at a time into gaseous silicon and solid carbon at high temperature.3,11−17 However, many observed experimental phenomena in the graphene/SiC growth process are difficult to explain using these mechanisms, including (a) graphene grown on SiC in atmospheric pressure Ar exhibits better morphology and higher carrier mobility,5 (b) graphene grown in an RF furnace in 10−3−10−5 Torr vacuum has better morphology and carrier mobility than that grown in UHV,4 and (c) the growth temperatures of graphene/SiC vary from one research group to another, by up to 300−400 °C.27 Further, a correlation of growth mechanisms and electronic properties of graphene/SiC has not been established. The simple thermal decomposition assumption does not explain the various structures formed during the evolution from a SiC lattice to graphene, such as the structure of carbon as Si vaporizes via SiC bilayer by bilayer, the structure of the defected SiC lattice before it collapses into the graphene layer, or the mechanism by which the defective SiC lattice collapses. The effects of chemical reactions in the growth chamber were not taken into account in the published studies.3,11−17 Lowpressure residual oxygen usually exists in high vacuum and even in UHV systems. The oxygen level can vary quite widely from one system to another. Additionally, commercial high-purity Ar usually contains subppm level oxygen, in which oxygen partial pressure is about 10−3 Torr in atmospheric pressure Ar. Other oxygen sources include the thin oxides on SiC surface and the desorbed oxygen and oxygen containing molecules on the sample stage at high temperatures. When low-pressure molecular oxygen is present, silicon oxycarbides (SiCxOy) can be formed by the reaction of SiC with O225 as in eq 1, which can be followed at higher temperatures by the decomposition of the SiCxOy into gaseous silicon, carbon monoxide and solid carbon by eq 2 through active oxidation28 at high temperatures: Figure 4. (a) C1s XPS of SiC after annealing at 1300 °C in 10−6−10−7 Torr for 30 min. (b) Si2p XPS of graphene/SiC grown at 1300 °C in 10−6−10−7 Torr for 30 min.
SiC(solid) + O2 (gas) → SiCx Oy (solid) + C(solid)
(1)
SiCx Oy (solid) → Si(gas) + CO(gas) + C(solid)
(2)
Low-pressure oxygen can also react with SiC directly via active oxidation as in eq 3:
lattice) compound with a chemical shift of the binding energy of Si2p in SiC.24 This SiCxOy peak is found in both the C1s and Si2p XPS for the sample annealed at 1300 °C, when graphene is not formed. The XPS results in Figure 4 have indicated that the graphene growth on SiC is a multistep process involving sp3 carbon and SiCxOy. Coupled with the AFM images and TEM images in Figures 1−3, the XPS results in Figure 4a,b are consistent with a structural evolution of graphene during growth from 1300 to 1500 °C. Therefore, the results from AFM, TEM, and XPS indicate that a defective SiC lattice layer is the intermediate structure that exists prior to the formation of graphene on our SiC surfaces. A “layer 0” or “buffer layer” has been reported at the interface between the graphene and the SiC, but the exact atomic structure of this buffer layer is unknown.7 The results and analysis of the defective SiC interface lattice layer prior to graphene growth observed in this study provide useful information toward understanding the atomic structure of the “buffer layer”, which could relate to the incomplete decomposition of the defective transition layer on graphene/ SiC interface. Further investigations are needed. Interfacial Defective Transition Layer in Graphene/SiC Growth. The presently proposed growth mechanisms for
SiC(solid) + O2 (gas) → SiO(gas) + C(carbon)
(3)
When oxygen is at a very low level, as in a UHV system, graphitic carbon is formed via simple Si vaporization by SiC thermal decomposition as in eq 4: SiC(solid) → Si(gas) + C(solid)
(4)
Since the molecular structures of silicon oxycarbides are not clear, the detailed process of the silicon oxycarbide decomposition is also unknown. However, it is very likely that oxygen replaces some Si sites in the SiC lattice.25 The supporting evidence of SiC oxidation involvements includes (a) elemental Si is not detected by XPS; (b) the number of graphene layers is somewhat self-limited, indicating graphene only grows near the surface SiC lattice; (c) the surface step edges are covered by graphene, which indicates that graphene is formed by decomposition of a thin layer covering the SiC surface, including on the step edges, not grown atomically on the lateral direction of the steps; (d) many groups have reported successful production of graphene/SiC structures where, generally, high-vacuum chambers with the presence of lowpressure residual oxygen and other oxygen containing 15345
dx.doi.org/10.1021/jp301996h | J. Phys. Chem. C 2012, 116, 15342−15347
The Journal of Physical Chemistry C
Article
composition in different growth chambers can lead to variations in growth conditions of the graphene/SiC structure. Since the defective layer, consisting of vacancies, silicon oxycarbides, and a defective SiC lattice, is considered as the intermediate layer for graphene growth in the proposed mechanism and the thickness of silicon oxycarbides is only a few nanometers thick, surface structures on the SiC are expected to play a major role in the structural uniformity of silicon oxycarbides layer. Therefore, controlling the uniformity of the defective SiC surface layer will be important to form large size and uniformed layer number graphene. The control and monitoring of the growth environment is likely to be the key for understanding and improving the structure and quality of graphene grown on SiC. This proposed growth mechanism for the graphene/SiC structure is able to explain graphene coverage on the step edges. Figure 6a shows a TEM of step coverage for graphene on SiC in which the graphene layer covers the steps like a carpet as is often observed in graphene/SiC structures.14,26 Since the surface reconstructions do not take place on the SiC step edges,
molecules have been used in most of the published reportsin fact, large growth temperature variations, up to 300−400 °C, are reported,27 and a possible reason for these variations in growth temperature is the difference of growth environment in the different chambers; and (e) the silicon oxycarbides are commonly found as an intermediate in the SiC oxidation process.24,25 Our previous work has shown that graphene can be grown on oxidized SiC20 and that low-pressure residual oxygen acts as an atomic catalyst in growth of metal-free CNTs/SiC.29 The oxygen catalysis on graphene is another topic beyond this study and will not be discussed further in this report. The experimental data show that at least three defective structures have contributed to the defective transition layer structures: vacancies, silicon oxycarbides, and a defective SiC lattice. The data from XPS and TEM analysis show the presence of the silicon oxycarbides and defected SiC lattices. The detection and role of vacancy sites near the SiC surface after annealing in vacuum were reported by Forbeaux et al.9,10 Based on the above experimental data and discussions, structural degradation and transitions of the interface at the atomic scale are important in the growth of graphene/SiC (0001) structures, and a mechanism was formulated and is shown in Figure 5. In this study, we have concluded that
Figure 5. Sketches of graphene growth mechanism on SiC via a defective surface layer.
graphene formation consists of a multistep process: Si removal by thermal decomposition and oxidation by low-pressure residual gases, collapsing of the near surface SiC lattice, conversion from sp3 to sp2 carbon, and increases in the degree of low-dimensional graphitization for graphene formation. Interpretations of Experimental Phenomena. The interface structural transitions in this study have explained the variations of graphene growth conditions in the literature.27 According to the proposed mechanism, the presence of residual oxygen and other oxygen-containing molecules impacts significantly on the formation of a thin defective surface layer on SiC as well as collapsing of the defective layer to grow graphene. Therefore, the residual gas pressure and gas
Figure 6. (a) Cross-section TEM image on surface step on graphene/ SiC grown at 1500 °C in 10−6−10−7 Torr for 30 min. (b) Sketches of graphene growth on step edges of SiC. 15346
dx.doi.org/10.1021/jp301996h | J. Phys. Chem. C 2012, 116, 15342−15347
The Journal of Physical Chemistry C
Article
(3) Hass, J.; de Heer, W. A.; Conrad, E. H. J. Phys.: Condens. Matter 2008, 20, 323202. (4) de Heer, W. A. MRS Bull. 2011, 36, 632−639. (5) Emtsev, K. V.; Bostwick, A.; Horn, K.; Jobst, J.; Kellogg, G. L.; Ley, L.; McChesney, J. L.; Ohta, T.; Reshanov, S. A.; Röhrl, J.; et al. Nat. Mater. 2009, 8, 203−207. (6) Weingart, S.; Bock, C.; Kunze, U.; Emtsev, K. V.; Seyller, Th.; Ley, L. Physica E 2010, 42, 687−690. (7) Hass, J.; Millán-Otoya, J. E.; First, P. N.; Conrad, E. H. Phys. Rev. B 2008, 78, 205424. (8) Muehlhoff, L.; Choyke, W. J.; Bozack, M. J.; Yates, J. T. J. Appl. Phys. 1986, 60, 2842−2853. (9) Forbeaux, I.; Themlin, J.-M.; Debever, J.-M. Phys. Rev. B 1998, 58, 16396−163406. (10) Forbeaux, I.; Themlin, J.-M.; Charrier, A.; Thilbaudau, F.; Debever, J.-M. Appl. Surf. Sci. 2000, 162−3, 406−412. (11) Borovikov, V.; Zangwill, A. Phys. Rev. B 2009, 80, 121406R. (12) Hupalo, M.; Conrad, E. H.; Tringides, M. C. Phys. Rev. B 2009, 80, 041401R. (13) Bolen, M. L.; Harrison, S. E.; Biedermann, L. B.; Capano, M. A. Phys. Rev. B 2009, 80, 115433. (14) Robinson, J.; Weng, X.; Trumbull, K.; Cavalero, R.; Wetherington, M.; Frantz, E.; LaBella, M.; Hughes, Z.; Fanton, M.; Snyder, D. ACS Nano 2010, 4, 153−158. (15) Norimatsu, W.; Kusunoki, M. Chem. Phys. Lett. 2009, 468, 52− 56. (16) Varchon, F.; Feng, R.; Hass, J.; Li, X.; Nguyen, B. N.; Naud, C.; Mallet, P.; Veuillen, J. Y.; Berger, C.; Conrad, E. H.; et al. Phys. Rev. Lett. 2007, 99, 126805. (17) Huang, H.; Chen, W.; Chen, S.; Wee, A. T. S. ACS Nano 2008, 2, 2513−2518. (18) Weng, X.; Robinson, J. A.; Trumbull, K.; Cavalero, R.; Fanton, M. A.; Snyder, D. Appl. Phys. Lett. 2010, 97, 201905. (19) Gao, X.; Chen, S.; Liu, T.; Chen, W.; Wee, A. T. S.; Nomoto, T.; Yagi, S.; Soda, K. Phys. Rev. B 2008, 78, 201404R. (20) Lu, W.; Boeckl, J. J.; Mitchel, W. C.; Crenshaw, T. R.; Collins, W. E.; Chang, R. P. H.; Feldman, L. C. J. Electron. Mater. 2009, 38, 731−736. (21) Meyer, J. C.; Kurasch, S.; Park, H. J.; Skaklova, V.; Künzel, D.; Groß, A.; Chuvilin, A.; Algara-Siller, G.; Roth, S.; Iwasaki, T.; et al. Nat. Mater. 2011, 10, 209−215. (22) Muller, D. A.; Kourkoutis, L. F.; Murfitt, M.; Song, J. H.; Hwang, H. Y.; Silcox, J.; Dellby, N.; Krivanek. Science 2008, 319, 1073−1076. (23) Colby, R.; Bolen, M. L.; Capano, M. A.; Stach, E. A. Appl. Phys. Lett. 2011, 99, 101904−101906. (24) Hornetz, B.; Michel, H.-J.; Halbritter, J. J. Vac. Sci. Technol., A 1995, 13, 767−771. (25) Ö nneby, C.; Pantano, C. G. J. Vac. Sci. Technol., A 1997, 15, 1597−1602. (26) Seyller, Th.; Emtsev, K. V.; Gao, K.; Speck, F.; Ley, L.; Tadich, A.; Broekman, L.; Riley, J. D.; Leckey, R. C. G.; Rader, O.; et al. Surf. Sci. 2006, 600, 3906−3911. (27) Lu, W.; Boeckl, J.; Mitchel, W. C. J. Phys. D: Appl. Phys. 2010, 43, 374004. (28) Song, Y.; Smith, F. W. Appl. Phys. Lett. 2002, 81, 3061−3063. (29) Lu, W.; Boeckl, J.; Mitchel, W. C.; Rigueur, J.; Collins, W. E. Mater. Sci. Forum 2006, 527−529, 1575−1577.
it is difficult for the current available growth mechanisms to explain why the surface reconstructions are not necessary for the growth of graphene on the step edges. From the viewpoint of structural transitions, as shown in Figure 6b, the defective layer on SiC is formed due to oxidation and/or decomposition, and there is no significant difference of the defective layer on the SiC terraces and on the step edges. Since graphene on both SiC terraces and step edges is formed through a defective layer, in our proposed mechanism, the lattice orientation difference on the terraces and step edges is not important to the structure of graphene. In other words, since graphene on SiC surface steps is not “epitaxial”, therefore epitaxial bonding and interactions with SiC lattice are not necessary for graphene growth on SiC (0001). The driving force for graphene growth is likely controlled by the thermodynamic energy of graphene covering the step edges and as well as on the terraces.
■
CONCLUSIONS An interface defective transition layer near the surface of the SiC substrate is believed to be an intermediate structure for graphene layer formation. This layer consists of SiOxCy, vacancies, and a defective SiC lattice. The transition layer is formed as Si evaporates from the SiC lattice at high temperatures via both thermal degradation and oxidation reactions enhanced by low-pressure residual oxygen and other oxygen-containing molecules interacting with SiC near the surface. The partially oxidized and structurally degraded SiC lattice layer is observed at temperatures lower than the graphene growth temperatures and then decomposes with increasing temperature, leading to graphene formation. From the viewpoint of structural transitions, growth of graphene/SiC (0001) from the SiC lattice to graphene from chemical reactions consists of multiple steps; Si removal by thermal decomposition and oxidation by low-pressure residual gases, collapsing of the near surface SiC lattice, conversion from sp3 to sp2 carbon, and increases in the degree of low-dimensional graphitization. The “buffer layer” on the graphene/SiC interface could be viewed as the incomplete decomposition of the defective/disordered transition layer. The mechanism is able to explain numerical experimental phenomena in the graphene/ SiC structural growth, such as graphene coverage of the step edges, growth environment effects, domain size, and variations of graphene growth conditions.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest. ⊥ Previous address: Department of Chemistry, Fisk University, Nashville, TN.
■
ACKNOWLEDGMENTS The authors acknowledge the support of AFOSR (Dr. Charles Lee) and of the Laboratory Director Funds (LDF) of Air Force Research Laboratory.
■
REFERENCES
(1) Berger, C.; Song, Z. M.; Li, T. B.; Li, X. B.; Ogbazghi, A. Y.; Feng, R.; Dai, Z.; Marchenkov, A. N.; Conrad, E. H.; First, P. N.; et al. J. Phys. Chem. B 2004, 108, 19912−19916. (2) Lin, Y.-M.; Dimitrakopoulos, C.; Jenkins, K. A.; Farmer, D. B.; Chiu, H.-Y.; Grill, A.; Avouris, Ph. Science 2010, 327, 662−662. 15347
dx.doi.org/10.1021/jp301996h | J. Phys. Chem. C 2012, 116, 15342−15347
