Graphene Formation on a 3C-SiC(111) Thin Film Grown on Si(110) Substrate∗

With its industrial adaptability, epitaxial graphene (EG), formed by a UHV annealing of SiC substrates, is attracting recent attention. While hexagonal SiC bulk substrates have been solely used for this purpose, benefits in use of 3C-SiC virtual substrate founded on Si substrates could be enormous. We have succeeded in fabricating a graphene film on a 3C-SiC(111) virtual substrate, which was preformed on a Si(110) substrate by gas-source molecular beam epitaxy using monomethyl silane. The geometrical matching in this configuration greatly suppresses the strain in the SiC film, which is related to this successful formation of graphene. [DOI: 10.1380/ejssnt.2009.311]


I. INTRODUCTION
The discovery of the single-layer graphene by Geim et al. [1] in 2004, together with their very primitive methodology of peeling the graphite, made a great impact in both academia and industry.Ever since the number of papers on this carbon monolayer has been skyrocketing.Trials to form graphene, however, can date back to 1975 when van Bommel et al. [2] annealed a SiC substrate at 1200 • C in high vacuum.Although they could not, despite their claim, actually form graphene with this method, annealing of SiC substrates, or epitaxial graphene (EG), has attracted increasing attention of researchers, and is now established as an alternative, more practical, method to fabricate a graphene sheet.Studies on graphene formation on 4H-SiC [3,4] or 6H-SiC [5][6][7] bulk substrates have been reported, as well as on device performance based on the material [8].
Since EG possesses higher adaptability to mass production as compared to the peeling method, it has a greater impact on industry.The drawbacks of EG, however, are the limited diameter of the available SiC substrates, and their low cost-effectiveness in the current price.If one can fabricate a graphene film onto a thin SiC film grown on a large-diameter Si wafer (SiC virtual substrate), its industrial impact would be enormous.In this study, we report one of the first successful formations of graphene on a SiC virtual substrate.
SiC has more than 200 polytypes.However, the only polytype that grows on Si substrate is the cubic (3C-)SiC.What is important in forming graphene onto 3C-SiC virtual substrate is that we must utilize a proper orientation of the 3C-SiC surface in order to accommodate with the 6-fold symmetry of the graphene crystal.In this context, 3C-SiC(111) surface is definitely the best choice.There is however a significant lattice mismatch between Si and 3C-SiC, which is as large as 20%, and it is probably for this reason that no reports have been ever made on the graphene formation on 3C-SiC(111)/Si(111) virtual substrate.
In this study, we have utilized Si(110) substrate, instead of Si(111), to grow 3C-SiC on it.The fact that a (111)oriented 3C-SiC film grows on the Si(110) substrate has been reported by Nishiguchi et al. [9] using propane and dichlorosilane.They explained the phenomenon by noticing a quasi-lattice-matching being realized in the configuration of (111)3C-SiC//(110)Si; [ 11 2]3C-SiC//[001]Si.Following their method, but with an additional aim to reduce the growth temperature, we have applied our gassource molecular-beam epixaty (GSMBE) method [10][11][12][13] to this 3C-SiC(111)/Si(110) system.By use of monomethyl silane (MMS) as a single source gas, we succeeded in reducing the growth temperture from 1300 • C [9] to 1000 • C [14].The half width of the XRD rocking curve, a measure of the film quality, has been reduced by a factor of 30% from that of the 3C-SiC(111)/Si(111) film.The film strain, as evaluated from the anisotropy (a − a ⊥ )/a between the lateral (a ) and the growthdirectional (a ⊥ ) lattice constants, is reduced by a factor of four as well.In this study, we have challenged the formation of graphene on this 3C-SiC(111) virtual substrate formed on the Si(110) substrate.

II. EXPERIMENTAL
GSMBE of 3C-SiC has been conducted by using MMS (Trichemical, 99.999%).Details of the apparatus are described in Ref. [12].Samples are B-doped, p-type Si(110) wafer, cut to 7 × 40 mm 2 size.After an ex-situ wet cleaning, they were transferred to the chamber, in which they were degassed, and flash-annealed at 1200 • C for several times to yield a clean surface.Sample heating was conducted by a resistive heating.After the in-situ cleaning, MMS was introduced to the chamber via leak valve at a pressure of 3.3 × 10 −3 Pa.Growth experiment consists of two stages: a bufferlayer formation conducted at 600 • C for 5 min [11] and a subsequent SiC growth at 1000 • C for 120 min.For this growth condition, the film thickness is typically 80 nm as we evaluate with a step profiler.After the growth, MMS was evacuated from the chamber, and the sample was annealed in vacuum at 1200 • C for 30 min.The sample surface was evaluated by a Nomarski microscope and Raman-scattering microscopy (Renishaw, Ar 514 nm).As a reference, a graphene (graphite) film transferred onto a 300-nm-thick SiO 2 film on Si(001) substrate with the peeling method has been evaluated as well.

III. RESULTS AND DISCUSSIONS
The top (bottom) spectrum in Fig. 1 shows the Raman spectrum from the graphene (graphite) film obtained by the peeling method.The spectrum designated as "graphite on SiO 2 " corresponds to a portion which looks dense in optical microscopy while the one as "FLG on SiO 2 " to a portion which looks shallow and is therefore identified as a few-layer graphene (FLG) on this 300-nm oxide [1].For both spectra, the G (1580 cm −1 ) and the 2D (2700 cm −1 ) bands can be observed.While the Gband positions mostly coincide each other, the position of the 2D band shows a red shift by ∼45 cm −1 as we move from graphite to FLG.The band becomes singlepeak as well.According to Ferrari et al. [15], a graphene film transferred onto SiO 2 films show a red shift in its 2D band by 45 cm −1 from that of graphite.We thus confirm formation of few-layer graphene in our sample prepared by the peeling method.
The spectrum designated as "FLG on VS" in Fig. 1 2600 2700 2800 2900 Raman shift (cm -1 ) Raman Intensity (arb.unit)FIG.2: Raman-scattering spectra around the 2D band.FLG on virtual substrate (VS) is shown to consists of single-and two-layer sheets.For comparison, a single-layer EG and a twolayer EG, both on 6H-SiC(0001) substrate [16] are shown.
is the one from the annealing of the 3C-SiC(111) virtual substrate (VS).The presence of the D band, in addition to the G and the 2D bands, indicates inclusion of defected graphite, similar to the case for EGs on SiC bulk substrate [16].What should be noted here is the fact that all the D, G, and 2D bands are blue-shifted from those on SiO 2 .Ni et al. [16] report similar blue shifts in the D, G, and 2D bands, and conclude that EGs on SiC substrates are subject to a compressive stress in the order of 2 GPa.They further found that the 2D-band position depends on the number of layers: a single-layer at 2715 cm −1 and a two-layer at 2736 cm −1 , which is shown in the bottom and the top spectrum in Fig. 2, respectively.The middle spectrum in Fig. 2 is from the "FLG on VG" shown in Fig. 1.As shown, the spectrum can be peak-separated into two components at 2715 and 2735 cm −1 , which are in excellent agreement with those from the single-and the two-layer graphenes on 6H-SiC.We therefore conclude that the C-related material formed on our annealed 3C-SiC(111)/Si(110) virtual substrate is of epitaxial graphene, most probably of a mixture of singleand two-layer graphene sheets.The narrower bandwidth in our sample suggests even better quality of the graphene.

IV. CONCLUSIONS
We have succeeded in forming a single-and a two-layer graphene sheets on a 3C-SiC virtual substrate, which is prepared by heteroepitaxy of 3C-SiC(111) on Si(110) substrate.The geometrical matching in this configuration http://www.sssj.org/ejssnt(J-Stage: http://www.jstage.jst.go.jp/browse/ejssnt/) e-Journal of Surface Science and Nanotechnology Volume 7 (2009) greatly suppresses the strain in the SiC film, which is related to this successful formation of graphene.This technology can be of enormous importance in realizing graphene-based devices on Si substrate, i.e., graphene-onsilicon (GOS) technology.