Interface Structure of Graphene on SiC for Various Preparation Conditions

In this thesis we study the preparation dependence of the interface structure of graphene on SiC. We compare epitaxial graphene grown in ultra-high vacuum (UHV), in an atmosphere of argon, and in a background of 10^-4 Torr of disilane. Graphene growth is studied on both the polar faces of SiC – the SiC(0001) surface, also known as the Si-face and the SiC(0001 ) surface, also known as the C-face. We find that the quality of graphene and the interface of graphene on the substrate depend on which face of SiC is used, and also what environment it was prepared in. Characterization using atomic force microscopy (AFM), low energy diffraction (LEED) and low energy electron microscopy (LEEM) reveals that on the C-face the interface structure prior to graphitization sensitively depends on the preparation conditions. On the Si-face the interface structure prior to graphitization does not change for any of the three environments, however the quality of the graphene formed shows an improvement when prepared in disilane or argon compared to UHV.

When SiC is heated in vacuum the Si atoms preferentially sublimate leaving behind C atoms that rearrange to form graphene. In our prior work we found that compared to the Si-face graphitization of the C-face in vacuum results in thicker graphene films with a larger distribution of thicknesses. A reason for this could be the different nature of the graphene-substrate interface in the two cases. On the Si-face graphene growth is mediated through an interface layer that displays a 6√3×6√3-R30 LEED pattern (this interface layer is also known as the "buffer layer") which acts as a template for graphene growth, while on the C-face no such buffer layer is formed. This buffer layer on the Si-face is known to consist of basically a graphene monolayer, but with some of the carbon atoms bonded to the underlying SiC.

Graphene produced on SiC in vacuum conditions is quite inhomogeneous with small domain sizes and (we refer to an area with a constant thickness of multilayer graphene as a "domain") and numerous pits. On the Si-face it has been found by many research groups including our own that graphitization in an atmosphere of argon results in large monolayer domains with an elimination of pits. The argon decreases the Si sublimation rate, thus increasing the temperature required for graphene formation. The higher graphitization temperature results in an improved morphology of the graphene film. Some researchers use a background of disilane instead of an atmosphere of argon, and in this thesis we report our results for graphitization of SiC in a disilane environment. On the Si-face we find an improvement in the morphology of disilane prepared graphene films compared to those prepared in vacuum, consistent with other researchers. In terms of the interface structure prior to graphitization no difference was found for graphene produced in UHV, argon or disilane. For all three environments the LEED pattern from the interface prior to graphitization displayed 6√3 x 6√3-R30 (6√3 for short) symmetry.

In an attempt to controllably form thin layers of graphene on the C-face we previously tried graphitizing in an atmosphere of argon, however that led to inhomogeneous islands of thick graphene forming over the surface. It was found that due to an unintentional oxidation of the surface during graphitization, the surface became resistant to graphitization. In this thesis we present results for graphitization of the C-face in a background of disilane, which to our knowledge has not been attempted before. We are able to form graphene films that are thin and uniform relative to those prepared in vacuum or argon. We demonstrate that by graphitizing in a background of disilane we avoid the unintentional oxidation that inhibits graphene formation on the C-face in argon.

For C-face samples prepared in disilane, prior to graphitization we observe a √43×√43±7.6° (√43 for short) in situ LEED pattern that has never been observed in vacuum prepared samples. This √43 pattern is found to disappear after air exposure. The ex situ LEEM reflectivity curves of such a disilane prepared sample show unique features not seen in any vacuum prepared sample. By analyzing the LEED pattern and the LEEM reflectivity curves we associate the unusual reflectivity curves we observe on the C-face with a buffer layer, analogous to the 63 layer that form on the Si-face. This buffer layer has the 43 symmetry due to bonding to the underlying SiC, but upon air exposure these bonds are broken (due to oxidation of the SiC) and the layer becomes "decoupled" from the SiC. This decoupling of the buffer layer on the C-face is analogous to what occurs upon oxidation or hydrogenation of the 63 layer on the Si-face. We believe that the √43 layer does not form in vacuum prepared samples due to kinetic limitations but is able to form in Si-rich environments (such as disilane or furnace grown graphene) as the graphitization takes place with the Si sublimation rate closer to equilibrium. The schematic shown below summarizes the differences between Si-face and C-face SiC/graphene interface structures, depending on preparation conditions (vacuum or Si-rich). The right most figure shows the main result of this work, which puts graphene formation on the C-face on a similar footing as for the Si-face since the "buffer" layer provides a template for the graphene growth.

A separate project discussed in this thesis is scanning tunneling microscopy/spectroscopy (STM/S) on epitaxial graphene on the Si-face. Two different studies were performed. In the first study we performed STM/S on a Si-face graphene sample in which a large fraction of the area was covered by a secondary disordered phase. The disordered phase showed a graphene-like spectrum with additional features that could arise from dangling bonds or defects. On the basis of additional data from AFM and Auger electron spectroscopy we argue that this secondary phase is similar to the nanocrystalline graphite (NCG) phase that we observe on C-face samples. In the second study we performed STS on a graphene sample that was functionalized by hydrogen. Functionalizing graphene changes the nature of its bonding and can open up a band gap in it. Our STS results indicate that no band gap opened up in the graphene, however we found the presence of additional states in the spectra that indicate the nature of the bonds in graphene had changed due to the hydrogen functionalization.

In the last study of this thesis we perform LEEM on graphene samples prepared on Cu foil. The samples were made in a chemical vapor deposition (CVD) chamber under different growth conditions. LEEM was used to measure the reflectivity curves and perform selected area diffraction on the samples. The reflectivity curves allowed us to determine the graphene thickness, and the selected area diffraction allowed us to determine the orientation of the graphene and whether it was single-crystal or not.