Graphene attracted a great deal of public interest because of its novel electrical characteristics, structural integrity, high electron mobility, and most important for my work, it was shown to be a room temperature ballistic transport material when it is epitaxially grown on SiC steps. However, there have been inconsistencies between graphene research groups studying the electronic properties of zig-zag (ZZ) edge graphene nanoribbons (GNR). While 2-probe variable geometry measurements demonstrate room temperature ballistic transport along ZZ-edge GNR (Baringhaus, 2014), direct band structure measurements using angle resolved photoemission spectroscopy (ARPES) find that conducting graphene does not grow on the 4H-SiC ZZ-edge sidewalls (Nevius, 2016). I have shown that these different results are due to the starting SiC polytype used as the graphene growth substrate. It has been typically assumed that the SiC polytype is not relevant to graphene growth on SiC, but the stability of SiC steps from which ZZ-edge ribbons grow is very different between polytypes (Nordell, 1999). The choice of SiC polytype determines whether or not ZZ-edge graphene ribbons are strongly or weakly bonded to the SiC trench sidewall. ZZ-edge ribbons grown on 4H-SiC are so strongly bonded to the substrate that there is no evidence of a linear Dirac cone in ARPES unless the graphene-substrate bonds are broken using a hydrogen intercalant. The resulting passivated ribbon is wide, spanning nearly the entire sidewall. In contrast, ZZ-edge ribbons grown on 6H-SiC readily show a modified Dirac cone with a pair of edge states close to the Fermi level. Analysis of the facet Dirac cone shape and valence bands imply that the sidewall graphene on 6H-SiC forms many narrow ribbons, unlike those grown on the 4H polytype. This is confirmed using STS measurements on the 6H-ZZ ribbons. Furthermore, 2-point transport measurements show that only the 6H-ZZ sidewall ribbons are ballistic conductors. If there is to be a carbon-based electronic revolution, understanding the characteristics of GNR grown on different polytypes and with different orientations will provide the foundation for that platform. In addition to my ribbon studies, I have also looked at the effects of a gate oxide grown with atomic layer deposition (ALD) on the semiconducting graphene buffer layer (BL) on SiC(0001). It is known that gate oxides do not strongly interact with the metallic graphene layer that grows above the BL, but the BL itself is highly reactive. The complex carbon-substrate bonding structure of this layer opens a band gap. Core-level XPS and valence-level ARPES show that the BL partially delaminates from the SiC after a thin aluminum oxide film is deposited, making it metallic rather than semiconducting. While voltage gating the BL is essential to produce flat graphene devices, I show that ALD oxides are not an ideal method because they destroy the unique semiconductive character of the BL.
Cite this work
Researchers should cite this work as follows:
MATIN Development Team