Kelvin Probe Microscopy Studies of Epitaxial Graphene on SiC(0001)

Abstract

Epitaxial graphene on SiC(0001) presents a promising platform for device applications and fundamental investigations. Graphene growth on SiC(0001) can produce consistent monolayer thickness on terraces and good electronic properties. In exfoliated graphene on SiO₂, random charged impurities in the SiO₂ surface are thought to be the dominant scatterers, explaining the observed transport properties as well as the spatial charge inhomogeneity seen in scanned-probe experiments. In contrast, the scattering mechanisms and charge distribution in epitaxial graphene remain relatively unexplored.

Here I use Kelvin probe microscopy (KPM) in ambient and UHV conditions to directly measure the surface potential of epitaxial graphene on SiC(0001). Ambient-environment KPM on graphene/SiC(0001) shows surface potential variations of only 12 meV. Taken together with transport measurements, the data suggest that the graphene samples in ambient are in the low-doped regime, near the minimum conductivity of roughly 4e<super>2</super>/h.

I am also able to use UHV KPM of graphene/ SiC(0001) to identify the discrete surface potentials of monolayer and bilayer graphene as well as the insulating interfacial carbon layer and bare SiC, correlated with scanning electron micrographs of the same location. The surface potential differences between monolayer and bilayer graphene and between IFL and monolayer graphene are both suggestive of low doping (&le;10<super>12</super> cm<super>-2</super>). The surface potentials of monolayer and bilayer graphene are relatively smooth, while the IFL and bare SiC, in contrast, showed larger variations in surface potential suggesting the presence of unscreened charged impurities present on the IFL that are later screened by the overgrown graphene. I model the potential variations for unscreened and graphene-screened charged impurities using the self-consistent theory of graphene developed by Adam et al. The results show that although surface potential variations are, as expected, larger in the IFL than in graphene, both surfaces display surface potential variations 10-40 times smaller than predicted by theory. While ambient electronic transport data and surface potential steps suggest our samples are only lightly doped (&le;10<super>12</super> cm<super>-2</super>), in a regime dominated by electron-hole puddles, we do not observe these puddles in UHV. The absence of puddles in UHV leaves the source of doping in these samples an open question.