High Resolution Microscopy of SiO2 and the Structure of SiO2-Supported Graphene

Abstract

Graphene has attracted great interest due to its exceptional electrical, mechanical, and chemical properties since its discovery in 2004. Since its first realization, the substrate of choice for graphene exfoliation has been Si wafer with approximately 300 nm thick SiO2 dielectric layer, because it allows 1) direct optical detection of monolayer flakes, and 2) a convenient back gate with dielectric for controlling carrier density in the graphene. However, the amorphous structure of SiO2 and its associated surface roughness has led to ongoing controversy in determining the structure of SiO2-supported graphene. The conductivity of graphene allows scanning tunneling microscopy (STM) to be used to measure its topography, generally allowing its structure to be atomically resolved. In contrast, the insulating SiO2 must be probed with atomic force microscopy (AFM), and this is often done using ambient tapping-mode AFM. STM measurements of graphene on SiO2 generally show greater roughness and finer corrugation than is seen in AFM measurements of SiO2, and this has been interpreted as evidence for “intrinsic” corrugation of the graphene. However, when the energetics of adhesion and elasticity are considered, the idea of intrinsic structure becomes quite controversial for graphene supported on a substrate. Here we show that UHV non-contact AFM (NC-AFM) measurement of SiO2 reveals structure unresolved in previous measurements, and shows both greater roughness and smaller lateral feature size than seen for graphene measured by STM. High-resolution measurement of the SiO2 topography enables an analysis based on the energetics of graphene bending and adhesion, showing that the graphene structure is highly conformal to the SiO2 beneath it. The topographies reported here contrast the atomically-flat crystalline surfaces used in benchmark NC-AFM measurements. They pose unique challenges for measurement resolution, and highlight the very different physical mechanisms which determine resolution in STM vs. NC-AFM. We discuss these issues and our recent efforts at quantitative modeling of the imaging process, with particular focus on the role of van der Waals forces and their contribution to the image signal.