Determining the number of graphene layers by Raman-based Si-peak analysis

Raman spectroscopy is the most used and versatile tool in the material characterization of graphene. Wide array of structural information, such as defects, doping, number of layers, stacking order and strain, are simultaneously encompassed in the Raman spectrum. This work focused in determining the number of layers using Raman. The main problem with the current Raman methods using graphene spectrum is that they often saturate and sense the stacking order, being limited to only few-layers of ABA-stacked graphene. This is especially problematic when analysing multilayer chemical vapour deposited graphene, which is often unintentionally randomly stacked.


In this work, the Si-peak analysis was developed to address these fundamental limitations. It relies on the silicon peak originating from the substrate, for which the absorption relates to the number of graphene layers. The normalized Sipeak intensity is obtained simultaneously with the Raman analysis of graphene on SiO2/Si-substrate. Model was tested using pristine graphene on 285 nm and 365 nm oxide, bare 0-400 nm oxide and random stacking. The calculated number of layers agreed with the thickness references and were insensitive to graphene properties unlike other tested Raman-based methods. Model was also robust against random stacking, even near the resonant rotational angle, where graphene Raman band changes become difficult to predict. The Si-peak analysis, being virtually immune to variation within graphene, is a promising non-destructive widely applicable method to determine the number of even arbitrarily stacked graphene layers up to 100.