Modelling Graphene Field-Effect Transistors

Abstract: Today, transistors with 20 nanometer (nm) channel length are in mass production and many researchers believe that we are reaching a limit with downsizing conventionally used silicon metal-oxide-semiconductor field-effect transistors (MOSFETs) [1]. To keep up with the trend of making the transistor smaller, new channel materials are studied, and graphene has come into the spotlight. Graphene became a serious contender mostly due to its high mobility, but other properties such as high velocity saturation and the two-dimensional (2D) nature of the material have gained more attention in recent years [2–4]. The first graphene field-effect transistor (GFET) was reported in 2004, since many transistors with graphene as a channel material have been successfully fabricated [3]. It is important to have accurate simulation models that showcase all the peculiar behaviours of GFETs. Even though several new models with high accuracy, have been presented in recent years, few theoretical explanations exist. This thesis work focuses greatly on the theory behind two different simulation models for GFETs. Several parameter approximations are investigated, with focus on the possibility of showcasing negative differential resistance (NDR). In conclusion, we can see that the drift-diffusion (DD) model show good agreement with data and showcases NDR, while the virtual source (VS) model is more unstable and does not give NDR. I hope this thesis can act as a knowledge base, to facilitate for future simulation models.