Establishing a List of Parameters and Their Empirical Calibration via Simulations of Some Experimental Graphene-Based Transistors

In the actual microelectronics context, transistors with alternative semiconductor materials, using elements from the fourth group, other than silicon or germanium—for example, carbon—have become materials of extreme interest in overcoming certain limitations experienced by traditional MOSFETs. Therefore, graphene field-effect transistors (GFETs) offer the advantage of ambipolar conduction, providing static transfer characteristics dominated by ON current, ION, both at positive gate bias for the electron accumulation channel and at negative gate bias for the hole accumulation channel. Between these two ON states, a Dirac point arises as a minimum current, representing the moment of compensation between electron and hole currents, which essentially defines the OFF current, IOFF. Due to this future development of GFET transistors, it is necessary to propose simulation techniques for optimizing these electronic devices. GFET simulations are present in the literature, but further calibrations are necessary, since graphene does not exist as a defined semiconductor material in the usual Technology Computer-Aided Design (TCAD) libraries. In addition, the properties of each graphene material differ from one experimental realization to another, depending on the technology applied by each author. This article starts with some experimental transfer characteristics measured for two types of fabricated GFET devices. Then, this study presents a few calibration techniques to simulate the graphene layer within GFET, aiming to fit the experimental transfer characteristics with the simulated curves while keeping some optimized technological sizes. The final goal of this article is to establish a list of GFET device parameters linked to the material properties and to estimate them through successive simulations, using empirical calibrations, for each type of experimental graphene material. This algorithm is not a validated procedure; instead, it indicates a set of sensitive GFET parameters, which allow for the adjustment of the shape and order of magnitude of the simulated current with the experimental one for given static characteristics.