Scalability of Atomic-Thin-Body (ATB) Transistors Based on Graphene Nanoribbons

TL;DR

This paper presents an analytical electrostatic model for atomic-thin-body transistors, applies it to graphene nanoribbon FETs, and discusses how geometry and dielectric choices influence device scaling and performance.

Contribution

It introduces a novel analytical model for electrostatic potential and scaling length in ATB transistors, specifically applied to graphene nanoribbon SB-FETs, highlighting new insights into device scaling limits.

Findings

Effective electrostatic scaling length depends on geometry and dielectric properties.

Thinner top oxide and high-k dielectrics improve electrostatic control.

Sub-10nm GNR SB-FETs can achieve subthreshold swing below 100mV/dec.

Abstract

A general solution for the electrostatic potential in an atomic-thin-body (ATB) field-effect transistor geometry is presented. The effective electrostatic scaling length, {\lambda}eff, is extracted from the analytical model, which cannot be approximated by the lowest order eigenmode as traditionally done in SOI-MOSFETs. An empirical equation for the scaling length that depends on the geometry parameters is proposed. It is shown that even for a thick SiO2 back oxide {\lambda}eff can be improved efficiently by thinner top oxide thickness, and to some extent, with high-k dielectrics. The model is then applied to self-consistent simulation of graphene nanoribbon (GNR) Schottky-barrier field-effect transistors (SB-FETs) at the ballistic limit. In the case of GNR SB-FETs, for large {\lambda}eff, the scaling is limited by the conventional electrostatic short channel effects (SCEs). On the other hand, for small {\lambda}eff, the scaling is limited by direct source-to-drain tunneling. A subthreshold swing below 100mV/dec is still possible with a sub-10nm gate length in GNR SB-FETs.