MC/DD study of metal grain induced current variability in a nanoscale InGaAs FinFET

Abstract

The on- and off-current variability due to TiN metal grain workfunction fluctuations in a 10.4 nm gate length In0.53Ga0.47As FinFET is analysed using two in-house simulation tools based on the finite element method: a 3D Drift-Diffusion device simulator and a 3D ensemble Monte Carlo simulator, that include quantum-corrections through the density gradient approach. The Id-Vg characteristics have been compared in the sub-threshold region against ballistic NEGF simulations, showing an excellent agreement. Monte Carlo simulations, considering a 〈100〉 channel orientation, show a larger on-current variability, over a 120% increase, compared with the results from Drift-Diffusion simulations. In this study, three different metal grain sizes (10, 7 and 5 nm) have been analysed. We have observed that the underestimation of the variability when using Drift-Diffusion simulations is increasing with a reduction in the grain size.