Optimization of electrothermal response of GAAFET using Taguchi's approach and an artificial neural network

Abstract

This work aims to propose a numerical optimization of the electrothermal behavior of a gate-all-around field effect transistor (GAAFET). The electrothermal model is composed of semiconductor equations coupled with heat conduction equations. The finite element method was adopted for the discretization of equations describing the electrothermal model. The effects of geometric and electrical parameters such as gate length (Lg), oxide thickness (tox), nanowire radius (R), trap density (Nt), ramp parameter (α),gate voltage (Vg), and drain voltage (Vd) were analyzed based on a Taguchi L27(37) orthogonal array. Optimization of the control factors was performed. Multiple linear and nonlinear regressions, as well as an artificial neural network, were used. Numerical analysis revealed that the overall optimal drain current and peak temperature device reach the maximum drain current with lower temperature. The analysis of variance shows that the maximum contribution (39 % for the drain current and 36 % for Tmax) is occupied by the ramp coefficient (α) while the other parameters have a small contribution (less than 3 %) to optimize the performance of the GAAFET transistor. The statistical analysis reveals that the developed artificial neural network model is efficient for the prediction of the electrical and thermal characteristics of the GAAFET transistor. The results of this study are promising for advancing our understanding of these innovative transistor technologies, enabling them to be efficiently integrated into the next generation of electronic products.