Abstract
A method for the numerical modeling of semiconductor devices is presented. The method is based on the classical Fourier approach. Discretization of the differential equations is achieved through trigonometric-series expansions in a Galerkin procedure. The use of fast Fourier transform techniques renders the algorithm computationally effective. The Fourier method assures infinite-order accurate solutions of the strongly nonlinear semiconductor equations. Excellent numerical conditioning and high efficiency of the method offer major advantages for simulation of complex nonlinear VLSI devices, where high accuracy, numerical stability and flexibility are of major concern. Properties of the method are demonstrated on application examples.

This publication has 3 references indexed in Scilit: