Noise reliability evaluation of nanoscale metal-oxide-semiconductor field-effect transistors based on Monte Carlo simulation and soft computing

In recent years, Monte Carlo simulation has been increasingly applied as a particle-based technique for microscopic device reliability analysis. This stochastic method solves physical and mathematical problems through statistical sampling and has become instrumental in noise characterization of nanoscale field-effect transistors. However, prior simulations of channel current noise were predominantly one-dimensional and neglected quantum effects arising from dimensional scaling. Therefore, this paper proposes a two-dimensional and three-dimensional Monte Carlo simulation current noise method for metal-oxide-semiconductor field-effect transistors (MOSFET) at the nanoscale. This method is based on the list method and an approximate method for quickly processing anisotropic scattering final states; it uses the successive over-relaxation (SOR) iteration method to iteratively solve the Poisson equation for the divided grid areas, and each node applies a multi-grid method to solve the Poisson equation at each grid point within the sub-area. At the same time, the effective potential method is used to quantum-correct each grid point, and the simulation tracks the accelerated flight and random scattering motion of a group of charged particles, keeping the load on all nodes always balanced. Benchmark results demonstrate that the Monte Carlo simulation framework not only reduces overall CPU time requirements but also validates the effectiveness of predicted MOSFET noise disturbance patterns. The conclusions exhibit excellent agreement with established theories and experimental data, thereby facilitating design and reliability analysis of nanoscale field-effect transistors