Parameter Extraction for Advanced MOSFET Model using Particle Swarm Optimization

In this paper, parameter extraction for PSP MOSFET model is demonstrated using Particle Swarm Optimization (PSO) algorithm. I-V measurements are taken on 65 nm technology NMOS devices. For the purpose of comparison, parameter extraction is also carried out using Genetic Algorithm (GA). It is shown that PSO algorithm gives better agreement between measurements and model in comparison to GA and with less computational effort. A novel “memory loss (ML)” operation is introduced in the PSO algorithm for the first time, which further improves algorithm efficiency. The PSO algorithm with ML operation has taken 81 minutes on average to extract parameters of two devices of channel lengths, 70 nm and 1 μm.