Finite-element modeling and optimization-based parameter extraction algorithm for NPT-IGBTs

Abstract

A finite-element, physics-based, non-punch-through (NPT) insulated gate bipolar transistor (IGBT) model is presented in this paper. The model's core is based on solving the ambipolar diffusion equation through a variational formulation, resulting in a system of ordinary differential equations (ODEs). The approach enables an easy implementation into a standard SPICE circuit simulator. The resulting system of ODES is solved as a set of (current controlled) RC nets describing charge carrier distribution in a low-doped zone. Other zones of the device are modeled with classical methods. This hybrid approach describes the device's dynamic and static behavior with good accuracy while maintaining low execution times. As physics-based models need a significant number of parameters, an automatic parameter extraction method has been developed. The procedure, based on an optimization algorithm (simulated annealing), enables an efficient extraction of parameters requiring some simple device waveform measurements. Experimental validation is performed. Results prove the usefulness of the proposed methodology for the efficient design of power circuits through simulation.