Prediction of Static Characteristic Parameters of an Insulated Gate Bipolar Transistor Using Artificial Neural Network

Breakdown voltage (BV), on-state voltage (V-on), static latch-up voltage (V-lu), static latch-up current density (J(lu)), and threshold voltage (V-th), etc., are critical static characteristic parameters of an IGBT for researchers. V-on and V-th can characterize the conduction capability of the device, while BV, V-lu, and J(lu) can help designers analyze the safe operating area (SOA) of the device and its reliability. In this paper, we propose a multi-layer artificial neural network (ANN) framework to predict these characteristic parameters. The proposed scheme can accurately fit the relationship between structural parameters and static characteristic parameters. Given the structural parameters of the device, characteristic parameters can be generated accurately and efficiently. Compared with technology computer-aided design (TCAD) simulation, the average errors of our scheme for each characteristic parameter are within 8%, especially for BV and V-th, while the errors are controlled within 1%, and the evaluation speed is improved more than 10(7) times. In addition, since the prediction process is mathematically a matrix operation process, there is no convergence problem, which there is in a TCAD simulation.

Automated summary

This paper proposes a multi-layer artificial neural network framework for predicting critical static characteristic parameters of an IGBT. By accurately fitting the relationship between structural parameters and characteristic parameters, the proposed scheme can generate characteristic parameters accurately and efficiently, while significantly improving the evaluation speed. Compared with TCAD simulation, the scheme achieves acceptable errors and eliminates the convergence problem.