MODELING OF POWER SEMICONDUCTOR DEVICES

Authors

Tanya Kirilova Gachovska, University of Nebraska-LincolnFollow

Date of this Version

Summer 8-10-2012

Document Type

Article

Citation

Gachovska, T.K. Modeling of power semiconductor devices. PhD dissertation, University of Nebraska-Lincoln, 2012

Comments

A DISSERTATION Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For Degree of Doctor of Philosophy, Major: Electrical Engineering, Under the Supervision of Professor Jerry Hudgins. Lincoln, Nebraska: December 2012

Copyright (c) 2012 Tanya Kirilova Gachovska

Abstract

One of the requirements for choosing a proper power electronic device for a converter is that it must possess a low specific on-resistance. The specific on-resistance of a bipolar device is related to the base width and doping concentration of the lightly doped drift region. This means that the doping concentration and the width of the low-doped base region in a bipolar device must be carefully considered to achieve a desired avalanche breakdown voltage and on-resistance. In order to determine the technological parameters of a semiconductor device, a one dimensional analysis is used to calculate the minimum depletion layer width, W_min, for a given breakdown voltage, V_BD, in Si, SiC and GaN p⁺n^-n⁺structures. Further investigation is done to determine the optimum width of associated depletion layers for different blocking voltages to achieve a minimal forward conduction voltage drop.

The complete one-dimensional model for calculation of the minimum depletion layer width, W_min, for a given breakdown voltage, V_BD, of a p⁺n^–n⁺ structure is developed and used to calculate the optimum width of the depletion layer for different blocking voltages to achieve a minimal forward drop. The results show that the calculations of the lightly doped drift region thicknesses, and associated breakdown voltages and forward voltage drops, lead to incorrect solutions when applied to high voltage p⁺n^–n⁺ structures using the simplified model equations for a p⁺n structure. These results also indicate a minimal impurity doping concentration for p⁺n^–n⁺ structures, below which little improvement in breakdown capability can be had. The analysis shows for example that optimization of the doping concentration to minimize V_F in a 5 kV Si diode could result in more than a 12% decrease in the forward drop, while for SiC and GaN this decrease is insignificant; typically less than 1%. Therefore, an optimization of the forward voltage drop by using the optimal doping concentration for corresponding breakdown voltages is necessary for proper design of a Si diode, while for wide band gap material devices this optimization is not necessary.

The second part of the dissertation presents a physics-based model of a SiC BJT and verification of its validity through experimental testing. The Fourier series solution is used to solve the ambipolar diffusion equation (ADE) in the transistor collector region. The model is realized using MATLAB^® and Simulink^®. The experimental results of static operation and also the simulated and experimental results of switching waveforms are given.

From the experimental and simulated results it is concluded that the model represents the static and switching characteristics of the SiC BJT quite well. From the experimental measurement results are calculated the switching losses of the BJT. The differences between simulated and measured switching losses during the turn-on and turn-off are 6.28% and 3.52%, respectively.

Adviser: Jerry L. Hudgins