Electro-thermal finite element analysis and verification of power module with aluminum wire

As a three-terminal power semiconductor device, an insulated gate bipolar transistor (IGBT) chip is characterized by its fast switching and low switching loss, explaining its extensive use in high-power electrical products. However, Joule heating may be induced during high current conditions, subsequently raising the temperature of the power module. Correspondingly, an uneven temperature distribution may decrease reliability and diminish the life of a power module in long-term operations. Therefore, this study investigates the temperature effect of current crowding and corresponding Joule heating on the IGBT chip. The electro-thermal finite element (FE) behavior of the power module is also investigated, based on a numerical analysis. The current densities of solder, IGBT chip, and aluminum pad are also examined to observe the current crowding effect. Analysis results indicate that the maximum current density occurs at the interface between aluminum pad and the aluminum wire. Additionally, the maximum current density may induce electromigration or failure behavior in long-term operations. These characteristics and temperature distribution can be analyzed during an electrical loading, based on electro-thermal FE analysis. Additionally, the test sample is designed to validate the temperature distribution in the simulation and experimental results. Moreover, the steady temperature of the test sample under electrical loading is determined using an infrared-detecting camera and T3ster. Comparing the simulation and experimental results demonstrates the reliability of the temperature validation method. Furthermore, the temperature from the electro-thermal FE analysis is treated as temperature loading in the thermo-mechanical FE analysis. Efforts are underway in our laboratory to investigate the temperature induced thermal stress. Results of this study demonstrate that numerical analysis can replace and reduce the number of experiments, as well as forecast the electrical, thermal, and mechanical behaviors of the power module efficiently. (C) 2014 Elsevier B.V. All rights reserved.