Statistical analysis and optimization of igbt manufacturing flow

Abstract

The use of computer simulation, design and optimization of power electronic devices formation technological processes can significantly reduce development time, improve the accuracy of calculations, choose the best options for implementation based on strict mathematical analysis. One of the most common power electronic devices is isolated gate bipolar transistor (IGBT), which combines the advantages of MOSFET and bipolar transistor. The achievement of high requirements for these devices is only possible by optimizing device design and manufacturing process parameters. Therefore, an important and necessary step in the modern cycle of IC design and manufacturing is to carry out the statistical analysis. Procedure of the IGBT threshold voltage optimization was realized. Through screening experiments according to the Plackett – Burman design, the most important input parameters (factors) that have the greatest impact on the output characteristic were detected. The coefficients of the approximation polynomial adequately describing the relationship between the input parameters and investigated output characteristics ware determined. Using the calculated approximation polynomial, a series of multiple, in a cycle of Monte Carlo, calculations to determine the spread of threshold voltage values at selected ranges of input parameters deviation were carried out. Combinations of input process parameters values were determined randomly by a normal distribution within a given range of changes. The procedure of IGBT process parameters optimization consist a mathematical problem of determining the value range of the input significant structural and technological parameters providing the change of the IGBT threshold voltage in a given interval. The presented results demonstrate the effectiveness of the proposed optimization techniques.