Application of Neural Network and Tabu Search to Find the Optimal Parameter of Wire Bonding Process in Semiconductor Packaging

• Main Authors: Kuei Hsien, 陳桂嫻
• Other Authors: Tung-Hsu Hou
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/07312144870693874514

碩士 === 國立雲林科技大學 === 工業工程與管理研究所碩士班 === 91 === With the prosperous development of high tech industry, the trend for electronics products not only has to be light, slimmer, and small, but also has to be multi-functional. The impact of this trend on the semiconductor industry is the in the increase in chip number and wafer size, and decrease in gate length. Therefore, the high density and high pincounts are required in the IC assembly process. In response to this trend, the wire bonding process of IC assembly has to apply the high precision wire bonding technology with fine pitch, and the 50 μm pad pitch will become the main product. Therefore, it is very important to find the optimal wire bonding process parameter of the 50μm pad. This research applies the back propagation nueral network to model the relationship between the wire bonding process parameters and output quality characteristics. Then a revised Tabu Search approach is used to find the optimal parameters for the process. The result of the revised Tabu Search is compared with that of the original Tabu Search. This research finds that the back propagation neural network can effectively model the relationship between the wire bonding process parameters and the output quality characteristics, and the revised Tabu search generates the better optimal process parameters with better efficiency than the original Tabu Search.