Department of Industrial Engineering and Engineering Management National Tsing Hua University 101, Kuang-Fu Road, Section 2, Hsinchu, Taiwan, 300, R.O.C.
Macronix International Co., Ltd.
As global competition continues to strengthen in the semiconductor industry, wafer fabs have been placing increasing importance on increasing die yield and reducing operation costs. Because of automatic manufacturing and information integration technologies, an increasingly large amount of raw data has been accumulated from various sources automatically or semi-automatically from day to day. Mining potentially useful information from large such database becomes very important in both research and application. However, little research has been done on manufacturing data of high-tech industry. In particular, due to the complex fabrication processes and the high cost of defects, using data mining approach to diagnosing defects in semiconductor manufacturing is a critical issue. We constructed a conceptual framework for data mining, proposed two methods for mining WAT data, and then applied them empirically in a fab. The results show the practical viability to assist the domain engineer in narrowing possible causes of manufacturing defects. This study concludes with discussions and remarks on future research directions.
Keywords:data mining, semiconductor manufacturing data, defect diagnosis, decision analysis, decision tree
Cite this article as: Chen-Fu Chien, Ting-Hao Lin, Cheng-Yung Peng and Shao-Chung Hsu, "Developing Data Mining Framework and Methods for Diagnosing Semiconductor Manufacturing Defects and an Empirical Study of Wafer Acceptance Test Data in A wafer FAB," Journal of the Chinese Institute of Industrial Engineers, 18, 37 -48 (2001).