부산대학교 도서관

Two classic experimental designs were used to develop the copper ball bonding process. A two-level factorial design was used to screen five process variables simultaneously and determine whether any of them had a significant effect on the process. A four-factor central composite design was used to optimize and map the bonding process window. The efficiency of using designed experiments allowed the development of the process and the achievement of a major milestone in only 15 experiments. The experimental designs which were used have wide application for process development and optimization. The concepts can be easily taught so that the experiments can become a standard for process technicians to follow. The use and analysis of these experimental designs are described.