Decision Support for IC Molding Parameter Settings Using Grey Relational Analysis and Neural Network

In order to be competitive in the semiconductor manufacturing industry, quality improvement and yield enhancement have received increasing attention. The research focuses on the molding process of Integrated Circuit (IC) assembly. The defects often occurred in molding process include hole, vein, crack, and floss. In order to raise the yield of molding process, the study applies the Taguchi method combined with grey relational analysis to find out the most appropriate molding parameters with multiple quality characteristics. The study further adopts a backpropagation neural network to estimate the optimal process parameters. Results show that the proposed approach can improve the quality of the molding process.