Run to run control based on adaptive cluster method for high-mix semiconductor manufacturing processes

In the semiconductor manufacturing process, products with different recipe will be produced on the same tool or products with same recipe will be produced on different tools due to the capacity of tools. Therefore, it is difficult to control a mixed run production process containing low production frequency products with high profit. We proposed a control scheme which combines adaptive k-means cluster method and run-to-run control to improve the performance of low frequency products in the mixed run process. Similar products could be classified into the same group adaptively and be controlled by a EWMA controller. The cluster controller is updated by large and small frequency products, so the low frequency products can be improved by the information from similar large frequency products. The proposed control scheme is also evaluated by an industrial application.