Distributed-based Hierarchical Clustering System for Large-scale Semiconductor Wafers

In this paper, we propose a Distributed-based Hierarchical Clustering System for Large-Scale Semiconductor Wafers (DHCSSW). By applying the bigdata framework to existing hierarchical clustering algorithm, the proposed system makes it feasible to cluster large-scale wafers with up to 320,000 wafers. To verify the performance of our approach, we used simulated wafer maps. The experimental results show that our system outperformed the existing hierarchical clustering in processing large-scale wafers, suggesting that currently used hierarchical clustering is insufficient in analyzing large-scale wafer maps. In addition, some failure patterns, which the existing approach is not able to detect, can be found with the DHCSSW. We anticipate that the DHCSSW will contribute in identifying the failure patterns in large-scale semiconductor wafers.