Principal component based k-nearest-neighbor rule for semiconductor process fault detection

Fault detection and classification (FDC) has been recognized in the semiconductor industry as an integral component of advanced process control (APC) framework in improving overall equipment efficiency (OEE). To explicitly account for the unique characteristics of the semiconductor processes, such as nonlinearity in most batch processes, multimodal batch trajectories due to product mix, the fault detection method based on the k-nearest-neighbor rule (FD-kNN) has been developed previously for fault detection in semiconductor manufacturing. However, because FD-kNN does not generate a classifier offline, it is computational and storage intensive, which could make it difficult for online process monitoring. To take the advantages of principal component analysis (PCA) in dimensionality reduction and FD-kNN in nonlinearity and multimode handling, a principal component based kNN (PC- kNN) is proposed. Two simulated examples and an industrial example are used to demonstrate the performance of the proposed PC-kNN method in fault detection.