A novel fault detection method for semiconductor manufacturing processes

In this paper, we present a novel fault detection method to address online monitoring problem of semiconductor manufacturing processes. To enhance the fault detection efficiency of existing k-nearest neighbor rule (kNN)-based methods, the principal component analysis (PCA) algorithm is employed to implement data dimension reduction and achieve features of high-dimensional data samples. In addition, to raise the fault detection accuracy for batch processes, the improved kNN algorithm based on the Mahalanobis distance is conducted on features of data samples. The proposed method is evaluated by extensive experiments with industrial examples. The experimental results illustrate great improvements on not only efficiency, but also accuracy. In particular, this method has real potential for monitoring semiconductor manufacturing processes reliably and in time.