Application of BP-neural networks in the FOCAL technique

FOCAL is an on-line measurement technique of the imaging parameters of a lithographic tool with high accuracy. These parameters include field curvature, astigmatism, best focus and image tilt. They can be acquired by the least-square algorithm from the alignment positions of the special marks on the exposed wafer. But the algorithm has some intrinsic limits which may lead to a failure of the curve fitting. This will influence the measurement accuracy of the imaging parameters obtained by FOCAL. Therefore, a more reliable algorithm for the FOCAL technique is needed. In this paper, the feed-forward back-propagation artificial neural network algorithm is introduced in the FOCAL technique, and the FOCAL technique based on BP ANN is proposed. The effects of the parameters, such as the number of neurons on the hidden-layer, the number of training epochs, on the measurement accuracy are analyzed in detail. It is proved that the FOCAL technique based on BP-ANN is more reliable and it is a better choice for measurement of the imaging parameters.