REAL-TIME, IN-SITU ELLIPSOMETRY SOLUTIONS USING ARTIFICIAL NEURAL-NETWORK PREPROCESSING

Development of real time, in-situ monitoring and control of thin film depositions using ellipsometrv requires both rapid data acquisition an d rapid processing. We recently developed a numerical solving method f ast enough to keep pace with data acquisition. Briefly, the method use s a very fast artificial neural network (ANN) to provide initial estim ates to a slower, more accurate variably damped least squares (VDLS) a lgorithm. The work here addresses a key question raised in the prior w ork: how the solution workload should best be shared between the ANN a nd VDLS for fast, accurate solutions. For Ni deposited on BK7 (borosil icate crown glass) substrates, ANN accuracy of 10% in d1 and d2, 0.1 i n n1 and 0.1 in k1 resulted in solutions generally under 10 iterations . Training of a network using 2000 data over 2000 presentations was su fficient to achieve this accuracy. Iteration contour plots of VDLS per formance combined with ANN target plots provided the necessary informa tion to determine the accuracy values required for proper operation wi th Ni on BK7 data.