ARTIFICIAL NEURAL-NETWORK EXPONENTIALLY WEIGHTED MOVING AVERAGE CONTROLLER FOR SEMICONDUCTOR PROCESSES

The linear exponentially weighted moving average (EWMA) controller has been shown to improve run-by-run process control for approximately li near processes that are subject to shifts or persistent drifts in the presence of noise. This work addresses the inability of the linear EWM A controller to adequately control processes that are poorly represent ed by such models. This issue is important to the success of the EWMA controller in semiconductor manufacturing where processes may be poorl y approximated with linear process models.-We address this issue by ou tlining an extension of the EWMA controller that utilizes an artificia l neural network (ANN) process model. The ANN model is dynamically upd ated using an EWMA of the biases in the ANN output layer. Recipe gener ation takes place by optimizing around the dynamic ANN model. We show that this framework improves on the linear EWMA controller for control ling higher order processes. Simulations show that this controller pro vides stable control for higher order processes that cause the linear EWMA controller to become unstable. In addition, we suggest that this architecture is robust in the face of model error and noise. This syst em allows the basic property of the EWMA controller, improved control with minimal added process noise, to be extended for use in higher ord er semiconductor processes. (C) 1997 American Vacuum Society.