NEURAL NETWORKS IN PLASMA PROCESSING

Over the last few years neural networks have been studied for potentia l applications in plasma processing. The focus of this article will be on two neural network models for complementary metal-oxide-semiconduc tor production. The models were developed with strict statistical cros s-validation and applied to developing a plasma gate etch controller a nd a plasma model of a contact etch process. For a gate etch controlle r, the process has been evaluated in a production environment and show n to improve the process variance and throughput. For a model of a con tact etch process we demonstrate that the model is limited by the inhe rent noise in the training data and that the direct current bias and e tch time are the key control factors that determine the product qualit y at the end of the etch. (C) 1996 American Vacuum Society.