A two-phase research project will be undertaken for which the objective is to use optical emission spectroscopy (OES) and in-situ tool data describing gas flow rates, rf power, temperature, pressure and dc bias voltage for real-time diagnosis and prognosis of malfunctions in plasma etch equipment. The diagnostic system will use evidential reasoning principals and neural networks to identify malfunctions from the process "signatures" inherent in the OES and tool data. The project will also involve developing a prognostic system capable of predicting potential equipment malfunctions in the processing of integrated circuit (IC) wafers. The prognostic system will use neural networks to construct time series models of the tool data and generate a malfunction alarm when the model predictions indicate the evolution of suspicious trends in the data. The detection, classification, and prediction capabilities could result in dramatic improvements in overall equipment effectiveness, reduction in downtime, and reduction of scrap in semiconductor manufacturing equipment. The successful development and implementation of the developed system can lead to significant productivity enhancement, which will be of great value to the semiconductor industry.
