In-Situ Leak Detection of Plasma Processing Chamber Using Neural Network and Optical Emission Spectroscopy

摘要

Monitoring a leak of plasma processing chamber is crucial to maintaining process quality and improving device yield and equipment throughput. A new technique to detect chamber leak is presented. This is accomplished by constructing a neural network model of optical emission spectroscopy (OES). Using OES, a total of 47 patterns were collected. A neural network model developed with OES pattern yielded accuracies of 1.08% and 1.58% for training and testing data, respectively. The appropriateness of neural network model was tested with the remaining OES patterns. The errors for leaky data were considerably larger, enough to be detectable. The performance of leak detection was evaluated more by applying cumulative sum (CUSUM) control chart to statistical mean, model prediction, and major radical intensity. The neural network model-based CUSUM was found the most effective to monitoring chamber leak.View full textDownload full textKeywordsCUSUM control chart, Detection, Leak, Model plasma process, Neural network, Optical emission spectroscopy, Time seriesRelated var addthis_config = { ui_cobrand: "Taylor & Francis Online", services_compact: "citeulike,netvibes,twitter,technorati,delicious,linkedin,facebook,stumbleupon,digg,google,more", pubid: "ra-4dff56cd6bb1830b" }; Add to shortlist Link Permalink http://dx.doi.org/10.1080/10426914.2010.492054