Towards a Complete Plasma Diagnostic System

TOWARDS A COMPLETE PLASMA DIAGNOSTIC SYSTEM by Dong Wu Zhao Doctor of Philosophy in Engineering-Electrical Engineering & Computer Sciences University of California at Berkeley Professor Costas J. Spanos, Chair We have set up a plasma diagnostic system with three sources of signals, OES, RF, and machine signals. CF2 OES lines 275 nm and 321 nm are found to be better than any other signals for poly-etch endpoint detection. In addition, excellent statistical models for wafer state prediction are obtained by linear stepwise regression on all available signals. A data exploration system, based on syntactical analysis, is developed for efficiently browsing of the data archive, allowing users unprecedented flexibility in examining the data both qualitatively and quantitatively. Two case studies of syntactic analysis for diagnostics are presented. Finally, the use of low frequency signals for plasma diagnostics in investigated. The syntactic method for analyzing the signals is proposed. __________________________ Professor Costas J. Spanos Dissertation Committee Chair