Modeling and robust control of a plasma deposition process

Current control practices within DC deposition processes concentrate on tight control over individual parameters e.g., current levels and gas flow rates, and overall process control is currently nonexistent. The modeling and robust control of the system in this paper, relating the current reference voltage to the plasma stagnation pressure, provided an initial step in developing process control. Data consisting of an input voltage and output stagnation pressure were collected about a nominal operating point. ARMAV models of increasing order were fitted to the data until the adequate system model was obtained. The modeling procedure resulted in a sixteenth order difference equation relating input to output. The model was reduced to a simple first order system by identifying the switching frequency of the power supply and arc-restrike phenomenon as the source of six complex conjugate pairs and identifying a single real root dominating the remaining response. The power supply inductance was credited as the source of the single root characterizing the system. A minimum mean squared control strategy was employed do develop a control equation for the reduced system model. While ultimately a control strategy would incorporate all the manipulable inputs and particle state parameters, the current/pressure modeling and control scheme illustrated the potential to develop real-time robust control of the deposition process.