Virtual metrology modeling for plasma etch operations

The objective of this paper is to present the utilization of information produced during plasma etching for the prediction of etch bias. A plasma etching process typically relies on the concentration of chemical species in reaction chambers over time, where each concentration depends on chamber pressure, gas flow rate, power level and other chamber and wafer properties. Plasma properties, as well as equipment factors are nonlinear and vary over time. In this work, we will use various statistical techniques to address challenges due to the nature of plasma data: high dimensionality, collinearity, overall non-linearity of system, variation of data structure due to equipment condition changing, etc.