Method for Semiconductor Process Optimization Using Functional of Spatial Variations and

We present a methodology for determining the opti- mal equipment settings for a processing step based on experiment designs and model-based optimization. Traditional techniques suffer from the choice of optimization metrics that fail to take into account the spatial effects on the wafer. The process yield becomes increasingly sensitive to spatial variations as the size of the wafer and processing complexity increase and the ratio of the size of the process chamber to the wafer decreases. Furthermore, some of the metrics like nonuniformity and selectivity, as defined today, cannot be modeled with the desired accuracy. In this article we present a set of alternate metrics to alleviate the above- mentioned problems. We include the results of our study of a plasma-assisted silicon nitride etch step.