Designing pulsed plasma etch processes by genetic algorithm

The complexity of modern semiconductor device fabrication has caused the parameter space of plasma process design to balloon to levels, which are untenable to navigate without algorithmic guidance. The degrees of freedom provided by the so-called process knobs alone present a substantial optimization challenge, one which is very costly to approach from a purely experimental perspective. Detailed simulations of plasma processes are often limited by sparse fundamental knowledge of species-surface site interactions, as well as the computational and experimental effort required to elucidate these relationships. Alternatively, phenomenological models can be used to establish guiding principles without the need for large data sets. Here, we describe the use of a genetic algorithm (GA), coupled with elementary plasma etch simulations, to optimize the time-ordering of plasma process steps. The goal of this approach is to provide rapid, intuitive guidance to reduce the size of the search space without the need to obtain or develop dense data sets. Our work demonstrates that GAs are capable of reliably finding solutions that optimize user-specified metrics. For example, continuous wave solutions are produced when a maximum vertical etch rate is desired, while pulsed solutions are produced when metrics such as mask selectivity are considered.