Fast species ranking for iterative species-oriented skeletal reduction of chemistry sets

Abstract

Abstract A fast algorithm is developed for ranking the species in a chemistry set according to their importance to the modeled densities of user-specified species of interest. The species ranking can be constructed for any set of user-specified plasma conditions, but here we focus predominantly on low-temperature plasmas, with gas temperatures between 300 and 1500 K covering the typical range of ICP and CCP plasma sources. This ranking scheme can be used to acquire insight into complex chemistry sets for modeling plasma phenomena or for a species-oriented reduction of the given chemistry set. The species-ranking method presented is based on a graph-theoretical representation of the detailed chemistry set and establishing indirect asymmetric coupling coefficients between pairs of species by the means of widely used graph search algorithms. Several alternative species-ranking schemes are proposed, all building on the theory behind different flavors of the directed relation graph method. The best-performing ranking method is identified statistically, by performing and evaluating a species-oriented iterative skeletal reduction on six, previously available, test chemistry sets (including O2–He and N2–H2) with varying plasma conditions. The species-ranking method presented leads to reductions of between 10 and 75% in the number of species compared to the original detailed chemistry set, depending on the specific test chemistry set and plasma conditions.