A Chemical Reduction Method based on Pathway Analysis

Numerical simulations are a great tool to study plasma physics. However, the nature of a plasma leads to the significant presence of a particle’s or a molecule’s (ro-vibrational) excited and/or ionized states. For a single atom, this might still be computable in reasonable time, but when a molecule or multiple different atoms are present the computational cost quickly skyrockets. The sheer amount of particles and reactions makes full-fledged plasma simulations infeasible. To still be able to model complex plasmas, chemical reduction must be applied. Chemical reduction decreases the complexity by reducing the number of species and/or reactions taken into account. This is done in a way such that the simulated physics can still be represented accurately and relevant for the application. The goal of chemical reduction is two-fold, on the one hand it achieves computational cost reduction, on the other hand it allows for better analysis of the underlying chemistry. In this work a new mechanism reduction method is developed based on the results of the pathway analysis [24] method. An extension of the pathway analysis implementation [26] within the PLASIMO [18] toolbox is created that allows for automated reduction of chemically complex plasmas in zero-dimensional models. The automated tool limits the required user effort and knowledge of the chemistry to a minimum. The pathway reduction method is capable of removing significant parts of the chemistry, in the form of reactions and species. This is done for both constant input power as well as pulsed input power simulations. Besides taking an extensive look at the effect of removing species and reactions on the accuracy of the reduced model, the reduction method is analysed in a number of ways, including a comparison to a naive method and a timescale analysis of the reduction’s effect on the stiffness of the system. Large steps have been made in developing and testing a new mechanism reduction method based on pathway analysis. An automated reduction tool is established and is a very promising basis for subsequent reduction work. An extensive look at possible future work is taken including a detailed proposal and some initial work for a side-by-side comparison between pathway reduction and the Intrinsic Low-Dimensional Manifold [21] method.