Capturing aromaticity in automatic mechanism generation software

Abstract

Better modeling of polycyclic aromatic hydrocarbon (PAH) formation has the potential to greatly aid in predicting soot formation. Automatic mechanism generation provides a powerful platform for comprehensively exploring reaction sequences implied by known chemistry and identifying the ones that are important in various systems. Reaction Mechanism Generator (RMG), a software developed on our group, has been successfully used to model a variety of chemistries, but has previously been unprepared to properly handle the challenge of aromatics. To accurately model PAH chemistry, several improvements were needed to core RMG algorithms: more robust resonance structure generation for PAHs and more flexible reaction generation algorithms for aromatics. Resonance structures are an integral part of the reaction generation algorithm, so representation accuracy is an important consideration. Clar structures were introduced as a concise and accurate representation for PAHs to replace Kekulé structures. This combined with a refactoring of the overall resonance algorithm also provided significant performance gains. Another major improvement was enabling reactivity of benzene bonds, which allows much greater flexibility in designing rate rules to differentiate aromatic and aliphatic reactions. As a test case, a model was generated for co-pyrolysis of iodonaphthalene and acetylene and compared with literature data, demonstrating the much-improved ability of RMG in modeling aromatic species.