Ab initio study of chain branching reactions in the combustion and atmospheric degradation mechanisms of hydrocarbons

Alexander C Davis, Purdue University


The increased interest in energy efficiency and the shift towards employing alternative fuels prompts the need for a deeper understanding of the chemistry of traditional and alternative fuel. Traditional fuels are comprised primarily of branched, unbranched and cyclic alkanes. During both their combustion and atmospheric decomposition mechanisms, these hydrocarbons form several key radical species including branched alkyl, n-alkyl, s-alkyl, alkylperoxy, alkoxy, and alkylcycloalkyl radicals. Each of these intermediate species can undergo a unimolecular isomerization reaction involving the migration of a hydrogen atom (H-migration) from one location along the chain to the radical site. These H-migration reactions essentially result in the redistribution of the radical site and lead to significant chain branching in the reaction mechanism. Although these systems have been studied and characterized previously, both experimentally and theoretically, the majority of the experimental design and analysis of those studies relies on a conceptual model that may not accurately account for the unique differences in the radical reactions being modeled. To assess the validity of the conceptual model, and various estimation methods used for H-migration reactions in hydrocarbon-based fuels, high level ab initio studies are performed to determine the kinetic parameters associated these reactions. To ensure completeness of the analysis, all H-migrations available to the eth-1-yl through oct-4-yl, 2-methylprop-1-yl through 6-methylhept-1-yl, methoxy through heptoxy, and ethylperoxy through hexylperoxy radicals, representing over 450 reactions, are investigated. The resulting ΔHrxns, activation energies, A-factors, tunneling coefficients, and overall rate constants are analyzed for patterns and trends, over a wide temperature range. Where found, explanations for possible causes for the patterns are provided. As a result, the failure of key assumptions employed in the traditional conceptual model is uncovered, and modifications to it are proposed. The results of this work will help to improve current understanding, and guide future development of fuel system studies and computer models.




Francisco, Purdue University.

Subject Area

Atmospheric Chemistry|Organic chemistry

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server