A probabilistic approach to reaction coordinate and rate constant modeling applied to epoxide ring-opening reactions

Abstract

The study will utilize a probabilistic reaction modeling method for ring-opening reactions of epoxide. In particular, to elucidate the reaction mechanism by the methods presented, focus will be placed on the nucleophillic attack of ethylene oxide by ammonia and its anion. This focus was chosen because of the potential to gain significant advantage in computational intensity required to model the epoxy-amino macromolecular curing reactions and resulting thermochemical and physical properties of the cured resin. The method employed utilizes the combinatorial probability that 1. Two molecules will approach a transition state with sufficient energy to drive reaction 2. Any reaction will occur for a given penetration into the potential energy surface. The concept of a transition state is relaxed to allow a dynamic probability that any reaction will proceed given a position on the intrinsic reaction coordinate (IRC) rather than searching for a specific transition state of theoretical reaction probability. 3. The reaction that occurs yields a desired stable or semi-stable molecular complex This study will focus on identifying possible stable and semi-stable products and corresponding rate constants. The technique developed here is novel in that it provides an unsupervised method to identify all structures corresponding to minima on the potential energy surface. The technique provides a pragmatic and efficient approach to sample a molecular system for different reaction mechanisms and provides a relative energy requirement to achieve these mechanisms with no presupposition of the mechanism, product, or transition state. It is possible from this data to derive rate constants for a reacting system, however, the rate constant derived for the EO/NH2 molecular system yielded significantly understated reaction probabilities and therefore rate constants.