Automative Quantum Chemical Microkinetic Simulations for Complex Chemical Reactions

    Today's demand for precisely predicting chemical reactions from first principles requires research to go beyond Gibbs' free energy diagrams and consider other effects such as concentration and quantum tunnelling. The present work introduces overreact, a novel Python package for propagating chemical reactions over time using data from electronic structure packages only. The overreact code infers all differential equations and parameters from a simple input that consists of a set of chemical equations and quantum chemistry package outputs for each chemical species. 

    Overreact is not only able to provide insight into concentration dependencies and comparison with experiment, but it also  serves as  a  benchmark  for  hypothesis  testing,  allowing  the  prediction  of chemical behaviour in a computationally affordable and precise way.  Kinetic models in such a framework can be progressively constructed, with a single pathway added at a time until a match with the experiment,  simplifying the workflow,  easing the computational burden and increasing efficiency. Such an incremental modeling protocol is arguably maximally efficient, as a minimal set of reactions lead to more easily rationalisable models.

    Applications that employ systems reported in the literature are presented, including gas-phase eclipsed-staggered isomerization of ethane, gas-phase umbrella inversion of ammonia, gas-phase degradation of methane by chlorine radical, and two solvation-phase reactions. Furthermore, Comments on a simple solvation-phase acid-base equilibrium. We show how it is possible to achieve reaction profiles and information matching experiments.