(134c) Automated Kinetics for Low-Temperature Oxidation Chemistry

We have recently developed an automated kinetics workflow code called KinBot,^{1, 2} which is able to explore and characterize relevant stationary points on reactive potential energy surfaces (PESs) of gas-phase chemical reactions. It casts the results in a form that readily allows the calculation of pressure- and temperature-dependent rate coefficients via external master equation (ME) solvers.

In this talk we are going to demonstrate the capabilities of the code and its new features on reactions that are important for the low-temperature oxidation of cyclopentane. We use KinBot to explore the cyclopentyl + O₂ and four cyclopentyl + O₂ + O₂ PESs up to CCSD(T)-F12/cc-pVTZ-F12//M06-2X/6-311++G(d,p) level, amended by off-line CASPT2-level calculations for the barrierless O₂ addition channels. We also automatically evaluate the uncertainty bounds of the calculated rate coefficients, which are originating from the uncertainties of the large number of ab initio parameters that enter the ME calculations.

Our calculations show that the key reaction pathways in low-temperature cyclopentane oxidation include internal H-transfer, cyclic-ether-formation and HO₂-elimination pathways, as shown in Scheme 1. We find that ring-opening pathways do not play a significant role, instead, ring-strain drives product branching in several cases. We compare the predictions of a ME-based kinetic model to experimental branching ratios and kinetic traces measured in low- and high-pressure cells using multiplexed photoionization mass spectrometry (MPIMS). The comparison spans the 10-7600 Torr pressure and 400-700 K temperature range, and includes propagated uncertainties. Analysis and interpretation of the MPIMS data are also aided by the automated exploration of cationic PESs to help identify species that do not form stable parent ions and instead fragment upon photoionization in the experiments.

1. R. Van De Vijver, A. L. Dewyer and J. Zádor, KinBot 2.0 v. https://github.com/zadorlab/KinBot, 2019.
2. R. Van de Vijver and J. Zádor, Comp. Phys. Comm., 2020, 248, 106947.