With a little help from our (AI) friend: A general transition state sampling method for tropospheric hydrogen abstraction reactions

Abstract

Due to their pivotal role in atmospheric processes, hydrogen abstraction reactions are vital in the study of tropospheric chemistry. In this work, we present an algorithm that generates a large number of chemically sound geometries for the optimization of transition states (TSs) of tropospheric bimolecular H abstraction reactions. The code, which was developed in the early stages with the help of artificial intelligence (AI), automatically detects the active H atoms of the molecule and can be used with the OH, Cl, and NO3 oxidants. Given the ubiquity of H transfer reactions in various fields, the algorithm was designed in general terms to facilitate easy adaptation to other oxidants, thus allowing for a broader range of applications. As a result of its use, a much greater number of TSs is predicted when compared with previous theoretical studies of six oxidation reactions. In addition to improving our understanding of the H-abstraction process, obtaining an increased number of TSs is also fundamentally important in calculating more accurate rate constants and atmospheric lifetimes for volatile organic compounds. The simplicity and significance of such a tool in the context of environmental chemistry should make it appealing to researchers of all backgrounds.