QCxMS2 – a program for the calculation of electron ionization mass spectra via automated reaction network discovery

Abstract

We present a new fully-automated computational workflow for the calculation of electron ionization mass spectra by automated reaction network discovery, transition state theory and Monte-Carlo simulations. Compared to its predecessor QCxMS [S. Grimme, Angew. Chem., Int. Ed., 52, 6306–6312] based on extensive molecular dynamics (MD) simulations, QCxMS2's more efficient approach of using stationary points on the potential energy surface (PES) enables the usage of accurate quantum chemical methods. Fragment geometries and reaction paths are optimized with fast semi-empirical quantum mechanical (SQM) methods and reaction barriers are refined at the density functional theory (DFT) level. This composite approach using GFN2-xTB geometries in combination with energies at the ωB97X-3c level proved to be an efficient combination. On a small but diverse test set of 16 organic and inorganic molecules, QCxMS2 spectra are more accurate than ones from QCxMS yielding on average a higher mass spectral matching of 0.700 compared to QCxMS with 0.622, and is more robust with a minimal matching of 0.498 versus 0.100. Further improvements were observed when both geometries and energies were computed at the ωB97X-3c level, yielding an average matching score of 0.730 and a minimal score of 0.527. Due to its higher accuracy and robustness while maintaining computational efficiency, we propose QCxMS2 as a complementary, more reliable and systematically improvable successor to QCxMS for elucidating fragmentation pathways and predicting electron ionization mass spectra of unknown chemical substances, e.g., in analytical chemistry applications. If coupled to currently developed improved SQM methods, QCxMS2 opens an efficient route to accurate, and routine mass spectra predictions. The QCxMS2 program suite is freely available on GitHub.