Enhancing automated reaction discovery with boxed molecular dynamics in energy space

Abstract

The rare event acceleration method BXDE is interfaced inthe present work with the automated reaction discovery methodAutoMeKin. To test the efficiency of the combined AutoMeKin-BXDEprocedure, the ozonolysis of -pinene is studied in comparison withstandard AutoMeKin. AutoMeKin-BXDE locates intermediates andtransition states that are more densely connected with each other andapproximately 50 kcal/mol more stable than those found with standardAutoMeKin. Other than the different density of edges between thenodes, both networks are scale-free and display small-worldproperties, mimicking the network of organic chemistry. Finally, whileAutoMeKin-BXDE finds more transition states than those previouslyreported for O3 + -pinene, the standard procedure fails to locatesome of the previously published reaction pathways using the samesimulation time of 2.5 ns. In summary, the mixed procedure is verypromising and clearly outperforms the standard simulation algorithmsimplemented in AutoMeKin. BXDE will be available in the next releaseof AutoMekin.