Modern computational quantum chemistry has the ability to characterize elementary molecular reaction steps at reasonably low cost. Using these techniques, one can hypothesize mechanisms for entire reaction sequences by chaining together several elementary steps and obtain reasonably accurate thermodynamic and kinetic descriptions. The success of this procedure, however, is highly dependent on chemical intuition: quantum chemical methods can only evaluate the energetics of the mechanism, but not the likelihood that the right mechanism was studied. In this presentation, I will show a new way to avoid this problem and automatically characterize reaction mechanisms. This technique starts with a reactant complex, generates many new structures that are one elementary step away (throwing step), and then uses string methods to locate the reaction path connecting each pair (rope relaxation). After kinetically feasible reaction steps are verified, the procedure is repeated until an entire reaction mechanism is completed. This technique avoids the requirement for significant amounts of chemical intuition as well as expensive searches over potential energy surfaces. The method is embarrassingly parallel, allowing full exploitation of modern computing clusters. A full description of the method as well as several examples will be given in the presentation.

Catching transition states by throwing ropes to stable intermediates

246th National Meeting (2013)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.