Studying organic reaction mechanisms using quantum chemical methods requires from the researcher an extensive knowledge of both organic chemistry and first-principles computation. The need for empirical knowledge arises because any reasonably complete exploration of the potential energy surfaces (PES) of organic reactions is computationally prohibitive. Taking inspiration from the field of artificial intelligence (AI), we developed a novel approach, termed Heuristically-Aided Quantum Chemistry (HAQC), that abstracts the empirical knowledge in terms of chemical heuristics---simple rules guiding the PES exploration---and combines them with structure optimizations using quantum chemical methods. In order to predict the most likely reaction mechanisms, HAQC first performs trial moves based on the heuristic rules (bond breaking and bond making in the simplest case) and subsequently accepts or rejects these moves using molecular energies from quantum chemical calculations. HAQC uses a consistent set of heuristic criteria to classify the proposed reaction paths of polar reactions (substitutions, additions, and eliminations) and pericyclic reactions (cyclizations, sigmatropic shifts, and cycloadditions) as kinetically feasible or infeasible. This approach accelerates mechanism prediction and mechanism discovery by avoiding costly transition-state optimizations and knowledge-based approaches. The flexible framework of HAQC opens intriguing opportunities to discover new reaction routes, to develop new approaches in drug and materials discovery, and to systematically apply machine learning techniques to organic synthesis.

Exploring reaction mechanisms with heuristics-aided quantum chemistry (HAQC)

254th National Meeting (2017)

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