Can the outcome of an aromatic edge migration be predicted on-the-fly?

Kinetic Monte Carlo (kMC) simulations of polycyclic aromatic hydrocarbon (PAH) growth require on-the-fly evaluation of equilibrium constants (K_eq) for aromatic edge migration reactions. Here we investigate whether K_eq values for five- and seven-membered ring migration reactions can be predicted from molecular structural descriptors with sufficient accuracy for kMC use. We show that no single structural index correlates adequately with K_eq, but that a combination of complementary descriptors-including HOMA-based aromaticity indices, graph-theoretical measures, and C-atom ring environment contributions-yields significantly more accurate correlations. We further introduce parametrization based on BOIA, a bond-order-based analogue of HOMA, and demonstrate that supplementing PM7/MOPAC energies with BOIA-derived structural descriptors and a group-additivity scheme yields a correlation coefficient of 0.95, with the test set quality closely matching that of the training set. These results suggest that the PM7/MOPAC approach is a promising candidate for on-the-fly K_eq prediction in kMC simulations of PAH growth.