Abstract

Quantitative structure–property relationships (QSPRs) have been developed and assessed for predicting the reorganization energy of polycyclic aromatic hydrocarbons (PAHs). Preliminary QSPR models, based on a combination of molecular signature and electronic eigenvalue difference descriptors, have been trained using more than 200 PAHs. Monte Carlo cross-validation systematically improves the performance of the models through progressive reduction of the training set and selection of best performing training subsets. The final biased QSPR model yields correlation coefficients q2 and r2 of 0.7 and 0.8, respectively, and an estimated error in predicting reorganization energy of ±0.014 eV.