The Q-e scheme is remarkably useful in the interpretation of the reactivity of a monomer in free-radical copolymerizations. To develop reliable quantitative structure-property relationship (QSPR) models for the prediction of Q and e parameters, density functional theory (DFT) calculations were carried out for 60 radicals with structure CH₃C¹H₂—C²HR³•|(C¹H₂＝C²HR³＋CH₃•→CH₃C¹H₂—C²HR³•), at the UB3LYP level of theory with 6-31G(d) basis set. Two sets of quantum chemical descriptors including atomic charges and frontier molecular orbital energies were used to develop artificial neural network (ANN) models for Q and e parameters, respectively. The final optimum ANN models were obtained by adjusting various parameters by trial-and-error, both of them have the network architectures of 3-5-1. The predicted Q and e parameters values are in good agreement with the experimental data, with test sets possessing correlation coefficients of 0.990 (rms＝0.269) for Q and 0.943 (rms＝0.331) for e. Furthermore, the external correlation coefficients of 0.980 for Q and 0.873 for e show that each of the two ANN models has true predictive ability. Thus, the present ANN models are accurate and reliable|calculating quantum chemical descriptors from radicals CH₃C¹H₂—C²HR³•|to predict Q and e values is feasible.

Key words: artificial neural network, free-radical copolymerization, Q-e scheme, QSPR, quantum chemical descriptor, radical