Prediction of the Preferable Mechanism of Nucleophilic Substitution at Saturated Carbon Atom and Prognosis of S_N1 Rate Constants by Means of QSPR

摘要

The nucleophilic substitution reactions constitute the best studied and important class of organic reactions. The ratio of the products of such reactions depends on the competition between the mono- and bimolecular mechanisms. Therefore, prediction of the preferable reaction mechanism and a priori evaluation of corresponding reaction rates are important tasks. Models for prediction of the nucleophilic substitution rate constants have been constructed for many reaction series. However, each of these models considers the effect of different parameters of reagents belonging to the same class of compounds. Such models show high correlation coefficients but cannot be thought of as universal [1, 2]. Therefore, it is necessary to construct a unified model that could allow one to adequately calculate the nucleophilic substitution rate constants, no matter what class of compounds the reagents belong to, and to predict the mechanism of such reactions (which has remained beyond the scope of machine learning problems). Previously [3], we have successfully applied the multicomponent QSPR method to predict the rate constants of S_N2 reactions. The aims of the present work are to create a classification model for determining the preferable mechanism of nucleophilic substitution reactions (S_N1 or S_N2) and to construct a universal model for predicting the S_N1 nucleophilic substitution rate constants.