QSPR modeling of adsorption coefficient Koc of alkyl(1-phenylsulfonyl) cycloalkane-carboxylates on soil and sediments using MLSER model and ab initio

Based on the modified theoretical linear solvation energy relationship (MTLSER) model and quantum chemical descriptors computed at HF/STO-3G, HF/LANL2DZ, B3LYP/LANL2D and B3LYP/6-31G* levels, respectively, different quantitative structure-property relationship (QSPR) models of adsorption coefficient (Koc) on soil and sediments of 28 alkyl(1-phenylsulfonyl) cycloalkane-carboxylates were developed. It is suggested that all the eight models developed in the present study have good correlation and relatively small error, in which the model with three variables, i.e. the lowest unoccupied molecular orbital (ELUMO), the most negative formal charge (q−) and molecular polarizability (α) as descriptors obtained from B3LYP/LANL2DZ level, exhibited the best optimum correlation (r2=0.9650, q2=0.9446), larger t values and the smallest error. It was also found that molecular polarizability is the most significant factor influencing log Koc, and the value of log Koc increases with increasing ELUMO, q− and α. The result showed that MTLSER models achieved from ab initio calculation are better than that from semi-empirical AM1 method.