Prediction of soil sorption coefficient of chemicals by molecular modeling technique

Abstract

Soil pollution is one of the pollutions caused by nature and human activities, which are mostly the use of insecticide and chemical fertilizer to make plants grow well and produce more products. However, these chemicals are absorbed by soil. In this research work, properties of chemical compound affecting soil sorption were analyzed, and model that can predict soil sorption coefficient of chemicals (KOC) was examined by using quantitative structure property relationship (QSPR) technique. Totally 1,327 compounds were selected, and their three-dimensional structures were constructed. Semi-empirical PM7 method was used for geometry optimization calculations. A total of 1,536 physicochemical properties were computed. Multiple linear regression was employed to analyze properties of compound affecting soil sorption and to find QSPR model. The obtained model has a good relationship and predictive ability with r2 = 0.762 and q2 = 0.761.