Chemical Structural Features Affecting the Soil Sorption of Sulfonylurea Herbicides

Abstract

ABSTRACT The contamination of groundwater by herbicides depends on their mobility in soil, which is related to the organic carbon to water partition coefficient, KOC. Because the log KOC values of widely used sulfonylurea herbicides do not correlate with easily accessible log P estimations, multivariate image analysis (MIA) descriptors were employed to build structure-property models to classify these compounds according to their soil sorption capability. In addition, a deeper analysis based on MIA contour maps of PLS regression coefficients and variable importance in projection scores was performed to obtain insight into the chemical features responsible for the log KOC behavior. A multiple linear regression model obtained from selected descriptors demonstrated high predictability (r2 = 0.95, q2 = 0.84, and r2 pred = 0.71) and also provided chemical insight into the log KOC values. It has been found that triazine rather than pyrimidine derivatives are less prone to leach out, but substituents attached to the sulfonyl group play an important role in modulating the log KOC values. Consequently, useful guidance for environmental risk assessment is provided for the development of new sulfonylurea herbicides.