ABSTRACT
Forty pesticides were selected in function of their chemical families and their physico-chemical properties to represent a wide range of pesticide properties. Adsorption of these pesticides was studied on two soils by batch experiments. The two soils differed largely in organic matter and calcite contents. Distribution coefficient Kd was determined for each pesticide on the two soils. Adsorption was higher for the soil having the highest organic matter content and the lowest calcite content. In order to identify pesticide properties governing retention, eight molecular descriptors were determined from three-dimensional (3D) structure of molecules. Class-specific quantitative structure properties relationship (QSPR) soil adsorption models using one and two parameters were developed from experimental Kd. Three properties seemed to influence most retention of pesticides: hydrophobicity, solubility, and polarisability. Models combining these properties were suggested and discussed.
Subject(s)
Pesticides/isolation & purification , Soil Pollutants/isolation & purification , Adsorption , Pesticides/chemistry , Soil/chemistry , Soil Pollutants/chemistry , SolubilityABSTRACT
The transfer of eleven phenylurea herbicides through soil columns was investigated in laboratory conditions in order to determine leaching properties in a calcareous soil. Elution curves with distilled water were plotted after herbicide application on the soil column. Phenylurea retention by the soil indicating interactions with soil can be classified as follows: fenuron < fluometron ≤ isoproturon = monuron < metoxuron < monolinuron < metobromuron < chlorotoluron < linuron = diuron < chlorbromuron. The number and nature of halogen atoms on the phenyl ring had an important influence on leaching. Retention was higher for molecules with higher number of halogen, and it was also higher for bromine than chlorine. Column elution experiments were compared to batch experiments from which the distribution coefficients K d were determined. According to Kendall correlation coefficients, parameter m/m 0 max from column experiments was relatively well linked to K d. In case of phenylurea, a linear relationship between K d and m/m 0 max was established.
Subject(s)
Herbicides/chemistry , Models, Chemical , Phenylurea Compounds/chemistry , Soil Pollutants/chemistry , Soil/chemistry , Calcium Carbonate/chemistry , Herbicides/analysis , Phenylurea Compounds/analysis , Soil Pollutants/analysisABSTRACT
The adsorption of 12 pesticides of the phenylurea family was studied by batch experiments in order to determine the adsorption coefficient, K d. The study was conducted in two soils chosen for their differences in organic matter and calcite contents. K d pesticide adsorption coefficients were higher for soil S1 than for soil S2 due to the presence of a higher organic matter content and a lower calcite content in soil S1. To identify pesticide properties governing retention, 18 molecular descriptors were considered. Class-specific quantitative structure-property relationship (QSPR) soil sorption models using one, two, and three descriptors were developed from our experimental data using linear regressions. One of the aims of this work was to check whether QSPR models that did not include literature values of K ow were able to predict K d coefficients in satisfactory agreement with our experimental data. The influence of the level of theory in determining K ow and polarisability predictors on the predictive performance of the model was also examined by comparing quantum chemistry and empirical (QikProp) approaches. The one-descriptor model using "quantum" polarisability α was found to perform almost as well as or better than the other models.
