Title : Drugs as an environmental problem chromatographic and computational studies of drugs partition between soil and water
The fate and transport of solutes in the environment depend on their physico-chemical properties such as lipophilicity, volatility, water solubility and ability to partition between soil and water. The soil-water partition coefficient Koc (normalized to the soil organic carbon content to reduce the differences among soils) is a very important parameter governing the fate of compounds in the soil-water compartment. It influences the chemicals’ mobility in soil, their environmental persistence and the processes of their removal in wastewater management facilities. Direct determination of Koc is based on studies of partitioning phenomena in biphasic soli-water systems using batch or continuous flow experiments. Such tests are, however, tedious and time-consuming and the results tend to be inconsistent due to experimental difficulties - incomplete separation of phases and volatilization, biological or chemical degradation of solutes. Alternatively, it is possible to predict soil – water partition using chromatographic descriptors obtained by liquid chromatography on octadecyl-, cyano-, diol-, ethyl-, trimethylammonium- or aminopropyl-modified silica or on sorbents containing immobilized humic acid or soil. Chromatographic methods of Koc determination are fast and relatively cheap, with the benefit of high reproducibility, especially when commercially available stationary phases are used.
Chromatographic retention factor log kIAM obtained from IAM HPLC chromatography with buffered, aqueous mobile phases and several calculated molecular descriptors obtained for a group of 175 structurally unrelated compounds were tested in order to generate useful models of solutes’ soil-water partition coefficient normalized to organic carbon log Koc. It was established that log kIAM is not sufficient as a sole predictor of the skin permeability coefficient. Simple, potentially useful models based on log kIAM and a selection of readily available, calculated descriptors were generated. The models proposed in the study were tested on a group of 50 compounds with known experimental log Koc values.