Modeling the toxicity and partitioning of chemicals using linear solvation energy relationships

The overall goal of this work is to develop mathematical models capable of predicting the toxicity and partitioning of diverse neutral organic contaminants in aquatic and terrestrial environments. A linear solvation energy relationship (LSER) partitioning model is proposed for compounds to multiple environmentally relevant phases: (1) target lipid for the prediction of toxicity to aquatic organisms, (2) particulate organic carbon, (3) dissolved organic carbon, and (4) quartz, for the quantification of sorption in soils and aquatic environments.;Previous models have largely relied on semi-empirical relationships with KOW, the octanol-water partition coefficient, to describe the partitioning to relevant environmental phases as part of the Equilibrium Partitioning (EqP) modeling framework. However, there are drawbacks to this approach. Multiple models using KOW as a descriptor are needed for different chemical classes. This method may result in incorrect classifications of compounds without requisite chemical intuition. Furthermore, fits to small datasets for chemical classes for which experimental data are sparse are generally not robust. Additionally, there is usually no contingency strategy for new contaminants that cannot be classified into predefined groups.;This work aims to remedy this with the use of LSERs that use the structural information from a chemical directly to calculate partitioning properties for a variety of neutral organic chemicals, and that describe the nature of the environmental phase as well. The less empirical LSER approach enables more robust environmental fate and transport predictions for chemicals for which there is a lack of experimental data, and is quick and simple in application.