Successful PhD Defense of Deedar NABI

Summary : Nonpolar chemicals are environmentally relevant because many of them are persistent in the environment, can bioaccumulate in organisms, and/or are toxic to biota including humans. Additionally, several such contaminants have long-range transport potential and can travel to pristine environments such as the Himalaya and the Arctic regions. Understanding fate, behavior and risk of such chemicals is challenging because the partitioning property data are not available for many of these compounds. Additionally, these chemicals may exist as complex mixtures in the environment. In order to overcome these challenges this thesis suggests a strategy whereby researchers can directly link the sample analysis activities with risk assessment activities, using comprehensive two-dimensional gas chromatography (GC×GC). GC×GC is capable of separating and quantifying thousands of chemicals in the complex mixtures. In this thesis, I show that GC×GC can be coupled to risk assessment, leading to tangible benefits for the environmental science community. In order to evaluate this approach, 40 diverse environmental partitioning and transport properties comprising of more than 1400 individual experimental property data were considered.
This thesis extends the domain of environmental applications of two-parameter GC×GC-based linear free energy relationships (LFERs) to a diverse set of nonpolar chemicals. In this thesis, “nonpolar” chemicals are defined as those that undergo negligible or weak hydrogen-bonding interactions with their surroundings. These include contaminant types comprising acyclic aliphatic, fused and bridged cyclic aliphatic, aromatic, and double-bonded hydrocarbon compounds that are unsubstituted or substituted with fluorine, chlorine, bromine, and/or iodine. Chapter 2 of this thesis demonstrates that only two independent dimensions of information are required to successfully describe the environmental partitioning properties of such contaminants. The corresponding chemical descriptors could be estimated accurately from first- and second-dimension retention indices produced by GC×GC. Consequently, GC×GC chromatogram retention information was used to successfully explain environmentally relevant partitioning properties for nonpolar analytes having boiling points
≤ 402 °C. The properties explored involve several environmentally relevant phases including air, water, octanol, hexadecane, natural organic matter, and biotic tissues. Regression statistics suggest that partitioning properties are predicted to near or within experimental error. Enthalpies of phase transfer needed to account for partitioning temperature dependence were also adequately estimated. The proposed approach has accuracy similar to that of the Abraham solvation model for the chemical test set considered here. When applied to nonpolar halogenated or hydrocarbon complex mixtures, this approach can be used to parameterize transport models or to screen for analytes having Long Range Transport Potential, Aquatic Bioaccumulation Potential, Arctic Contamination Potential, and other characteristic partitioning behaviors.
Chapter 3 of this thesis also makes a link between GC×GC retention information of nonpolar analytes and exposure- and effect-related properties. This chapter describes the estimation of partition coefficients, exchange rate constants and diffusion constants for several relevant biotic and abiotic phases. The GC×GC retention indices were used to predict the partition coefficients for several passive samplers in air, water and biotic tissues for a diverse set of nonpolar compounds. Partitioning to biotic phases such as mussels, membrane and storage lipids, serum and muscle proteins, and carbohydrates were successfully explained using GC×GC retention indices. GC×GC-based LFER was also successfully evaluated for baseline toxicity of nonpolar contaminants. These properties were predicted within the range of reported experimental error. Diffusion coefficients, kinetic exchange coefficients, and salting-out coefficient were also successfully estimated using GC×GC retention indices. This chapter demonstrates that the estimated parameters are also appropriate to parameterize the diffusion-based models used to describe the time dependence of exchange processes. Finally, this chapter proposes an approach to estimate several in vitro and in vivo partitioning properties such as milk-water, liver-blood, adipose-plasma, skin-blood, lung-plasma, muscle-plasma, and muscle-blood partition coefficients.
In a dynamic environmental system, understanding bioconcentration kinetics is essential for an ecological risk assessment of contaminants. For organisms, chemical exchange processes are mainly controlled by chemical and biological parameters. Chapter 4 of this thesis evaluates the hypothesis that the uptake and depuration rate constants for fish can be explained in terms of fish weight and multiple intermolecular interaction descriptors. Using Abraham solute descriptors and fish weight, variance in depuration rate constants of diverse neutral organic chemicals was explained successfully with a root mean square error (RMSE) value of 0.52 log units and R2 of 0.82. Solute size and the hydrogen-bonding trait were found to be the dominant parameters controlling the exchange process. Uptake rate constants were predicted with relatively poor statistics, which indicated that factors other than physicochemical properties and fish size were imparting considerable variance to the data. Approaches based on the GC×GC retention indices and diffusion constants were also able to explain the variability in bioconcentration kinetics, exhibiting comparable statistics. As a means to understand bioconcentration kinetics, the approach developed here could be used as an alternative to animal testing.
Polychlorinated n-alkanes (PCAs) are notoriously complex halogenated nonpolar mixture. Work in Chapter 5 shows that even with GC×GC short chain PCAs could not be resolved down to individual congeners. Consequently, Chapter 5 proposes a pragmatic approach of linking the GC×GC separation of PCAs into groups and subgroups to their environmental properties using GC×GC-based LFERs. This chapter also discusses the sorption behavior of compounds of varying hydrophobicities. In order to have minimum losses to headspace and glass wall sorption, a new method was evaluated where the starting concentrations of the analytes on polydimethylsiloxane (PDMS) chips were varied and aqueous volumes were kept constant. The PCA mixture was spiked with four probe compounds (pentachloroethane, hexachloroethane, PCB 101 and mirex) to evaluate the methodology. However, due to technical issues with the GC×GC instrument and low aqueous concentrations, PDMS-water partition coefficients were determined only for 12 chemicals. Sorption isotherms for the PDMS-water system indicated that the approach adopted in this study was free from significant artifacts. For the dissolved organic matter system, sorption data did not fit to any of 4 types of linear models evaluated in this study, and sorption curves appeared to follow a Langmuir adsorption model.
In summary, this thesis revealed that GC×GC encoded almost the same amount of partitioning property information for nonpolar contaminants as the 5 to 6 parameters of the Abraham solvation model. GC×GC thus creates a link between environmental sample analysis and property estimation that can help us to reach many environmental objectives more quickly, including risk assessment, precautionary measures, remediation, and decision support for policy and management.