PhD Thesis Defence Presentations - Panagiota Siachouli

        Atmospheric aerosols influence climate, air quality, and human health through their effects on radiation balance, cloud formation, and heterogeneous chemistry. Organic compounds constitute a major fraction of aerosols and play a central role in determining their physicochemical properties. Most of these compounds exhibit a wide range of functional groups and complex molecular architectures, forming a chemically rich organic landscape that remains only partially understood. Investigating their physicochemical properties is therefore essential for understanding their role in atmospheric processes. Glass transition temperature (T_g) and self-diffusion coefficient (D) influence the phase state of organic compounds, thereby affecting key processes such as gas-particle partitioning, heterogeneous chemistry, and long-range transport.

        In this work, Molecular Dynamics (MD) simulations were employed to investigate T_g and D for a diverse set of atmospherically relevant organic compounds, including ketones, alcohols, carboxylic acids, and multifunctional ones. T_g was determined from changes in density and non-bonded energy during cooling simulations, with results compared against available experimental data and semi-empirical equations. For a few compounds, the method of segmental relaxation was also utilized. Self-diffusion coefficients at 298 K were either directly obtained from simulations or, for compounds with melting points above this temperature, estimated via Arrhenius extrapolation from higher-temperature simulations. For both properties, the dataset encompassed a broad range of organic compounds with varied molecular weights, functional group identities, and molecular architectures, enabling the identification of systematic trends.

        In the first part of the thesis, the framework of predicting T_g via MD simulations is established. The scope was to examine the influence of functional groups, carbon chain length, and molecular architecture on T_g for a set of atmospherically relevant organic compounds. The MD simulations agreed within 20% with the available experimental measurements, while consistently overpredicting T_g. T_g was found to be most sensitive to functional group type, following the hierarchy -COOH > -OH > -C=O. Increasing the number of carbon atoms increased T_g. No clear correlation between oxygen to carbon (O:C) ratio and T_g was observed, whereas molecular weight was positively correlated with the increase of T_g. Molecular architecture was also found to influence T_g, with non-aromatic cyclic compounds showing higher T_g than linear or branched counterparts. 

        In the second part of the thesis, the T_g framework was extended to a broader and more diverse set of organic compounds, and the MD results were used to develop a T_g parameterization based on molecular characteristics. Both the MD predictions and the parameterization linked T_g to functional group identity and number, molecular architecture, and elemental composition. The simulations confirmed the functional group hierarchy -COOH > -OH > -C=O, a trend maintained even when multiple functional groups coexisted within a molecule. Cyclic structures consistently exhibited higher T_g values than their linear counterparts, and T_g increased with the number of carbon atoms.  The parameterization reproduced the MD simulation results within 10% and, after applying a scaling factor of 0.9, achieved agreement within 10% with the available experimental measurements. The robustness of the parameterization was evaluated through a leave-one-out approach, which identified the functional groups and molecular features that most strongly influence T_g and highlighted compounds whose limited representation leads to greater variability in predicted contributions.

        In the third part, MD simulations were carried out to investigate D of atmospherically relevant organic compounds at 298 K. The study began with simple monofunctional compounds—ketones, alcohols and carboxylic acids—before extending to multifunctional ones that had previously been examined for their T_g. Predicted D values agreed with available experimental measurements within a factor of 1.9 and with prior MD studies within a factor of 1.7. Diffusivity decreased with increasing carbon chain length, molecular weight, and functional group multiplicity. The hierarchy of sensitivity was found to be: -C=O > -OH > -COOH, consistent with their T_g behavior and reflecting the importance of hydrogen bond network formation in controlling molecular mobility. Multifunctional compounds exhibited the same sensitivity trend, with compounds containing carboxyl-hydroxyl functional group combinations showing the lowest D values. The spatial arrangement of functional groups also influenced D, with closely positioned groups producing the strongest mobility constraints and the lower predicted values. Molecular weight was positively correlated with reduced D values, whereas elemental composition, such as oxygen to carbon ratio showed poor predictive potential.

        Overall, these results provide a molecular-level understanding of the molecular features influencing T_g and D in organic aerosol components, offering predictive capability for compounds lacking measurements and supporting improved representation of aerosol phase state and molecular mobility in atmospheric processes and climate models.