On the glass transition temperature of organic compounds via molecular dynamics simulations

The organic fraction of atmospheric particulate matter contains tens of thousands of complex compounds that have one or more functional groups. Quantifying the physical and chemical properties of each of these compounds experimentally is challenging and time consuming. The glass transition temperature, Tg, is one of these properties since it can help determine the phase state of aerosols in different parts of the atmosphere. This phase state influences gas-particle partitioning of semi-volatile compounds, the timescales of diffusion inside the particle, water uptake, as well as the rates and kinetics of heterogeneous reactions and oxidation that take place on aerosols. Experimental Tg determination can be demanding, because of the challenges presented by the synthesis and purification of the corresponding organic compounds.
Molecular Dynamics (MD) simulations have the advantage of detailed prediction of the desired properties on the molecular level with relatively low cost compared to actual experiments. In our work, we implement MD simulations to determine Tg of various organic compounds. Although Tg of organic compounds has been examined experimentally, the discrepancies in the bibliography are vast not only between experiments but also between experimental and predicted values, derived from theoretical or semi-empirical proposed equations. In the current work we focus on organic compounds of atmospheric interest, and we investigate in detail the contributions of the various functional groups to Tg. The investigated organic compounds vary in the carbon chain length as well as in the number and type of functional groups (i.e., hydroxyl and carboxylic groups). The Tg is determined by applying different cooling rates over a wide temperature span for several independent initial configurations of the examined organic molecules in bulk phase and by analyzing properties such as the density and the energy of non-bonded interactions. The results of the molecular simulations are compared with available experimental data in the bibliography, and theoretical or empirical Tg predictions.