Predicting the Self-Diffusion Coefficient of Atmospheric Organic Aerosol Components via Molecular Dynamics Simulations.

Molecular dynamics (MD) simulations have been used to predict self-diffusion coefficients of atmospherically relevant organic compounds in the particulate phase at room temperature. We have used three approaches depending on the mobility of the compound. The self-diffusion coefficient D of the most mobile species (monoketones, monoalcohols and a few monocarboxylic acids) was obtained directly from MD simulations at 298 K. For less mobile species (oxomalonic acid, tartronic acid, cis-pinonic acid and dihydroxyacetone) MD simulations at higher temperatures were used together with Williams-Landel-Ferry (WLF) or Vogel-Fulcher-Tammann (VFT) extrapolation to obtain their self-diffusion coefficient D down to room temperature. Finally, the self-diffusion coefficient of compounds such as malonic acid, adipic acid, azelaic acid, tricarballylic acid, 3-methyl-1,2,3-butanecarboxylic acid (MBTCA) and 2-oxoadipic acid, which are too immobile a few degrees below their melting point, was evaluated only at 420 K for comparison purposes. The simulations showed that D is strongly influenced by the presence of functional groups. For the monofunctional compounds examined in this study, mobility decreases in the order: -COOH > -OH > -CO for monofunctional compounds. Functional-group multiplicity and proximity of highly polar groups further reduce mobility through the formation of dense hydrogen-bond networks and compact molecular packing. Global descriptors such as molecular weight and elemental ratios showed meaningful correlations with D only within specific chemical families. Functional-group identity, coexistence and molecular topology provide a more informative basis for interpreting diffusivity in atmospherically relevant organic compounds.