Molecular Dynamics Simulation of Diffusion and O 2 Dissolution in Water Using Four Water Molecular Models

This work evaluates four popular 3-site water models, TIP3P, SPC/E, SPC, and F3C, in predicting temperature-dependent O 2 diffusion and dissolution and self-diffusion in liquid water for electrochemical devices, such as fuel cells, electrolyzers and metal-air batteries, using molecular dynamics simulations. We report that the TIP3P and SPC models overestimate O 2 diffusivity, though qualitatively describe its temperature dependence, while the SPC/E and F3C models show a good prediction of O 2 diffusivity with the former accurately predicting its activation energy. The TIP3P, SPC, and F3C models overestimate the O 2 solubility with the TIP3P even incorrectly predicting its temperature dependence. The SPC/E model shows the best performance in predicting both the O 2 diffusivity (<15% error) and solubility (<16% error) in the temperature range. The radial distribution functions of O atoms are present and show that lower peaks, i.e. more intermolecular space among water molecules for the random walks, leads to higher prediction of diffusivities. Comparison with literature experimental data is summarized.