Influence of thermostat on droplet spreading in molecular dynamics simulations

Abstract

Molecular dynamic (MD) simulation has been extensively utilized to study droplet spreading, where the choice of thermostatting method can have a significant impact on the reliability of results. In this study, we perform simulations of droplet spreading across diverse scenarios, elucidating the reliability of thermostatting fluid (TF). Two widely employed thermostats, namely, the Nosé-Hoover (NH) algorithm and the Langevin (LGV) algorithm, are considered. Our findings reveal that while the NH thermostat is suitable for non-volatile droplet simulations in a vacuum where the flow inside the droplet is less impacted, it struggles in scenarios involving violent gas adsorption. Due to the global nature of NH algorithm, the temperature of adsorbed film cannot be regulated effectively, leading to suppressed adsorption. The distorted surface energy subsequently impacts the spreading dynamics. In contrast, the LGV algorithm, when applied by deactivating the thermostat in the spreading directions, yields accurate results across a broader range of scenarios, including non-volatile droplets spreading in a vacuum and in a gas environment. Nevertheless, TF schemes reveal their limitations when dealing with volatile droplets, where the heat transfer between the liquid and the solid wall significantly influences spreading. In TF systems, energy for evaporation comes directly from the thermostat, which leads to distorted temperature profiles of fluids, affecting the evaporation rate. In summary, special care should be taken when thermostatting fluid directly, especially in scenarios involving volatile droplets and violent adsorption.