Molecular dynamics study of ice melting on a silver plate

Abstract

Ice fall incidents, such as detached ice chunks from bridge stay cables, pose not only serious safety hazards to pedestrians and vehicles below but also significant serviceability issues, as bridge closures required for inspection or ice removal can lead to costly disruptions. The current trend of climate change exacerbates this kind of hazard. The availability of a reliable engineering tool, such as an accurate numerical model grounded in nanoscale melt-front physics, is imperative to provide a clear insight into the ice detachment mechanism and develop effective de-icing solutions. Classical molecular dynamics simulations are conducted in the current study to investigate the melting of an ice cube in an atomically flat silver slab. The TIP4P/ice water model is adopted and the simulation is conducted in canonical ensemble with a layer-resolved Langevin thermostat. The phase evolution is tracked via the averaged tetrahedral order parameter, while systematically varying five controls: the depth of heated layers, the silver substrate thickness, the ice thickness, the lateral confinement, and the ice crystal contact orientation (basal vs prism). Results show that melting is controlled primarily by the substrate temperature; variations in heat-conducting-layer count had a minor influence and converged to similar end states. Doubling the ice thickness increases the melt time approximately by three times, whereas relaxing periodic boundaries reshapes the melt into domes or spreading films. Presenting the basal plane instead of a prism plane accelerates loss of crystalline order. Collectively, the simulations yield a numerically consistent set of parameters that not only advances the existing knowledge of nanoscale ice melting simulation but can also be transferred to continuum-scale de-icing simulations, enabling accurate modeling of melt-induced ice detachment from structural components, such as bridge stay cables.