Thermostating strategies and interaction potentials in interfacial water simulations: are they really needed?

Understanding the molecular behavior of water near solid surfaces is critical for advancing simulations in catalysis, energy storage, and nanoscale heat transfer. Here, we reveal how the choice of thermostating strategy and interaction potential fundamentally alters the structural and dynamical properties of interfacial water. Using classical molecular dynamics (MD) simulations of SPC/E water near a Pt (111) surface, we compare four distinct thermal control schemes in combination with two interaction models: the Zhu-Philpott (ZP) potential and a fitted Lennard-Jones (LJ) approximation. Our results demonstrate that thermostating both the water and the solid yields the most realistic layering, orientation, and thermal equilibration in the first hydration layer (∼0.5 nm from the surface). Surprisingly, simulations using only water thermostats but with a frozen or non-thermostated wall produce overstructured or thermally biased interfacial profiles. Furthermore, we show that a carefully optimized LJ potential can approximate ZP’s accuracy in density layering but fails to replicate angular orientation. The study points out the situations where simplified models are valid, and where detailed interactions and thermal effects must be included, offering both new knowledge and useful rules for improving simulation accuracy.