Extended MARTINI Water Model for Heat Transfer Studies

Abstract: The computationally efficient classical MARTINI model is extended to simulate heat transfer simulations of water. The current MARTINI model, variations of it and other coarse grain water models focus on reproducing the thermodynamic properties below room temperature, hence making them unsuitable for studying high temperature simulations especially evaporation at 100 {\deg}C. In this work, the MARTINI model is reparametrized using a combination of Genetic Algorithm, Artificial Neural Network and Nelder-Mead optimization technique to match the phase equilibrium properties of water. The reparametrized model (MARTINI-E) accurately reproduces density, enthalpy of vaporization and surface tension at 100{\deg}C and outperforms other leading coarse grain water models. The model is also validated using the energy conservation and enthalpy change due to latent heat in a lamellar system. This new water model can be used for simulating phase change phenomena, thin film evaporation and other energy transport mechanisms accurately.