(257c) A Novel Multiscale Approach for Prediction of Thermodynamics Properties Based on Distribution of Molecular Conformations

The conformation distribution of flexible molecules may have a significant influence on its physical properties and phase behaviors. However, conventional thermodynamic models, such as a volumetric equation of state (EOS) that relates the pressure, volume, temperature and composition (PVTx) of a fluid, do not include such molecular details when describing thermodynamic properties and phase equilibria. In this work, we proposed a novel multiscale approach that provide quantitative predictions of phase behaviors for mixtures involving conformationally flexible species. In this method, the electrostatic interaction between molecules determined based on first principle implicit solvation calculation and the dispersion interactions obtained from molecular dynamic simulation are used as input to the PR+COSMOSAC EOS for describing the PVTx of a mixture. Furthermore, the transition of a species in different conformations are considered as chemical reactions with the equilibrium constant determined from the G4 method. Once the necessary parameters are obtained from different simulations, subsequent phase equilibrium predictions can be achieved within milliseconds. We validated this approach using 1,2-dichloroethane (DCE), which can exist in gauche and trans conformations. Our results show that this approach provides not only an accurate prediction for the vapor-liquid equilibrium of pure (DCE) and its mixture with other chemicals, but also a quantitative description of conformation distribution of DCE in these systems. This novel method can be very useful for the prediction of thermodynamic properties of fluids with explicit inclusion of molecular conformations.