Transferable Coarse-Grained Models of Liquid-Liquid Equilibrium Using Local Density Potentials Optimized with the...

Area of Science:

• Computational chemistry
• Materials science
• Chemical engineering

Background:

• Bottom-up coarse-grained (CG) models are crucial for simulating large molecular systems.
• Predicting fluid phase equilibria with CG models is challenging due to poor transferability.
• Existing CG models often neglect multibody interactions by relying solely on pair potentials.

Purpose of the Study:

• To develop transferable single-site CG models for liquid mixtures.
• To enhance the accuracy of CG models in predicting phase equilibria and structural properties.
• To incorporate many-body effects into CG models cost-effectively.

Main Methods:

• Supplementing traditional CG pair interactions with local density potentials.
• Developing CG models for benzene-water solutions using intra- and interspecies local density potentials.
• Evaluating model transferability using structural metrics like pair correlation functions and density profiles.

Main Results:

• The proposed CG models demonstrate impressive transferability in structural metrics across composition space.
• Local density potentials effectively capture many-body effects in liquid mixtures.
• The new models outperform traditional pair-only CG representations in terms of transferability.

Conclusions:

• Local density potentials offer a general and improved approach for CG modeling of complex phase equilibria.
• The method shows significant promise for enhancing the predictive power of CG simulations for liquid mixtures.
• Further refinement may be needed for complex thermodynamic properties like interfacial tension.