Dual approach for effective potentials that accurately model structure and energetics.

Area of Science:

• Computational chemistry
• Molecular modeling
• Statistical mechanics

Background:

• Coarse-grained (CG) models simplify complex molecular systems by reducing degrees of freedom.
• Effective potentials in CG models often lack transferability and accurate energy estimations due to entropic contributions.

Purpose of the Study:

• To develop a novel dual-potential approach for creating more accurate and transferable CG effective potentials.
• To model both free energies and potential energies effectively using CG configurations.

Main Methods:

• Proposed a dual-potential variational approach combining structure-based and energy-based principles.
• Applied the method to 1-site coarse-grained models of water and methanol.
• Utilized simulations with the structure-based potential for configuration sampling and postprocessing with the energy-based potential for energy estimation.

Main Results:

• Accurate sampling of configuration space was achieved using the structure-based potential.
• Average atomic energies were accurately estimated by postprocessing simulations with the energy-based potential.
• The difference between the two potentials provided a qualitatively accurate prediction of the temperature dependence.

Conclusions:

• The dual-potential approach offers improved accuracy and transferability for coarse-grained models.
• This method effectively separates the modeling of configurational thermodynamics and energetics.
• It provides a pathway to more reliable molecular simulations of complex systems.