Multiscale Coarse-Graining with Effective Polarizabilities: A Fully Bottom-Up Approach

Area of Science:

• Computational Chemistry
• Materials Science
• Statistical Mechanics

Background:

• Coarse-grain (CG) models simplify complex systems for longer simulations by reducing detail.
• Standard CG models often omit electrostatic interactions, crucial for many material properties.
• Existing methods struggle to accurately represent charge and dipole moments in CG models.

Purpose of the Study:

• To develop a bottom-up CG model preserving center of mass (CoM), charge, and dipole of atomic clusters.
• To introduce a novel fractional mapping approach for CG site representation.
• To validate the CG model's accuracy in reproducing electrostatic properties and statistical mechanics.

Main Methods:

• Developed a CG representation using charged dimers for atomic clusters.
• Derived a formal mapping of microscopic coordinates to CG representation.
• Employed fractional mapping of atomistic particles to CG dimer sites.
• Validated the method on nitromethane (submolecular and molecular resolution) and water clusters.

Main Results:

• The CG model accurately reproduces CoM-CoM radial distribution functions and dipole distributions.
• Submolecular nitromethane model shows transferability to simulations with external fields.
• Single-dimer nitromethane model matches dipole-dipole correlation but deviates in dielectric constant.
• Water simulations show minor discrepancies in dipole-dipole correlation due to force decomposition.

Conclusions:

• The novel fractional mapping CG method effectively preserves essential electrostatic properties.
• The model shows promise for simulating systems where electrostatics are critical.
• Further refinement is needed to address discrepancies in dynamic properties like dielectric constants.