Polarizable water models from mixed computational and empirical optimization.

Area of Science:

• Computational Chemistry
• Physical Chemistry
• Materials Science

Background:

• Accurate molecular mechanics (MM) models are crucial for simulating complex liquids.
• Polarizable molecular mechanics (PMM) models offer improved accuracy but require careful parameterization.
• Optimizing PMM parameters for realistic liquid simulations remains a challenge.

Purpose of the Study:

• To develop and validate a novel hybrid computational and empirical approach for optimizing PMM model parameters.
• To create accurate PMM models for water, specifically the TLνP models (ν = 3, 4, 5).
• To assess the predictive power of the developed models for various thermophysical properties of water.

Main Methods:

• Hybrid quantum mechanics/molecular mechanics (QM/MM) calculations combining density functional theory (DFT) with PMM.
• Parameter optimization using molecular dynamics (MD) simulations under defined thermodynamic conditions.
• Development of TLνP water models incorporating Gaussian inducible dipoles and Buckingham potentials.

Main Results:

• Successfully optimized TLνP PMM models for water (ν = 3, 4, 5) with accurate gas-phase properties.
• Models demonstrated high fidelity in reproducing target observables like potential energy, pressure, and radial distribution functions.
• Predicted properties including dielectric constant, viscosity, and heat capacity showed good agreement with experimental data.

Conclusions:

• The proposed hybrid approach effectively optimizes PMM parameters for complex liquids.
• The developed TLνP water models represent a significant advancement in simulating water properties.
• Further refinements can address remaining deficiencies, enhancing the models' applicability.