Development of a coarse-grained water forcefield via multistate iterative Boltzmann inversion

TL;DR

This paper presents a new coarse-grained water model developed using multistate iterative Boltzmann inversion, which accurately reproduces key properties of water and offers advantages over previous single-state models.

Contribution

The paper introduces a multistate iterative Boltzmann inversion approach for developing coarse-grained water forcefields, improving accuracy and applicability over prior single-state models.

Findings

Accurately reproduces water density and structural correlations at 305 K.

Demonstrates stability of a liquid droplet at physiological conditions.

Shows reduced crystallization tendency compared to previous models.

Abstract

A coarse-grained water model is developed using multistate iterative Boltzmann inversion. Following previous work, the k-means algorithm is used to dynamically map multiple water molecules to a single coarse-grained bead, allowing the use of structure-based coarse-graining methods. The model is derived to match the bulk and interfacial properties of liquid water and improves upon previous work that used single state iterative Boltzmann inversion. The model accurately reproduces the density and structural correlations of water at 305 K and 1.0 atm, stability of a liquid droplet at 305 K, and shows little tendency to crystallize at physiological conditions. This work also illustrates several advantages of using multistate iterative Boltzmann inversion for deriving generally applicable coarse-grained forcefields.