Accurate Thermophysical Properties of Water using Machine-Learned Potentials

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hilpert, Tobias, Kresse, Georg
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908795855175680
author Hilpert, Tobias
Kresse, Georg
author_facet Hilpert, Tobias
Kresse, Georg
contents Simulating water from first principles remains a significant computational challenge due to the slow dynamics of the underlying system. Although machine-learned interatomic potentials (MLPs) can accelerate these simulations, they often fail to achieve the required level of accuracy for reliable uncertainty quantification. In this study, we use MACE - an equivariant graph neural network architecture that has been trained using an extensive RPBE-D3 database - to predict density isobars, diffusion constants, radial distribution functions, and melting points. Although equivariant MACE models are computationally more expensive than simpler architectures, such as kernel-based potentials (KbPs), their significantly lower total energy errors allow for reliable thermodynamic reweighting with minimal bias. Our results are consistent with those of previous studies using KbPs; however, equivariant models can be validated against the ground-truth density functional theory (DFT) ensemble, providing a critical advantage. These findings establish equivariant MLPs as robust and reliable tools for investigating the thermophysical properties of water with DFT-level accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2601_21103
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Accurate Thermophysical Properties of Water using Machine-Learned Potentials
Hilpert, Tobias
Kresse, Georg
Chemical Physics
Soft Condensed Matter
Simulating water from first principles remains a significant computational challenge due to the slow dynamics of the underlying system. Although machine-learned interatomic potentials (MLPs) can accelerate these simulations, they often fail to achieve the required level of accuracy for reliable uncertainty quantification. In this study, we use MACE - an equivariant graph neural network architecture that has been trained using an extensive RPBE-D3 database - to predict density isobars, diffusion constants, radial distribution functions, and melting points. Although equivariant MACE models are computationally more expensive than simpler architectures, such as kernel-based potentials (KbPs), their significantly lower total energy errors allow for reliable thermodynamic reweighting with minimal bias. Our results are consistent with those of previous studies using KbPs; however, equivariant models can be validated against the ground-truth density functional theory (DFT) ensemble, providing a critical advantage. These findings establish equivariant MLPs as robust and reliable tools for investigating the thermophysical properties of water with DFT-level accuracy.
title Accurate Thermophysical Properties of Water using Machine-Learned Potentials
topic Chemical Physics
Soft Condensed Matter
url https://arxiv.org/abs/2601.21103