NEP-MB-pol: A unified machine-learned framework for fast and accurate prediction of water's thermodynamic and transport properties
Abstract
Water's unique hydrogen-bonding network and anomalous properties pose significant challenges for accurately modeling its structural, thermodynamic, and transport behavior across varied conditions. Although machine-learned potentials have advanced the prediction of individual properties, a unified computational framework capable of simultaneously capturing water's complex and subtle properties with high accuracy has remained elusive. Here, we address this challenge by introducing NEP-MB-pol, a highly accurate and efficient neuroevolution potential (NEP) trained on extensive many-body polarization (MB-pol) reference data approaching coupled-cluster-level accuracy, combined with path-integral molecular dynamics and quantum-correction techniques to incorporate nuclear quantum effects. This NEP-MB-pol framework reproduces experimentally measured structural, thermodynamic, and transport properties of water across a broad temperature range, achieving simultaneous, fast, and accurate prediction of self-diffusion coefficient, viscosity, and thermal conductivity. Our approach provides a unified and robust tool for exploring thermodynamic and transport properties of water under diverse conditions, with significant potential for broader applications across research fields.
AI-Generated Overview
-
Research Focus: The study presents NEP-MB-pol, a machine-learned framework aimed at accurately predicting the thermodynamic and transport properties of water, addressing challenges posed by its unique hydrogen-bonding network and properties.
-
Methodology: The framework employs a neuroevolution potential (NEP) trained on extensive many-body polarization (MB-pol) reference data, along with path-integral molecular dynamics and quantum-correction techniques to account for nuclear quantum effects, allowing for simultaneous computations of various properties.
-
Results: NEP-MB-pol successfully reproduces experimentally measured structural, thermodynamic, and transport properties of water (such as self-diffusion coefficient, viscosity, and thermal conductivity) across a broad temperature range, demonstrating a high level of accuracy.
-
Key Contribution(s): The framework integrates high-level physics-based accuracy with machine-learning efficiency, enabling extensive simulations and quantitative predictions of various properties of water without relying on empirical fitting to experimental data.
-
Significance: NEP-MB-pol represents a significant advancement in computational models for predicting water properties, showing the ability to achieve both accuracy and efficiency, which has been a longstanding challenge in the field.
-
Broader Applications: The NEP-MB-pol framework is adaptable and could be extended to model water behavior under extreme thermodynamic conditions, as well as applicable across diverse research fields such as physical chemistry, materials science, and environmental studies.