Structure and Dynamics of the Magnetite(001)/Water Interface from Molecular Dynamics Simulations Based on a Neural...

Area of Science:

• Materials Science
• Surface Chemistry
• Computational Chemistry

Background:

• The magnetite/water interface is prevalent in natural and technological systems.
• Molecular-level understanding of its structure and dynamics is limited.
• Accurate simulation methods are needed to explore this interface.

Purpose of the Study:

• To develop an efficient neural network potential (NNP) for the magnetite/water system.
• To investigate the structural and dynamical properties of water on the magnetite (001) surface.
• To elucidate water molecule behavior across various coverage regimes.

Main Methods:

• Developed a Behler-Parrinello neural network potential (NNP) trained on density functional theory data.
• Performed extensive molecular dynamics (MD) simulations of the magnetite (001) surface.
• Simulated a wide range of water coverages, from single molecules to bulk water.

Main Results:

• Identified novel ground states for low-coverage water on the Subsurface Cation Vacancy (SCV) model.
• Observed distinct layering in the water density profile at the surface.
• Quantified anisotropic diffusion of water molecules on the SCV model.
• Gained qualitative insights into water molecule dissociation mechanisms.

Conclusions:

• The developed NNP enables efficient and accurate simulations of the magnetite/water interface.
• Water molecules exhibit complex structural arrangements and anisotropic diffusion at the magnetite surface.
• The study provides a foundation for understanding interfacial phenomena in magnetite-based systems.