Understanding the Anomalous Diffusion of Water in Aqueous Electrolytes Using Machine Learned Potentials

Area of Science:

• Physical Chemistry
• Computational Chemistry
• Materials Science

Background:

• Anomalous diffusion of water in salt solutions, observed experimentally, is not accurately reproduced by traditional molecular dynamics (MD) simulations.
• This discrepancy arises from the inadequate representation of ion-water interactions in standard force fields.

Purpose of the Study:

• To develop and validate a novel computational approach for simulating anomalous diffusion in aqueous salt solutions.
• To elucidate the microscopic mechanisms behind ion-specific effects on water transport properties.

Main Methods:

• Utilized machine-learned atomic potentials (MLPs) trained on density functional theory (DFT) data.
• Performed atomistic MD simulations of aqueous cesium iodide (CsI) and sodium chloride (NaCl) solutions.

Main Results:

• MLP-based MD simulations successfully reproduced experimentally observed thermodynamic, structural, dynamical, and transport properties.
• The simulations accurately captured the varied trends in water diffusivities across different salt concentrations and ion types.
• Analysis revealed that CsI enhances water diffusion, while NaCl exhibits a net retardation due to competing ion effects.

Conclusions:

• Machine-learned potentials provide a robust method for accurately simulating ion-water interactions and anomalous diffusion.
• The study unraveled the molecular-level origins of differing water transport behaviors in CsI and NaCl solutions.
• This approach opens new avenues for predicting and understanding the properties of electrolyte solutions.