Developing machine-learned potentials to simultaneously capture the dynamics of excess protons and hydroxide ions in...

Area of Science:

• Computational chemistry
• Physical chemistry
• Materials science

Background:

• Proton and hydroxide ion transport in water is vital for many chemical and biological processes.
• Accurate simulation requires computationally expensive ab initio molecular dynamics and path-integral methods.
• Current methods face limitations in achieving necessary time and length scales for converged transport properties.

Purpose of the Study:

• To develop and apply machine-learned potentials (MLPs) for simulating excess proton and hydroxide ion transport in water.
• To achieve ab initio accuracy at a significantly reduced computational cost.
• To enable multi-nanosecond simulations for analyzing transport mechanisms and properties.

Main Methods:

• Development of machine-learned potentials (MLPs) for excess protons and hydroxide ions.
• Simulations performed at generalized gradient approximation and hybrid density functional theory levels of accuracy.
• Utilized both classical and path-integral simulations over multiple nanoseconds.
• Analysis of proton transfer events and ion diffusion coefficients.

Main Results:

• MLPs successfully reproduced ab initio trends for proton and hydroxide ion transport.
• Achieved converged diffusion coefficients for both excess protons and hydroxide ions.
• Simulations provided insights into the role of hypercoordination in hydroxide ion transport.
• Confirmed asymmetry in diffusion between excess protons and hydroxide ions.

Conclusions:

• MLPs offer a computationally efficient alternative to ab initio methods for simulating proton and hydroxide ion transport.
• The developed MLPs enable accurate modeling of nuclear quantum effects and bond dynamics.
• This work advances the understanding of ion transport mechanisms in aqueous systems and their implications.