Reactive molecular dynamics models from ab initio molecular dynamics data using relative entropy minimization.

Area of Science:

• Computational Chemistry
• Chemical Physics
• Materials Science

Background:

• Simulating proton transfer in water is crucial for understanding chemical reactions and biological processes.
• Accurate molecular dynamics (MD) force fields are essential for reliable simulations.
• Developing efficient yet accurate reactive force fields remains a challenge.

Purpose of the Study:

• To develop novel multiscale reactive molecular dynamics (MS-RMD) models for the hydrated excess proton in water.
• To validate the accuracy of the developed MS-RMD models against high-level ab initio molecular dynamics (AIMD) data.
• To demonstrate a new method for rapidly generating accurate and efficient reactive MD force fields.

Main Methods:

• Development of two new MS-RMD models using data from AIMD simulations.
• Utilizing an algorithm based on relative entropy minimization for model development.
• Validation of MS-RMD models by comparing potential of mean force and radial distribution functions with AIMD results.

Main Results:

• The developed MS-RMD models accurately reproduce the potential of mean force along the proton transfer coordinate.
• Radial distribution functions from MS-RMD simulations faithfully match those obtained from AIMD.
• The relative entropy minimization algorithm successfully generated accurate and efficient reactive MD force fields.

Conclusions:

• The new MS-RMD models provide a computationally efficient and accurate representation of the hydrated excess proton in water.
• The relative entropy minimization approach is a powerful tool for developing high-performance reactive force fields.
• These findings advance the simulation capabilities for studying proton transfer phenomena in aqueous systems.