Genetic Algorithm Driven Force Field Parameterization for Molten Alkali-Metal Carbonate and Hydroxide Salts

Area of Science:

• Chemistry
• Materials Science
• Geochemistry

Background:

• Molten alkali-metal carbonates and hydroxides are crucial in molten carbonate fuel cells and Earth's geochemistry.
• Molecular simulations offer a method to study these systems under extreme conditions, bypassing experimental challenges.

Purpose of the Study:

• To develop accurate classical force fields for molten alkali-metal carbonates and hydroxides.
• To enable reliable molecular simulations of these systems for predicting liquid chemical potentials.

Main Methods:

• Utilized a genetic algorithm to fit *ab initio* molecular dynamics-computed densities and radial distribution functions.
• Incorporated experimental enthalpies of formation into the fitting process.
• Derived new classical force fields based on liquid phase structure and energetics.

Main Results:

• Developed classical force fields that accurately predict liquid chemical potentials for molten alkali-metal carbonates and hydroxides.
• Ensured accurate liquid phase structure and energetics through the chosen fitting properties.
• Observed that predicted dynamics, while slower than experimental, generally maintain correct trends across systems.

Conclusions:

• The newly parametrized force fields provide a reliable tool for simulating molten alkali-metal carbonates and hydroxides.
• These force fields can be extended to molten carbonate-hydroxide mixtures using standard combining rules.
• This work facilitates the study of geochemistry and fuel cell technology through accurate molecular simulations.