Efficient treatment of long-range electrostatics in charge equilibration approaches

Area of Science:

• Computational Materials Science
• Quantum Chemistry
• Condensed Matter Physics

Background:

• Accurate modeling of electrostatic interactions is crucial for materials simulations.
• Previous methods relied on Slater-type orbitals (STOs) for charge shielding.
• Integrating advanced charge equilibration methods can enhance simulation accuracy and efficiency.

Purpose of the Study:

• To present a novel charge equilibration method utilizing real-space Gaussians.
• To integrate this method into the Electrode package of the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS).
• To demonstrate the method's consistency with previous approaches and its potential for machine-learning force fields.

Main Methods:

• Developed a charge equilibration method based on real-space Gaussian charge densities.
• Implemented the method within the LAMMPS software, leveraging its particle-mesh Ewald approach.
• Fitted Gaussian charge distributions to Slater-type orbital repulsion to ensure a smooth transition.

Main Results:

• Optimized Gaussian widths for O, Si, and Ti species.
• Achieved results consistent with STO-based methods for SiO2 polymorphs.
• Demonstrated convergence towards electronegativity equalization method (EEM) results for Ti/TiOx interfaces in the narrow Gaussian limit.

Conclusions:

• The Gaussian-based charge equilibration method provides a viable and accurate alternative to STO-based methods.
• The implementation is compatible with efficient simulation packages like LAMMPS.
• This method shows promise for enhancing machine-learning force fields that incorporate long-range electrostatics.