Minimal Basis Iterative Stockholder: Atoms in Molecules for Force-Field Development

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Atomic partial charges are crucial for force-field models, often derived from complex atoms-in-molecules (AIM) methods.
• Advanced models incorporate electron cloud distribution and atomic multipoles for improved accuracy.

Purpose of the Study:

• To introduce a novel, computationally efficient method for defining electrostatic force fields.
• To develop a concise approximation of electron density for simplified Coulomb interaction evaluation.

Main Methods:

• The Minimal Basis Iterative Stockholder (MBIS) method expands an approximate "pro-density" in a minimal basis of atom-centered Slater functions.
• Parameters are optimized by minimizing Kullback-Leibler divergence against reference electron densities from electronic structure calculations.
• An iterative algorithm with linear-scaling computational cost refines the pro-density.

Main Results:

• MBIS demonstrates effectiveness in modeling electrostatic interactions, outperforming 14 other AIM methods.
• The method is suitable for supramolecular systems and condensed phases due to its scalability.
• MBIS can be used as a density-fitting technique and for rescaling atomic polarizabilities.

Conclusions:

• MBIS provides an efficient and accurate approach to electrostatic modeling in computational chemistry.
• Its linear-scaling computational cost enables applications to large and complex molecular systems.
• The method enhances the development of more sophisticated force-field models for diverse chemical applications.