Development and Implementation of Atomically Anisotropic First-Principles Force Fields: A Benzene Case Study.

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Physical Chemistry

Background:

• π-interactions are crucial in chemical and biological systems.
• Aromatic molecules pose challenges for accurate force field development due to complex intermolecular forces and anisotropic electron densities.
• Existing *ab initio* force fields often lack accuracy for bulk properties or phase transferability.

Purpose of the Study:

• To develop an accurate and transferable *ab initio* force field for aromatic interactions.
• To demonstrate the utility of atomic-level anisotropy in molecular modeling.
• To introduce a flexible computational tool for anisotropic intermolecular interactions.

Main Methods:

• Development of the MASTIFF (*ab initio* Molecular Anisotropic Site-Site Force Field) model.
• Incorporation of atomically anisotropic descriptions of intermolecular interactions.
• Creation of an OpenMM plugin for customizable anisotropic functional forms in molecular dynamics (MD) simulations.

Main Results:

• The MASTIFF force field for benzene shows high accuracy for liquid phase properties.
• The model demonstrates transferability to gas and solid phases, outperforming existing models.
• A computationally efficient OpenMM plugin was developed for general use in MD simulations requiring nonspherical atomic features.

Conclusions:

• Atomic-level anisotropy is essential for next-generation *ab initio* force field development.
• The MASTIFF force field provides a robust and accurate model for aromatic interactions across different phases.
• The developed OpenMM plugin facilitates the application of anisotropic models in diverse MD simulations.