Development of the Quantum Inspired SIBFA Many-Body Polarizable Force Field: Enabling Condensed Phase Molecular Dynamics Simulations

Abstract

We present the extension of the SIBFA (Sum of Interactions Between Fragments Ab initio Computed many-body polarizable force field to condensed phase Molecular Dynamics (MD) simulations. The Quantum-Inspired SIBFA procedure is grounded on simplified integrals obtained from localized molecular orbital theory and achieves full separability of its intermolecular potential. It embodies long-range multipolar electrostatics (up to quadrupole) coupled to a short-range penetration correction (up to charge-quadrupole), exchange-repulsion, many-body polarization, many-body charge transfer/delocalization, exchange-dispersion and dispersion (up to C10). This enables the reproduction of all energy contributions of ab initio Symmetry-Adapted Perturbation Theory (SAPT(DFT)) gas phase reference computations. The SIBFA approach has been integrated within the Tinker-HP massively parallel MD package. To do so all SIBFA energy gradients have been derived and the approach has been extended to enable periodic boundary conditions simulations using Smooth Particle Mesh Ewald. This novel implementation also notably includes a computationally tractable simplification of the many-body charge transfer/delocalization contribution. As a proof of concept, we perform a first computational experiment defining a water model fitted on a limited set of (SAPT(DFT)) data. SIBFA is shown to enable a satisfactory reproduction of both gas phase energetic contributions and condensed phase properties highlighting the importance of its physically-motivated functional form.

References in corpus (2)

• On the accuracy of the MB-pol many-body potential for water: Interaction energies, vibrational frequencies, and classical thermodynamic and dynamical properties from clusters to liquid water and ice
• Tinker-HP : Accelerating Molecular Dynamics Simulations of Large Complex Systems with Advanced Point Dipole Polarizable Force Fields using GPUs and Multi-GPUs systems

Cited by in corpus (7)

• Force-Field-Enhanced Neural Network Interactions: from Local Equivariant Embedding to Atom-in-Molecule properties and long-range effects
• Routine Molecular Dynamics Simulations Including Nuclear Quantum Effects: from Force Fields to Machine Learning Potentials
• Accurate Deep Learning-aided Density-free Strategy for Many-Body Dispersion-corrected Density Functional Theory
• Improving Condensed Phase Water Dynamics with Explicit Nuclear Quantum Effects: the Polarizable Q-AMOEBA Force Field
• FeNNol: an Efficient and Flexible Library for Building Force-field-enhanced Neural Network Potentials
• Velocity Jumps for Molecular Dynamics
• The Q-AMOEBA (CF) Polarizable Potential