A coupled polarization-matrix inversion and iteration approach for accelerating the dipole convergence in a...

Area of Science:

• Computational Chemistry
• Molecular Dynamics
• Quantum Chemistry

Background:

• Accurate simulation of condensed phase systems requires precise calculation of induced dipoles.
• Polarizable intermolecular potential functions (PIPF) are crucial for modeling molecular interactions.
• Existing methods like direct iterative approaches struggle with convergence when including all intramolecular interactions.

Purpose of the Study:

• To develop a robust and efficient method for calculating induced dipoles in condensed phase systems.
• To address the convergence issues encountered in direct iterative methods for PIPF.
• To accelerate the computation of molecular polarization in simulations.

Main Methods:

• Formulation of a coupled polarization-matrix inversion and iteration (CPII) algorithm.
• Reformulation of the Thole interaction dipole model using molecular block matrices.
• Application of the CPII method to systems of water, NMA, and an alanine pentapeptide.

Main Results:

• The CPII method successfully achieved rapid convergence of dipole induction polarization in all tested systems.
• The direct iterative approach failed to converge when intramolecular interactions were fully included.
• CPII demonstrated reduced computational costs compared to direct iteration by decreasing iteration steps.

Conclusions:

• The CPII method provides a stable and efficient solution for calculating induced dipoles using PIPF.
• CPII overcomes numerical stability problems associated with including all intramolecular interactions in the Thole model.
• This advancement enables more accurate and faster molecular dynamics simulations of polarizable systems.