Mutually polarizable QM/MM model with in situ optimized localized basis functions

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Quantum Mechanics

Background:

• Accurate modeling of complex molecular systems requires sophisticated computational methods.
• Existing quantum-mechanical/molecular mechanics (QM/MM) approaches have limitations in accurately describing the QM/MM interface.
• The need for improved QM/MM methods that handle charge penetration and overpolarization is critical.

Purpose of the Study:

• To develop and validate an advanced QM/MM approach with in situ optimization of localized orbitals.
• To address challenges of overpolarization and unphysical charge transfer at the QM/MM interface.
• To improve the accuracy of QM/MM energetics, particularly for hydrogen bonds and solvation shells.

Main Methods:

• Implemented in situ optimization of localized orbitals within a QM/MM framework.
• Utilized ONETEP linear-scaling density functional theory for the quantum subsystem.
• Employed the AMOEBA polarizable force field for the classical subsystem with multipolar electrostatics.
• Developed a total energy minimization scheme for the coupled QM/MM system.
• Introduced a modified repulsive interaction to mitigate charge penetration and overpolarization.

Main Results:

• The in situ optimized basis functions enhance QM/MM interface accuracy but require careful handling of charge penetration.
• A novel repulsive interaction term effectively resolves overpolarization and unphysical charge transfer issues.
• The validated method accurately models hydrogen bonds, water-ion pairs, and solvation shells.
• Explicit polarization in the MM component improves agreement with all-quantum calculations.
• The QM/MM approach allows for minimal QM regions while maintaining accurate energetics.

Conclusions:

• The developed QM/MM method with in situ orbital optimization provides a robust and accurate approach for molecular simulations.
• The strategies to counteract overpolarization are effective, leading to reliable energetic predictions.
• This model enables efficient and accurate modeling of complex chemical systems with explicit polarization.