Multiple time step integrators in ab initio molecular dynamics.

Area of Science:

• Computational Chemistry
• Molecular Dynamics Simulations
• Quantum Chemistry

Background:

• Multiple time-scale algorithms enhance molecular dynamics (MD) efficiency by exploiting time-scale separation in chemical systems.
• Existing methods are well-established for empirical potentials but limited for ab initio MD due to challenges in splitting the potential energy surface.
• Efficiently handling the fast and slow varying components of ab initio potentials is crucial for advancing simulation accuracy and speed.

Purpose of the Study:

• To develop and present novel schemes for efficient time-scale separation in ab initio molecular dynamics.
• To enable larger time steps in ab initio MD simulations without compromising accuracy or energy conservation.
• To significantly accelerate ab initio MD calculations for complex chemical systems.

Main Methods:

• Developed two new schemes for time-scale separation in ab initio calculations.
• Scheme 1: Fragment decomposition of the molecular system.
• Scheme 2: Range separation of the Coulomb operator within the electronic Hamiltonian.

Main Results:

• Demonstrated successful application to water clusters and a solvated hydroxide ion.
• Achieved outer time steps of 2.5 fs, comparable to empirical potential methods.
• Enabled bond breaking and reformation during dynamics with large time steps.
• Observed computational speedups of up to 4.4x compared to standard Born-Oppenheimer ab initio MD.
• Maintained excellent energy conservation and accuracy.

Conclusions:

• The presented schemes effectively enable multiple time-scale separation in ab initio MD.
• These advancements significantly accelerate ab initio MD simulations, making them more practical for complex chemical problems.
• The methods offer a substantial improvement in computational efficiency for quantum chemistry simulations.