A Rejection Scheme for Off-Lattice Kinetic Monte Carlo Simulation

Area of Science:

• Computational materials science
• Atomistic simulations
• Chemical physics

Background:

• Off-lattice kinetic Monte Carlo (KMC) simulations require calculating rates for all possible atomic moves.
• This involves extensive energy landscape searches for saddle points, which is computationally intensive.

Purpose of the Study:

• To introduce a novel KMC algorithm for efficient off-lattice simulations.
• To reduce the computational cost associated with atomistic simulations of material systems.

Main Methods:

• Developed a rejection scheme replacing true rates with estimates based on atom-specific nearest-neighbor bond counts.
• Implemented a localized saddle point search focused on individual atoms to accept/reject transitions.
• Tested the algorithm on a growing two-species nanocluster model.

Main Results:

• The new KMC algorithm significantly reduces computation time.
• Achieved a 90% reduction for clusters of ~55 particles and 96% for clusters of ~65 particles.
• Performance improvement scales with the number of particles in the system.

Conclusions:

• The proposed KMC algorithm offers a substantial performance boost for off-lattice simulations.
• This method enables faster atomistic modeling of complex systems like nanoclusters.
• The approach is particularly beneficial for large systems where computational cost is a bottleneck.