Nucleation free-energy barriers with Hybrid Monte-Carlo/Umbrella Sampling

Area of Science:

• Computational materials science
• Chemical physics
• Thermodynamics

Background:

• Accurate computation of nucleation free-energy barriers is crucial for understanding phase transitions.
• Molecular dynamics (MD) simulations are powerful tools for studying these phenomena.
• Hybrid Monte Carlo (HMC) and Umbrella Sampling (US) offer advanced methods for free-energy calculations.

Purpose of the Study:

• To evaluate nucleation free-energy barriers using a combination of MD, HMC, and US.
• To compute the crystallization barrier of NaCl from its melt.
• To assess the performance and convergence of HMC for different simulation parameters.

Main Methods:

• Hybrid Monte Carlo (HMC) simulations combined with Umbrella Sampling (US).
• Assessment of HMC performance based on time-steps and MD steps within cycles.
• Exploration of a 'non-Metropolised' HMC version and approximations using isothermal-isobaric MD trajectories.

Main Results:

• HMC convergence is independent of time-step for potential energies and densities.
• HMC acceptance ratio is highly dependent on time-step, achieving near 100% with typical MD time-steps.
• The HMC/US method, including its approximations, accurately computes NaCl nucleation free-energy barriers, showing excellent agreement with reported values.

Conclusions:

• The HMC/US technique provides a robust and accurate method for calculating nucleation free-energy barriers.
• Optimizing time-steps in HMC is critical for simulation performance and acceptance ratios.
• Approximations to HMC/US maintain accuracy, suggesting flexibility in computational approaches.