Systematic coarse-graining in nucleation theory.

Area of Science:

• Physical Chemistry
• Computational Materials Science
• Statistical Mechanics

Background:

• Atomistic Monte Carlo (MC) simulations are standard for nucleation rate predictions.
• Existing methods often use a truncated state space, considering only nucleus size as an order parameter.
• This approach leads to significant deviations (3-5 orders of magnitude) from molecular dynamics (MD) simulations for Lennard-Jones argon.

Purpose of the Study:

• To formulate a nonequilibrium statistical mechanics of nucleation in an extended state space.
• To explain the discrepancy between MC and MD simulations for nucleation rates.
• To introduce a more comprehensive theoretical framework for nucleation phenomena.

Main Methods:

• Formulation of nucleation theory in an extended state space, including internal energy and nucleus momentum.
• Comparison of predictions from the extended model with existing MC simulations and MD data (Diemand et al., 2013).
• Analysis of the emergent properties of nuclei within the extended framework.

Main Results:

• The extended state space model successfully reconciles the differences between MC and MD simulations.
• A natural definition for nucleus temperature emerges, consistent with the McGraw and LaViolette fluctuation law.
• The framework allows for the extension of existing nucleation theories to include additional order parameters.

Conclusions:

• The standard truncated state space is insufficient for accurate nucleation rate predictions.
• Incorporating internal energy and momentum into the state space significantly improves theoretical models.
• The developed extended framework offers a more robust and versatile approach to nucleation studies.