Overcoming timescale and finite-size limitations to compute nucleation rates from small scale Well Tempered Metadynamics simulations

TL;DR

This paper presents a method combining enhanced sampling and finite-size corrections to accurately compute nucleation rates from small-scale simulations, enabling realistic timescale predictions for condensation processes.

Contribution

It introduces a novel approach that overcomes timescale and finite-size limitations in molecular dynamics simulations of nucleation, bridging the gap to real-world conditions.

Findings

Successfully computed nucleation times of hours at realistic supersaturation.

Demonstrated approach on argon condensation, validating the method.

Bridged the gap between small-scale simulations and real physical nucleation rates.

Abstract

Condensation of a liquid droplet from a supersaturated vapour phase is initiated by a prototypical nucleation event. As such it is challenging to compute its rate from atomistic molecular dynamics simulations. In fact at realistic supersaturation conditions condensation occurs on time scales that far exceed what can be reached with conventional molecular dynamics methods. Another known problem in this context is the distortion of the free energy profile associated to nucleation due to the small, finite size of typical simulation boxes. In this work the problem of time scale is addressed with a recently developed enhanced sampling method while contextually correcting for finite size effects. We demonstrate our approach by studying the condensation of argon, and showing that characteristic nucleation times of the order of magnitude of hours can be reliably calculated, approaching realistic supersaturation conditions, thus bridging the gap between what standard molecular dynamics simulations can do and real physical systems.