Properties of liquid clusters in large-scale molecular dynamics nucleation simulations

Raymond Angélil, Jürg Diemand, Kyoko K. Tanaka, Hidekazu Tanaka

Research output: Contribution to journalArticlepeer-review

31 Citations (Scopus)


We have performed large-scale Lennard-Jones molecular dynamics simulations of homogeneous vapor-to-liquid nucleation, with 109 atoms. This large number allows us to resolve extremely low nucleation rates, and also provides excellent statistics for cluster properties over a wide range of cluster sizes. The nucleation rates, cluster growth rates, and size distributions are presented in Diemand et al. [J. Chem. Phys. 139, 74309 (2013)], while this paper analyses the properties of the clusters. We explore the cluster temperatures, density profiles, potential energies, and shapes. A thorough understanding of the properties of the clusters is crucial to the formulation of nucleation models. Significant latent heat is retained by stable clusters, by as much as ΔkT = 0.1ε for clusters with size i = 100. We find that the clusters deviate remarkably from spherical - with ellipsoidal axis ratios for critical cluster sizes typically within b/c = 0.7 ± 0.05 and a/c = 0.5 ± 0.05. We examine cluster spin angular momentum, and find that it plays a negligible role in the cluster dynamics. The interfaces of large, stable clusters are thinner than planar equilibrium interfaces by 10%-30%. At the critical cluster size, the cluster central densities are between 5% and 30% lower than the bulk liquid expectations. These lower densities imply larger-than-expected surface areas, which increase the energy cost to form a surface, which lowers nucleation rates.

Original languageEnglish
Article number074303
JournalJournal of Chemical Physics
Issue number7
Publication statusPublished - 2014 Feb 21
Externally publishedYes

ASJC Scopus subject areas

  • Physics and Astronomy(all)
  • Physical and Theoretical Chemistry

Fingerprint Dive into the research topics of 'Properties of liquid clusters in large-scale molecular dynamics nucleation simulations'. Together they form a unique fingerprint.

Cite this