Theory of crystal nucleation of glass-forming liquids: Some new developments

Abstract

Experiments on crystal nucleation are predominantly interpreted theoretically in terms of classical nucleation theory (CNT). This theory relies on thermodynamic concepts developed by Gibbs. Their application implies that the bulk properties of critical clusters governing nucleation are widely identical to those of the newly evolving macroscopic phases. Size effects are incorporated into CNT exclusively by a curvature dependence of the surface tension. This method works well but only down to temperatures near to the maximum of the steady-state nucleation rate. This maximum (at (Formula presented.)) is correlated with the glass transition temperature, (Formula presented.), that is, (Formula presented.). For lower temperatures, significant deviations between theoretical predictions and experimental data are observed. We describe here different methods how a curvature dependence of the surface tension can be introduced to describe crystal nucleation correctly in the range (Formula presented.). Problems occurring at (Formula presented.), denoted sometimes as “breakdown of CNT,” are shown to be caused by the glass transition of the liquid. Modifications of CNT are advanced resolving them and giving also the possibility of interpretation of a variety of further experimental data in crystal nucleation (including hysteresis effects in cooling and heating, nucleation flashes in heating) in terms of CNT.