| In the classical view, the nucleation of a new phase is a stochastic process, involving fluctuations to surmount an energy barrier. Over recent years good progress has been made in understanding the nucleation of solids in melts at very high large supercooling. This presentation, in contrast, focuses on the opposite extreme of very small supercooling. Nucleation is then exclusively on heterogeneities, and we show that it is best analyzed as an athermal process in which the number of nucleation events is determined by temperature, independent of time. This process is deterministic, not stochastic. Two examples are chosen to show the relevance of athermal nucleation of solidification: the development of microstructures in cast metals and the control of freezing in living systems. In solidifying metals, controlled nucleation is very inefficient, with grains forming on only 0.1 to 1% of the added particles. The inefficiency arises from interactions competing over a range of length scale, and the importance of particle size distribution is highlighted. Biogenic ice-nucleating agents, important in a wide range of systems, operate in the athermal nucleation regime, and this explains both the development of their form and the limitations on their operation. In both cases, attention for future work must focus on the analysis of the interactions between the liquid and the nucleant substrate. Recent work (modeling and microscopical observations) on adsorption and layered ordering in the liquid will be reviewed. |
|