Droplet undercooling effect

Seed Crystals. Crystals inside droplets may occasionally stick out of the surface over several nanometers. If such a drop encounters another one by Brownian motion (see Section 13.2.1), the protruding crystal may occasionally pierce the surface of that droplet. If the latter is still fully liquid, the crystal may act as a seed and induce crystallization. This has been observed to occur in emulsions of hexadecane in water, where part of the drops were solid and part liquid (undercooled). It is a slow process, for instance taking two weeks for completion. It has been calculated that about one in 107 or 108 encounters was effective in such a case. 

The results of Bartell on undercooled small droplets (discussed in Section 2.1 above) raise real questions about the existence of a spinodal singularity, and especially about the type of analysis used by Pruppacher. They show that nucleation can be avoided in liquid water all the way down to -73°C, with no sign of an anomaly at -45°C (however, a small droplet has a high enough internal pressure that any anomaly may be suppressed by that effect alone). Recent simulations have also raised questions about the spinodal hypothesis. Poole et al. simulated water at low temperatures and over a range of pressures to explore the phase diagram. They suggested that the... 

---