Bright-field microscopic images of fluff-like water crystals. Typical appearance of nucleating water crystals showing an asymmetric brush-like crystal growth in trans-DCE with 80 ppm water cooled to 213 K. Scale bar 50 mm.
Detailed view of hair ice aka frost beard taken at Mount Maxwell, Salt Spring Island, British Columbia, Canada (reproduced with permission).
The theoretical understanding of crystal nucleation has not improved much since the work of Ostwald and Gibbs a century ago. An exception is the work by Daan Frenkel (Science 277, 1975 (1997)), on the role of critical density fluctuations on crystallisation, but these principles are often applied within chemical engineering without a real understanding of the applicability of the underlying chemical physics.
In a publication in the journal ChemComm, we describe a study of the role of liquid-liquid phase separation on crystallisation by using mixtures of two liquids (a solute liquid and a solvent liquid) and cooling them very rapidly until they undergo liquid-liquid demixing. The mixtures are chosen such that the solute liquid will freeze immediately on demixing.
In mixtures of water with non-polar solvent liquids, we find that this gives rise to the growth of surprisingly non-symmetrical wires of crystalline water, which leads to the formation of fluff balls. When a water ice crystal grows in an environment low in free water molecules (such as in the cold air of the atmosphere), it forms hexagonal snowflake patterns caused by the limitation in the rate of diffusion of the free water molecules. In fact, there are currently 121 known categories of aqueous solid precipitation particles, none of which is shaped like a wire or fluff.
There is only one little-known wire-like form of water ice, known as Haareis. Haareis is formed when a template (such as the bark of a dead tree or porous soil) is just above 0° C while the air above is just below. However, in the results described here, there is no external physical template. Thus, we believe that here the spinodal decomposition of water and the non-polar solvent-liquid is acting as an effective template for crystal growth. We believe that such a liquid phase-separation induced crystal templating might be a general phenomenon with wider applicability.