6. Ice crystal growth and evaporation

The dynamic diffusion chamber described earlier is ideal for growth of crystals under controlled temperature, supersaturation and air velocity. Supersaturation and air velocity are equivalent for a dendrite growing under ventilated conditions, so the effect of sudden change of velocity is equivalent to change of supersaturation. Sudden change of temperature can be accomplished simply by moving the vertical position of a crystal growing on a filament in the chamber. Hallett and Knight (1994) discussed the symmetry of arms grown on either side of each of the six branches of a dendrite grown near --15C. It is inferred that environmental fluctuations along a dendrite fall path are required to initiate side arms symmetrically; under uniform growth conditions the position of side arms is random. Increase of velocity leads to a faceted plate sprouting to a dendrite at a critical value of velocity, depending on supersaturation. The role of habit is manifest in a specific way for other crystal shapes which influence fall velocity, with a maximum for crystals with equal length in principal crystallographic directions (Redder and Fukuta, 1991). This is expressed as a functional relation between the dimensionless Reynolds and Best numbers, (ratio of inertial to viscous forces and a combination of Reynolds number and drag coefficient to compute terminal fall velocity).

Further insight into the ice crystal economy of clouds has come from experiments in evaporation of ice under subsaturated flow in a diffusion chamber. Crystals are first grown under various conditions to give selected shapes and sizes, and then evaporated under reverse flow under known subsaturation. Significant break-up occurs for crystals columns/dendrites below 80% relative humidity (Dong et al, 1994).

Studies of the nature of the surface layer of ice have been made through SO take up rates and by electrical effects. The first technique uses ice or dilute solution spheres frozen in liquid nitrogen as packing for a column, through which SO is passed until break-through occurs, (Conklin and Bales 1993; Conklin et al 1993). Absorption can be measured in a surface layer which is estimated as in the range 0.003 to 0.03 m thick, at --60C to 0.5 m to 3 m at --1C. Results are applied to SO deposition to snow surfaces and could be applicable to snow scavenging. Dong and Hallett (1992) grew droplets and crystals in a diffusion chamber and found that at temperatures above --4C ice and supercooled water acquired a positive charge whereas below this temperature ice acquired a negative charge, while the supercooled drops still charge positively down to --8C. Thus evidence is strong that the surface of ice is in a somewhat disordered state, influenced by ambient saturation and the presence of impurities, as well as possibly defects in the ice lattice. A highly disordered lattice will have some properties of a liquid layer. A technique for studying this, such as a tunneling microscopy, may be possible.