On several recent mornings my windshield is covered with frost. Part of this is a very fine white film made up of microscopic crystals, interspersed with a few much larger (visible structure) crystals every 6 inches or so. Oddly, every large crystal is surrounded by a clear area free of the white film that uniformly covers the rest of the glass. I am guessing the large crystal attracts the vapor in the immediate area, and so it grows at the expense of the smaller crystals. This seems logical (else how would the larger crystal grow larger than the others?). However, the clear area is perhaps 3mm wide… this seems a long way for such an attraction to extend.
What gives?
It’s not attraction, it’s diffusion, it’s “staining.” Well, actually it’s the reverse, it’s “unstaining.”
Any of the invisible water-gas which touches the ice surface will stick. But then those water molecules are no longer wandering around, so the region of air near the ice surface is depleted of water molecules relative to regions farther out. The water molecules diffuse through the air slowly, so if an ice surface is swallowing all molecules which touch it, then it takes awhile for more molecules in the distant air to wander in and re-fill the air which has lost its water molecules.
Remember, at distances this close to a solid surface the “boundary layer effect” is very strong and wind is negligible, so the air acts almost like jelly or syrup. Evaporated water suspended in the air acts like a “stain.” Imagine that water molecules were visible as red gas in the air. A crystal which trapped all the red molecules touching it would soon start to “emit” an expanding halo of non-redness. It’s just the opposite of a crystal of red dye placed onto jello: the dye becomes surrounded by an expanding region of redness, with the reddest part next to the dye. An ice crystal does it backwards, growing rather than shrinking, and it becomes surrounded with an expanding region of “dryness”, with the dryest air hovering adjacent to the ice surface.