Melting ice with salt - how does it start?

I know that salt reduces the melting point of ice. I can see how a little bit of salty water can melt ice to water which can dissolve even more salt. But how does it even start? There is ice, which is hard. There are grains of salt, which are hard. So how does the salt enter the ice, to begin with?

• Well, there must be points where one hard body touches the other. That's where it all starts. – Ivan Neretin Dec 13 '17 at 7:31
• – Nilay Ghosh Dec 13 '17 at 7:32
• Well, a surface of for example ice is not ice|air its more like ice|lessdenseice|chaoticice|water|air if you are in a normal household setting. – Fl.pf. Dec 13 '17 at 7:33
• @IvanNeretin The way you formulated it that's how pregnancy starts, too:) To the topic, there is solid-state chemistry existing for a good reason; also, ice is considered to have a pronounced "surface melting" effect. – andselisk Dec 13 '17 at 7:34
• Well, surface molecules wiggle harder than those in the bulk. Sometimes they jump into the air and return in a short while, or in a long while, or never. In effect, a brick of ice is covered by a few molecular layers (or more) of confined water which is neither quite liquid nor quite solid. – Ivan Neretin Dec 13 '17 at 11:35

I believe that Ivan's comments are on the right track. Generally any melting phenomenon to do with ice begins with the fact that the surface of ice is disordered. By this I mean that ice on the whole is a periodic crystal like other solids, but at the surface there are imperfections, primarily due to the extra conformational freedom afforded to the surface molecules. This distortion then propagates down quite a few layers [1]. Ref. [1] discusses the properties of this surface layer which is liquid-like, and points out that this layer exists even down to quite low temperatures.

If these surface molecules "see" something charged (each ion that is) like $\ce{NaCl}$, then the surface molecules already have the conformational freedom to solvate the ions. Additionally, there is an energetic benefit to this because the hydrogen bonds between these ions and water are stronger than the water-water hydrogen bonds.

In general, melting phenomena are extremely complicated and the question of how melting "begins" is has not really been answered completely or satisfactorily (can provide refs. later if you want). This case is a bit different though because it is really a solvation phenomenon, which is adequately explained by the liquid-like layer at the surface.

[1]: Beaglehole, D., & Nason, D. (1980). Transition layer on the surface on ice. Surface science, 96(1-3), 357-363.

One possible mechanism: water evaporates from the ice, condenses onto salt pellets, forms salt solution which can then directly take up more ice and salt. Next thing you know, the road is covered in a corrosive solution that's getting all over the underside of your car.

• Plausible as it sounds, I don't believe this is the main mechanism. – Ivan Neretin Dec 13 '17 at 11:37
• Hence, "one possible mechanism". Would like to see, and probably +1, your answer. – Oscar Lanzi Dec 13 '17 at 12:27