Slippery ice

Why is ice slippery when most other cold solids are not?

Cold, hard and smooth are not the same as slippery. Polished granite at minus five degrees is cold, hard and smooth, and you can walk on it; ice at the same temperature will put you on your back. So the explanation cannot be hardness or polish. It must be something ice does that granite does not.

The textbook answer for most of a century was pressure: your weight, concentrated on a small contact, melts the ice, and you slide on the water. A satisfying story. Shall we check whether the numbers support it?

They do not, and it is worth seeing by how much. Ice is unusual in that squeezing it lowers its melting point — but only by about 0.007 degrees Celsius per atmosphere. A skate blade may reach some tens of atmospheres, which buys a fraction of a degree. Yet people skate happily at minus ten and minus twenty. The mechanism is real; it is simply orders of magnitude too weak.

A second objection needs no arithmetic at all. Ice is dangerously slippery under a flat shoe, a car tyre, a bare palm — contacts whose pressures are far too low for the story even to begin.

What else could it be? Consider a molecule at the very top surface. In the bulk, each water molecule is locked in a lattice by hydrogen bonds reaching in every direction. At the surface there is nothing above; roughly half its bonds are missing. Is it held as firmly as one buried inside? Plainly not. So we should expect the outermost layers to be disordered and mobile even while the bulk is solidly frozen, and more so as the temperature climbs toward melting. This is surface premelting. Faraday guessed at something like it in the mid-nineteenth century, and modern surface-sensitive experiments and simulations have confirmed a mobile, liquid-like film on ice well below zero. Its thickness grows with temperature — less than a molecular layer when very cold, up to something on the order of tens of nanometres near melting, though estimates differ substantially between techniques.

So there is a lubricant present before you arrive. Now add what happens once you move: sliding dissipates energy as heat exactly where the surfaces touch, and ice needs very little warming to melt, so a moving contact generates its own film, more of it the faster you go. Which matters more? That is genuinely still argued. The premelted layer explains why ice is slippery when you stand still and frictional heating clearly matters at speed, while recent work suggests the mobile layer behaves less like bulk water than like a weakly bound, viscous film — which would change how the two combine. Anyone quoting a clean split is ahead of the evidence.

Now the comparison the question actually asked for. Look at how close each solid is to its own melting point. Ice at minus ten sits at about 96% of its melting temperature on the absolute scale; iron or quartz at that same air temperature sits at roughly a seventh of theirs. Everyday cold, for water, is a hair below melting; for almost every other solid we meet, it is deep-frozen.

Three refinements. First, pressure melting is not wrong, merely minor: a small real contribution under a sharp blade, and nothing under a shoe.

Second, slipperiness is not a fixed property of ice but depends on temperature, speed and load. Ice is most treacherous within a few degrees of melting and becomes markedly less slippery as it gets very cold; below roughly minus thirty, sledge runners and skates both drag noticeably. That temperature dependence is itself evidence for the premelted layer, whose thickness follows the same trend.

Third, the point about other solids generalises. Surface premelting is not unique to water — metals show it too — but it appears only near a material's own melting point, which for most solids is nowhere near a winter's day.

• Why water expands on freezing, and what that says about the hydrogen-bonded lattice.
• How curling stones are steered, and why a swept patch of ice behaves differently.
• Why premelting also governs frost heave, glacier sliding, and how snowflakes stick together.

Every term the collection defines is gathered in the glossary.