Ice Isn't Slippery for the Reason You Were Taught

For most of the twentieth century, schoolbooks gave one of two reasons ice is slippery. Either the pressure of a skate blade locally melted the ice, or the friction of the blade did. Both are wrong, and both have been wrong long enough for the actual answer to be settled.

A skater exerts roughly a few megapascals on the ice through their blade. The pressure-melting curve for water sits well above that — you'd need closer to a hundred megapascals to drop the melting point a single degree. Real ice rinks run between −5 °C and −10 °C. The blade doesn't melt the ice with its weight. Frictional heating helps a little above −2 °C, but skaters glide easily at −20 °C, where heat-of-friction can't account for it either.

The answer, sharpened by a 2018 paper in Journal of Physical Chemistry Letters by Bart Weber and colleagues at AMOLF, the University of Amsterdam and the Max Planck Institute, is that the top layer of an ice surface is just different from the bulk. Molecules at the surface have only two of the three or four hydrogen bonds they'd have inside the lattice. Those weakly bound molecules don't sit still; they roll across each other.

Weber's group measured steel-on-ice friction across a 100-degree range and found it tracked the surface mobility almost exactly. The activation energy of friction (about 11.5 kJ/mol) matched the activation energy of those rolling molecules in molecular-dynamics simulations of the ice-air interface. Below about −80 °C the surface stiffens up, and ice is no longer slippery — boots squeak on it the way they squeak on dry rubber. The narrow window where ice slides easily is the window where the top molecules can roll. Skating temperature is the inside of that window.