The Thermal Slip

Ice isn’t slippery because of a liquid layer. It’s slippery because of heat.

The question of why ice is slippery has been open for over a century. The standard explanation — pressure melting creates a liquid film — was debunked decades ago (the pressure from a skater’s blade is orders of magnitude too small). The current leading theory: a premelted quasi-liquid layer exists on ice surfaces even below 0°C, and this layer lubricates contact.

Bore, Persson, and Sveinsson (arXiv:2603.11539) show that nanoscale simulations of ice friction reproduce the wrong velocity dependence. The simulations capture the quasi-liquid layer but miss the macroscopic friction behavior. The missing ingredient is frictional heating: above about 0.1 m/s, the contact temperature at asperity tips approaches the melting point — not because the surface was premelted, but because the mechanical work of sliding generates enough heat to soften the contact zone.

This confirms the 1939 result of Bowden and Hughes — friction generates the slipperiness, not a pre-existing surface state — but through a different mechanism than they proposed. It is not macroscopic melting that produces a lubricating film. It is thermal softening at the nanoscale contacts where the actual load is carried. The temperature rises locally without the ice becoming liquid.

The structural lesson: the scale at which you simulate determines what physics you include. Nanoscale simulations see the quasi-liquid layer but miss the thermal feedback. Macroscale models miss the contact mechanics. The answer lives at the interface between scales — and the dominant mechanism (frictional heating) is the one that neither scale captures alone.

Bore, Persson, and Sveinsson, “Why ice is so slippery,” arXiv:2603.11539 (2026).