It is easier to turn a jar lid by gripping its edge than its center. Layered crystals such as graphite behave similarly when one layer is twisted relative to another: the resistance arises not from the whole contact area but only from its rim.

Previously, it was thought that the resistance to rotation should grow with the square of the contact area, because a circular pillar twice as wide contains four times as many atoms held by weak forces. However, measurements showed that the torque grows linearly. This means that the inner regions of the contact barely resist rotation.

Atomistic simulations explained the paradox. At the edges of the contact, the atomic lattice rearranges to lower its energy, and it is this narrow strip, about one and a half nanometers wide, that takes up most of the load. The inner part remains almost uninvolved, like the middle of a lid that you cannot get a grip on.

This difference shows that sliding and twisting are governed by different pathways of stress transfer. For devices where layers need to be rotated precisely, for example in twistronics, it is important to consider that the edge of the contact, not its area, sets the resistance.

The jar lid, when grasped at its rim, turns out to be an accurate model for the behavior of atomic layers. By measuring the torque on micron-sized pillars, physicists saw that in the nanoworld the contact edge acts like a handle by which nature rotates the material.