A physicist applauds evidence for the quantum spin Hall effect.

I have been fascinated by the ballistic (collisionless) motion of charge carriers in solids since the start of my career. In practice this motion is often impeded by unavoidable impurities in the solid. But when it works, the charge carriers maintain their quantum properties while dissipating a minimum amount of energy.

Applying a strong magnetic field perpendicular to a two-dimensional conducting layer can accomplish the feat. Then, the quantum Hall effect kicks in, forcing the charges to the edges of the sample where they skip along in so-called 'chiral edge channels'. Backward scattering is virtually eliminated because that would require the charges to find a way to the opposite edge, where charges move in the opposite direction.

Recently, Laurens Molenkamp of the University of Würzburg in Germany and his colleagues took a step towards verifying the quantum spin Hall effect (M. König et al. Science 318, 766–770; 2007). This is where chiral edge channels form spontaneously in semiconductor insulators with peculiar electronic structures — namely, where the valence band is energetically higher than the conduction band because of the strong spin-orbit interaction between electron spins and electron velocities. This means that spin-up electrons are carried only by edge channels moving in one direction and spin-down elections are carried by edge channels moving in the opposite direction.

Molenkamp's team used a thin layer of mercury telluride sandwiched between two layers of mercury cadmium telluride. Because measuring spin current is difficult, they recorded the conductance of this middle layer to verify the ballistic transport that characterizes edge-channel transport. It was quantized, as predicted.

With further verification, the finding could lead to low-power devices based on the transport of spins rather than charges. Thus a quirk in the scientific field I have always loved might find a practical application.

Discuss this paper at http://blogs.nature.com/nature/journalclub