Despite extensive searches for alternatives to silicon-based transistors, nothing yet comes close to usurping silicon’s place in the computer-chip industry. Moreover, although miniaturization of silicon electronics cannot continue forever, it is possible that there may be other ways in which it can manipulate and use the properties of electrons. Spin-based electronics, or spintronics, provides one such possibility through the use of aligned electron spin currents instead of charge currents. To add to this, Andrei Bernevig and co-workers1 now propose that the orbital Hall effect could produce orbital currents to keep silicon ahead of the pack.

In the classical Hall effect, opposite charges accumulate at the edges of a conductor in crossed magnetic and electric fields, as a result of the Lorentz force on the electrons. This sets up a voltage difference across the sample that can be measured using electrical transport techniques. Detecting spin accumulation requires more work. In an applied electric field, spins in a semiconductor can be deflected to opposite edges if there is a strong enough interaction between spin and orbital motion (spin–orbit coupling) — this sort of plays the role of the Lorentz force in the Hall effect. Polarized light hitting the polarized spins will be rotated, and this is a measurable effect2.

Unfortunately, this technique does not work in silicon because the spin–orbit coupling is too weak. But consider hole-doped silicon, in which the p-orbitals can be populated. In an electric field along the y direction, the holes will flow along x, with their orbital local moments polarized along z. Holes flowing along +x will tend to populate the px+ipy orbitals, giving a net orbital polarization along +z. Similarly, holes flowing along –x will lead to orbital polarization along –z. Bernevig et al. believe that the resulting accumulation of oppositely polarized orbitals will affect circularly polarized light in different ways, and so can be detected like the spin Hall effect.

As the orbital current does not carry any charge, there would be no heat dissipation — the bane of silicon chips. This sounds promising, but the effect must first be verified experimentally. Then we can talk about orbitronics.