Polarization separates spin currents from charge currents

Researchers at the University of Toronto (Toronto, Ont., Canada) are exploring all-optical spin currents in which electrons are driven around and sorted using light in accordance with their spins, rather than having bias voltages drag them around by their charges. "Spin-tronics," as the concept is called, is expected to provide new methods for data storage and processing and is also considered a first step toward building a solid-state quantum computer. The team induced an electronic current by optically manipulating spins without any bias voltage several years ago, according to Ph.D. student Ravi Bhat. Most recently team members have completed a mathematical proof and are in the process of demonstrating a pure spin current devoid of any net charge displacement.¹

In spin-tronics, the voltage that propels electrons through a charge carrier is replaced by a phase difference between two coherent light beams. For instance, the University of Toronto researchers used two different wavelengths of light to elevate electrons from the valence band into the conduction band of a gallium arsenide (GaAs) semiconductor material. Based on the 1.5-eV bandgap of GaAs, the researchers illuminated the material with a 1.55-eV signal at a wavelength of 775 nm and a 0.8-eV signal at 1550 nm. The higher-energy beam elevated an electron to the conduction band with a single photon, while the lower-energy beam required two photons. The quantum mechanical interference between the two possible electron paths induced a current in the conduction band without the need for bias voltage.

"The interference pattern is what gives us the current," Bhat said. "Depending on the phase between those two colors we get a spin current going one way or the other."

The phase difference between the two beams can be shifted experimentally by a variety of methods, which include inserting a glass slide in the beam path that has a different index of refraction for each of the two wavelengths, or simply changing the relative path length of the two beams. The resultant spin-current direction depends on the sine of the phase difference.

"If you shift the phase by pi then the current switches direction," Bhat said. "And if you switch it by one-half pi the current will go to zero."

Even though this type of spin current does not require a bias voltage, it is still not considered a pure spin current because of the net charge movement. The recent mathematical demonstration and upcoming experiment, however, involve a spin current without net charge movement that is induced by manipulating the relative polarization of the two beams.

When both beams have the same circular polarization, a spin-polarized current occurs in which all of the electrons of a particular spin (either up or down) move in one direction, dictated by the phase difference between the beams. What the researchers have shown mathematically and hope to prove in the upcoming experiment is that when the two beams are linearly polarized perpendicular to each other, the net charge movement goes to zero. Electrons with up-spins are expected to go in one direction and electrons with down-spins will go in the opposite direction, thereby inducing a spin current without an electrical current (see figure).

The Toronto experiments have produced pure spin currents all-optically, which is a new process, according to Bhat. All-optical spin currents are expected to enable rapid manipulation by ultrafast-laser pulses.

REFERENCE

1. R. D. R. Bhat and J. E. Sipe, Phys. Rev. Lett. 85, 25 (Dec. 18, 2000).