A Bright Future for Spintronics

A novel combination of magnetic and semiconductor materials could lead to optical and electronic devices that are faster and lower-powered.

The desire to build smaller, faster, cheaper electronics has prompted a number of researchers to try using the "spin" of an electron in transistors. These "spintronic" transistors could be highly energy-efficient and do more computation than traditional transistors in a smaller space. In addition, in optoelectronic applications, lasers and light-emitting diodes that take advantage of the spin of electrons could increase the data-carrying capacity of light.

But one of the key hurdles in this emerging field is that the magnetic and semiconductor materials needed to make a spintronic device are notoriously incompatible.

Now researchers at the Ohio University and Ohio State University have developed a magnetic-semiconductor system that, based on initial tests, looks like it could be an advance. Arthur Smith, professor of physics at Ohio University, and his colleagues have successfully grown manganese gallium, a magnetic metal, on gallium nitride, a common semiconductor that is used to make blue lasers and LEDs, and to amplify radio frequency signals.

The researchers say that the spacing of the atoms in the material layers is a near-ideal match, creating a smooth interface between the layers, and thereby increasing the chances of producing a workable spintronics device. Without a clean interface, says Smith, when electrons travel across the barrier between the metal and semiconductor, they can lose their original spin, ruining the device. Moreover, their new system maintains its magnetic properties at room temperature, says Smith. Many potential spintronic materials work well only at extremely cold temperatures, although recent developments have produced some room-temperature materials (see "A New Spin on Computing").

Although further testing is needed to confirm that electrons will maintain their spin characteristics while traveling from metal to semiconductor, Smith says these early tests are encouraging. "We think there's a good chance that it'll work pretty well," he says.

Electronic systems that use the spin of an electron--a quantum mechanical property that comes in two varieties: up or down--would work similarly to today's transistors, but have several advantages. Presently, electrical current alone is responsible for the logic functions in circuits. Current flowing through a transistor represents a 1; the absence of current, a 0. If the spin of an electron could be controlled, a "spin up" electron could represent a 1, and "spin down" a 0.

Unlike electrical current, spin can be maintained even if the power is off, and a spintronic circuit would use less power because a current wouldn't need to be constantly applied. This is why companies such as Freescale Semiconductor are exploring spin-based solid-state memory (see "A Better Memory Chip").