By Geoff Brumfiel

Low-power computer chips that don't rely on an electrical current to handle data have just come a big step closer.

The key result, unveiled today in Nature1, sounds deceptively simple: scientists have injected electrons into silicon at room temperature and set a majority of them spinning in the same direction.

But the experiment "is a real breakthrough", says Jaroslav Fabian, a theoretical physicist at the University of Regensburg, Germany. "Silicon is entering in a big way into spintronics."

Conventional electronics uses the electron's charge to move and process information. Spintronics -- or spin-transport electronics -- offers an alternative, relying on the electron's magnetism, or spin, to encode information.

Charge can be stored and is easy to manipulate using electric fields, properties that have enabled electrical engineers to develop ever-smaller chips over the past four decades. "But there are now increasing concerns that this progress may come to a halt," says Ron Jansen, a physicist at the University of Twente in Enschede, the Netherlands. As chips shrink they also become more complex and operate at higher speeds, driving an exponential growth in the amount of power needed to move electrons around them. Unless an alternative can be found, Jansen says, electronic devices may soon become too power-hungry to be practical.

Jansen and others believe that the answer may lie in spin. Rather than moving charges around, spin-based devices would simply have to flip the direction of an electron's internal bar magnet.

Single-layer success

The difficulty with spintronics is that although electron spins are readily aligned in magnetic materials, they are less well-behaved in the semiconductors used in the electronics industry. To build a spintronic device, researchers must be able to move spin-aligned electrons from magnetic to semiconductor materials without losing that alignment.

Previous experiments had managed the feat at ultra-cold temperatures, or by using exotic semiconductors such as gallium arsenide. But now Jansen and his colleagues have successfully injected spin electrons en masse into everyday silicon at room temperatures.

The team began with a magnetic nickel-iron alloy used in the read head of hard-disk drives, and an ordinary slab of silicon. Between the two they sandwiched an ultra-thin layer of aluminium oxide, roughly a nanometre thick. The aluminium oxide acts as an insulator, but when a voltage is applied some electrons are able to quantum-mechanically tunnel from the magnetic material into the silicon. The aluminium-oxide interface allows some spin directions to pass through more easily than others, creating a net excess of spins pointing in the same direction.

The key to the team's success was the single layer of aluminium oxide. Previous experiments had used two or more such layers, which stifled the flow of spin-aligned electrons. A single, thin coating between the magnetic material and the silicon allowed the electrons to flow smoothly at room temperature.

The simplicity and reliability of the technique are likely to make it a new standard in the field, says Shinji Yuasa, a researcher at the National Institute of Advanced Industrial Science and Technology in Tsukuba, Japan.

But there are still a few more steps before spintronics can come to fruition, says Jansen. Crucially, researchers must develop reliable ways to flip spins around once they are in silicon. Still, "The building blocks are there," he says. "Now it's just a matter of building something."