Germanium Revived from the Spintronics Graveyard
In the spin Hall effect (SHE), an electric current flowing through a material creates a flow of electron spin in the perpendicular direction. Now researchers have observed a surprisingly strong SHE in the semiconductor germanium, a finding that could ease the way to building spintronics devices, which would work by controlling electron spin, rather than charge. The research team also observed spin accumulation over long distances, a property especially useful for spintronics components having more complexity. Germanium may be more practical than other spin Hall materials because it interfaces more easily with silicon, the dominant material of today’s electronics.
Typical electronic circuits involve net flows of electric charge, but not of spin, the angular momentum carried by electrons. Spin currents can arise, for example, if spin-up and spin-down electrons tend to move in different directions. Spintronic devices would manipulate such currents for improved computing and electronic performance.
For several decades, researchers have explored the SHE as a means to create and control spin currents. Compound semiconductors such as gallium arsenide and indium gallium arsenide have been the materials of choice, partly because they transform charge currents into spin currents more efficiently than pure materials such as germanium. Germanium might otherwise be a good choice, because it is compatible with silicon, for which there is a large manufacturing infrastructure. But experiments, as expected, have only detected very small spin currents generated through the SHE in germanium.
Although germanium generates spin currents poorly, such currents, once created, tend to last much longer than in other materials. This effect might compensate for the low efficiency and produce a large spin accumulation, suspected Federico Bottegoni and colleagues from the Polytechnic University of Milan. They decided to take another look at germanium, focusing specifically on the accumulated spin density, rather than on the spin current.
The team fabricated a 100-by-250-micrometer rectangle of germanium doped with phosphorous on a silicon surface. They then drove a current in the long direction and sought to measure the SHE by detecting the accumulated spin density vs position across the short direction—spin-up electrons should accumulate at one edge and spin-down at the opposite edge. The team produced a spin-density map using a technique called magneto-optical Kerr microscopy, which is based on detecting the rotation in the polarization of light sent through the material.
To their surprise, the results showed a spin density accumulated through the SHE nearly 100 times larger than that reported for a similar sample of indium gallium arsenide and comparable to the highest value measured for gallium arsenide. The observed effect in germanium, as team member Marco Finazzi points out, is also more pronounced in showing accumulation over a longer distance. The experiments found a variation in spin density over a distance of 80 micrometers, nearly 100 times farther than seen previously in other semiconductors. The researchers suggest that this long-distance spin effect could make the SHE in germanium especially useful in making so-called multiterminal spintronics devices, such as low-power transistors.
The next step, says Finazzi, will be to build the SHE into a device able to do useful logic operations, the building blocks of digital computing. “The aim of spintronics is a device where the spin of the electrons flowing into a channel would be electrically controlled,” he says. “Despite numerous efforts, no one has yet been able to demonstrate a working device.”
“This is a very timely and important contribution,” says spintronics theorist Igor Zutic of the State University of New York at Buffalo, who notes that germanium has largely been forgotten in recent spintronics research. “I would expect to see important proof-of-principle demonstrations in one or two years,” he says. Zutic says that in today’s devices, the metallic connections between different parts of a circuit produce significant heat and represent one of the main bottlenecks to continued device miniaturization. He believes that semiconductor spin current contacts could be a possible alternative.
This research is published in Physical Review Letters.
–Mark Buchanan
Mark Buchanan is a freelance science writer who splits his time between Abergavenny, UK, and Notre Dame de Courson, France.