A New Spin on Silicon Chips
Today’s computers rely on moving and storing electronic charge in semiconductors. They ignore another property of electrons known as spin. Manipulating an electron’s spin, as opposed to manipulating its charge, is faster and takes much less energy. That means electronic circuits that store and process data using an electron’s spin would make computers faster, smaller, and more energy efficient.
While spintronic devices are easy to make using magnetic metals, to do so using semiconductors is challenging. So far, researchers have made spintronic devices from gallium arsenide, but making them from the far cheaper silicon has been difficult. Ian Appelbaum and his colleagues at the University of Delaware have now made the first silicon-based spintronic device, which they describe in this week’s Nature.
Appelbaum says that silicon-based spintronics could be easily incorporated into present-day integrated circuits. Also, theory says that electron spins survive for a long time in silicon.
Electron spins come in two directions: “up” and “down.” In conventional charge-based electronics, electrons’ spins randomly fluctuate. But in a spintronics device, an up or down spin could represent a “1” or “0”.
Some forms of spintronics are already used in computers. For instance, the read heads in high-capacity computer hard drives use a metal-based spintronic device called a spin valve. The valve contains a nonmagnetic metal layer sandwiched between two magnetic layers, one of which has a fixed magnetic-field direction. As the read head travels over the 1s and 0s stored as magnet fields on the disk, the fields in the two magnetic layers flip back and forth, aligning and misaligning. When the magnetic fields are aligned, electrons with spins in the same direction flow through the device, representing a bit 1.
The trouble with making silicon spintronic devices has been measuring spin direction, says Paul Crowell, who does spintronics research at the University of Minnesota, in Minneapolis. There are ways to inject electrons with aligned spins into silicon, but without being able to measure spin in the material, one can’t know whether the electrons maintain their spin in silicon, let alone control their spin. In gallium-arsenide devices, one can use light to measure spin, Crowell says, but this is not possible in silicon, an area in which “one has to fly completely blind.”
To detect the spin coming out of silicon, Appelbaum and his colleagues made a detector with a unique layered structure: a nickel-iron layer on top of a copper layer, which is deposited on a silicon substrate.
A 10-micrometer-thick silicon layer lies on top of the detector, and above that comes the electron-injecting part of the chip. In the injector, highly energized electrons pass through a magnetic iron layer, which filters out all electrons with a spin down. Spin-up electrons pass through the 10-micrometer silicon layer and go to the detector. In the detector, if the nickel-iron layer’s magnetic-field direction matches the spin direction, electrons go through to the silicon substrate, leading to a small current. But if the researchers flip the direction of the magnetic field in the nickel-iron layer, there is no current.
The key is the detector’s complex layered structure, which the researchers make using a special technique to deposit silicon on top of the nickel-iron layer. “It’s a very ingenious scheme to electrically generate and transport spins in silicon, [to] electrically detect the spins, and doing all of this on a chip,” says David Awschalom, a physics professor who studies semiconductor spintronics at the University of California, Santa Barbara.
Others believe that the work is an experimental demonstration of a principle but is not very practical. “What this paper shows is that spin can survive 10 microns, which is pretty neat,” says Stuart Parkin, director of the spintronics science and application center at IBM’s Almaden Research Center in San Jose, CA. “From an application point of view, it doesn’t really tell us how to make an interesting, useful device.”
One major issue is the tiny current output of the device, Parkin says. The researchers put three milliamperes in, and the output is in picoamperes, which is too small to be useful. Another problem is the special technique that the researchers use to make the device’s layered structure. This method is complicated and not at all compatible with current silicon fabrication, says Parkin.
But, says Crowell, this work marks the first time that spin has been measured in silicon, and that’s a great start toward silicon-based spintronics.
Geoffrey Hinton tells us why he’s now scared of the tech he helped build
“I have suddenly switched my views on whether these things are going to be more intelligent than us.”
ChatGPT is going to change education, not destroy it
The narrative around cheating students doesn’t tell the whole story. Meet the teachers who think generative AI could actually make learning better.
Meet the people who use Notion to plan their whole lives
The workplace tool’s appeal extends far beyond organizing work projects. Many users find it’s just as useful for managing their free time.
Learning to code isn’t enough
Historically, learn-to-code efforts have provided opportunities for the few, but new efforts are aiming to be inclusive.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.