PORTLAND, Ore. Future semiconductors encoding bits on the spin of electronscalled spintronicspromise to meld the advantages of optical-like polarization with silicon's ubiquitous infrastructure. The first step toward that goalelectronic injection and detection of spin-polarized electronshas been demonstrated for the first time by Ian Appelbaum's research group at the University of Delaware (Newark).
"We have demonstrated the world's first injection of spin-polarized electrons into silicon with completely electronic detection," said Appelbaum, an assistant professor in the Electrical and Computer Engineering department.
Appelbaum's prototype chip layers aluminum atop a spin polarization filter composed of a ferromagnetic film of cobalt-iron, through which electrons are injected into a silicon transport layer. Below that, a second ferromagnetic film of nickel-iron performs the electronic detection of the spin-polarized electrons.
"We use hot electron transport through a ferromagnetic thin film in order to do the spin filtering," said Appelbaum. "As unpolarized electrons pass through the thin film of cobalt iron, one orientation of spins is scattered away so that there are more of the orthogonal polarization that couple with the conduction states of the silicon transport layer."
Beneath the silicon, the second nickel-iron ferromagnetic layer has a variable spin orientation controlled by an external magnetic field, thereby enabling detection of the spin-polarized electrons by modulating them with the magnetic field.
"We modulate the injected electrons with external magnetic fieldthe same way that two polaroid filters can be rotated to modulate the light passing through them," explained Appelbaum.
To provide conclusive proof that the injected electrons were polarized, Appelbaum's group also performed spin precession measurements by changing the magnitude of the magnetic field, resulting in oscillations in the total number of electrons detected.
"So far our best prototype chips have demonstrated spin polarization into silicon of nearly 40 percent, which is greater than any other research group," said Appelbaum.
Subsequent demonstration chips have improved on the total amount of injected current by moving the ferromagnet away from the base of the tunnel junction at the injector side and putting it into the emitter, so that the spin polarization comes from the inherent polarization of the ferromagnet's equilibrium spin balance instead of the non-equilibrium spin balance previously created through hot electron transport.
Appelbaum's group now plans to work toward real spintronics circuitry by first increasing the polarization percentage of injected electrons, then by characterizing the effects of common semiconductor practices such as doping.
"Injection, transport and detection are the barest essentials for any semiconductor spintronics device," said Appelbaum. "Now we want to build upon our demonstration so we can do something useful in a real spintronics circuit."
In the pursuit of real spintronic semiconductor circuits, Appelbaum's group is currently studying the effects of doping, as well as searching for methods to manipulate the orientation of spins as they are injected. Their goal is complete electrical control of spin with 100 percent of the injected electrons at the desired orientation; however, practical consideration will probably involve a trade-off that achieves less than 100 percent polarization.
"In order to get higher percentages you have to make the ferromagnetic film thicker, but the thicker the film the less current you get through it," said Appelbaum. "How close we get to 100 percent remains to be seenso far we have improved injected polarization by five factors of two. Now if we can get one more factor of two, then we think we will be ready to start building real spintronics circuitry."