FEATURES ‘Theorists Could Add New Spin(tronics) to Transisters The hunt is on for successors to electronic transistors By Mike Ghenu, The Varsity (University of Toronto) el muna SY i ‘ TORONTO (CUP)—You never change a winning game, goes the old tennis adage—but always a losing one. The electronic transistor, a workhorse in the electronic industry for the past five decades, no longer seems at the top of its game. A transistor is much like the sluice gate that controls the flow of water through a channel. When a current is applied to one of its terminals, opening the gate, electronic signals can pass through. Arranged in sequence, transis- tors can be made to carry out logical operations. This makes them the ubiqui- tous building blocks of all electronic devices. But as the number of transistors that electronics manufacturers stuff onto a piece of silicon wafer soar, circuit designers have started to run up against fundamental physical constraints pre- venting them from cramming in yet more. One of the main culprits is heat, which requires additional cooling of sess microchips, pushing up their cost. Seeking to add more tricks to the transistor’s game, researchers are looking to spintronics, a new branch of electron- ics, for help. While in traditional elec- tronics information is transmitted by controlling the amount of charge pass- ing through a circuit, spintronics uses one of the electron’s more esoteric prop- erties—the quantum angular momen- tum, or spin-to code and process infor- mation. The concept of spin is difficult to fathom, since it is a vector quantity. Roughly speaking, you can envision an electron’s spin as a little bar magnet. The magnet’s arrangement determines whether the electron has spin “up” or spin “down’’—its direction determines the spin’s direction. Putting a new spin on things Unlike electric currents, which are flows of electrons, spin currents are flows of electrons with their spins ori- ented in a particular direction. If devices could be made that transmit information using such currents, the world of elec- tronics could be turned on its head. Not only would spintronic devices process information more quickly and efficiently, say scientists, but they would use a lot less energy and they could packed even more tightly than transistors currently are. A lot of that is still spin, though. Researchers are still struggling to gener- ate sufficiently strong spin currents, according to a paper published in January in the journal Science by Dr. Prashant Sharma, a researcher at Argonne National Laboratory in the US. But a paper recently published in Physical Review Letters, by Ali Najmaie, Dr. John Sipe, and Dr. Eugene Sherman, three University of Toronto quantum physicists, proposes a way to do just that. They have shown, theoretically, that it is possible to create spin currents in a gallium arsenide crystal—a common semi-conductor material—using laser pulses. These pulses, which deliver a lit- tle kick of energy to electrons in the crystal, cause an electron’s spin to flip to a specific orientation, a process called “Raman scattering.” “Light comes in at a particular colour and goes out the other side of the semi- conductor at a slightly different colour,” said Najmaie, the lead author. The slight change in colour accounts for the energy loss photons incur from delivering their punch of energy to electrons. The geometry of the semi-conductor crystal segregates electrons according to their spin direction. One upshot of Raman scattering is that a lot less energy gets deposited in the semi-conductor crystal. “It’s three degrees of magnitude or so lower than previous proposals,” Najmaie noted. Experimentation to application Sharma, the author of the Science paper, said the idea seems a viable one, but with one limitation. “The direction of spin polarization remains confined to a particular plane of the crystal. As a result, the full potential of a spin cur- rent—to use the vector nature of the spin to transmit information—is not uti- lized,’ he commented in an email. The University of Toronto researchers say they will now work with experimentalist colleagues to try to gen- erate spin currents in the lab. “We’re cre- ating, in some sense, puddles of spins,” said Sipe. “The idea is that device physi- cists will figure out how to take these puddles of spins and use them in an information processing scheme.” The ball is in their court.