In June 1929 a paper by the 23-year-old Nevill Mott appeared in the Proceedings of the Royal Society of London. As Mott noted in his introduction, theoretical arguments and empirical evidence supported the notion that electrons have an intrinsic magnetic moment, or spin. “The question arises,” he wrote, “has the free electron ‘really’ got a magnetic moment, a magnetic moment that we can by any conceivable experiment observe?”
Mott’s question is subtler than it might first appear. If you turn to the paper’s appendix, you’ll find what Niels Bohr told Mott: The uncertainty principle forestalls any attempt to distinguish an electron’s intrinsic magnetic moment from the magnetic field that arises from its motion. But, as Mott (shown here) demonstrates in his paper, it is possible to quantify the intrinsic magnetic moment because it turns out the probability that an electron scatters off an atomic nucleus in a given direction depends on the orientation of the electron’s spin.
Forty-two years after Mott’s paper was published, Mikhail Dyakonov and Vladimir Perel of the Ioffe Institute in Leningrad found a similar effect in semiconductors. According to their theoretical analysis, an electric field applied along a strip of semiconductor drives electrons to scatter off impurities in a spin-dependent way: Those with up spins veer to one side, while those with down spins veer to the other side.
Dyakonov and Perel’s paper did not attract much attention. Indeed, in 1999 Jorge Hirsch reproduced the analysis without either himself or—one assumes—his paper’s reviewers being aware of its Russian antecedent. He called the phenomenon the spin Hall effect.
The effect that Dyakonov, Perel, and Hirsch predicted depends on the presence of extrinsic impurities. But in 2003 two groups of theorists—Shuichi Murakami, Naoto Nagaosa, and Shou-Cheng Zhang; Jairo Sinova, Allan MacDonald, and their collaborators—independently proposed that a spin Hall effect could arise intrinsically when spin–orbit coupling of electrons to the lattice atoms acts with the applied electric field to change the semiconductor’s band structure.
Controlling electron spins through electric fields is technologically enticing. Murakami, Nagaosa, and Zhang wrote in the 2003 Science paper:
Principles found here could enable quantum spintronic devices with integrated information processing and storage units, operating with low power consumption and performing reversible quantum computation.
What’s more, because electric fields ultimately arise from static charges, they constitute a finer, faster, and more convenient means to control spins than do magnetic fields, which ultimately arise from moving charges.
Weak then strong
I did not become acquainted with the spin Hall effect’s history through deep, broad study of 20th-century physics. Rather, in 2005 I wrote a news story about the effect’s experimental verification. David Awschalom, his graduate students Yuichiro Kato and Roberto Myers, and Art Gossard detected the effect through a spin-dependence in the polarization of reflected light. Independently, Jörg Wunderlich, Bernd Kästner, Sinova, and Tomas Jungwirth looked instead for the circularly polarized light emitted by an LED when spin-polarized electrons and holes recombine.
Both experiments were tours de force of ingenuity and execution. They needed to be. The spin Hall effect in gallium arsenide, the material the two groups used, is weak—too weak, perhaps, form the basis of a industrially significant technology. Nevertheless, the spin Hall effect has been one of the past decade’s most fruitful areas of research. There’s now a quantum spin Hall effect, a spin Hall effect for light, and an inverse spin Hall effect.
Now comes a paper that reports a spin Hall effect of technologically interesting magnitude. In the 4 May issue of Science, Cornell University’s Robert Buhrman and his collaborators report the discovery of what they call a massive spin Hall effect in the brittle, semimetal β phase of tantalum. The effect, which works at room temperature, is strong enough to flip the spins in an adjacent ferromagnet.
That spin-flipping ability could form the basis of nonvolatile computer memory—that is, memory that isn’t wiped clean whenever you turn off the power. In fact, to demonstrate β-tantalum’s promise, Buhrman’s team built nanodevices whose active components consisted of few-nanometer-thick layers of β-tantalum, cobalt iron boron (a ferromagnet), and magnesium dioxide (to improve the ferromagnet’s performance).
The prototype device did indeed work, at room temperature, as a magnetization switch. The switching current Ic is higher than that of magnetic tunnel junctions (MTJs), which use spin-polarized currents to perform the switching. However, Buhrman anticipates that with routine optimization Ic could be reduced to the point that spin Hall devices would compete with MTJs.
Mott’s 1929 paper, his first, is characteristic of his early interest in atomic and nuclear physics. In 1933 he took up a professorship at the University of Bristol, and from then on he devoted himself to the field in which he received a Nobel prize: condensed matter physics. Mott died in 1996 at the age of 90—too soon, unfortunately, to learn what his theoretical paper in atomic physics had begotten 83 years later.