Historical research guide · 0337 · 2010
DIRD Metallic Spintronics
DIA / AAWSAP contractor
In one minute
The Defense Intelligence Agency commissioned this review for its Advanced Aerospace Weapon System Applications programme and issued it on 23 March 2010. Its subject is spintronics in metals: using an electron’s spin, rather than only its charge, to carry and process information. The author builds the field out of two effects that are each other’s mirror image. Giant magnetoresistance, which won the 2007 Nobel Prize in Physics, is the large drop in a magnetic sandwich’s resistance when a field lines its magnetic layers up — the effect that made modern hard-drive read heads possible. Spin-transfer torque is the reverse: force a dense enough current through a tiny constriction and the electrons hand their angular momentum to the magnet, flipping it or setting it precessing at tens of gigahertz. The report explains the mechanism, the experiments and the devices that follow — racetrack memory, spin-torque magnetic RAM, a nanoscale current-tuned microwave oscillator — and argues that what aerospace needs from all this is radiation-hard, nonvolatile electronics that barely warm up.
Connection
Chapter 1 reads the AAWSAP reference set as one government stocktake of what an advanced craft would need, and this is the volume on the electronics that would have to survive the trip. Chapter 13 keeps arriving at the same conclusion this report reaches independently: the limit is not the idea but the material, and here it is spin — a quantum property of the electron — being turned into an engineering control knob.
From the source
Giant magnetoresistance is the large drop in a magnetic multilayer’s resistance when an applied field aligns the magnetisations of its ferromagnetic layers, and the accepted explanation is Mott’s two-current model: below the Curie temperature spin-up and spin-down electrons carry current almost independently and in parallel, and each is scattered far more strongly by a layer magnetised against it. Research on the effect was awarded the 2007 Nobel Prize in Physics.Section 1, p. i; Section 2.1, pp. 1-3; Figure 2
GMR was the first spintronic effect to reach mass production. A spin valve — two ferromagnetic layers separated by a thin nonmagnetic spacer, one of them pinned by exchange bias to an adjacent antiferromagnet — reads the stray field of a magnetic domain on a hard disk, and its responsiveness is what enabled very high areal packing densities. Sputtered cobalt-copper multilayers gave room-temperature magnetoresistance 13 times greater than the permalloy sensors then in use.Section 2.1, p. 2; Section 2.2, p. 3
Spin-transfer torque is the inverse effect and rests on conservation of angular momentum: an electron crossing into a magnet has its spin reoriented, and by Newton’s third law the magnet takes an equal and opposite torque. One electron does nothing, so the effect needs current densities above 10^12 amperes per square metre, forced through a constriction — a mechanical point contact under 10 nanometres, a lithographic nanopillar, a nanowire — and it was first seen by Tsoi and colleagues in a cobalt-copper multilayer in 1998, two years after Slonczewski and Berger predicted it.Section 3.1, pp. 4-7; Figures 3 and 5
Spin-transfer torque turns a direct current into an alternating one: the driven precession of the free layer modulates the device’s resistance, so a dc input becomes an ac output whose frequency is tuned by field and current from a few gigahertz to more than 100 gigahertz. Linewidths as narrow as 2 megahertz have been demonstrated, giving quality factors over 18,000 in a nanoscale oscillator.Section 3.3, pp. 10-11
The same torque moves magnetic domain walls along a wire, which is the basis of racetrack memory: a ferromagnetic nanowire carries data as a pattern of domains, current pulses shift the whole pattern past fixed read and write elements, and U-shaped wires standing up out of the chip plane would give storage density above solid-state flash with far better read and write performance than a hard drive. Luc Berger predicted the effect in 1978.Section 3.4, pp. 12; Figure 11
Antiferromagnetic spintronics is the report’s open question. The predicted antiferromagnetic GMR was not found in sputtered antiferromagnet-spacer-antiferromagnet multilayers, probably because transport in such imperfect films is diffusive rather than ballistic. But in four separate experiments a high current shifted the exchange-bias switching field in opposite directions for opposite current polarity — which rules out Joule heating, since heating would shift it the same way both times — and that asymmetry is taken as evidence of spin torque acting on an antiferromagnet.Sections 4.1 and 4.2, pp. 13-16; Figure 12
The way in
https://documents2.theblackvault.com/documents/dia/AAWSAP-DIRDs/DIRD_10_Metallic_Spintronics.pdfHow to cite it
DIA / AAWSAP contractor (2010) DIRD Metallic Spintronics https://documents2.theblackvault.com/documents/dia/AAWSAP-DIRDs/DIRD_10_Metallic_Spintronics.pdf