Magnetic skyrmions

In magnetic systems, a wide variety of magnetic structures with distinct topologies exist and can exhibit quasiparticle-like behavior.^{Reference Braun19} Well-studied examples are magnetic vortices,^{Reference Cowburn, Koltsov, Adeyeye, Welland and Tricker13,Reference Shinjo, Okuno, Hassdorf, Shigeto and Ono14} stabilized by long-range dipolar interactions that exhibit collective motion governed by the dynamics of the individual spins forming the vortices.^{Reference Choe, Acremann, Scholl, Bauer, Doran, Stohr and Padmore20,Reference Novosad, Fradin, Roy, Buchanan, Guslienko and Bader21} Recently, magnetic skyrmions^{Reference Mühlbauer, Binz, Jonietz, Pfleiderer, Rosch, Neubauer, Georgii and Böni22} have garnered increasing interest. These are topologically distinct spin structures, in which the spins point in all three-dimensional directions, as is schematically shown in Figure 2a. Magnetic skyrmions were discovered^{Reference Mühlbauer, Binz, Jonietz, Pfleiderer, Rosch, Neubauer, Georgii and Böni22} in bulk materials with chiral Dzyaloshinskii–Moriya interactions (DMI), which is an antisymmetric exchange interaction given by the vector cross product of two neighboring spins. At temperatures close to the magnetic ordering temperature and moderately high magnetic fields, the competition between the DMI (favoring a helical magnetic structure) and direct interactions with externally applied magnetic fields (favoring a preferred magnetization direction) results in a compromise, where tubes of magnetization form that are aligned antiparallel to that of the surrounding matrix. Unlike magnetic vortices, which occur in geometrically confined structures, magnetic skyrmions can form in extended bulk materials. Often, they form a well-defined hexagonal lattice similar to vortex lattices in superconductors. Just like superconducting vortices, skyrmions in magnetic systems are mobile and can be moved with very small electric currents.^{Reference Schulz, Ritz, Bauer, Halder, Wagner, Franz, Pfleiderer, Everschor, Garst and Rosch23,Reference Everschor-Sitte and Sitte24}

Figure 2.

Magnetic skyrmions. (a) Schematic showing a skyrmion spin structure as expected from interfacial Dzyaloshinskii–Moriya interactions in magnetic multilayers. The color is used to emphasize the different out-of-plane magnetization components. (b) Magneto-optic Kerr effect (MOKE) imaging of the domain structure in a Ta/CoFeB/TaO_x trilayer patterned into a 60-μm-wide wire with a 3-μm-wide constriction. With a small magnetic field (${B_ \bot }$) of –0.5 mT applied perpendicular to the film plane, only isolated band (left) and bubble domains (right) are visible. (c) MOKE imaging after a current pulse of j _e= 5 × 10⁵ A/cm² has been applied. This results in the transformation of band domains on the left into magnetic skyrmions on the right. (a) Adapted with permission from Reference 24. © 2014 AIP Publishing LLC. (b–c) Adapted with permission from Reference 15. © 2015 AAAS.

The efficient manipulation of skyrmions with small electric currents has two origins. First, electrons that move in a skyrmion spin texture experience a virtual magnetic field (exceeding 100 T for skyrmions with diameters in the 10 nm range and below) that results in their transverse deflection and gives rise to the topological Hall effect.^{Reference Everschor-Sitte and Sitte24,Reference Neubauer, Pfleiderer, Binz, Rosch, Ritz, Niklowitz and Boni25} At the same time, a back-action results in motion of the skyrmions. The second aspect that facilitates skyrmion motion is decreased pinning due to their compact structure, which allows them to move around defects. Since skyrmions can be small (few nm) and easily manipulated electrically, they are potentially useful for information technologies.^{Reference Fert, Cros and Sampaio26}

To realize the ambitious goal of utilizing magnetic skyrmions for information technologies, efficient generation and annihilation of individual skyrmions on demand is needed. For instance, in thin FePd on Ir(111), skyrmions can be generated and annihilated via spin-polarized currents at cryogenic temperatures (8 K).^{Reference Romming, Hanneken, Menzel, Bickel, Wolter, von Bergmann, Kubetzka and Wiesendanger16} For practical applications, however, generation of skyrmions at room temperature and demonstration of high mobility is needed. This has recently been demonstrated in Ta/CoFeB/TaO_x multilayers, as shown in Figure 2b.^{Reference Jiang, Upadhyaya, Zhang, Yu, Jungfleisch, Fradin, Pearson, Tserkovnyak, Wang, Heinonen, te Velthuis and Hoffmann15} The perpendicular anisotropy in this material results in extended magnetic band domains, as seen on the left-hand side of Figure 2b. At the same time, a heavy element (e.g., Ta) layer, in combination with a thin ferromagnetic (FM) layer with perpendicular anisotropy, may stabilize domain walls with well-defined chirality due to interfacial DMI.^{Reference Emori, Bauer, Ahn, Martinez and Beach27,Reference Ryu, Thomas, Yang and Parkin28} These can be very efficiently moved due to the spin Hall effect.^{Reference Hoffmann29} In the device shown in Figure 2b, this electric current-driven domain-wall motion is used to push the band domain through the constriction. Upon exiting from the constriction, the band domain expands laterally due to laterally inhomogeneous spin transfer torque giving rise to an instability, generating individual skyrmion bubbles (see Figure 2c).

This mechanism is similar to Rayleigh–Plateau instabilities in surface-tension-dominated fluid flows, which result in droplet formation (i.e., for a dripping water faucet).^{Reference Eggers30} Interestingly, the critical electrical current for the motion of these synthetically produced skyrmions is two orders of magnitude smaller than for regular domain wall motion, indicating that indeed the peculiar structure of skyrmions results in significantly decreased pinning.^{Reference Jiang, Upadhyaya, Zhang, Yu, Jungfleisch, Fradin, Pearson, Tserkovnyak, Wang, Heinonen, te Velthuis and Hoffmann15}

Superconducting ratchets

Superconducting vortices interacting with artificial pinning arrays have attracted attention for over two decades^{Reference Baert, Metlushko, Jonckheere, Moshchalkov and Bruynseraede31,Reference Martin, Velez, Nogues and Schuller32} and yielded unique phenomena such as vortex ratchets. In a superconducting ratchet, a periodic alternating current is rectified, giving rise to a DC voltage. The topology that produces this effect consists of a superconducting (Nb, for instance) film on top of a rectangular array of triangular pinning sites, as shown in Figure 3.^{Reference Villegas, Savel’ev, Nori, Gonzalez, Anguita, García and Vicent33} Upon de-pinning under a symmetric AC driving current, vortices acquire a net velocity that depends on their number (i.e., on the perpendicular applied magnetic field). Up to three vortices per unit cell of the array are trapped in the triangles and acquire a net velocity in the same direction (red arrows) when de-pinned (see Figure 3a–b). For higher-vortex densities, the excess vortices become interstitial (blue), and “feel” the opposite asymmetry via their interaction with the (red) pinned vortices (see Figure 3c–d). Under an AC driving force, interstitials de-pin first and acquire a net velocity opposite (blue arrow) to the pinned ones (red). The reversed rectification holds until a sufficiently intense drive de-pins vortices trapped by triangles, producing positive rectification. Thus, the system behaves as a very unique electrical rectifier, in which an injected AC current leads to a DC voltage whose sign depends on the applied magnetic field and the magnitude of current.

Figure 3.

Superconducting vortex ratchets. (a–d) Sketches of the vortex configuration and net direction of motion (depicted with red and blue arrows) in Nb deposited over a magnetic nano-triangle pinning array, for different vortex densities. Vortices pinned on the triangles are shown in red and interstitial vortices are shown in blue spheres. The distance between bottom sides of the triangles (white line) is 746 nm. (e–h) Net vortex velocity (〈v〉) as a function of the AC drive amplitude (F _L) for the situations depicted in (a–d). n indicates the number of vortices per unit cell of the array. A positive net velocity is observed upon de-pinning of the (red) vortices at sufficiently high drives. In the presence of interstitial vortices (blue), for n > 3, their motion yields a negative rectification at low drives. The relative weight of the negative/positive rectification depends on the ratio of interstitial/pinned vortices. Adapted with permission from Reference 33. © 2003 AAAS.

Over the past decade, vortex ratchets have been profusely studied (e.g., seeking new types of asymmetric pinning [nonuniform arrays, constrictions, surface barriers], high-frequency effects).^{Reference Van de Vondel, Gladilin, Silhanek, Gillijns, Tempere, Devreese and Moshchalkov34,Reference Wordenweber, Hollmann, Schubert, Kutzner and Panaitov35} Recent developments in the area, including reconfigurable energy landscapes,^{Reference Crassous, Bernard, Fusil, Bouzehouane, Le Bourdais, Enouz-Vedrenne, Briatico, Bibes, Barthelemy and Villegas36,Reference Trastoy, Malnou, Ulysse, Bernard, Bergeal, Faini, Lesueur, Briatico and Villegas37} vortex manipulation through magnetic force microscopy,^{Reference Auslaender, Luan, Straver, Hoffman, Koshnick, Zeldov, Bonn, Liang, Hardy and Moler38} and single-vortex imaging,^{Reference Auslaender, Luan, Straver, Hoffman, Koshnick, Zeldov, Bonn, Liang, Hardy and Moler38,Reference Embon, Anahory, Suhov, Halbertal, Cuppens, Yakovenko, Uri, Myasoedov, Rappaport, Huber, Gurevich and Zeldov39} have provided key ingredients and paved the way toward novel vortex ratchet applications. At the same time, vortices on artificial pinning arrays constitute a much broader research area whose recent developments and exciting prospects cannot be detailed here. Theory^{Reference Ray, Reichhardt, Jankó and Reichhardt40} and experiments^{Reference Guénon, Rosen, Basaran and Schuller41} have shown that nonperiodic arrays provide sizable superconducting critical current enhancements over a larger field range than those obtained with periodic arrays. Thus, ordered, nonperiodic arrays are probably the route toward new developments in the field.

FM heterostructures

FM heterostructures are classic examples in which mesoscopic phenomena appear due to competition between various magnetic length scales. The typical magnetic domain wall width is larger than several tens of nm in many FM materials. It is very unusual and unexpected to find a small (<10 nm) domain wall, especially in a typical soft ferromagnet, such as Permalloy or yttrium iron garnet. A classic interesting example from the fundamental and applied points of view, where very short domain walls can be engineered, is the one formed by combinations of magnetically soft/hard bilayers.^{Reference Fullerton, Jiang, Grimsditch, Sowers and Bader42–Reference Guo, Zhang and Wang45} In particular, an FM in intimate contact with an AFM^{Reference Morales, Li, Petracic, Batlle, Schuller, Olamit and Liu46} or a soft magnetic material in contact with a hard one exhibits these interesting effects. Due to the interfacial interactions in FM/AFM (“exchange biased”) or soft/hard (“FM spring”) heterostructures, a short mesoscopic domain wall (<10 nm) can form, as shown in Figure 4a. As a consequence, the FM reversal becomes asymmetric and reversible as shown in Figure 4b.^{Reference Morales, Basaran, Villegas, Navas, Soriano, Mora, Redondo, Batlle and Schuller47} In these systems, the magnetization reversal initiates with the nucleation of a quasi-Bloch wall, where the magnetization rotates in the soft FM in a plane parallel to the interface. This results in a spiral spin configuration that motivates the expression “exchange spring.” As the reversed field increases, the quasi-Bloch wall in the soft FM is compressed against the hard one, and the interfacial hard spins also increasingly rotate. The soft spins rotate back into alignment with the hard phase if the reverse field is removed.

Figure 4.

Exchange spring. (a) Depicted depth profile of the magnetization in the soft ferromagnetic layer during reversal. Colored arrows highlight the different rotation amounts of the moments from the pinning direction. (b) Example hysteresis curves of Py/FeF₂ bilayer exchange bias system that exhibits asymmetric and reversible behavior due to exchange spring formation. Note: M, magnetization; M _S, saturation magnetization; μ₀, magnetic permeability of free space; H, external magnetic field; t _Py, thickness of permalloy layer. Adapted with permission from Reference 47. © 2015 American Physical Society.

Nucleation and evolution of the quasi-Bloch walls during exchange-spring reversal have been characterized using a magneto-optical indicator,^{Reference Gornakov, Nikitenko, Shapiro, Shull, Jiang and Bader48} polarized neutron reflectivity,^{Reference O’Donovan, Borchers, Majkrzak, Hellwig and Fullerton49} coherent nuclear resonant scattering,^{Reference Rohlsberger, Thomas, Schlage, Burkel, Leupold and Ruffer50} x-ray resonant magnetic scattering,^{Reference Choi, Jiang, Pearson, Bader and Liu51} layer-resolved conversion-electron Mössbauer spectroscopy,^{Reference Uzdin, Vega, Khrenov, Keune, Kuncser, Jiang and Bader52} and magnetotransport.^{Reference Basaran, Morales, Guénon and Schuller53}

It is important to note that the formation of mesoscopic quasi-Bloch walls are mainly governed by the interfacial coupling between soft/hard or FM/AFM, which depends on the microstructural details of the interface and sample morphology. This can produce many rich phenomena^{Reference Nogués and Schuller54} related to roughness,^{Reference Kuch, Chelaru, Offi, Wang, Kotsugi and Kirschner55,Reference Nogués, Lederman, Moran, Schuller and Rao56} grain size,^{Reference Saerbeck, Zhu, Lott, Lee, LeClair, Mankey, Stampfl and Klose57–Reference Pakala, Huai, Anderson and Miloslavsky59} crystallinity,^{Reference Nogués, Moran, Lederman, Schuller and Rao60,Reference Jungblut, Coehoorn, Johnson, aan de Stegge and Reinders61} interlayer diffusion,^{Reference Kim, Lee, Song, Park, Yang and Min62} and defects.^{Reference Mougin, Mewes, Jung, Engel, Ehresmann, Schmoranzer, Fassbender and Hillebrands63–Reference Basaran, Saerbeck, de la Venta, Huckfeldt, Ehresmann and Schuller65} Furthermore, these properties depend on temperature and magnetic field history of the sample.^{Reference Nogués, Lederman, Moran and Schuller66,Reference Li, Eisenmenger, Miller and Schuller67} Therefore, general magnetic properties on these heterostructures are attributable to several extrinsic and intrinsic parameters, which become very important at the mesoscale.^{Reference Dennis, Borges, Buda, Ebels, Gregg, Hehn, Jouguelet, Ounadjela, Petej, Prejbeanu and Thornton68–Reference Nogués, Sort, Langlais, Skumryev, Suriñach, Muñoz and Baró70}

Triplet superconductivity

The superconducting (S) proximity effect—penetration of the superconducting condensate into a nonsuperconducting material in contact with a superconductor—is long-ranged (up to ∼microns) for clean metals. However, it is usually short-ranged for FMs, because the exchange field tends to align the electron spins breaking apart conventional opposite-spin singlet Cooper pairs.^{Reference Buzdin71} This occurs within a length scale ξ_F ∼nm, with ξ_F$&#x003D; \sqrt {\hbar D/2{E_{{\rm{ex}}}}}$ for diffusive systems (D is the electronic diffusion constant, E _ex is the exchange splitting, and $\hbar$ is the reduced Planck constant) and ξ_F = v _F/2E _ex for ballistic systems (v _F is the Fermi velocity).

Long-range proximity effects have been experimentally observed at specific S/FM interfaces^{Reference Visani, Sefrioui, Tornos, Leon, Briatico, Bibes, Barthelemy, Santamaria and Villegas12,Reference Keizer, Goennenwein, Klapwijk, Miao, Xiao and Gupta72–Reference Kalcheim, Kirzhner, Koren and Millo76} and explained by the generation of unconventional equal-spin triplet pairs. These are immune to the exchange field and can propagate over long distances into the FM.^{Reference Bergeret, Volkov and Efetov77–Reference Eschrig79} While the salient features of conventional S/FM proximity effects have received attention for two decades,^{Reference Buzdin71} equal-spin triplets currently attract the spotlight. Although triplets were earlier evoked to explain the coexistence of ferromagnetism and superconductivity in certain bulk materials,^{Reference Saxena, Agarwal, Ahilan, Grosche, Haselwimmer, Steiner, Pugh, Walker, Julian, Monthoux, Lonzarich, Huxley, Sheikin, Braithwaite and Flouquet80} artificial S/FM hybrids have only recently opened the door to deeper understanding and manipulation of this unconventional superconducting state.

Open research topics in this area include an understanding of specific systems (e.g., cuprate/manganite interfaces)^{Reference Visani, Sefrioui, Tornos, Leon, Briatico, Bibes, Barthelemy, Santamaria and Villegas12,Reference Kalcheim, Kirzhner, Koren and Millo76,Reference Egilmez, Robinson, MacManus-Driscoll, Chen, Wang and Blamire81} and the search for new S/FM combinations and novel mechanisms. Theory has recently shown that spin-orbit coupling^{Reference Bergeret and Tokatly82} can promote equal-spin triplet pairing, a possibility that opens an approach to manipulate the triplet condensate. The dynamic coupling between the equal-spin triplet condensate and the FM magnetization is another appealing perspective. Theory shows that a time-varying magnetization influences the triplet condensate generation due to the nonequilibrium spin accumulation induced at the interface.^{Reference Houzet83,Reference Yokoyama and Tserkovnyak84} Conversely, an equal-spin triplet supercurrent is predicted to affect the magnetization via spin-torque.^{Reference Linder, Brataas, Shomali and Zareyan85,Reference Pugach and Buzdin86} However, no experimental studies exist in which the triplet condensate is manipulated with stimuli such as microwaves. These fundamental prospects, in addition to the potential of triplet superconductivity for “superconducting spintronics,” make this emergent topic an exciting area of research.

Magnon BEC

Interestingly, in thin films, the competition between short-range exchange and long-range dipolar interactions can give rise to nonmonotonic magnon dispersion with minima at finite wave vectors, schematically shown in Figure 5a.^{Reference Nowik-Boltyk, Dzyapko, Demidov, Berloff and Demokritov87} These minima can be utilized to explore collective magnon states, which can form a BEC even at room temperature, as initially demonstrated a decade ago^{Reference Demokritov, Demidov, Dzyapko, Melkov, Serga, Hillebrands and Slavin88} based on earlier theoretical predictions.^{Reference Kalafati and Safonov89} In this seminal experiment, a very dense gas of magnons was excited via parametric pumping to an energy well above the magnon dispersion minimum. During parametric pumping, a microwave photon generates two magnons with opposite wave vectors, as shown schematically in Figure 5a. Subsequently, four magnon-scattering processes redistribute the energy, while maintaining the magnon density. Once the magnon density exceeds a critical value, the BEC forms at the dispersion curve minimum, as experimentally measured with spatially resolved Brillouin light-scattering.^{Reference Nowik-Boltyk, Dzyapko, Demidov, Berloff and Demokritov87,Reference Demokritov, Demidov, Dzyapko, Melkov, Serga, Hillebrands and Slavin88,Reference Hillebrands and Zhu90,Reference Serga, Tiberkevich, Sandweg, Vasyuchka, Bozhko, Chumak, Neumann, Obry, Melkov, Slavin and Hillebrands91} The coherence between the two components of the BEC with opposite wave vectors produces a standing wave pattern directly revealing the magnon wavelength at the dispersion minima.

Figure 5.

Bose–Einstein condensed (BEC) magnons. (a) Schematic illustration of the magnon dispersion in a magnetic thin film. The competition between short-range exchange and long-range dipolar interactions gives rise to two global minima in the dispersion at opposite wave vectors. Magnons excited at higher energy can subsequently create two coherent BECs at the wave vectors –k _BEC and k _BEC, where the magnon dispersion has two minima. (b) Measurement of the spatial distribution of BEC magnons in a Y₃Fe₅O₁₂ thin film. The two magnon condensates can interfere, which gives rise to a standing wave pattern. This can be directly probed with spatially resolved Brillouin light-scattering measurements along the two lateral directions (y and z), where the scattered intensity (indicated by the color code on the right side of the graph) is proportional to the condensate density. Dashed areas and circles indicate the position of topological defects in the condensate. Adapted with permission from Reference 87. © 2012 Macmillan Publishers Ltd.

The generation of room-temperature magnon BECs opens up exciting opportunities for the exploration of macroscopic quantum coherent phenomena at temperatures far above those studied in any superconductors or liquid He. For example, Josephson effects, between two superconducting condensates coupled via a weak link, have been key to high-precision measurements, as well as the practical implementations of quantum computation. Interestingly, although Josephson effects have been studied in superconducting systems for about half a century, only recently has this concept been extended to other BEC systems such as atomic gases.^{Reference Levy, Lahoud, Shomroni and Steinhauer92} A recent theoretical prediction claims that similar Josephson effects should exist for magnon BECs.^{Reference Nakata, van Hoogdalem, Simon and Loss93} One main challenge is to generate magnon BECs with spatially modulated phase relationships. Toward this end, establishing magnon BEC via local charge^{Reference Bender, Duine and Tserkovnyak94} or heat currents^{Reference Bender, Duine, Brataas and Tserkovnyak95} may offer a possible solution. In fact, it was suggested very recently that temperature gradients within magnon BECs may be accompanied by magnon supercurrents.^{Reference Clausen, Bozhko, Vasyuchka, Melkov, Hillebrands and Serga96} Other approaches of spatially tailoring magnon properties include local field modulations.^{Reference Vogt, Schultheiss, Jain, Pearson, Hoffmann, Bader and Hillebrands97,Reference Vogt, Fradin, Pearson, Sebastian, Bader, Hillebrands, Hoffmann and Schultheiss98}

Superconducting spintronics

Conventional spintronics combines information storage in FMs (memory) and information transport by spin-polarized electrons. The archetypal device (spin valve) consists of two FM electrodes connected by a nonmagnetic channel. The relative orientation of the electrodes’ magnetization (parallel/antiparallel) yields electrical-resistance states (low/high, respectively), due to spin-dependent electron scattering. The superconducting analogues of the spin valve generally consist of S/FM hybrids that show resistive switching driven by the magnetic history,^{Reference Gu, You, Jiang, Pearson, Bazaliy and Bader99–Reference Banerjee, Smiet, Smits, Ozaeta, Bergeret, Blamire and Robinson107} using the dependence of the critical temperature (T _C) on the FM magnetization. However, many superconducting “spin valves” do not truly exploit spin-dependent transport, rather, they use a variety of different mechanisms, from FM stray magnetic field^{Reference Villegas, Li and Schuller102,Reference Visani, Metaxas, Collaudin, Calvet, Bernard, Briatico, Deranlot, Bouzehouane and Villegas104,Reference Stamopoulos, Manios and Pissas108,Reference Zhu, Cheng, Boone and Krivorotov109} or exchange field^{Reference Gu, You, Jiang, Pearson, Bazaliy and Bader99,Reference Li, Roschewsky, Assaf, Eich, Epstein-Martin, Heiman, Munzenberg and Moodera105} effects on the superconductor, to the proximity effect—either the conventional (short-range singlet)^{Reference Moraru, Pratt and Birge101} or triplet^{Reference Banerjee, Smiet, Smits, Ozaeta, Bergeret, Blamire and Robinson107,Reference Jara, Safranski, Krivorotov, Wu, Malmi-Kakkada, Valls and Halterman110,Reference Flokstra, Cunningham, Kim, Satchell, Burnell, Curran, Bending, Kinane, Cooper, Langridge, Isidori, Pugach, Eschrig and Lee111} (see example in Figure 6). Only in some cases^{Reference Pena, Sefrioui, Arias, Leon, Santamaria, Martinez, te Velthuis and Hoffmann100,Reference Miao, Yoon, Santos and Moodera103} does resistive switching arise from spin accumulation in the superconductor due to quasiparticle injection from the ferromagnet. Regardless of the governing mechanism, resistance switching is usually much larger than in conventional spin valves. However, as T _C variations are minute, the “spin-valve” effect is observed only in a very narrow temperature range near the intrinsic T _C (see Figure 6c).

Figure 6.

Superconducting spintronics. (a) Illustration of a triplet superconducting spin valve. The superconductor (Nb) is sandwiched between two ferromagnetic (FM) (Ho/Py bilayers). Ho is a conical magnet that behaves as a spin-mixer to generate an equal-spin triplet state. (b) The hysteresis loop for a spin valve with a 5-nm-thick layer of Ho on either side of Nb shows that the exchange bias exerted by the FeMn layer on the lower Py layer allows for parallel/antiparallel magnetization states of the Py layers. (c) The spin-valve magnetoresistance (R) shows hysteresis, with a resistive switching within the field range in which the Py layers magnetization is labeled antiparallel. In the case of antiparallel magnetization, some of the triplet pairs are not able to flow from one FM electrode into the opposite one, which results in higher resistance. The effect is observed only in a narrow (few mK) temperature range across the superconducting transition (T _c ∼6.90 K). The curves correspond to two measurements at different temperatures within the range of 6.98 to 6.90 K. Note: Py, Permalloy; M, magnetization; M _S, saturation magnetization; μ₀, magnetic permeability of free space; H, applied magnetic field. Adapted with permission from Reference 107. © 2014 Macmillan Publishers Ltd.

A conceptual leap uses superconductivity to improve “information transport” above the conventional “spin valves” (which work as a memory). For this, superconductors (instead of metals or semiconductors) are used as the spin-transport channel. Interestingly, experiments have repeatedly shown that the spin-diffusion lengths in superconducting metals are significantly longer below T _C than above T _C,^{Reference Poli, Morten, Urech, Brataas, Haviland and Korenivski112–Reference Wakamura, Hasegawa, Ohnishi, Niimi and Otani114} due to different spin-relaxation mechanisms. Equal-spin triplet superconductivity can be used to induce “spin-polarized superconductivity” in the FM, as opposed to creating a spin accumulation in the superconductor. Therefore, information would be carried by spin-polarized superconducting pairs in the FM, which leads to important implications.

In contrast to the nonequilibrium spin accumulation produced by spin injection, equal-spin triplet superconductivity is an equilibrium state induced by proximity in the FM without any external stimuli (such as electrical current, microwaves, or light). This should allow for novel, simpler device architectures. Also, certain quantum coherence effects (e.g., nonlocality and Josephson effects) provide possibilities^{Reference Eschrig115,Reference Linder and Robinson116} that are absent in conventional spintronics. Recent realizations seeking that potential include magnetic-history switchable Josephson junctions.^{Reference Banerjee, Robinson and Blamire117} Of course, there are quite a few challenges ahead. For instance, devising methods that efficiently discriminate and “read” the polarization of equal-spin triplet states in FMs, when both spin-up and spin-down triplet pairs coexist. A possible way is the use of half-FM metals, in which only one equal-spin polarization is possible. Another standing challenge that could attract future attention is the transfer of the equal-spin triplets from the FM into another material (e.g., a nonmagnetic metal) where equal-spin triplets could stay phase coherent over long distances.