Perspectives on Spintronics: Professor Sergio O. Valenzuela

During the last two decades, there has been a renewed interest in the research of spin physics by electrical means in the solid state community, yielding a variety of spectacular phenomena. The interest is motivated by the quest to understand basic physical principles underlying the electron and nuclear spin interactions and by possible technological applications. In conventional electronics, information can be represented, manipulated and transported in the form of the electron charge but the spins are ignored. In spin-based electronics, or spintronics, the goal is the active manipulation of spin degrees of freedom for practical use. Early developments in the field are reviewed in [1].

The term spintronics emerged in the 1990´s following experiments in the 1980´s that led to fundamental breakthroughs and applications. These experiments had the common feature of using ferromagnets to inject and detect spins, which remarkably led to the discovery of the giant magnetoresistance (GMR) effect that quickly resulted in the miniaturization of the recording heads of hard-disk drives, and earned their discoverers, Fert and Grünberg, the 2007 Nobel Prize in Physics [2,3].

The advent of magnetic tunnel junctions (MTJ) [4] and proposals of semiconducting spintronic devices [5] further invigorated the research in the field and led to new applications, now in the random access memory (RAM) market. In 2006, the first magnetic RAM (MRAM) chip was commercialized and is nowadays used in military and aviation applications. Its intrinsic speed, low power consumption, and durability, make MRAM a leading candidate as a “universal memory” promising to replace flash, SRAM and DRAM. The main drawback of MRAM is its low integration density. The memory element is an MTJ, which consists of two ferromagnetic electrodes separated by a thin insulator. One of the electrodes has a fixed magnetization and the other can be controlled by an external magnetic field. The readout can be done by measuring the resistance of the MTJ. However, as the device is scaled down in size, the presence of magnetic fields from neighboring elements can lead to false writes. This problem could be solved by new developments in spin torque technology that would eliminate the need of external magnetic fields [6]. Here, the switching results from angular momentum transfer from a spin polarized current generated by the fixed ferromagnetic electrode that exerts a torque on the magnetization of the second “free” electrode.

Spintronics is thus a highly dynamic research field where current applications continuously feed from fundamental research, where spin torque effects being the latest discovery that may soon reach commercialization. In this article, we feature a few emerging topics that are related to current research at the Catalan Institute of Nanotechnology (ICN) in Barcelona. The topics range from spin Hall and Rashba spin-orbit based torques, to energy harvesting involving spins, to topological insulators. We briefly mention the new concepts, their advantages and possible technological impact.

Spin Hall and Rashba Spin Orbit Torques

As mentioned above, spin torque technologies are gaining relevance on spintronic applications. Along these lines, recent advances on spin Hall and Rashba spin-orbit based torques can further improve the performance of spin torque based RAM.

Analogous to the conventional Hall effect, where a charge current in the presence of a magnetic field results in charge accumulation at the edge of the sample, the spin Hall effect refers to the generation of spin accumulation as a result of a charge current in the presence of spin-orbit interaction. Such process is represented schematically in Fig. 1a. There, conduction electrons are scattered by imperfections in the crystal (impurities or defects), where spin orbit interaction causes a spin asymmetric effect. The spin Hall effect and its reciprocal effect (Fig. 1b), where a spin current generates charge accumulation, are closely related to the anomalous Hall effect but appear in a nonmagnetic conductor.

Fig. 1. a). Spin Hall effect.  Spin accumulation is induced at the edges of the sample due to spin-orbit interaction when a pure charge current j is applied. The transverse voltage is zero as no charge imbalance is induced.
b) Spin-current induced Hall effect or reciprocal spin Hall effect. A pure spin current js is injected. Due to spin-orbit interaction a transverse charge current, and an associated voltage, is induced.

The spin Hall effect was predicted by Dyakonov and Perel in 1971 [7] but it was not directly observed until 2004 [8]. The first experiments used optical methods for detection; fully electrical measurements of these effects had to wait another two years for their realization [9]. The first electrical measurements focused on the reciprocal effect. The spin current was injected using a ferromagnetic source in a non-local geometry [9]. Because spin-up and spin down currents circulating in opposite directions are deflected to the same edge, charge accumulation is built up and a voltage can be measured (Fig. 1b). Fig. 2 shows a scanning electron microscope image of a device used for these measurements. Using a similar structure, the effect was later on studied in heavy metals, such as Au or Pt, in order to maximize the effect of spin-orbit interaction [10]. Such studies also included the effect of (heavy) impurities in a nonmagnetic metallic matrix [11,12].

The spin Hall effect could represent a straightforward method to generate spin currents without the need of ferromagnets as part of a spin Hall transistor [13]. Such spin currents can also be used to exert a torque in a ferromagnet [14] to flip the free layer in an MRAM cell. A related effect involves the spin-orbit Rashba effect, where the accumulation of spins in one direction is due to intrinsic properties of the structure and not to impurities. Current-driven spin torque induced by the Rashba effect was recently reported in Pt/Co/AlO_x structures lacking inversion symmetry [15]. Large effective magnetic fields of the order of 1 T per 10⁸ A cm⁻² underscore the relevance of these results.

Fig. 2. Scanning electron microscope  image of a sample showing the Hall cross  (right end) and spin injector in Fig. 1 b (wide light strip). The ferromagnet on the left (narrow light strip) is used as a reference to study spin injection and spin transport properties (after Ref. [9]).

Energy Harvesting and Spin Caloritronics

Energy harvesting is the process by which energy is captured from external (free) sources (thermal, wind, solar energy, or ambient electromagnetic noise) and used in autonomous devices. In recent years, these ideas are gathering a great deal of attention in the spintronics community with the aim at controlling spin currents by means of heat currents or noise and vice versa. Indeed, when applied to heat transport, it has led to a new sub-field of spintronics, coined spin caloritronics [16]. Amongst recent discoveries in the field we find the spin-Seebeck effect in metals [17], insulators [18] and semiconductors [19], the demonstration of spin injection driven by heat flow [20], and the realization of spin ratchets [21].

The spin-Seebeck effect refers to the generation of spin currents in a ferromagnet subject to a temperature gradient. It was first reported in permalloy (NiFe) using Pt strips as spin detectors that relied on the reciprocal spin Hall effect [17]. Recent experiments using insulators reporting similar results [18] have shown that the spin Seebeck effect might be radically different from the charge Seebeck effect, which is not present in insulators. It is still not clear the origin of the spin currents, although recent proposals point to out-of-equilibrium magnon dynamics. A temperature difference between magnons and electrons in the ferromagnet would drive a spin current into Pt as a consequence of spin pumping [18].

Although the detected signals are so far small and hardly useful for applications, the observations are relevant because of the possibility of electric voltage generation from heat-flow in an insulator. Thermoelectric generation usually relies on the Seebeck effect, which is the generation of electric voltage as a result of a temperature gradient. This is a feature that is efficient in electric conductors, which limits its application. Charge carriers that participate in the Seebeck effect also carry heat, making it difficult to build a thermal gradient and incrementing the complexity of device design. This problem would be readily circumvented by the spin Seebeck insulators [18].

Another means to generate pure spin currents involves a new concept: the so-called spin ratchet. A pure spin ratchet is by definition a system able to produce pure spin currents without the transport of charge. It generates useful work from a signal or perturbation (for example, an ac voltage or environmental noise of electric or magnetic origin, etc) when combined with intrinsic asymmetry, often realized by a ratchet potential. The ratchet potential makes it easier for, say, spin-up electrons to flow to the right and for spin-down electrons to flow to the left. The idea is represented schematically in Fig. 3.

There have been a number of proposals for spin ratchets, using mesoscopic semiconductors and non-uniform magnetic fields, asymmetric periodic structures with Rashba spin-orbit interaction, and double-well structures combined with local external magnetic fields and resonant tunnelling [22]. The first experimental realization, however, is based on a different mechanism based on a single electron transistor with a superconducting island and normal leads [21]. The main requirements for the spin ratchet effect to be observed are rather simple: i) a small-volume thin superconducting island, ii) a Zeeman-induced splitting imposed by an applied magnetic field, and iii) an asymmetric tunnelling to the metal electrodes.

Fig. 3. Schematic potential in a rocking spin ratchet. The potential is reversed along the motion direction for spins with opposite orientation. The asymmetry leads to easy motion of spin-up electrons to the right and spin-down electrons to the left.

This demonstration opens a new avenue in the research of low dissipation spintronics. The drawback of the single electron transistor spin ratchet is that it only works at temperatures below the critical temperature of the superconductor. However, at such low temperatures, it operates at the single electron level and is relevant for fundamental research. Given its simple fabrication, it can, for example, be used to initialize and read out the state of future spin-based quantum bits or to identify the spin orientation of single electrons in a test of the Einstein-Podolsky-Rosen paradox with spin-entangled electrons [21].

Graphene Spintronics

Scalability makes graphene a viable material alternative for developing “beyond CMOS” nanoelectronics [23]. Devices have already demonstrated cut-off frequencies above any known systems with an extrapolated value of 100 GHz. Charge excitations close to the Fermi level can be described as massless relativistic particles obeying a Dirac equation, whereas a new degree of freedom reflecting sublattice degeneracy appears in the electronic states as a pseudospin. Because of the resulting symmetry, electronic states turn out to be highly insensitive to disorder and very large charge mobilities are expected. The mobility could be as large as 10⁵cm²V^-1s^-1 at room temperature for carrier densities of 10¹² cm^-2. This compares extremely well with silicon, which has mobilities two orders of magnitude lower.

Graphene, and carbon-based nanostructures in general, are also attractive for spintronics because of their carrier concentration tunability and low spin-orbit and hyperfine interactions, which should lead to long spin coherence times. Gate control of spin conduction is of high interest for multifunctional spintronic devices, which could offer a true capability for efficient spin manipulation, and a broad spectrum of spintronic applications, including memories, transistors, and logic gates.

Several groups [24] have reported successful spin injection and detection experiments at room temperature in poor quality graphene with spin relaxation lengths in the range of a few micrometers. These results are very promising, although it is yet unclear what limits the spin lifetime as theoretical estimations for clean graphene predict spin relaxation lengths at least one order of magnitude larger [24].

The future of graphene-based spintronic applications will depend on a thorough understanding and clever use of spin injection, as well as the innovative manipulation of the spin degrees of freedom, which could include chemical functionalization or proximity induced ferromagnetism. The exploration of the interface and intrinsic properties offered by this material will be a prerequisite to bring it to a mainstream spintronic technology.

Dissipationless Spintronics and Topological Insulators

Topological insulators are a new topological state of matter that was discovered just five years ago. It was first predicted and discovered in two-dimensional (2D) systems and later on in 3D materials [25]. In the 3D case, the surface state (or interface with an ordinary insulator) can be described by a 2D massless Dirac fermion with a dispersion forming a Dirac cone with the crossing point located at the time reversal invariant momentum k = 0. The degeneracy at k = 0 and the surface metallic states are protected by time inversion symmetry. Electrons travelling on such a surface state are weakly sensitive to impurity scattering and their spins have opposite orientation at momenta k and -k (Fig. 4).

Fig. 4. In the simplest 3D topological insulator the dispersion relation is described by a single Dirac cone and the Fermi circle encloses a single Dirac point. The states are spin polarized, and there is a unique correspondence between the electron momenta and their spin.

Notably, topological insulators can be predicted within the simple framework of the band theory of solids and more than fifty compounds have already been proposed [25]. Their nontrivial topology originates from a spin orbit-induced inversion of the usual ordering of the conduction and valence bands. Theoretical predictions have been confirmed in a series of groundbreaking experiments in HgTe quantum wells, Bi_xSb_1-x alloys, and Bi₂Se₃ and Bi₂Te₃ crystals [25].

Measurements in HgTe [26] have demonstrated dissipationless current flow at the edges of the sample, whereas the use of Hall currents to switch the magnetization of ferromagnets may help eliminate power dissipation in magnetic recording technologies [27]. Although all of the reported results are from experiments well below room temperature, interest in topological insulators will certainly keep growing as novel device architectures are proposed to take advantage of the extraordinaire properties of these materials.

Final Remarks

There have been outstanding advances in spintronics in the last two decades that combine scientific and technological breakthroughs. As giant magnetoresistance opened the way to control (spin) transport through ferromagnets, new materials and innovative device structures are shaping a new paradigm for information technology. The progress has been very rapid with commercial applications within ten years of fundamental physics discoveries.

During this time, the scope of spintronics has become very broad making it impossible to feature here all of the topics in the field, which also include spin manipulation in semiconductors, diamond NV centers and quantum computing, to mention a few [28]. In this short article, we chose to feature a few emerging topics that are gathering great attention. Some appear to be closer to the marketplace (spin Hall and Rashba spin torques) than others (spin-related energy harvesting, graphene spintronics, topological insulators). However, they are a small sample that shows how spintronics is constantly fed by scientific breakthroughs and, at the same time, strengthening its synergy with industry, putting the field in prime situation for increased technological impact in the near and long term.

Acknowledgements

We acknowledge financial support from MICINN (MAT2010-18065) and the European Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement NANOFUNCTION n°257375.

References

[1] I. Žutiæ, J. Fabian, and S. Das Sarma, Rev. Mod. Phys. 76, 323 (2004). S. O. Valenzuela, Int. J. Mod. Phys. B, 23, 2413 (2009).

[2] A. Fert, Rev. Mod. Phys. 80, 1517 (2008).

[3] P. A. Grünberg, Rev. Mod. Phys. 80, 1531 (2008).

[4] J. S. Moodera et al., Phys. Rev. Lett. 74, 3273 (1995). T. Miyazaki and N. Tezuka, J. Magn. Magn. Mater. 139, L231 (1995).

[5] S. Datta and B. Das, Appl. Phys. Lett. 56, 665 (1990).

[6] J. A. Katine, E. E. Fullerton, J. Magn. Magn. Mat. 320, 1217 (2008). J. Z. Sun and D. C. Ralph, J. Magn. Magn. Mater. 320, 1227 (2008).

[7] M. I. Dyakonov and V. I. Perel, JETP Lett. 13, 467 (1971). M. I. Dyakonov and V. I. Perel, Phys. Lett. 35A, 459 (1971).

[8] Y. K. Kato, R. C. Myers, A. C. Gossard, and D. D. Awschalom, Science 306, 1910 (2004). J. Wunderlich, B. Kaestner, J. Sinova, and T. Jungwirth, Phys. Rev. Lett. 94, 047204 (2005).

[9] S. O. Valenzuela and M. Tinkham, Nature 442, 176 (2006). S. O. Valenzuela and M. Tinkham, J. App. Phys.101, 09B103 (2007).

[10] T. Kimura, Y. Otani, T. Sato, S. Takahashi, and S. Maekawa, Phys. Rev. Lett. 98, 156601 (2007). L. Vila, T. Kimura, and Y. Otani, Phys. Rev. Lett. 99, 226604 (2007).

[11] Y. Niimi, M. Morota, D. H. Wei, C. Deranlot, M. Basletic, A. Hamzic, A. Fert, Y. Otani, Phys. Rev. Lett. 106, 126601 (2011).

[12] G. Y. Guo, S. Maekawa, and N. Nagaosa, Phys. Rev. Lett. 102, 036401 (2009). A. Fert and P.M. Levy Phys. Rev. Lett. 106, 157208 (2011).

[13] J. Wunderlich, et al., Science 330, 1801 (2010)

[14] L. Q. Liu, T. Moriyama, D. C. Ralph, and R. A. Buhrman, Phys. Rev. Lett. 106, 036601 (2011).

[15] I. M. Miron et al., Nature Mater. 9, 230, (2010). A.Manchon, and S. Zhang, Phys. Rev. B 79, 094422(2009).

[16] “Spin Caloritronics” Sol. State. Comm. 150, Issues 11-12, Pages 459-552 (2010) Edited by G.E.W Bauer, A.H. MacDonald and S. Maekawa

[17] K. Uchida et al. Nature 455,778 (2008).

[18] K. Uchida et al. Nature Mater. 9, 894 (2010).[

[19] C. M. Jaworski et al. Nature Mater. 9, 898 (2010).

[20] A. Slachter, F. L. Bakker, J-P. Adam and B. J. van Wees Nature Phys. 6, 879 (2010).

[21] M. V. Costache and S. O. Valenzuela, Science 330, 1645 (2010).

[22] M. Scheid, D. Bercioux, K. Richter, N. J. Phys. 9, 401 (2007). S. Smirnov, D. Bercioux, M. Grifoni, K. Richter, Phys. Rev. Lett. 100, 230601 (2008). M. Scheid, A. Lassl, K. Richter, EPL 87, 17001 (2009).

[23] A. K. Geim and K. S. Novoselov, Nature Mater. 6, 183 (2007).

[24] N. Tombros et al., Nature 448, 571 (2007). M. Ohishi et al., Jpn. J. Appl. Phys. 46, L605 (2007). S. Cho, Y.-Fu Chen, and M. S. Fuhrer, Appl. Phys. Lett. 91, 123105 (2007). N. Tombros, at al. Phys. Rev. Lett. 101, 046601 (2008) C. Józsa, et al., Phys. Rev. Lett. 100, 236603 (2008). H. Goto et al., Appl. Phys. Lett. 92, 212110 (2008). W. Han et al., Appl. Phys. Lett. 94, 222109 (2009). W. Han et al., Phys. Rev. Lett. 102, 137205 (2009). C. Józsa, et al. Phys. Rev. B: Rap. Comm. 80, 241403(R) (2009). K. Pi et al, Phys. Rev. Lett. 104, 187201 (2010). W. Han et al., Phys. Rev. Lett. 105, 167202 (2010).

[25] For a review see M. Z. Hasan and C. L. Kane, Rev. Mod. Phys. 82, 3045 (2010). X-L. Qi and S-C.Zhang arXiv:1008.2026v1 [cond-mat.mes-hall].

[26] A. Roth et al., Science 325, 294 (2009).

[27] I. Garate and M. Franz, Phys. Rev. Lett. 104, 146802 (2010).

[28] D. D. Awschalom and M.E. Flatté, Nature Phys. 3, 153 (2007).