Session C3: Antiferromagnetic Spintronics

Electrical switching of an antiferromagnet

Louis N\'eel pointed out in his Nobel lecture that while abundant and interesting from theoretical viewpoint, antiferromagnets did not seem to have any applications. Indeed, the alternating directions of magnetic moments on individual atoms and the resulting zero net magnetization make antiferromagnets hard to control by tools common in ferromagnets. Strong coupling would be achieved if the externally generated field had a sign alternating on the scale of a lattice constant at which moments alternate in AFMs. However, generating such a field has been regarded unfeasible, hindering the research and applications of these abundant magnetic materials. We have recently predicted that relativistic quantum mechanics may offer staggered current induced fields with the sign alternating within the magnetic unit cell which can facilitate a reversible switching of an antiferromagnet by applying electrical currents with comparable efficiency to ferromagnets. Among suitable materials is a high Néel temperature antiferromagnet, tetragonal-phase CuMnAs, which we have recently synthesized in the form of single-crystal epilayers structurally compatible with common semiconductors. We demonstrate electrical writing and read-out, combined with the insensitivity to magnetic field perturbations, in a proof-of-concept antiferromagnetic memory device. References: [1] J. Zelezny, et al., Phys. Rev. Lett. 113, 157201 (2014). [2] P. Wadley, et al., Nat. Commun. 4, 2322 (2013). [3] P. Wadley et al. http://arxiv.org/abs/1503.03765. [4] T. Jungwirth, X. Marti, P. Wadley, J. Wunderlich, http://arxiv.org/abs/1509.05296.   

\textbf{Interconnections between magnetic state and transport currents in antiferromagnetic Sr}$_{\mathbf{2}}$\textbf{IrO}$_{\mathbf{4}}$

Interconnections between magnetic state and transport currents in ferromagnetic (F) heterostructures are the basis for spintronic applications, e.g. tunneling magnetoresistance and spin-transfer torque phenomena provide a means to read and write information in magnetic memory devices like STTRAM. Similar interconnections were proposed [1] to occur in systems where F-components are replaced with antiferromagnets (AFM). We demonstrated experimentally the existence of such interconnections in antiferromagnetic Mott insulator Sr$_{2}$IrO$_{4}$: first, we found [2] a very large anisotropic magnetoresistance (AMR) which can be used to monitor (read) the magnetic state of AFM; second, we demonstrated [3] the feasibility of reversible resistive switching driven by high-density currents/high electric fields which can be used for writing in AFM memory applications. These results support the feasibility of AFM spintronics where antiferromagnets are used in place of ferromagnets. This work was supported in part by C-SPIN, one of six centers of STARnet, a Semiconductor Research Corporation program, sponsored by MARCO and DARPA, and by NSF grants DMR-1207577, DMR-1265162 and DMR-1122603. [1] A. S. N\'{u}\~{n}ez et al., Phys. Rev. B 73, 214426 (2006); [2] C. Wang et al., Phys. Rev. X 4, 041034 (2014); [3] C. Wang et al, PRB 92, 115136 (2015).   

\textbf{Spin-Hall effects in metallic antiferromagnets}.

Materials possessing new parameters for efficient and tunable spin Hall effects are being explored, among which antiferromagnets have become one of the most promising candidates. Two distinct properties of antiferromagnets are the microscopic spin magnetic moment ordering and the intrinsic anisotropy. Thus the natural question arises whether these two unique features of antiferromagnets can become new degrees of freedom for tuning their spin Hall effects. We performed experimental studies using spin pumping and inverse spin Hall detection on prototypical CuAu-I-type metallic antiferromagnets, PtMn, IrMn, PdMn, and FeMn, in which we observed increasing spin Hall effects for the alloys with heavier elements included\footnote{ W. Zhang \textit{et al}, Phys. Rev. Lett. 113, 196602 (2014); Phys. Rev. B 92, 144405 (2015). }. In particular, PtMn shows a large spin Hall effect that is comparable to Pt. We also demonstrated that the spin transfer torques from the antiferromagnets are large enough to excite ferromagnetic resonance of an adjacent ferromagnetic layer. We conclude that the sign and magnitude of the spin Hall effects in these antiferromagnets are determined by the atomic spin-orbit coupling of the heavy elements (e.g. Pt and Ir) as well as the large spin magnetic moments of Mn. In addition, by using epitaxial growth, we investigated the influence of the different crystalline and magnetic orientations on the anisotropic spin Hall effects of these antiferromagnets. Most of the experimental results were further corroborated by first-principles calculations, which determine the intrinsic spin Hall effect contribution and suggest pronounced anisotropies. Thus metallic antiferromagnets may become an active component for manipulating spin dependent transport properties in spintronic concepts\footnote{ This work was done in collaboration with: M. Benjamin Jungfleisch, Frank Freimuth, Joseph N. Sklenar, Wanjun Jiang, John E. Pearson, Yuriy Mokrousov, John B. Ketterson, and Axel Hoffmann}.   

Spin Transport by Collective Spin Excitations

We report studies of angular momentum transport in insulating materials. Our measurements reveal efficient spin pumping from high wavevector $k$ spin waves in thin film $\rm Y_3Fe_5O_{12}$ (YIG): spin pumping is independent of wavevector up to $k \sim 20 \, \mu \rm m^{-1}$ [1]. Optical detection of YIG FMR by NV centers in diamond reveals a role for spin waves in this insulator-to-insulator spin transfer process [2]. Spin transport is typically suppressed by insulating barriers, but we find that fluctuating antiferromagnetic correlations enable efficient spin transport at nm-scale thicknesses in insulating antiferromagnets, even in the absence of long-range order, and that the spin decay length increases with the strength of the antiferromagnetic correlations [3,4]. [1] S.A. Manuilov, C.H. Du, R. Adur, H.L. Wang, V.P. Bhallamudi, F.Y. Yang and P.C. Hammel, Applied Physics Letters 107 042405 (2015); [2] C.S. Wolfe, V.P. Bhallamudi, H.L. Wang, C.H. Du, S. Manuilov, R.M. Teeling-Smith, A.J. Berger, R. Adur, F.Y. Yang and P.C. Hammel, Physical Review B Rapid Communication 89 180406 (2014); [3] H.L.Wang, C.H. Du, P.C. Hammel and F.Y. Yang, Physical Review Letters 113 097202 (2014); [4] H.L.Wang, C.H. Du, P.C. Hammel and F.Y. Yang, Physical Review B 91 220410 (2015).   

Mechanism of spin current transfer through antiferromagnetic dielectrics

The mechanisms of spin current (SC) transfer are well-studied in both metallic systems, where SC is carried mostly by spin-polarized electrons, and in ferromagnetic (FM) dielectrics, where propagating spin waves (magnons) are responsible for the spin transfer. The possibility of SC transfer through {\em antiferromagnetic dielectrics} (AFMD) is much less investigated, although recent experimental studies by H.~Wang {\it et al.} [H.~Wang {\it et al.}, Phys. Rev. Lett. {\bf 113}, 097202 (2014)] demonstrated extraordinary high efficiency of SC transfer in tri-layer FM-AFMD-Platinum (YIG-NiO-Pt) systems measured by the inverse spin Hall effect (ISHE). Perhaps the most unexpected result of these studies was that, with the increase of the thickness of the AFMD layer, the ISHE voltage, first, {\em increased}, and, then, exponentially decayed with the characteristic decay length of $\lambda \sim 10$~nm. Moreover, the excitation frequency, equal to the ferromagnetic resonance (FMR) frequency of the YIG layer, was rather low compared to the frequencies of the antiferromagnetic resonance in the AFMD, which rules out the eigenmodes of the AFMD layer as potential carriers of the spin current. Here we propose a possible mechanism of SC transfer through the AFMD with a biaxial anisotropy, which explains all previous experimental findings and opens a new way of manipulating spin currents using anisotropic AFMD materials. We show, that spin current can be carried by {\em evanescent} AFMD modes non-resonantly excited at the FM-AFMD interface. The decay length of the evanescent modes is defined by the AFMD anisotropy and determines the SC penetration depth into the AFMD. Furthermore, the anisotropy of the AFMD leads to the coupling between the spin subsystem and the crystal lattice of the AFMD, which makes possible exchange of angular momentum between these subsystems. We demonstrate that, under certain realistic conditions, the angular momentum flows from the lattice to the spin subsystem, in which case the AFMD layer acts as a {\em spin current amplifier}. The enhancement or the suppression of the spin current by the AFMD lattice depends on the phase shift between the two evanescent AFMD modes and, thus, can be controlled by the method of excitation.