Figure - available from: Nature Communications
This content is subject to copyright. Terms and conditions apply.
![Voltage-controlled perpendicular magnetic anisotropy
a Schematic illustration of the mechanism of the voltage-driven frequency tuning via voltage-controlled magnetic anisotropy (VCMA). The green and gray arrows represent the directions of spin-orbit torque and damping torque, respectively. b The resonance frequency (fres\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${f}_{{{{{{\rm{res}}}}}}}$$\end{document}) as a function of in-plane magnetic field (B∣∣\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${B}_{{{{{{\rm{| }}}}}}{{{{{\rm{| }}}}}}}$$\end{document}) for high PMA (blue) and low PMA (red). c Magnetization (M\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$M$$\end{document}) versus out-of-plane magnetic field (Bz\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${B}_{z}$$\end{document}) of the Co/Ni film. d Anomalous Hall resistance (RH\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${R}_{{{{{{\rm{H}}}}}}}$$\end{document}) curves of the Co/Ni sample for sequentially applied gate voltages of +3 V, +5 V, −3 V, and −5 V, respectively.](publication/361652015/figure/fig1/AS:1172866552872960@1656644585018/Voltage-controlled-perpendicular-magnetic-anisotropy-a-Schematic-illustration-of-the.png)
Voltage-controlled perpendicular magnetic anisotropy a Schematic illustration of the mechanism of the voltage-driven frequency tuning via voltage-controlled magnetic anisotropy (VCMA). The green and gray arrows represent the directions of spin-orbit torque and damping torque, respectively. b The resonance frequency (fres\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${f}_{{{{{{\rm{res}}}}}}}$$\end{document}) as a function of in-plane magnetic field (B∣∣\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${B}_{{{{{{\rm{| }}}}}}{{{{{\rm{| }}}}}}}$$\end{document}) for high PMA (blue) and low PMA (red). c Magnetization (M\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$M$$\end{document}) versus out-of-plane magnetic field (Bz\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${B}_{z}$$\end{document}) of the Co/Ni film. d Anomalous Hall resistance (RH\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${R}_{{{{{{\rm{H}}}}}}}$$\end{document}) curves of the Co/Ni sample for sequentially applied gate voltages of +3 V, +5 V, −3 V, and −5 V, respectively.
Source publication
Spin Hall nano-oscillators (SHNOs) exploiting current-driven magnetization auto-oscillation have recently received much attention because of their potential for neuromorphic computing. Widespread applications of neuromorphic devices with SHNOs require an energy-efficient method of tuning oscillation frequency over broad ranges and storing trained f...
Citations
... The nano-scale dimensions of SHNOs facilitate high-frequency operations and allow for significant miniaturization of the devices [60,61]. SHNOs are characterized by their wide frequency tunability [27,43,62], nanoscale size [61,63], and strong non-linear behavior, which are highly desirable for unconventional computing applications due to their easily accessible free ferromagnetic layer [64]. ...
... Oscillator arrays, on the other hand, rely on mutual synchronization and SHIL for the construction of the artificial spin state; with couplings being mediated by dipolar fields, propagating exchange waves or electrical current. Control over these couplings in STNOs and SHNOs include thermal excitation [33,47], voltage controlled magnetic anisotropy [42][43][44]62,129] and STT using memristive gates [130][131][132]. Although very enticing, spintronic oscillator-based IMs have only been reported based on small array measurements [71] and network simulations [133,134], with control over their individual couplings and biases remaining an engineering challenge yet to be resolved. ...
The ever increasing demand for computational power combined with the predicted plateau for the miniaturi-zation of existing silicon-based technologies has made the search for low power alternatives an industrial and scientifically engaging problem. In this work, we explore spintronics-based Ising machines as hardware computation accelerators. We start by presenting the physical platforms on which this emerging field is being developed, the different control schemes and the type of algorithms and problems on which these machines outperform conventional computers. We then benchmark these technologies and provide an outlook for future developments and use-cases that can help them get a running start for integration into the next generation of computing devices.
... The nano-scale dimensions of SHNOs facilitate high-frequency operations and allow for significant miniaturization of the devices [58,59]. SHNOs are characterized by their wide frequency tunability [27,60,43], nanoscale size [59,61], and strong non-linear behavior, which are highly desirable for unconventional computing applications due to their easily accessible free ferromagnetic layer [62]. ...
... Oscillator arrays, on the other hand, rely on mutual synchronization and SHIL for the construction of the artificial spin state; with couplings being mediated by dipolar fields, propagating exchange waves or electrical current. Control over these couplings in STNOs and SHNOs include thermal excitation [47,33], voltage controlled magnetic anisotropy [42,43,44,121,60] and STT using memristive gates [122,123,124]. Although very enticing, spintronic oscillator-based IM have only been reported based on small array measurements [68] and network simulations [125,126], with control over their individual couplings and biases remain an engineering challenge yet to be resolved. ...
The ever increasing demand for computational power combined with the predicted plateau for the miniaturization of existing silicon-based technologies has made the search for low power alternatives an industrial and scientifically engaging problem. In this work, we explore spintronics-based Ising machines as hardware computation accelerators. We start by presenting the physical platforms on which this emerging field is being developed, the different control schemes and the type of algorithms and problems on which these machines outperform conventional computers. We then benchmark these technologies and provide an outlook for future developments and use-cases that can help them get a running start for integration into the next generation of computing devices.
... Various emerging applications, including reconfigurable spin-wave logic circuits [10,11], unconventional computing [12], and Ising machines [13], rely on these advances. Multiple novel mechanisms have been explored to generate and amplify PSWs [14,15], such as current induced spin-transfer torque (STT) [16][17][18] and spin-orbit torque (SOT) [19][20][21][22][23]. Nano-constriction spin Hall nano-oscillators (SHNOs) with perpendicular magnetic anisotropy (PMA) [24,25] are a particularly promising approach, as they are easy to fabricate [26,27], CMOS compatible [28], strongly voltage tunable [29][30][31][32], and known for their superior mutual synchronization at various length scales and dimensions [33][34][35]. ...
Spin-orbit torque can drive auto-oscillations of propagating spin wave (PSW) modes in nano-constriction spin Hall nano-oscillators (SHNOs). These modes allow both long-range coupling and the potential of controlling its phase—critical aspect for nano-magnonics, spin wave logic, and Ising machines. Here, we demonstrate PSW-driven variable-phase coupling between two nano-constriction SHNOs and study how their separation and the PSW wave vector impact their mutual synchronization. In addition to ordinary in-phase mutual synchronization, we observe, using both electrical measurements and phase-resolved μ-Brillouin Light Scattering microscopy, mutual synchronization with a phase that can be tuned from 0 to π using the drive current or the applied field. Micromagnetic simulations corroborate the experiments and visualize how the PSW patterns in the bridge connecting the two nano-constrictions govern the coupling. These results advance the capabilities of mutually synchronized SHNOs and open up new possibilities for applications in spin wave logic, unconventional computing, and Ising Machines.
... However, future efficient integrated systems will use amorphous magnetic materials where the dynamics is not dominated by pinning, as demonstrated in Ref. 36. Furthermore, non-volatile resonance frequency tuning through magnetic anisotropy, as demonstrated in Ref. 37. ...
... To reduce pinning effect, we can, for instance, use amorphous materials as proposed in Ref. 36. To achieve need non-volatile weights, we need novel resonance frequency tuning, for instance, as proposed in Ref. 37. Second, we need to have MTJs functioning at wide ranges of frequencies. ...
Extracting information from radio-frequency (RF) signals using artificial neural networks at low energy cost is a critical need for a wide range of applications from radars to health. These RF inputs are composed of multiple frequencies. Here, we show that magnetic tunnel junctions can process analog RF inputs with multiple frequencies in parallel and perform synaptic operations. Using a backpropagation-free method called extreme learning, we classify noisy images encoded by RF signals, using experimental data from magnetic tunnel junctions functioning as both synapses and neurons. We achieve the same accuracy as an equivalent software neural network. These results are a key step for embedded RF artificial intelligence.
... Previous works in the conventional spintronics field have proved that voltage-controlled magnetic anisotropy is a highly energy-efficient approach to control magnetization [30], e.g., magnetization reversal and procession, compared to the current-based approach. In addition, the in-plane configuration of SHNOs can easy to achieve both current-and voltage-based collaborative control of nonlinear dynamics in three-terminal SHNOs [31][32][33]. ...
We experimentally study the dynamical modes excited by current-induced spin-orbit torques and the electrostatic gating effect in a three-terminal planar nanogap spin Hall nano-oscillator (SHNO) with a moderate interfacial perpendicular magnetic anisotropy (iPMA). Both quasilinear propagating spin-wave and localized ”bullet” modes are achieved and controlled by varying in-plane magnetic field and dc current. The minimum linewidth shows a linear dependence on the actual temperature of the active area, consistent with the single-mode dynamics described by the nonlinear theory of spin-torque nano-oscillator. The observed electrostatic gating tuning oscillation frequency arises from voltage-controlled magnetic anisotropy and threshold current of SHNO. The latter is related to voltage modification of the nonlinear damping and the interfacial spin-orbit coupling. In contrast to previously observed two-mode coexistence in Py/Pt-based SHNOs with a negligible iPMA, a single coherent spin-wave mode can be achieved by selecting the ferromagnet layer with a suitable iPMA, where nonlinear mode coupling in this nanogap SHNO can be diminished by bringing in the PMA field to compensate the easy-plane shape anisotropy. Moreover, the simulations demonstrate that the experimentally observed current-driven auto-oscillation modes are closely associated with the nonlinear damping and mode coupling, which can be tuned by the effective PMA coefficient. The demonstrated nonlinear mode coupling mechanism and electrical control approach of spin-wave modes could provide the clue to facilitate the implementation of the mutual synchronization map for neuromorphic computing applications in SHNO array networks.
... Novel effects such as voltage-controlled magnetic anisotropy can be used to increase spin-charge conversion and thus energy efficiency. Non-volatile tuning of the synaptic weights can be achieved by a similar mechanism through modification of the resonance frequencies 12,25,47 . Alternatively, binary non-volatile synapses exploiting vortex states have been demonstrated 48 . ...
... Higher frequencies are critical for both energy consumption and processing speed. While spintronic oscillators reliably exhibit frequencies in the 10 GHz range 47,57 , there are prospects for reaching several tens of gigahertz 58,59 or even terahertz 60,61 . ...
Spintronic nano-synapses and nano-neurons perform neural network operations with high accuracy thanks to their rich, reproducible and controllable magnetization dynamics. These dynamical nanodevices could transform artificial intelligence hardware, provided they implement state-of-the-art deep neural networks. However, there is today no scalable way to connect them in multilayers. Here we show that the flagship nano-components of spintronics, magnetic tunnel junctions, can be connected into multilayer neural networks where they implement both synapses and neurons thanks to their magnetization dynamics, and communicate by processing, transmitting and receiving radiofrequency signals. We build a hardware spintronic neural network composed of nine magnetic tunnel junctions connected in two layers, and show that it natively classifies nonlinearly separable radiofrequency inputs with an accuracy of 97.7%. Using physical simulations, we demonstrate that a large network of nanoscale junctions can achieve state-of-the-art identification of drones from their radiofrequency transmissions, without digitization and consuming only a few milliwatts, which constitutes a gain of several orders of magnitude in power consumption compared to currently used techniques. This study lays the foundation for deep, dynamical, spintronic neural networks.
... Large-amplitude EP persistent dynamics have been theoretically studied in spin-transfer-torque nanopillar devices 30,31 but have yet to be explored in SOT devices, such as SHOs. Here we report the experimental realization of a nanowire EP-SHO based on a Pt|FM bilayer, where FM is a Co|Ni superlattice [35][36][37][38][39][40] . The EP-SHO dynamics are achieved via tuning the Co|Ni perpendicular magnetic anisotropy (PMA) and the magnetic shape anisotropy of the nanowire to manufacture an easy plane defined by the wire axis and the film normal (xz-plane) as shown in Fig. 2a. ...
Spin Hall oscillators (SHOs) based on bilayers of a ferromagnet (FM) and a non-magnetic heavy metal (HM) are electrically tunable nanoscale microwave signal generators. Achieving high output power in SHOs requires driving large-amplitude magnetization dynamics by a direct spin Hall current. Here we present an SHO engineered to have easy-plane magnetic anisotropy oriented normal to the bilayer plane, enabling large-amplitude easy-plane dynamics driven by spin Hall current. Our experiments and micromagnetic simulations demonstrate that the easy-plane anisotropy can be achieved by tuning the magnetic shape anisotropy and perpendicular magnetic anisotropy in a nanowire SHO, leading to a significant enhancement of the generated microwave power. The easy-plane SHO experimentally demonstrated here is an ideal candidate for realization of a spintronic spiking neuron. Our results provide an approach to design of high-power SHOs for wireless communications, neuromorphic computing, and microwave assisted magnetic recording.
... 34 The mutual synchronization of these oscillators in a chain can be explored for bioinspired computing and beyond 9,23,35 where each oscillator behaves as a neuron. Combined with voltage 26,27,36 and/or memristive 25 control of synchronization (synaptic weights), these large chains can be used to locally or globally control the coupling between oscillators (neurons). The present work also serves as a stepping stone in the direction toward further scaling mutual synchronization to much larger square or rectangular arrays well beyond the previously demonstrated 64 oscillators. ...
Mutual synchronization of N serially connected spintronic nano-oscillators boosts their coherence by N and peak power by N2. Increasing the number of synchronized nano-oscillators in chains holds significance for improved signal quality and emerging applications such as oscillator based unconventional computing. We successfully fabricate spin Hall nano-oscillator chains with up to 50 serially connected nanoconstrictions using W/NiFe, W/CoFeB/MgO, and NiFe/Pt stacks. Our experiments demonstrate robust and complete mutual synchronization of 21 nanoconstrictions at an operating frequency of 10 GHz, achieving line widths <134 kHz and quality factors >79,000. As the number of mutually synchronized oscillators increases, we observe a quadratic increase in peak power, resulting in 400-fold higher peak power in long chains compared to individual nanoconstrictions. While chains longer than 21 nanoconstrictions also achieve complete mutual synchronization, it is less robust, and their signal quality does not improve significantly, as they tend to break into partially synchronized states.
... Previous works in the conventional spintronics field have proved that voltage-controlled magnetic anisotropy is a highly energy-efficient approach to control magnetization [30], e.g., magnetization reversal and procession, compared to the current-based approach. In addition, the in-plane configuration of SHNOs can easy to achieve both current-and voltage-based collaborative control of nonlinear dynamics in three-terminal SHNOs [31][32][33]. ...
We experimentally study the dynamical modes excited by current-induced spin-orbit torque and its electrostatic gating effect in a 3-terminal planar nano-gap spin Hall nano-oscillator (SHNO) with a moderate interfacial perpendicular magnetic anisotropy (IPMA). Both quasilinear propagating spin-wave and localized "bullet" modes are achieved and controlled by varying the applied in-plane magnetic field and driving current. The minimum linewidth shows a linear dependence on the actual temperature of the active area, confirming single-mode dynamics based on the nonlinear theory of spin-torque nano-oscillation with a single mode. The observed electrostatic gating tuning oscillation frequency arises from voltage-controlled magnetic anisotropy and threshold current of SHNO via modification of the nonlinear damping and/or the interfacial spin-orbit coupling of the magnetic multilayer. In contrast to previously observed two-mode coexistence degrading the spectral purity in Py/Pt-based SHNOs with a negligible IPMA, a single coherent spin-wave mode with a low driven current can be achieved by selecting the ferromagnet layer with a suitable IPMA because the nonlinear mode coupling can be diminished by bringing in the PMA field to compensate the easy-plane shape anisotropy. Moreover, the simulations demonstrate that the experimentally observed current and gate-voltage modulation of auto-oscillation modes are also closely associated with the nonlinear damping and mode coupling, which are determined by the ellipticity of magnetization precession. The demonstrated nonlinear mode coupling mechanism and electrical control approach of spin-wave modes could provide the clue to facilitate the implementation of the mutual synchronization map for neuromorphic computing applications in SHNO array networks.
... This can be achieved with VCMA-driven parametric excitations, as already proposed for spintronic oscillators with nano-constrictions and MTJ with two terminals [27,28]. Our calculations predict the robustness of the binarization technique and that it can be achieved in a range of frequencies close to two times the self-oscillation frequency. ...
The commercial and industrial demand for the solution of hard combinatorial optimization problems push forward the development of efficient solvers. One of them is the Ising machine which can solve combinatorial problems mapped to Ising Hamiltonians. In particular, spintronic hardware implementations of Ising machines can be very efficient in terms of area and performance, and are relatively low-cost considering the potential to create hybrid CMOS-spintronic technology. Here, we perform a comparison of coherent and probabilistic paradigms of Ising machines on several hard Max-Cut instances, analyzing their scalability and performance at software level. We show that probabilistic Ising machines outperform coherent Ising machines in terms of the number of iterations required to achieve the problem s solution. Nevertheless, high frequency spintronic oscillators with sub-nanosecond synchronization times could be very promising as ultrafast Ising machines. In addition, considering that a coherent Ising machine acts better for Max-Cut problems because of the absence of the linear term in the Ising Hamiltonian, we introduce a procedure to encode Max-3SAT to Max-Cut. We foresee potential synergic interplays between the two paradigms.
