Binding events through the mutual synchronization of spintronic nano-neurons.

The brain naturally binds events from different sources in unique concepts. It is hypothesized that this process occurs through the transient mutual synchronization of neurons located in different regions of the brain when the stimulus is presented. This mechanism of 'binding through synchronization' can be directly implemented in neural networks composed of coupled oscillators. To do so, the oscillators must be able to mutually synchronize for the range of inputs corresponding to a single class, and otherwise remain desynchronized. Here we show that the outstanding ability of spintronic nano-oscillators to mutually synchronize and the possibility to precisely control the occurrence of mutual synchronization by tuning the oscillator frequencies over wide ranges allows pattern recognition. We demonstrate experimentally on a simple task that three spintronic nano-oscillators can bind consecutive events and thus recognize and distinguish temporal sequences. This work is a step forward in the construction of neural networks that exploit the non-linear dynamic properties of their components to perform brain-inspired computations.

Large voltage-induced coercivity change in Pt/Co/CoO/amorphous TiOx structure and heavy metal insertion effect.

There is urgent need for spintronics materials exhibiting a large voltage modulation effect to fulfill the great demand for high-speed, low-power-consumption information processing systems. Fcc-Co (111)-based systems are a promising option for research on the voltage effect, on account of their large perpendicular magnetic anisotropy (PMA) and high degree of freedom in structure. Aiming to observe a large voltage effect in a fcc-Co (111)-based system at room temperature, we investigated the voltage-induced coercivity (Hc) change of perpendicularly magnetized Pt/heavy metal/Co/CoO/amorphous TiOx structures. The thin CoO layer in the structure was the result of the surface oxidation of Co. We observed a large voltage-induced Hc change of 20.2 mT by applying 2 V (0.32 V/nm) to a sample without heavy metal insertion, and an Hc change of 15.4 mT by applying 1.8 V (0.29 V/nm) to an Ir-inserted sample. The relative thick Co thickness, Co surface oxidation, and large dielectric constant of TiOx layer could be related to the large voltage-induced Hc change. Furthermore, we demonstrated the separate adjustment of Hc and a voltage-induced Hc change by utilizing both upper and lower interfaces of Co.

Giant charge-to-spin conversion in ferromagnet via spin-orbit coupling.

Converting charge current into spin current via the spin Hall effect enables efficient manipulation of magnetization by electrical current. However, its geometrical restriction is a serious obstacle to device applications because it prevents switching of perpendicular magnetization in the absence of an external field. To resolve this issue, ferromagnetic materials have attracted attentions because their time reversal asymmetry induces magnetic-dependent charge-to-spin conversion that removes this restriction. Here, we achieved a large enhancement of magnetic-dependent charge-to-spin conversion by clarifying its mechanism. Through layer thickness dependence of the conversion efficiency, we revealed a coexistence of interfacial and bulk contributions to the magnetic-dependent charge-to-spin conversion. Moreover, the interfacial contribution to charge-to-spin conversion is found to be dominant and can be controlled via interfacial band engineering. The efficiency of charge-to-spin conversion in ferromagnet was found to be an order larger than that of other materials with reduced symmetry.

Control of the stochastic response of magnetization dynamics in spin-torque oscillator through radio-frequency magnetic fields.

Neuromorphic computing using spintronic devices, such as spin-torque oscillators (STOs), has been intensively studied for energy-efficient data processing. One of the critical issues in this application is stochasticity in magnetization dynamics, which limits the accuracy of computation. Such stochastic behavior, however, plays a key role in stochastic computing and machine learning. It is therefore important to develop methods for both suppressing and enhancing stochastic response in spintronic devices. We report on experimental investigations on control of stochastic quantity, such as the width of a distribution of transient time in magnetization dynamics in vortex-type STO. The spin-transfer effect can suppress stochasticity in transient dynamics from a non-oscillating to oscillating state, whereas an application of a radio-frequency magnetic field is effective in reducing stochasticity on the time evolution of the oscillating state.

Generation of charge current from magnetization oscillation via the inverse of voltage-controlled magnetic anisotropy effect.

It is well known that oscillating magnetization induces charge current in a circuit via Faraday's law of electromagnetic induction. New physical phenomena by which magnetization dynamics can produce charge current have gained considerable interest recently. For example, moving magnetization textures, such as domain walls, generates charge current through the spin-motive force. Here, we examine an entirely different effect, which couples magnetization and electric field at the interface between an ultrathin metallic ferromagnet and dielectric. We show that this coupling can convert magnetic energy into electrical energy. This phenomenon is the Onsager reciprocal of the voltage-controlled magnetic anisotropy effect. The effect provides a previously unexplored probe to measure the magnetization dynamics of nanomagnets.

Influence of flicker noise and nonlinearity on the frequency spectrum of spin torque nano-oscillators.

The correlation of phase fluctuations in any type of oscillator fundamentally defines its spectral shape. However, in nonlinear oscillators, such as spin torque nano-oscillators, the frequency spectrum can become particularly complex. This is specifically true when not only considering thermal but also colored 1/f flicker noise processes, which are crucial in the context of the oscillator's long term stability. In this study, we address the frequency spectrum of spin torque oscillators in the regime of large-amplitude steady oscillations experimentally and as well theoretically. We particularly take both thermal and flicker noise into account. We perform a series of measurements of the phase noise and the spectrum on spin torque vortex oscillators, notably varying the measurement time duration. Furthermore, we develop the modelling of thermal and flicker noise in Thiele equation based simulations. We also derive the complete phase variance in the framework of the nonlinear auto-oscillator theory and deduce the actual frequency spectrum. We investigate its dependence on the measurement time duration and compare with the experimental results. Long term stability is important in several of the recent applicative developments of spin torque oscillators. This study brings some insights on how to better address this issue.

Voltage-Driven Magnetization Switching Controlled by Microwave Electric Field Pumping.

We study the dynamic switching properties of a nanomagnet under microwave electric field pumping. The periodic modulation of an anisotropy field induced by microwave electric field pumping efficiently excites the uniform magnetization oscillation, allowing for precise control of magnetization switching. Accurate shaping of the pumping voltage waveform also enables us to investigate the transient reaction of magnetization to the relative phase difference of the pumping signal. We demonstrate both experimentally and theoretically the existence of a dead angle in which the uniform oscillation of magnetization is inhibited even though the microwave frequency itself satisfies the conditions of parametric excitation. Our results provide an energy-efficient way of manipulating ultrafast magnetization dynamics in nanomagnetic devices.

Role of non-linear data processing on speech recognition task in the framework of reservoir computing.

The reservoir computing neural network architecture is widely used to test hardware systems for neuromorphic computing. One of the preferred tasks for bench-marking such devices is automatic speech recognition. This task requires acoustic transformations from sound waveforms with varying amplitudes to frequency domain maps that can be seen as feature extraction techniques. Depending on the conversion method, these transformations sometimes obscure the contribution of the neuromorphic hardware to the overall speech recognition performance. Here, we quantify and separate the contributions of the acoustic transformations and the neuromorphic hardware to the speech recognition success rate. We show that the non-linearity in the acoustic transformation plays a critical role in feature extraction. We compute the gain in word success rate provided by a reservoir computing device compared to the acoustic transformation only, and show that it is an appropriate bench-mark for comparing different hardware. Finally, we experimentally and numerically quantify the impact of the different acoustic transformations for neuromorphic hardware based on magnetic nano-oscillators.

Tunnel magnetoresistance angular and bias dependence enabling tuneable wireless communication.

Spin-transfer torques (STTs) can be exploited in order to manipulate the magnetic moments of nanomagnets, thus allowing for new consumer-oriented devices to be designed. Of particular interest here are tuneable radio-frequency (RF) oscillators for wireless communication. Currently, the structure that maximizes the output power is an Fe/MgO/Fe-type magnetic tunnel junction (MTJ) with a fixed layer magnetized in the plane of the layers and a free layer magnetized perpendicular to the plane. This structure allows for most of the tunnel magnetoresistance (TMR) to be converted into output power. Here, we experimentally and theoretically demonstrate that the main mechanism sustaining steady-state precession in such structures is the angular dependence of the magnetoresistance. The TMR of such devices is known to exhibit a broken-linear dependence versus the applied bias. Our results show that the TMR bias dependence effectively quenches spin-transfer-driven precession and introduces a non-monotonic frequency dependence at high applied currents. This has an impact on devices seeking to work in the 'THz gap' due to their non-trivial TMR bias dependences.

Microwave amplification in a magnetic tunnel junction induced by heat-to-spin conversion at the nanoscale.

Heat-driven engines are hard to realize in nanoscale machines because of efficient heat dissipation1. However, in the realm of spintronics, heat has been employed successfully-for example, heat current has been converted into a spin current in a NiFe|Pt bilayer system2, and Joule heating has enabled selective writing in magnetic memory arrays3. Here, we use Joule heating in nanoscale magnetic tunnel junctions to create a giant spin torque due to a magnetic anisotropy change. Efficient conversion from heat dynamics to spin dynamics is obtained because of a large interfacial thermal resistance at an FeB|MgO interface. The heat-driven spin torque is equivalent to a voltage-controlled magnetic anisotropy4,5 of approximately 300 fJ V-1 m-1, which is more than twice the value reported in a (Co)FeB|MgO system6,7. We demonstrate an electric microwave amplification gain of 20% in a d.c. biased magnetic tunnel junction as a result of this spin torque. While electric d.c. power amplification in spintronic devices has been realized previously8, the microwave amplification was limited to relatively small amplification gains (G = radiofrequency output voltage/radiofrequency input voltage) and has never exceeded 1 (refs 9-13). A magnetic tunnel junction driven by radiofrequency spin transfer torque using ferromagnetic resonance enabled a relatively large gain of G ≈ 0.55 (ref. 12). Furthermore, radiofrequency spin waves were tuned by the spin transfer effect14,15. The heat-driven giant spin torque in the FeB|MgO16,17 magnetic tunnel junction, which shows a large magnetization precession and resistance oscillation under a d.c. bias, overcomes the above limitations and provides a gain larger than 1.

Vowel recognition with four coupled spin-torque nano-oscillators.

In recent years, artificial neural networks have become the flagship algorithm of artificial intelligence1. In these systems, neuron activation functions are static, and computing is achieved through standard arithmetic operations. By contrast, a prominent branch of neuroinspired computing embraces the dynamical nature of the brain and proposes to endow each component of a neural network with dynamical functionality, such as oscillations, and to rely on emergent physical phenomena, such as synchronization2-6, for solving complex problems with small networks7-11. This approach is especially interesting for hardware implementations, because emerging nanoelectronic devices can provide compact and energy-efficient nonlinear auto-oscillators that mimic the periodic spiking activity of biological neurons12-16. The dynamical couplings between oscillators can then be used to mediate the synaptic communication between the artificial neurons. One challenge for using nanodevices in this way is to achieve learning, which requires fine control and tuning of their coupled oscillations17; the dynamical features of nanodevices can be difficult to control and prone to noise and variability18. Here we show that the outstanding tunability of spintronic nano-oscillators-that is, the possibility of accurately controlling their frequency across a wide range, through electrical current and magnetic field-can be used to address this challenge. We successfully train a hardware network of four spin-torque nano-oscillators to recognize spoken vowels by tuning their frequencies according to an automatic real-time learning rule. We show that the high experimental recognition rates stem from the ability of these oscillators to synchronize. Our results demonstrate that non-trivial pattern classification tasks can be achieved with small hardware neural networks by endowing them with nonlinear dynamical features such as oscillations and synchronization.

Scaling up electrically synchronized spin torque oscillator networks.

Synchronized nonlinear oscillators networks are at the core of numerous families of applications including phased array wave generators and neuromorphic pattern matching systems. In these devices, stable synchronization between large numbers of nanoscale oscillators is a key issue that remains to be demonstrated. Here, we show experimentally that synchronized spin-torque oscillator networks can be scaled up. By increasing the number of synchronized oscillators up to eight, we obtain that the emitted power and the quality factor increase linearly with the number of oscillators. Even more importantly, we demonstrate that the stability of synchronization in time exceeds 1.6 milliseconds corresponding to 105 periods of oscillation. Our study demonstrates that spin-torque oscillators are suitable for applications based on synchronized networks of oscillators.

Neural-like computing with populations of superparamagnetic basis functions.

In neuroscience, population coding theory demonstrates that neural assemblies can achieve fault-tolerant information processing. Mapped to nanoelectronics, this strategy could allow for reliable computing with scaled-down, noisy, imperfect devices. Doing so requires that the population components form a set of basis functions in terms of their response functions to inputs, offering a physical substrate for computing. Such a population can be implemented with CMOS technology, but the corresponding circuits have high area or energy requirements. Here, we show that nanoscale magnetic tunnel junctions can instead be assembled to meet these requirements. We demonstrate experimentally that a population of nine junctions can implement a basis set of functions, providing the data to achieve, for example, the generation of cursive letters. We design hybrid magnetic-CMOS systems based on interlinked populations of junctions and show that they can learn to realize non-linear variability-resilient transformations with a low imprint area and low power.

Neuromorphic computing with nanoscale spintronic oscillators.

Neurons in the brain behave as nonlinear oscillators, which develop rhythmic activity and interact to process information. Taking inspiration from this behaviour to realize high-density, low-power neuromorphic computing will require very large numbers of nanoscale nonlinear oscillators. A simple estimation indicates that to fit 108 oscillators organized in a two-dimensional array inside a chip the size of a thumb, the lateral dimension of each oscillator must be smaller than one micrometre. However, nanoscale devices tend to be noisy and to lack the stability that is required to process data in a reliable way. For this reason, despite multiple theoretical proposals and several candidates, including memristive and superconducting oscillators, a proof of concept of neuromorphic computing using nanoscale oscillators has yet to be demonstrated. Here we show experimentally that a nanoscale spintronic oscillator (a magnetic tunnel junction) can be used to achieve spoken-digit recognition with an accuracy similar to that of state-of-the-art neural networks. We also determine the regime of magnetization dynamics that leads to the greatest performance. These results, combined with the ability of the spintronic oscillators to interact with each other, and their long lifetime and low energy consumption, open up a path to fast, parallel, on-chip computation based on networks of oscillators.

A magnetic synapse: multilevel spin-torque memristor with perpendicular anisotropy.

Memristors are non-volatile nano-resistors which resistance can be tuned by applied currents or voltages and set to a large number of levels. Thanks to these properties, memristors are ideal building blocks for a number of applications such as multilevel non-volatile memories and artificial nano-synapses, which are the focus of this work. A key point towards the development of large scale memristive neuromorphic hardware is to build these neural networks with a memristor technology compatible with the best candidates for the future mainstream non-volatile memories. Here we show the first experimental achievement of a multilevel memristor compatible with spin-torque magnetic random access memories. The resistive switching in our spin-torque memristor is linked to the displacement of a magnetic domain wall by spin-torques in a perpendicularly magnetized magnetic tunnel junction. We demonstrate that our magnetic synapse has a large number of intermediate resistance states, sufficient for neural computation. Moreover, we show that engineering the device geometry allows leveraging the most efficient spin torque to displace the magnetic domain wall at low current densities and thus to minimize the energy cost of our memristor. Our results pave the way for spin-torque based analog magnetic neural computation.

Controlling the phase locking of stochastic magnetic bits for ultra-low power computation.

When fabricating magnetic memories, one of the main challenges is to maintain the bit stability while downscaling. Indeed, for magnetic volumes of a few thousand nm(3), the energy barrier between magnetic configurations becomes comparable to the thermal energy at room temperature. Then, switches of the magnetization spontaneously occur. These volatile, superparamagnetic nanomagnets are generally considered useless. But what if we could use them as low power computational building blocks? Remarkably, they can oscillate without the need of any external dc drive, and despite their stochastic nature, they can beat in unison with an external periodic signal. Here we show that the phase locking of superparamagnetic tunnel junctions can be induced and suppressed by electrical noise injection. We develop a comprehensive model giving the conditions for synchronization, and predict that it can be achieved with a total energy cost lower than 10(-13) J. Our results open the path to ultra-low power computation based on the controlled synchronization of oscillators.

Self-Injection Locking of a Vortex Spin Torque Oscillator by Delayed Feedback.

The self-synchronization of spin torque oscillators is investigated experimentally by re-injecting its radiofrequency (rf) current after a certain delay time. We demonstrate that the integrated power and spectral linewidth are improved for optimal delays. Moreover by varying the phase difference between the emitted power and the re-injected one, we find a clear oscillatory dependence on the phase difference with a 2π periodicity of the frequency of the oscillator as well as its power and linewidth. Such periodical behavior within the self-injection regime is well described by the general model of nonlinear auto-oscillators including not only a delayed rf current but also all spin torque forces responsible for the self-synchronization. Our results reveal new approaches for controlling the non-autonomous dynamics of spin torque oscillators, a key issue for rf spintronics applications as well as for the development of neuro-inspired spin-torque oscillators based devices.

Extremely Coherent Microwave Emission from Spin Torque Oscillator Stabilized by Phase Locked Loop.

Spin torque oscillator (STO) has been attracting a great deal of attention as a candidate for the next generation microwave signal sources for various modern electronics systems since its advent. However, the phase noise of STOs under free running oscillation is still too large to be used in practical microwave applications, thus an industrially viable means to stabilize its oscillation has been strongly sought. Here we demonstrate implementation of a phase locked loop using a STO as a voltage controlled oscillator (VCO) that generates a 7.344 GHz microwave signal stabilized by a 153 MHz reference signal. Spectrum measurement showed successful phase locking of the microwave signal to the reference signal, characterized by an extremely narrow oscillation peak with a linewidth of less than the measurement limit of 1 Hz. This demonstration should be a major breakthrough toward various practical applications of STOs.

High domain wall velocities via spin transfer torque using vertical current injection.

Domain walls, nanoscale transition regions separating oppositely oriented ferromagnetic domains, have significant promise for use in spintronic devices for data storage and memristive applications. The state of these devices is related to the wall position and thus rapid operation will require a controllable onset of domain wall motion and high speed wall displacement. These processes are traditionally driven by spin transfer torque due to lateral injection of spin polarized current through a ferromagnetic nanostrip. However, this geometry is often hampered by low maximum wall velocities and/or a need for prohibitively high current densities. Here, using time-resolved magnetotransport measurements, we show that vertical injection of spin currents through a magnetic tunnel junction can drive domain walls over hundreds of nanometers at ~500 m/s using current densities on the order of 6 MA/cm(2). Moreover, these measurements provide information about the stochastic and deterministic aspects of current driven domain wall mediated switching.

Quasi-omnidirectional electrical spectrometer for studying spin dynamics in magnetic tunnel junctions.

We report an omnidirectional electrical spectroscopy setup for studying the spin dynamics in a nanoscale magnet. It has a measureable solid angle range comprising about 50% of the total range and allows the magnetoresistance and spin-torque diode signal to be measured simultaneously at any angle to the magnetization. This setup can provide detailed information about the spin-wave resonance modes excited in a nanoscale magnet.