Search

Magnonics aims to utilize magnons (quanta of spin-waves) to process and transfer digital and analog information, promising high throughput, low power computing for the post-CMOS era. Any such future magnonic circuits will require spin wave signal sources and amplifiers. We propose that acoustic pumping of spin waves provides a mechanism to implement these functions in efficient and highly localized devices. In support of this, we have developed a general theoretical model of linear and parametric magneto-acoustic interactions, covering all possible polarizations of acoustic waves and spin wave modes. The model combines the predictive power of analytical techniques with numerical micromagnetic simulations and is thus well-suited for the design of complex physical devices. Based on this, we determine the configurations most amenable to spin wave generation and amplification. As an experimental prototype we demonstrate an acoustically-pumped amplifier for spin-waves. Our device consists of an yttrium-iron-garnet (YIG) film grown on a gallium gadolinium garnet (GGG) substrate, with a bulk acoustic waves (BAW) transducer fabricated on the top of the GGG substrate. We show experimentally that the amplitude of the propagating spin-waves increases with the application of the BAW. Moreover, this scheme can be used as a signal correlator, where the modulated spin-waves and acoustic waves serve as signal inputs and the resulting modulation of the amplified spin wave serves as the output.

We have studied an infinite Hubbard chain with a spin-orbit coupling term, submitted to a uniform magnetic field as well as local phonons. By means of a Lang-Firsov transformation, we show that an effective interacting fermion model emerges. Moreover, a self-consistent mean-field theory of this model, formulated in terms of thermal Green's functions, shows that a BCS term emerges, thus leading to a superconducting phase transition at zero temperature. We find analytical expressions for the phase boundary, that agree well with exact numerical diagonalization of the Hamiltonian.

Spin torque magnetic memory (ST-MRAM) is currently under intense academic and industrial development, as it features non-volatility, high write and read speed and high endurance. However, one of its great challenge is the probabilistic nature of programming magnetic tunnel junctions, which imposes significant circuit or energy overhead for conventional ST-MRAM applications. In this work, we show that in unconventional computing applications, this drawback can actually be turned into an advantage. First, we show that conventional magnetic tunnel junctions can be reinterpreted as stochastic “synapses” that can be the basic element of low-energy learning systems. System-level simulations on a task of vehicle counting highlight the potential of the technology for learning systems. We investigate in detail the impact of magnetic tunnel junctions’ imperfections. Second, we introduce how intentionally superparamagnetic tunnel junctions can be the basis for low-energy fundamentally stochastic computing schemes, which harness part of their energy in thermal noise. We give two examples built around the concepts of synchronization and Bayesian inference. These results suggest that the stochastic effects of spintronic devices, traditionally interpreted by electrical engineers as a drawback, can be reinvented as an opportunity for low energy circuit design.

When electrons in a magnetic metal are driven far from equilibrium via ultrafast heating of the electrons, the magnetic order undergoes radical changes within tens of femtoseconds due to massive flows of energy and angular momentum between electrons, spins, and phonons. In ferrimagnetic metals such as GdFeCo, ultrafast optical heating can deterministically reverse the magnetization in less than a picosecond. In this talk, I describe our experimental work to gain a better understanding of how energy is exchanged between electrons, phonon, and spins in a magnetic metal following ultrafast heating. We use time-resolved measurements of the magneto-optic Kerr effect to record the response of ferro- and ferri-magnetic metals to heating via ultrafast optical or electrical pulses. Picosecond electrical pulses are generated with photoconductive Auston switches. By comparing the magnetic dynamics that result from electrical vs. optical heating, we identify differences in the rate of energy transfer to phonons from thermal vs. nonthermal electrons. We also find that both optical and electrical heating are effective for ultrafast switching of ferrimagnetic metals. We observe deterministic, repeatable ultrafast reversal of the magnetization of a GdFeCo thin film with a single sub-10 ps electrical pulse. The magnetization reverses in ~10 ps, which is more than one order of magnitude faster than other electrically controlled magnetic switching mechanisms.

The manipulation of magnetization by an electric current can be obtained thanks to the spin-transfer torque (STT) phenomenon. The two main mechanisms that give rise to this phenomenon are the spin filtering torque (SFT) and the spin-orbit torque (SOT). The former is obtained in multilayer systems by passing a spin-polarized current through the multilayers, whereas the latter is achieved through a direct transfer of angular momentum from the crystal lattice through the spin–orbit interaction. In this work, we study the influence of these two mechanisms on the magnetization dynamics of a three-terminal spin-torque oscillator. The device is composed of either a spin-valve (SV) or a magnetic tunnel junction (MTJ) on top of a Pt wire. The system can be excited either by SFT or by SOT depending on whether the current is applied through the magnetic structure (SV or MTJ) or through the Pt wire. Finally, each device is compared to a reference sample where the Pt wire is replaced by a Cu wire. Therefore, no SOT is expected in this second set of devices. In this presentation, we compare the different types of devices in order to understand the transport in the three terminal devices and study the role of the different mechanisms (SFT and STT) on the magnetization dynamics. Finally, we propose a method to compare the spin transfer efficiency of the two mechanisms on a single device.

Interfacial spin-orbit torques (SOTs) enable manipulation of the magnetization through an in-plane charge current, which has drawn increasing attention for spintronic applications. In the search for material systems that can provide efficient SOTs, much work has been focused on polycrystalline ferromagnetic metal/non-magnetic metal bilayers [1-3]. Here the current flows in the non-magnetic layers and induces a torque at the interfaces via the spin Hall effect of the non-magnetic layer [1], the Rashba effect at the interface [2] or spin-momentum locking when a topological insulator is involved [3]. In this presentation, we report the observation of robust SOT occuring in a well characterized single crystalline Fe/insulating GaAs (001) interface at room temperature where the SOT is caused by the lack of space inversion symmetry at the interface. We find that the magnitude of the interfacial SOT per unit charge current density is comparable in strength with that in ferromagnetic metal/non-magnetic metal systems. This large magnitude also allows for the observation of spin-to-charge current conversion at the interface, which is known as spin-galvanic effect [4]. The results suggest that single crystalline Fe/GaAs interfaces may enable efficient magnetization manipulation through purely electric means [5]. References [1] L. Liu, et al., Science 336, 555 (2012). [2] I. M. Miron, et al., Nature 476, 189 (2011). [3] A. R. Mellnik, et al., Nature 511, 449 (2014). [4] S. D. Ganichev, et al., Nature 417, 153 (2002). [5] L. Chen, et al., Nature Commun. 7, 13802 (2016).

In recent years, Spin Orbit interaction as a source of spin current has been widely used through the physics of Spin Hall Effect (SHE). The peculiar symmetry of SHE allows creating a spin accumulation at the interface between a spin-orbit metal and a magnetic insulator that could lead to a net pure spin current flowing from the metal into the insulator. This spin current will induce a torque on the magnetization and eventually could drive it into steady motion. As a ferromagnetic insulator with a very low Gilbert damping, Yttrium Iron Garnet (YIG) is a very promising candidate to investigate pure spin current phenomena. Only very recently, with the developments in preparation of high-quality nanometer-thick YIG films, the implementation of insulator-based spin-torque devices became practically feasible. Here, we report on the excitation of auto-oscillations in microstructures of YIG(20nm)\Pt(8nm) driven by Spin Orbit torque (SOT). By injection of a dc current in the adjacent Pt layer, we have been able to prove that the SOT due to SHE is sufficiently strong and efficient to drive the YIG magnetization dynamics at frequencies that closely follow the ferromagnetic resonance mode. These auto-oscillations have been detected either inductively using a spectrum analyzer or directly observed using micro-focus Brillouin Light Scattering. Furthermore, we achieved an efficient control of spin waves attenuation length in a YIG waveguide using the SHE in the sub-critical regime i.e. below the auto-oscillations threshold. We believe that our finding pave the path to active magnonics devices made out of YIG films.

In this presentation we will review the spintronic terahertz (THz) phenomena such as ultrafast transient demagnetization [1], generation of spin waves via THz Zeeman torque [2,3], Mott scattering [4], and ultrafast inverse spin-Hall effect [5]. [1] E. Beaurepaire, G. M. Turner, S. M. Harrel, M. C. Beard, J.-Y. Bigot, and C. A. Schmuttenmaer, "Coherent terahertz emission from ferromagnetic films excited by femtosecond laser pulses," Appl. Phys. Lett. 84, 3465 (2004) [2] T. Kampfrath, A. Sell, G. Klatt, A. Pashkin, S. Mahrlein, T. Dekorsy, M. Wolf, M. Fiebig, A. Leitenstorfer, and R. Huber, "Coherent terahertz control of antiferromagnetic spin waves," Nature Photonics 5, 31 (2011) [3] Z. Jin, Z. Mics, G. Ma, Z. Cheng, M. Bonn, and D. Turchinovich, "Single-pulse terahertz coherent control of spin resonance in the canted antiferromagnet YFeO3, mediated by dielectric anisotropy", Phys. Rev. B 87, 094422 (2013) [4] Z. Jin, A. Tkach, F. Casper, V. Spetter, H. Grimm, A. Thomas, T. Kampfrath, M. Bonn, M. Klaui, and D. Turchinovich, "Accessing the fundamentals of magnetotransport in metals with terahertz probes", Nature Physics 11, 761 (2015) [5] T. Seifert, S. Jaiswal, U. Martens, J. Hannegan, L. Braun, P. Maldonado, F. Freimuth, A. Kronenberg, J. Henrizi, I. Radu, F. Freimuth, E. Beaurepaire, Y. Mokrousov, P.M. Oppeneer, M. Jourdan, G. Jakob, D. Turchinovich, L. M. Hayden, M. Wolf, M. Munzenberg, M. Klaui, and T. Kampfrath, "Efficient metallic spintronic emitters of ultrabroadband terahertz radiation", Nature Photonics 10, 483 (2016)

In the last five years, Artificial Intelligence has made striking progress, and now defeating humans at subtle strategy games, such as Go, and even Poker. However, these algorithms are running on traditional processors which have a radically different architecture than the biological neural networks they are inspired from. This considerably slows them down and requires massive amounts of electrical power, more than ten thousand times what the brain typically need to function. This energy dissipation is not only becoming an environmental issue, but it also sets a limit to the size of neural networks that can be simulated. We are at a point where we need to rethink the way we compute, and build hardware chips directly inspired from the architecture of the brain. This is a challenge. Indeed, contrarily to current electronic systems, the brain is a huge parallel network closely entangling memory and processing. In this talk, I will show that, for building the neuromorphic chips of the future, we will need to emulate functionalities of synapses and neurons at the nanoscale. I will review the recent developments of memristive nano-synapses and oscillating nano-neurons, the physical mechanisms at stake, and the challenges in terms of materials. Finally, I will present the first achievements of neuromorphic computing with novel nanodevices and the fascinating perspectives of this emerging field.