The motion of domain walls due to the spin torque generated by coherent carrier transport is of considerable interest for the development of spintronic devices. We model the charge and spin transport through domain walls in ferromagnetic semiconductors for various systems. With an appropriate model Hamiltonian for the spin– dependent potential, we calculate wavefunctions inside the domain walls which are then used to calculate transmission and reflection coefficients, which are then in turn used to calculate current and spin torque. Starting with a simple approximation for the change in magnetization inside the domain wall, and ending with a sophisticated transfer matrix method, we model the common π wall, the less–studied 2π wall, and a system of two π walls separated by a variable distance. We uncover an interesting width dependence on the transport properties of the domain wall. 2π walls in particular, have definitive maximums in resistance and spin torque for certain domain wall widths that can be seen as a function of the spin mistracking in the system — when the spins are either passing straight through the domain wall (narrow walls) or adiabatically following the magnetization (wide walls), the resistance is low as transmission is high. In the intermediate region, there is room for the spins to rotate their magnetization, but not necessarily all the way through a 360 degree rotation, leading to reflection and resistance. We also calculate that there are widths for which the total velocity of a 2π wall is greater than that of a same–sized π wall. In the double–wall system, we model how the system reacts to changes in the separation of the domain walls. When the domain walls are far apart, they act as a spin–selective resonant double barrier, with sharp resonance peaks in the transmission profile. Brought closer and closer together, the number and sharpness of the peaks decrease, the spectrum smooths out, and the domain walls brought together have a transmission spectrum with many of the similar features from the 2π wall. Looking at the individual walls, we find an interesting interaction that has three distinct regimes: 1) repulsion, where the left wall moves to the left and the right wall to the right; 2) motion together, where the two walls both move to the right, even at the same velocity for one special value of separation; and 3) attraction, where the left wall moves to the right and the right wall moves to the left. This speaks to a kind of natural equilibrium distance between the domain walls. This is of major interest for device purposes as it means that stacks of domain walls could be self–correcting in their motions along a track. Much experimental work needs to be done to make this a reality, however.

In this dissertation, we provide an accurate description of spin transport in magnetic textures and in particular, we investigate in detail, the nature of spin torque and magnetic damping in such systems. Indeed, as will be further discussed in this thesis, the current-driven velocity of magnetic textures is related to the ratio between the so-called non-adiabatic torque and magnetic damping. Uncovering the physics underlying these phenomena can lead to the optimal design of magnetic systems with improved efficiency. We identified three interesting classes of systems which have attracted enormous research interest (i) Magnetic textures in systems with broken inversion symmetry: We investigate the nature of magnetic damping in non-centrosymmetric ferromagnets. Based on phenomenological and microscopic derivations, we show that the magnetic damping becomes chiral, i.e. depends on the chirality of the magnetic texture. (ii) Ferromagnetic domain walls, skyrmions and vortices: We address the physics of spin transport in sharp disordered magnetic domain walls and vortex cores. We demonstrate that upon spin-independent scattering, the non-adiabatic torque can be significantly enhanced. Such an enhancement is large for vortex cores compared to transverse domain walls. We also show that the topological spin currents owing in these structures dramatically enhances the non-adiabaticity, an effect unique to non-trivial topological textures (iii) Antiferromagnetic skyrmions: We extend this study to antiferromagnetic skyrmions and show that such an enhanced topological torque also exist in these systems. Even more interestingly, while such a non-adiabatic torque inuences the undesirable transverse velocity of ferromagnetic skyrmions, in antiferromagnetic skyrmions, the topological non-adiabatic torque directly determines the longitudinal velocity. As a consequence, scaling down the antiferromagnetic skyrmion results in a much more efficient spin torque.

No description available.

Les nano-oscillateurs à transfert de spin (STNO) sont des candidats prometteurs pour la réalisation de composants radiofréquence (RF) intégrés, du à leur taille nanométrique, l'importante gamme de fréquences de base qu'ils peuvent couvrir, ainsi qu'à leur accordabilité autour de ces fréquences de base. Le signal RF est obtenu grâce à l'effet de transfert de spin (STT) qui donne lieu à une oscillation non-linéaire de l'aimantation dans un élément magnétorésistif. Jusqu'ici, ces excitations ont été examinées dans le cadre d'une couche magnétique isolée, c'est-à-dire sans prendre en compte le couplage entre couches. Cependant, nombreux aspects du spectre d'excitation ne peuvent pas être expliqués si l'on considère une couche isolée. Dans cette thèse nous nous attacherons à répondre à la question importante du couplage dynamique entre couches magnétiques dans un nanopilier magnétorésistif, afin de développer une meilleure compréhension des spectres d'excitation, et en particulier la dépendance en courant et champ magnétique appliqué des caractéristiques du pic d'émission, telles que la largeur de raie et la fréquence. Une première étude est réalisée pour un système composé de deux couches ferromagnétiques, couplées entre elles par le couplage RKKY (ce système est appelé un ferrimagnétique synthétique (SyF)). Le couplage induit des différences importantes dans la dépendance en courant de la fréquence par rapport aux excitations d'une couche isolée. Ces différences sont expliquées par l'important couplage dynamique RKKY. Une seconde étude prend en considération une interaction plus complexe, ayant lieu dans un nano-pilier STNO standard basé sur jonctions tunnel ou vannes de spin. Ce dispositif est composé d'un SyF ainsi que d'une couche libre(FL) magnétique, séparés par une fine couche métallique ou isolante. Pour ce système, en plus du couplage dynamique RKKY propre au SyF, nous prenons en compte le couplage dynamique généré par le champ dipolaire ainsi que le spin-torque mutuel (MSTT) entre la couche libre et le SyF. Ce couplage multiple donne lieu à deux signatures distinctes. La première est l'apparition d'un « saut » dans le spectre d'excitation dû à l'hybridation des modes SyF and FL dans le régime atténué. Le second est dû à l'interaction entre les excitations en régime entretenu, éventuellement via leurs composantes harmoniques, avec les excitations en régime atténué. Cette interaction donne lieu à des discontinuités dans la dépendance fréquence – champ, ce lorsque les excitations FL sont prédominantes. Il est intéressant de noter que cela mène à des régions ou la largeur de raie est diminuée. De plus, lorsque les excitations SyF sont prédominantes, la largeur de raie est diminuée par rapport aux cas ou les excitations FL sont prédominantes. Partant de ces observations, nous proposons une structure plus complexe, où un seconde couche de type SyF remplace la couche libre. Les résultats obtenus par une combinaison d'expériences, de simulations numériques et d'analyse analytique, montrent le rôle important des interactions dynamiques dans un nano-pilier. Ils ouvrent de nouvelles voies pour la conception de configurations STNO qui mèneront à des améliorations des performances du signal ainsi synthétisé. / Spin-torque nano-oscillators (STNOs) are promising candidates for integrated radiofrequency (RF) components due to their nanoscale size, the large range of base frequencies that can be covered, as well as the large achievable tuning ranges around the base frequency. The RF signal is obtained due to the spin transfer torque (STT) generating a non-linear magnetization oscillation in a magnetoresistive device. In the past, these excitations were investigated using the picture of a single (or independent) layer. However, many features of the excitation spectra observed experimentally in nanopillar devices cannot be explained considering a single layer. In this thesis we address the important question on the dynamical coupling between the magnetic layers inside a magneto-resistive nanopillar device, to gain a better understanding of the excitation spectra, i.e. the dependence of the frequency and the linewidth on current and applied magnetic field. A first study is realized for a coupled system, composed by two ferromagnetic layers, coupled by the interlayer RKKY coupling (so called Synthetic Ferrimagnet SyF). Due to the coupling the frequency dependence versus current is different as compared to excitations of a single layer. This is explained by the strong dynamical RKKY coupling. A second study considers a more complex interaction, occurring within standard STNO nanopillar spin valves or tunnel junctions. They are composed by a SyF separated by a metallic or insulating spacer respectively from the single free layer (FL). For this system we take into account besides the RKKY coupling within the SyF, also the dynamical dipolar field coupling and the mutual spin torque (MSTT) between the SyF and the free layer. We find two definite signatures arising from this coupling. The first is a gap in the steady state excitation spectra that is due to the hybridization of the SyF and FL modes in the damped regime. The second is the possibility of the spintorque driven excitation or its harmonics with the damped modes leading to discontinuities in the frequency field dependence when the free layer is dominantly excited. Interestingly this leads to a region of reduced linewidth. Furthermore for SyF layer dominated excitations, the linewidth is lower than in the FL dominated excitations. From these observations we propose a more complex structure, composed by two SyF layers where the single FL is replaced by a SyF. The results obtained by a combination of experiments, numerical simulations and analytical analysis, demonstrate the important role of the dynamic interactions in nanopillar STNOs and provide routes for designing novel STNO configurations that should lead to improved microwave performances.

Understanding magnetization reversal is very important in designing high density and high data transfer rate recording media. This research has been motivated by interest in developing new nonvolatile data storage solutions as magnetic random access memories - MRAMs. This dissertation is intended to provide a theoretical analysis of static and dynamic magnetization switching of magnetic systems within the framework of critical curve (CC). Based on the time scale involved, a quasi-static or dynamic CC approach is used. The static magnetization switching can be elegantly described using the concept of critical curves. The critical curves of simple uncoupled films used in MRAM are discussed. We propose a new sensitive method for CC determination of 2D magnetic systems. This method is validated experimentally by measuring experimental critical curves of a series of Co/SiO2 multilayers systems. The dynamics switching is studied using the Landau-Lifshitz-Gilbert (LLG) equation of motion. The switching diagram so-called dynamic critical curve of Stonerlike particles subject to short magnetic field pulses is presented, giving useful information for optimizing field pulse parameters in order to make ultrafast and stable switching possible. For the first time, the dynamic critical curves (dCCs) for synthetic antiferromagnet (SAF) structures are introduced in this work. Comparing with CC, which are currently used for studying the switching in toggle MRAM, dCCs show the consistent switching and bring more useful information on the speed of magnetization reversal. Based on dCCs, better understanding of the switching diagram of toggle MRAM following toggle writing scheme can be achieved. The dynamic switching triggered by spin torque transfer in spin-torque MRAM cell has been also derived in this dissertation. We have studied the magnetization's dynamics properties as a function of applied current pulse amplitude, shape, and also as a function of the Gilbert damping constant. The great important result has been obtained is that, the boundary between switching/non-switching regions is not smooth but having a seashell spiral fringes. The influence of thermal fluctuation on the switching behavior is also discussed in this work.

Cette thèse comporte deux parties. L'objectif de la première partie était l'implémentation d'une méthode de résolution de l'équation dynamique de Landau-Lifshitz-Gilbert fondée sur l'approximation des éléments finis. Deux approches ont été présentées: la première reposant sur une formulation avec contrainte et la seconde mettant en œuvre des fonctions tests dans le plan tangent à l'aimantation. Seule la seconde approche reproduit la dynamique obtenue en différences finies sur des cas tests. Dans la seconde partie, le but était d'étudier le déplacement de parois de domaines magnétiques sous l'effet d'un champ magnétique ou d'un courant polarisé en spin dans des systèmes à anisotropie perpendiculaire. Il a été nécessaire d'introduire dans l'équation dynamique les termes dus au transfert de spin. Des systèmes idéaux et des systèmes avec différents types de défauts ont été étudiés. Les résultats numériques ont été comparés avec les données expérimentales disponibles.

Les nano-oscillateurs à transfert de spin (STNO) sont des candidats prometteurs pour la réalisation de composants radiofréquence (RF) intégrés, du à leur taille nanométrique, l'importante gamme de fréquences de base qu'ils peuvent couvrir, ainsi qu'à leur accordabilité autour de ces fréquences de base. Le signal RF est obtenu grâce à l'effet de transfert de spin (STT) qui donne lieu à une oscillation non-linéaire de l'aimantation dans un élément magnétorésistif. Jusqu'ici, ces excitations ont été examinées dans le cadre d'une couche magnétique isolée, c'est-à-dire sans prendre en compte le couplage entre couches. Cependant, nombreux aspects du spectre d'excitation ne peuvent pas être expliqués si l'on considère une couche isolée. Dans cette thèse nous nous attacherons à répondre à la question importante du couplage dynamique entre couches magnétiques dans un nanopilier magnétorésistif, afin de développer une meilleure compréhension des spectres d'excitation, et en particulier la dépendance en courant et champ magnétique appliqué des caractéristiques du pic d'émission, telles que la largeur de raie et la fréquence. Une première étude est réalisée pour un système composé de deux couches ferromagnétiques, couplées entre elles par le couplage RKKY (ce système est appelé un ferrimagnétique synthétique (SyF)). Le couplage induit des différences importantes dans la dépendance en courant de la fréquence par rapport aux excitations d'une couche isolée. Ces différences sont expliquées par l'important couplage dynamique RKKY. Une seconde étude prend en considération une interaction plus complexe, ayant lieu dans un nano-pilier STNO standard basé sur jonctions tunnel ou vannes de spin. Ce dispositif est composé d'un SyF ainsi que d'une couche libre(FL) magnétique, séparés par une fine couche métallique ou isolante. Pour ce système, en plus du couplage dynamique RKKY propre au SyF, nous prenons en compte le couplage dynamique généré par le champ dipolaire ainsi que le spin-torque mutuel (MSTT) entre la couche libre et le SyF. Ce couplage multiple donne lieu à deux signatures distinctes. La première est l'apparition d'un " saut " dans le spectre d'excitation dû à l'hybridation des modes SyF and FL dans le régime atténué. Le second est dû à l'interaction entre les excitations en régime entretenu, éventuellement via leurs composantes harmoniques, avec les excitations en régime atténué. Cette interaction donne lieu à des discontinuités dans la dépendance fréquence - champ, ce lorsque les excitations FL sont prédominantes. Il est intéressant de noter que cela mène à des régions ou la largeur de raie est diminuée. De plus, lorsque les excitations SyF sont prédominantes, la largeur de raie est diminuée par rapport aux cas ou les excitations FL sont prédominantes. Partant de ces observations, nous proposons une structure plus complexe, où un seconde couche de type SyF remplace la couche libre. Les résultats obtenus par une combinaison d'expériences, de simulations numériques et d'analyse analytique, montrent le rôle important des interactions dynamiques dans un nano-pilier. Ils ouvrent de nouvelles voies pour la conception de configurations STNO qui mèneront à des améliorations des performances du signal ainsi synthétisé.

The impact of the spin degree of freedom on the transport properties of electrons traveling through magnetic materials has been known since the pioneer work of Mott [1]. Since then it has been demonstrated that the spin angular momentum plays a key role in the scattering process of electrons in magnetic multilayers. This role has been emphasized by the discovery of the Giant Magnetoresistance in 1988 by Fert and Grunberg [2, 3]. Among the numerous applications and effects that emerged in mesoscopic devices two mechanisms have attracted our attention during the course of this thesis: the spin transfer torque and the spin Hall effects. The former consists in the transfer of the spin angular momentum from itinerant carriers to local magnetic moments [4]. This mechanism results in the current-driven magnetization switching and excitations, which has potential application in terms of magnetic data storage and non-volatile memories. The latter, spin Hall effect, is considered as well to be one of the most fascinating mechanisms in condensed matter physics due to its ability of generating non-equilibrium spin currents without the need for any magnetic materials. In fact the spin Hall effect relies only on the presence of the spin-orbit interaction in order to create an imbalance between the majority and minority spins. The objective of this thesis is to investigate the impact of disorder on spin dependent transport phenomena. To do so, we identified three classes of systems on which such disorder may have a dramatic influence: (i) antiferromagnetic materials, (ii) impurity-driven spin-orbit coupled systems and (iii) two dimensional semiconducting electron gases with Rashba spin-orbit coupling. Antiferromagnetic materials - We showed that in antiferromagnetic spin-valves, spin transfer torque is highly sensitive to disorder, which prevents its experimental observation. To solve this issue, we proposed to use either a tunnel barrier as a spacer or a local spin torque using spin-orbit coupling. In both cases, we demonstrated that the torque is much more robust against impurities, which opens appealing venues for its experimental observation. Extrinsic spin-orbit coupled systems - In disordered metals accommodating spin orbit coupled impurities, it is well-known that spin Hall effect emerges due to spin dependent Mott scattering. Following a recent prediction, we showed that another effect coexists: the spin swapping effect, that converts an incoming spin current into another spin current by "swapping" the momentum and spin directions. We showed that this effect can generate peculiar spin torque in ultrathin magnetic bilayers. Semiconductors spintronics - This last field of research has attracted a massive amount of hope in the past fifteen years, due to the ability of coherently manipulating the spin degree of freedom through interfacial, so-called Rashba, spin-orbit coupling. However, numerical simulations failed reproducing experimental results due to coherent interferences between the very large number of modes present in the system. We showed that spin-independent disorder can actually wash out these interferences and promote the conservation of the spin signal. In the course of this PhD, we showed that while disorder-induced dephasing is usually detrimental to the transmission of spin information, in selected situation, it can actually promote spin transport mechanisms and participate to the enhancement of the desired spintronics phenomenon.

Les nano-oscillateurs à transfert de spin sont des composants radiofréquences magnétiques non-linéaires, nanométrique, de faible consommation en énergie et accordables en fréquence. Ce sont aussi potentiellement des candidats prometteurs pour l’élaboration de larges réseaux d’oscillateurs couplés. Ces derniers peuvent être utilisés dans les architectures neuromorphiques qui nécessitent des assemblées très denses d’unités de calcul complexes imitant les neurones biologiques et comportant des connexions ajustables entre elles. L’approche neuromorphique permet de pallier aux limitations des ordinateurs actuels et de diminuer leur consommation en énergie. En effet pour résoudre des tâches cognitives telles que la reconnaissance vocale, le cerveau fonctionne bien plus efficacement en terme d’énergie qu’un ordinateur classique. Au vu du grand nombre de neurone dans le cerveau (100 milliards) une puce neuro-inspirée requière des oscillateurs de très petite taille tels que les nano-oscillateurs à transfert de spin. Récemment, une première démonstration de calcul neuromorphique avec un unique nano-oscillateur à transfert de spin a été établie. Cependant, pour aller au-delà, il faut démontrer le calcul neuromorphique avec plusieurs nano-oscillateurs et pouvoir réaliser l’apprentissage. Une difficulté majeure dans l’apprentissage des réseaux de nano-oscillateurs est qu’il faut ajuster le couplage entre eux. Dans cette thèse, en exploitant l'accordabilité en fréquence des nano-oscillateurs magnétiques, nous avons démontré expérimentalement l'apprentissage des nano-oscillateurs couplés pour classifier des voyelles prononcées avec un taux de reconnaissance de 88%. Afin de réaliser cette tache de classification, nous nous sommes inspirés de la synchronisation des taux d’activation des neurones biologiques et nous avons exploité la synchronisation des nano-oscillateurs magnétiques à des stimuli micro-ondes extérieurs. Les taux de reconnaissances observés sont dus aux fortes accordabilités et couplage intermédiaire des nano-oscillateurs utilisés. Enfin, afin de réaliser des taches plus difficiles nécessitant de larges réseaux de neurones, nous avons démontré numériquement qu’un réseau d’une centaine de nano-oscillateurs magnétiques peut être conçu avec les contraintes standards de nano-fabrication. / Spin-torque nano-oscillators are non-linear, nano-scale, low power consumption, tunable magnetic microwave oscillators which are promising candidates for building large networks of coupled oscillators. Those can be used as building blocks for neuromorphic hardware which requires high-density networks of neuron-like complex processing units coupled by tunable connections. The neuromorphic approach allows to overcome the limitation of nowadays computers and to reduce their energy consumption. Indeed, in order to perform cognitive tasks as voice recognition or image recognition, the brain is much more efficient in terms of energy consumption. Due to the large number of required neurons (100 billions), a neuromorphic chip requires very small oscillators such as spin-torque nano-oscillators to emulate neurons. Recently a first demonstration of neuromorphic computing with a single spin-torque nano-oscillator was established, allowing spoken digit recognition with state of the art performance. However, to realize more complex cognitive tasks, it is still necessary to demonstrate a very important property of neural networks: learning an iterative process through which a neural network can be trained using an initial fraction of the inputs and then adjusting internal parameters to improve its recognition or classification performance. One difficulty is that training networks of coupled nano-oscillators requires tuning the coupling between them. Here, through the high frequency tunability of spin-torque nano-oscillators, we demonstrate experimentally the learning ability of coupled nano-oscillators to classify spoken vowels with a recognition rate of 88%. To realize this classification task, we took inspiration from the synchronization of rhythmic activity of biological neurons and we leveraged the synchronization of spin-torque nano-oscillators to external microwave stimuli. The high experimental recognition rates stem from the weak-coupling regime and the high tunability of spin-torque nano-oscillators. Finally, in order to realize more difficult cognitive tasks requiring large neural networks, we show numerically that arrays of hundreds of spin-torque nano-oscillators can be designed with the constraints of standard nano-fabrication techniques.

<div>Neuromorphic computing has gained tremendous interest because of its ability to overcome the limitations of traditional signal processing algorithms in data intensive applications such as image recognition, video analytics, or language translation. The new computing paradigm is built with the goal of achieving high energy efficiency, comparable to biological systems.</div><div>To achieve such energy efficiency, there is a need to explore new neuro-mimetic devices, circuits, and architecture, along with new learning algorithms. To that effect, we propose two main approaches:</div><div><br></div><div>First, we explore an energy-efficient hardware implementation of a bio-plausible Spiking Neural Network (SNN). The key highlights of our proposed system for SNNs are 1) addressing connectivity issues arising from Network On Chip (NOC)-based SNNs, and 2) proposing stochastic CMOS binary SNNs using biased random number generator (BRNG). On-chip Power Line Communication (PLC) is proposed to address the connectivity issues in NOC-based SNNs. PLC can use the on-chip power lines augmented with low-overhead receiver and transmitter to communicate data between neurons that are spatially far apart. We also propose a CMOS '<i>stochastic-bit</i>' with on-chip stochastic Spike Timing Dependent Plasticity (sSTDP) based learning for memory-compressed binary SNNs. A chip was fabricated in 90 nm CMOS process to demonstrate memory-efficient reconfigurable on-chip learning using sSTDP training. </div><div><br></div><div>Second, we explored coupled oscillatory systems for distance computation and convolution operation. Recent research on nano-oscillators has shown the possibility of using coupled oscillator networks as a core computing primitive for analog/non-Boolean computations. Spin-torque oscillator (STO) can be an attractive candidate for such oscillators because it is CMOS compatible, highly integratable, scalable, and frequency/phase tunable. Based on these promising features, we propose a new coupled-oscillator based architecture for hybrid spintronic/CMOS hardware that computes multi-dimensional norm. The hybrid system composed of an array of four injection-locked STOs and a CMOS detector is experimentally demonstrated. Energy and scaling analysis shows that the proposed STO-based coupled oscillatory system has higher energy efficiency compared to the CMOS-based system, and an order of magnitude faster computation speed in distance computation for high dimensional input vectors.</div>